Cover by Gerald Rice
www.cnckitsplus.com
I
Copyright © 2011
All rights reserved. No part of this book shall be reproduced or transmitted in any form or by any
means (i.e., electronic, mechanical, magnetic, photographic), including photocopying, recording or
by any known information storage and retrieval system, without prior written permission of the
publisher. No patent liability is assumed with respect to the use of the information contained herein.
Although every precaution has been taken in the preparation of this book, the publisher and author
assume no responsibility for errors or omissions; neither is any liability assumed for damages
resulting from the use of the information contained herein. This material is subject to change without
notice.
ISBN 978-0-9767596-2-1
Written by Benjamin Fleming
Printed in the United States of America
Publishing date: 2011
1st Printing
II
EDM Resources
EDM PCB & Kit
No point-to-point or detailed wiring description is given in this book, though complete schematics of
the various circuits are shown. For those who may feel uncomfortable wiring from a schematic, a
commercially produced Printed Circuit Board (PCB) is available. As time and funds permit, it is
hoped that a complete kit of materials or a stuffed board may be available for purchase.
For the most up-to-date information on the cost and availability of boards, books and kits visit
WWW.homebuiltedmmachines.com
Should you have a question, contact me at the Email address below:
homebuiltedmmachines@gmail.com
Yahoo group:
A Yahoo group has been formed to assist builders of the Pulse as well as RC design.
http://groups.yahoo.com/group/EDMHomeBuilders
III
Disclaimer and License Agreement
The author, Benjamin Fleming, is NOT a professional engineer, nor has he had any formal training in
the design of EDM (Electrical Discharge Machine/ing) technology or apparatus.
The designs found herein require the use of voltages and currents known to
KILL PEOPLE!
The said author, as well as the distributors of this publication, assume no liability for injury to
person(s) or property that may result from the building, testing, or operations of designs described in
this publication. The intent of this publication is for
EDUCATIONAL PURPOSES ONLY
License agreement:
By purchasing Build a Pulse EDM Machine book, it entitles the purchaser the right to build an
EDM of this design for personal use only.
The design outlined in this text is not to be built and sold for gain or profit without agreement
from the author, Benjamin Fleming. All violators of this agreement will be aggressively
prosecuted.
In other words, you cannot go into business building this machine without my approval. If you want
to do that, manufacture and sell your own design.
IV
What is Electrical Discharge Machining?
Electrical Discharge Machining (EDM) is a process that uses electrical discharges emitted from an
electrode to erode an electrically conductive material. As a result, it is possible to erode or “burn” the
reverse shape of the electrode into the workpiece.
An EDM system is comprised of a dielectric reservoir (or storage tank), a work tank, a dielectric
pump and filter, and the servo control and generator, also known as a power supply. Also included
are associated wiring, plumbing and fixturing.
The workpiece is placed in the dielectric tank and affixed to a metal plate in the tank. The tank is
filled with a hydrocarbon dielectric, which ionizes (i.e., becomes conductive) in the presence of an
electrical field. The dielectric fluid breaks down electrically after a short ionization period. The
electric field is created by applying a voltage between the electrode and the workpiece (known as the
gap). Since in EDM the gap is very small (a couple of thousandths of an inch at most), the applied
voltage does not need to be very large for an EDM “spark” to occur. Typical operating gap voltages
for EDM machines are in the range of 20-50VDC; the ionization voltage may be quite a bit higher.
The servo control system maintains the appropriate separation of the electrode and workpiece as
determined by the operator or computer. The gap voltage feeds back to the servo control system so
that the proper separation of the electrode and workpiece can be maintained. As each spark is
discharged from the electrode (the cutting tool) to the workpiece, a small amount of metal is
vaporized and a crater is left in the workpiece. The dielectric fluid quickly cools the vaporized metal.
The solidified metal partials, known as swarf, are removed from the work area by the circulating of
the dielectric oil; the filtering system traps and holds the swarf.
The EDM burning process is repeated thousands (even millions) of times each second as the
workpiece is “eroded away”. As a result of millions of electrical sparks melting small quantities of
metal, it is possible to erode various shapes into any material that will conduct electricity; even some
semiconductor materials may be machined via this process. Generally speaking, the hardness of the
workpiece has little effect on the burn; this is one of the biggest advantages of the EDM process.
V
Table of Contents
Chapter 1
Introduction
I am in a Quandary
What do you Want in an EDM
It is More than a Power Supply
Some Inspiration
1-1
1-1
1-2
1-2
1-3
Chapter 2
A Little EDM History
2-1
Chapter 3
Overview of Mechanical & Electrical Components
Mechanical
Electrical
3-1
3-1
3-2
Chapter 4
The Power Supply
The Power Circuit
The Timer
Control and Motor Driver Circuit
How Does a Window Comparator Work?
Conditions at Startup
Short-Circuit Condition
What to Do When the Current Stops Flowing
The Peak Detector
Manual Up/Down and Dither Circuit
4-1
4-2
4-2
4-3
4-4
4-4
4-4
4-5
4-5
4-6
Chapter 5
Building the Power Supply
Soldering Tips
Construction Notes
Power-Supply Housing
Case Stiffeners
Front Panel & Power-Supply Layout
Power-Supply Layout
Major Components and Mounting
Wiring the High-Voltage Section
Power-Resistor Sizing and Mounting
Finish High-Voltage Wiring
PCB Location & Heat Sink Selection
Wiring the PCB
5-1
5-1
5-2
5-3
5-4
5-6
5-5
5-6
5-7
5-8
5-11
5-12
5-12
VI
EDS Issues
PCB-to-Panel Connections
5-12
5-13
Chapter 6
The Servo
DC Gear Motor vs. Stepper
Servo Motor Selection for the Ram
The Lead Screw
Removing Shaft Endplay
Attaching the Lead Screw to the Gearmotor Shaft
Pin Method
Glue Method
The Slide
Slide Assembly
Final Slide Adjustment
Mounting Arm
Mounting Stud
Construction of the Head Bracket
Front Cover
The Finished Head
Head Wiring Options
Wiring the Head to the Power Supply
ON/OFF Motor Control
ON, OFF and UP limit Switch
ON, OFF, UP and DOWN Limit Switch
ON, OFF, UP, DOWN Limit and Depth-Control Switch
6-1
6-2
6-2
6-3
6-3
6-3
6-3
6-4
6-5
6-6
6-6
6-7
6-7
6-8
6-10
6-10
6-11
6-11
6-12
6-12
6-12
6-13
Chapter 7
Indexing
Indexer Chuck Overview
Indexer Chuck with no Metal Insert
Metal-Inserted Indexer Chuck
The Metal Insert
Misc. Indexer Chuck Photos
Electrode Manufacturing Fixture Block
Electrode Holders
The Squaring Tool
7-1
7-2
7-2
7-4
7-4
7-5
7-6
7-7
7-8
VII
Chapter 8
Depth Control
8-1
Chapter 9
Finish the Job
9-1
9-1
9-1
9-2
9-2
Building a Dedicated, or Add-on, EDM
Overview of the Work Tank and XY Movement
The Sherline Conversion
XY Tables and Selection
Chapter 10
Dielectric & Filter Selection
How Flammable is the Dielectric?
Dielectric & Materials Compatibility
Filters
Automotive Spin-on Filters
Automotive Filter Mounts/Filters
Diesel Filters
Water Filters
Filter Bags/Prefilter
5-Gallon Bucket Bags
10-1
10-2
10-2
10-3
10-3
10-4
10-4
10-4
10-4
10-5
Chapter 11
Tank & Pump Selection
Other Storage/Work Tank Options
Pump Selection
The Pump Dilemma
Submergible Pump
Non-submergible Pump Options
11-1
11-1
11-3
11-4
11-4
11-5
Chapter 12
The Back Slide
Frame Work Overview
Back Slide Rail
CRS WAYS
Tubular Frame Option
Wood Deck Option
Final Assembly of the Slide and WAYS
12-1
12-2
12-2
12-3
12-3
12-4
12-6
VIII
Chapter 13
Tank Assembly
Work Tank Support and Mounting Feet
Work Plate
Plumbing the Tanks
Dielectric Flow Control
Other Valve Options
Flexible Flushing Line
Variable-Speed Motor
Work Tank Dielectric Level Control
Pre-filter (Bag Filter)
13-1
13-1
13-2
13-3
13-5
13-6
13-6
13-7
13-7
13-7
Chapter 14
The Spark
What Happens in the Gap?
Duty Cycle
Frequency of the Spark
Metal Removal Rates
DC Arcing
14-1
14-1
14-3
14-4
14-4
14-5
Chapter 15
Electrodes
Factors Contributing to Reduced Electrode Wear
Electrode Polarity
Volumetric Wear
End Wear
Corner Wear
Electrode Wear Test
Overcut
Recast Layer
Electrode Materials
Metallic Electrodes in General
Brass
Copper
Copper/Silver Tungsten
Graphite Electrodes
Copper Graphite
Graphite Machining Tips
Suppliers of EDM Graphite
15-1
15-1
15-2
15-3
15-3
15-3
15-3
15-4
15-4
15-5
15-5
15-5
15-6
15-6
15-7
15-8
15-8
15-9
IX
Chapter 16
Flushing
EDM Flush Fittings
Jet Flushing
Through-electrode Pressure Flushing
Sealed Flush
Pot Flushing
Suction Flushing
Trepanning
16-1
16-1
16-2
16-2
16-3
16-3
16-4
16-4
Chapter 17
Your First Burn
17-1
Chapter 18
Jewelry Manufacturing and EDM Tips
Using Tokens and Coins as an Electrode
Using a Copper Earring as an Electrode
Pressed Arbor
Glued Arbor
Glued-together Electrodes
Upside Down and Sideways
Carbide Insert Tool Holder
Multiple Electrode Burns
Broken Screw and Bolt Removal Techniques
18-1
18-1
18-2
18-3
18-3
18-3
18-3
18-3
18-4
18-4
Chapter 19
Glossary
19-1
Chapter 20
Reference Materials
20-1
Parts List
21-1
21-1
21-3
Electronic
Non-Electronic
X
Preface
Many of you have heard this story, but please bear with me for those not familiar with it. My
introduction to the world of EDM came at a yard sale in the early 1980’s! I purchased two boxes of
old Popular Science and Popular Mechanics magazines. In one of the magazines, the basic EDM
theory was explained, along with details of how to build a very simple, manually operated EDM. I
built the machine and really became obsessed with the EDM process, but the task of manually
advancing the electrode (by listening to the sound of the burn) was no fun! The EDM process is a
slow one, so a person could spend lots of time just twisting a lead screw a few degrees at a time!! I
saw the need for an automated control of the electrode.
From that meager beginning, attempts were made to find additional information (This was pre-World
Wide Web; imagine that!) on the control systems and power supplies of EDM equipment. One EDM
company engineer (from ELOX) told me that “the kind of information you are seeking is proprietary
and only actively worked on by about 20 people in the U.S.” After that comment, I was determined to
design an EDM that would automatically position the electrode while I was doing other things in the
shop.
Several months passed and I was at a Radio Shack store, looking through electronics books; I read a
description of a window comparator circuit, and wondered if this circuit might be used to control the
position of the electrode. I build a window comparator circuit and connected it to a relay to control a
DC motor. After much experimentation, I was able to develop a system that provided some
automated control of the electrode; that was an exciting day in Charleston, South Carolina!!
In the briefest terms, that is how I came about developing my first automated Resistive Captive (RC)
EDM machine. Along the way I have attempted to get those with more electronic knowledge than I
had (I am a machinist by trade) involved in this project; I have spoken with at least several hundred
electronics hobbyists, as well as more than a few electrical and electronic engineers in these past
thirty years. Some have gotten involved for a brief time, but none have stayed around to ever see any
designs end in a productive way.
In about 2005 I decided to write a book on how to build the automated RC EDM machine. In the
summer of 2006, I completed The EDM How-To Book, which many of you have read and used to
build the machine described. There is much to be said for the simplicity and ruggedness of the RC
EDM power supply, but it was not a true EDM in the eyes of many. I was constantly asked if I could
design a pulse EDM for more precision work.
I, too, have wanted a true pulse machine but did not have the money to purchase one, so I set about
to design and build it. I have worked more hours and spent more dollars than I want to think about
trying to figure out how to get one to work. This book is the accumulation of many hundreds of hours
of experimenting and learning. If I paid myself for the hours invested, I don’t think I would ever
break even, but it has been a learning and growing experience.
XI
Chapter 1
I am in a Quandary
You don’t have to be crazy to spend a large part of your life
trying to figure out how to build an Electrical Discharge
Machine (EDM) pulse machine but it helps. The photo on
the left shows my workshop as it often looks; I am sure
OSHA would have shut me down long time ago if they only
knew. Sometimes it looked much worse so if something in
this book doesn’t quite make sense you will understand.
As I start writing this book I am in a quandary as to how to
proceed. The audience that I am attempting to address has
very different backgrounds and skills. They are, for the
most part, people with electronics and machinist backgrounds. Until the introduction of computers in
the machine shop, the two usually had little in common. There are many in the electrical/electronic
field who see EDM as a fascinating journey into the little known field of metal working. The metal
workers and machinists are often intimidated by anything to do with electronics, but thrill at the
ability to cut heretofore extremely difficult materials. Then there is a new, rapidly growing group of
people interested in computer control of machines, tools and robots, which in most cases is a group
that bridges the gap between the two previously mentioned groups. The question arises: Which group
should I cater to, or how do I address the various groups fairly?
From an electronic perspective designing a simple pulse machine was a challenge for me. I have
come to understand that SIMPLE IS HARD! Even though it is simple compared to most commercial
pulse EDM machines it is still complex for the non-electrical/electronic person. I feared this would
discourage most “machinist types” before starting the project. The “electronics and computer types,”
on the other hand, may enjoy the “thrill of the chase” in building from scratch. In an attempt to level
the playing field, a Printed Circuit Board (PCB) is available see page III (in the front of the book). By
utilizing the PCB it’s possible for even the novice electronics person to overcome the wiring
complexity of a vast part of the machine. Now my fear is that the electronic types might be
discouraged by the complexity of the mechanical design.
For those more electronically inclined, schematics of the EDM are presented in the text, with a very
limited discussion as how to build from the schematics. A person knowledgeable of electrical and
electronic schematics and skill with point-to-point wiring techniques should be able to successfully
wire the power supply.
On the mechanical side, this is not a “cut and paste” book for building the EDM. If there is one thing
I know about most machinists, it is that they are a creative and talented bunch. Usually all that’s
needed is a napkin sketch or a photograph to be off and running. In most cases I give drawings and
photos of what I built, but frequently I suggest several optional ideas and let you decide what best fits
your needs, skill level, and budget.
1-1
One of the biggest problems I faced on the mechanical side is those persons with access to a full
machine shop who wants a very professionally designed mechanical machine. Then there are those
with a hacksaw and a drill press and minimal metal working skills who also want to build a machine.
I have access to a complete machine shop (including a Pulse-type EDM ) and the skill to build.
I could build a very professional machine, but what percentage of the readership would be able to
duplicate it? Not many, I would suspect. How do I satisfy the various groups?? I am not sure I have
that answer, so I have tried to hit a middle-of-the-road approach and hope that the two groups
understand the position I am in as an author/writer and make the deviations necessary to build a
machine that will fit their needs and skill level.
What do you Want in an EDM?
Before beginning to build a machine you should ask yourself: “what is the intended purpose of
building a pulse EDM?” If all you are wanting is the ability to burn out the occasional broken tap or
drill perhaps a much simpler, easier-to-operate, and much less expensive machine, a Resistor
Capacitor RC-type EDM will serve the purpose. Such a machine is described in my previous book,
The EDM How-To Book. For those who are new to EDM the difference between the two is that the
RC type machine uses a capacitor, discharged across the gap to remove metal. The pulse machine
does not discharge a capacitor; instead the spark is controlled by a timer and power Mosfets.
One of the main advantages of a pulsed supply is decreased electrode wear (in most cases) and
increased cutting speed. As with most things, this comes at the cost of increased power supply
complexity, which is more prone to failure and much more expensive to build.
With a pulse machine, work such as that shown in the photo to the left and on the books back cover is
possible (also see page 18-2). To achieve such results, the Chuck E. Cheese token was burned into
tool steel for 20 minutes using a 200us ON time. Then, for 25 minutes, the burn ON time was
decreased to 25us. The image was then buffed with jeweler’s rouge for about 10 minutes.
If the need is for a “tap burner” and electrode wear is not a
serious issue, then I suggest building the simpler RC supply. If
there is a need for doing accurate EDM work, such as squaring
corners in milled pockets or machining difficult shapes in steel or
other exotic metals, or finer finish work such as jewelry work,
then a pulse machine is most likely exactly what you are looking
for. On the other hand, if you love to build and you are looking
for a different kind of project this may just be the ticket. Not many machine shops can boast of a true
pulse EDM setting on the work bench for peanuts on the dollar.
It is More Than a Power Supply
A true EDM is more than just the power supply; a complete system is needed to perform exacting
work. This includes the pulse power supply with automatic dithering capabilities, an electronic edge
finder for ease of set-up, and a precision ram with electrode-indexing capabilities. The ram should
have burn-depth control, while the dielectric tank must provide flushing as well as a dielectric
1-2
filtering system and a method to hold the workpiece. Rapid filling and emptying the work tank are
needed as well, to speed the process.
If you purchased my previous book and built the servo ram and dielectric tank, there is good news:
You are well on the way to having an advanced EDM system! The previous ram and dielectric tank
will work seamlessly with the new Pulse EDM power supply. All that’s needed is the pulse power
supply, though some may wish to upgrade the servo ram with the advanced capabilities suggested
herein. For those who do not have a copy of previous book, fear not; this book is a complete standalone book for building a complete EDM system.
Some Inspiration
I am including a few photos of work done by readers of
The EDM How-To Book. The photo on the left is a superb job done by
Alan on a PCB. This board appears to include most of the electronics
for the RC machine, except for the transformers and the power
resistors.
The photo on the right is a blackfaced cabinet that Bob designed and manufactured. In my opinion,
it is one of the best looking EDMs I have seen.
On the right, you can see that Jim did a very nice job on a tank design
(which is incomplete in this picture) that uses a vertical column to
attach the servo ram to. This makes for a very compact EDM, but as
with most things, there is a tradeoff (See page 11-1 & 2).
A talented Mr. Cary (combining various
ideas gleaned from several sources) built an
extremely precise and rugged EDM ram,
which is shown in the photo to the left. This
head uses a stepper motor to position the electrode. Such a system requires
a completely different motor driver than is outlined in the previous book or
in this text. Even though it does not follow my design, I could not pass up
an opportunity to display his talents.
Rich designed a ram that incorporates a
ball clutch (See the photo on the right).
This permits the lead screw (a precision
ball screw in this design) to
mechanically disengage should binding occur. This feature
mechanically protects the drive motor and the motor driver chip.
Others have followed Rich’s design ideas and built similar clutch
arrangements.
1-3
Chapter 2
A Little EDM History
In the 1770’s, an English scientist, Joseph Priestley, discovered the eroding effect of electrical
discharges. More than 150 years passed before some practical use was made of the discovery. The
Russians are usually given credit for discovering the EDM process in 1943, though there are some
justifiable reasons to suspect that they might not have been the first (see “Cutting Metal By
Electricity,” patent number 416,873, Dec. 10, 1889, issued to Mr. Benjamin C. Tilghman of Pa.).
The process was first called Spark Erosion and is still known by that term in some quarters; more
recently it is referred to as Electrical Discharge Machining (EDM). During an investigation of how to
suppress the erosion of electrical switch contacts, the Soviet scientists B. R. Lazarenko and N. I.
Lazarenko had an idea. Why not use the destructive effect of the electrical charge in a controlled
manner to machine metals (being developed for the war effort), which were proving difficult to
machine with conventional machine tools? They developed a working spark erosion machine, where
an electrical discharge takes place between two conductors in a dielectric liquid bath.
The Lazarenkos developed what early on came to be called a Lazarenko circuit. Today it is referred
to as a relaxation circuit (RC). It was used for a while in the building of generators for EDM or Spark
Erosions machines as they are referred to in Europe. This concept is used even today in some
machines to obtain extremely fine surface finishes, as well as in Micro (micron range) EDM
applications. Soon after the first RC generators were in operation, the never-ending research was on
for a better power supply. The search was for a system that would cut faster, be more efficient, and
produce less wear on the electrode and more wear on the workpiece.
During the search, circuits were developed that used a motor generator in place of the DC power
supply; they were called Rotary-Impulse Generators (RIGs), and they used a capacitor as well as a
diode in the charging and discharging circuit. These high-frequency AC generators produce
unidirectional pulses. The RIG was/is capable of removing metal at very high rates but, in doing so,
produced very rough surface finishes. The operating frequency was low and had minimal
adjustability, making fine surface finishes nearly impossible.
The next step in advancement was a modification of the RC circuit design. The resistor in the RC
circuit was removed and a series of vacuum tubes (connected in parallel) replaced the resistor though
the discharge capacitor in parallel with the gap, which remained as in the original RC circuit. The
bank of vacuum tubes became, in effect, a variable resistor that could be turned on and off depending
on the flow of current in the gap. When the cutting capacitor was discharged, a sensing circuit turned
the vacuum tubes on and charged the capacitor, and then the discharge/charge cycle repeated. The
sensing circuit had a limited ability to detect when a short occurred in the gap. If a short was detected,
current flow was stopped until the short was cleared. This was a great improvement over the RC
design but is not what we think of today when describing a pulse EDM.
In the next phase of evolution of EDM power supplies, it was discovered that it was not necessary to
sense when current stopped flowing in the gap to turn the tubes on and charge the capacitor. In this
design, the only connection that the tubes had with the gap was to interrupt all current flow in case of
2-1
a shorted condition in the gap. The grids of the parallel tube bank were
connected to a fixed-frequency oscillator, the discharge capacitor still
remained in parallel with the gap, and the tubes turned on and off,
charging this capacitor. These new circuits with vacuum tubes were
called Controlled-Pulse Circuits (CPC). It was discovered that the CPC
power supplies removed metal faster and electrode wear was reduced.
Yet the search for ever-higher metal removal rates, lower electrode
erosion, and more control was the cry from those in the industry. Things
have not changed much; that is still the goal. We appear to always be
looking for better and faster horses.
The electrical characteristics of tubes and the arc gap are not compatible.
Tubes are high-voltage, low-current devices; the gap requires large
currents at relative low voltages. Therefore, large numbers of vacuum
tubes were required to deliver the needed current. The photo on the right
is an EDM power supply that I came across in storage in Illinois. Note
the bank of tubes in the lower section of the power supply. During
operation, these tubes are protected by closing the two doors.
To get around the tube incompatibility problem, circuits were developed using a small number of
tubes switching on and off the high-voltage (primary) side of a step-down transformer. For those who
may not know, a transformer supplied with pulsing DC on the primary side generates AC on the
secondary side; thus, it requires a diode to rectify the alternating current. The high current-demanding
gap was connected to the low-voltage (high-current) secondary side of the transformer. This design
reduced the number of tubes required, but the use of a transformer restricted the high frequencies
desired for certain machining conditions.
It was not until the invention of the solid state switch (transistor) that an electronic switch was
available that was compatible with the gap demands. Transistors function at high efficiencies for the
gap conditions of low voltage and high current. Using a solid-state switch does not require the use of
a capacitor in parallel with the gap, as did most tube designs. The transistor, as well as power mosfets
and IBGTs, have been used as high-current, low-voltage switches in EDM power supplies in recent
years. The power supply described in this book uses fast-switching power mosfets.
With the advent of solid-state switches came the ability to precisely control the current from pulse to
pulse under varying conditions. Other benefits include the ability to greatly
reduce the DC arcing problem, as well as that to increase the metal removal rate
and reduce the erosion of the electrode under certain machining conditions.
From my research, little appears to have been preserved from a historical aspect
in terms of who did what in the development of the pulse power supplies, but I
have been able to uncover a little information, as found below.
Agie, a Swiss company, was one of the earlier companies to experiment with EDM processes. In
1954 they presented their first die-sinking EDM to the world (see photo above). In 1969, Agie
became the first EDM company to incorporated NC (Numerical Control) controls.
2-2
Charmilles, also a Swiss company, became
interested in the EDM process in 1952 and
formed a machine/tool division in 1953 to
specialize in the design and production of EDM.
They debuted an EDM machine in Milan in
1954-55 at a machine/tool exhibition. The
drawing at left is one I came across while
researching for this book; it is of the Eleroda D1
and is dated 1956. It has lots of handwritten
notation, some in English. This drawing is of an
RC machine that I have been told was
commercially sold, though some later versions of
the same machine evidently had a modified RC
design as they were making steps toward a pulse machine. There are still operational Eleroda D1s
around; from time to time they appear on line auctions! In 1959, at an exhibition in Paris, they
presented the first fully transistorized generator. Charmilles claims to have sold over 10,000 EDMs in
the U.S. and over 40,000 around the world since 1954.
One of the earliest players on the EDM scene in the U.S. was the ELOX Company. At the dawn of
EDM development, it was believed that the metal removal process was through ELectrical
Oxidization; hence the company was called ELOX. ELOX was the largest U.S. manufacturer of
EDM. Today many ELOX machines are found in industries not only in the U.S., but throughout the
world. For many years, the term “ELOX” was used in the U.S., instead of “EDM”
A note of interest; ELOX was purchased by Agie in about 1988, and Charmilles joined forces with
Agie in 1996. In 1999, Agie Charmilles bought the Brazilian EDM manufacturer Engemaq. As of
July 2007, GF AgieCharmilles announced the merger of its former product brands (along with
Mikron) under the single brand GF AgieCharmilles.
There are manufacturers of EDM equipment from the Pacific Rim but I have not been able to track
any of their developments. I do know that Mitsubishi did not market EDM products in North America
until 1983. That year they introduced two CNC ram machines.
In the early 1960’s, as EDM Power Supplies were switching from the RC design to tube-type
supplies (the forerunners of today pulse supplies), high-quality graphite (a spinoff of the U.S. space
program) was introduced in the U.S. that proved to work very well in EDM electrode applications. As
a result, many job shops (in the U.S. in particular) switched from using metallic electrodes to graphite
ones. For whatever reason, graphite has never been accepted in other parts of the world like it has in
the U.S., and metallic electrodes continued to be preferred in many countries.
The concept of orbiting the electrode (moving the electrode in a circular path around a center point)
was introduced sometime in the early to mid-70’s. At first its main application was simply a method
for correcting for undersized electrodes, but soon it was discovered that orbiting improved flushing,
which means that surface finishes improved, cutting times went down and production went up. Today
full CNC control of ram (sinker) machines is common.
2-3
Chapter 3
Overview of Mechanical & Electrical Components
Mechanical
A bit of good news: If you purchased The EDM How-To Book, and built either of the servocontrolled rams described (see photos below ), they will work with the pulse machine without any
modifications, even though they were mainly intended for work such as removal of broken drills and
taps or in applications where precision EDM work is not required. Neither ram has the rigidity, nor
the electrode-indexing capability, for doing the precision EDM work described in this book.
If you truly want to take advantage of the capabilities of the pulse machine, a more advanced and
precision servo ram is required. The new ram designs are based on the earlier designs but are greatly
enhanced.
The dielectric tank and filtering system described in The EDM How-To Book (above right) will
perform its duties flawlessly with no major changes. For those who have not built the dielectric tank
and filtering system, a complete description of a similar system is given.
In the how-to book I suggest that the EDM be installed
on a drill press or a Milling machine. That is still a
very good idea, especially if you have access to such
machine tools. The disadvantage, of course, is that,
most of the time, the mill/drill press cannot be utilized
for EDM-only work, meaning that the machine needs
to be set up and taken down frequently.
I describe how to build a self-contained bench-top
EDM machine (photo at left), which includes a storage
tank and XY table and a work tank all stacked on each
other. The work tank may be easily removed (from the
XY table) and positioned on a milling machine or drill
press if so desired.
3-1
Electrical
I have done everything that I can think of to keep the circuits as simple as possible and yet produce a
powerful and useful pulse EDM machine. The circuit consists of a timer to generate the needed
pulses for cutting and a modified window comparator circuit configured to control the DC servo.
In this text, I will not give instructions for hand-wiring point-to-point connections of the control and
pulse sections of the power supply. Those with some electronics wiring experience may choose to
wire their own supply. The schematic for the PCB may be found on page 4-1; the remaining
schematics are found on pages 5-11 & 5-13. The wiring for the servo head is on page 6-13
For those wishing to keep things as easy as possible a PCB is available (see page III) that greatly
simplifies the wiring. The board requires soldering; if you have never soldered, it is not difficult but,
to ensure best results, please read through the “Soldering” section starting on page 5-1.
The power supply breaks down into three basic parts.
The first section is the high-voltage and -current sections that deliver the brute force needed to
remove metal. In the photos below most of the high-voltage circuit is found across the back and down
the right hand side of the chassis.
The second section is the asymmetrical timing circuitry and the electronic switches (Hexfets) that
turns ON and OFF at precise intervals to deliver the voltage and current to the workpiece.
The third section is the control circuit, which is responsible for making the decisions when to advance
and retract the electrode to ensure a smooth and efficient burn takes place.
The latter two circuits, for the most part, are located on the PCB, which is located in the middle of the
chassis near the front panel.
3-2
Chapter 4
The Power Supply
The design of the power supply (and the control circuit in particular) has not been a simple matter for me.
Evidently that is also the case for others, as I am not aware of any existing plans for a do-it-yourself EDM
pulse sinker (ram) on the market at this time.
On the following pages I describe the operation of the power supply in as simple, non-technical terms as I
know how. I am sure that readers with electronics backgrounds will cringe at my descriptions. I do take a few
liberties, which are not completely correct (in “electronics speak”) in order to simplify the explanation;
Electronic Guys, please bear with me so those less educated (in terms of electronic knowledge) might
understand what is going on. The operation of each section of the power supply is explained.
Below is the schematic of the available PCB as of the writing of this book. A hand-wired version should
work well, but a description of the wiring sequence for such is not found in this text. Though this power
supply works well, you
should know that the power
supply is an ongoing
process of experimentation
and improving. I am
always looking for ways to
improve the performance of
the machine. Faster cuts and
better surface finishes are
constantly the goal.
You can chase that faster,
better rabbit around and
around and never catch it
because there is always
another way to build that
might prove to be better. I
finally decided to stop
experimenting and finish
the book. You have no idea
how many times this circuit
has changed.
The resulting circuit is not
perfect, but it works very
well. Sometimes you just
have to shoot the
experimenter & tinkerer and
go to print; otherwise, it is
never going to happen.
Even though this chapter is early in the book, it is actually the last chapter written. This ensured that the most
current version of the power supply as of the printing is depicted. Know that if you purchase a PCB, the
circuit may be different from what is found in this set of plans.
4-1
The Power Circuit
When the Hexfets (Q3-8) turn ON (see drawing below), current flows from the power supply (rectifier D20
and the large C20 capacitor) through the current-controlling power resistor/s R40 (and R41 through R43 if
they are switched on? in the circuit with their respective switch), through the SW15 and SW16 switches, and
to the gap. If conditions in the gap are good, a spark jumps across the gap (the EDM process) and returns to the
negative side of the power supply via D21 and the six parallel Hexfet switches (PCB-mounted).
Current flows as long as the
Hexfets are ON and the gap
conditions are such to permit
current to flow. After a
predetermined time, the timer
switches OFF. This OFF time is
necessary, to deionize the gap and
prepare for the next cycle.
Resistors R40-47 and the gap
control the amount of current that
flows in the circuit. CB1 is a
Circuit Breaker.
The Timer
This particular pulse EDM has a free-running timer, i.e., it turns ON/OFF at a set rate irrespective of
conditions in the gap. This is not necessarily the best or most efficient design, but it keeps the circuit simple
and low-cost, and works well for most home or small shop applications. There are many older commercial
machines operating today with similar or even simpler types of pulse-control circuit designs.
The heart of the switching circuit is based on one of the most popular IC (integrated circuit) chips ever
manufactured, the 555 timer. The most common wave form used in the EDM process is an asymmetrical
square wave, i.e., the ON time is longer than the OFF time (most of the time).
Adjustment of front-panel control
potentiometers (pot). R35 (or R35A-F
see page 4-3) controls the length of
time Q3-8 are turned ON; this is
known as the DURATION.
Adjustment of R34 controls the length
of the OFF time; this is known as the
INTERVAL.
DURATION equates to ON time;
INTERVAL equates to OFF time.
Virtually all times in EDM work are expressed in microseconds (1 micro second = 1 millionth of a second).
The timing cycle is based on the time it takes to charge (or discharge) capacitor C7. Since C7 is a fixed value,
4-2
the length of time that the (Q3-8) Hexfets are switched ON (Duration) or OFF (Interval) is determined by the
value of resistors R34 and R35, i.e., high resistance, long charge time; low resistance, short charge time.
A variable resistor (pot.) may be used, as the previous schematic illustrates, but I
prefer the ON time resistor (R35) to be a series of fixed values. These fixed
resistors (R35, A-F) may be wired on a multi-position gang switch for ease of ON
time selection (see schematic at left). By using fixed resistors, a known ON time
is assured with each click of the switch. There may be as many fixed steps as you
wish to wire; six are illustrated here. The output connect to TB 13&14 is? in place
of the pot.
In the previous schematic, R35A-25 indicates the resistor value (not shown) that will
deliver a 25us ON time. The value of this resistor may be calculated but it will vary
depending on the tolerance of C7 and the resistor(s). Shown at right is the actual value of
the resistors required in my particular power supply to deliver the specific ON times
from 25us to 400us; yours should be close to these values.
I found the particular values by using a pot. and looking at the waveform on an oscilloscope while
adjusting the pot. Once the targeted ON time was reached, the actual resistance value of the pot. was
measured. Depending on the level of accuracy required, you could use a smaller resistor in series with a
trimmer pot. to get the value or use the closest standard value resistor. In this particular pulse machine, getting
an exact ON time is not critical so using the nearest standard resistor or several fixed-value resistors wired in a
series and or in parallel will be fine. The main thing is at least “ball park” the ON time. Using a pot.? for R35
makes it extremely difficult to have any idea which ON time is being generated. Knowing the ON time is
important when setting low-wear conditions, as described on page 15-1.
Knowing the length of the OFF time is not critical, and the use of a potentiometer is permissible (though the
OFF times could also be selected in fixed steps as well). If affixing OFF times in steps, do not go much
beyond about 50us, which would equate to a maximum resistor value of about 5K for R34. The maximum
interval of 50us should be more than enough OFF time for most any situation you might encounter.
The output signal from U2, via pin 3, switches on and off transistors Q1&2. These two transistors are
configured such that they are able to properly switch the Hexfets ON and OFF.
Control and Motor Driver Circuit
The control circuit is similar in design to what I used in The EDM How-To Book for the RC machine but with
needed modification to make it work for a pulse supply. These modifications took me a L-O-N-G time to
figure out, yet in the end it was a very simple solution: I have learned more often than not that
Simple is Hard.
The control circuit makes the decision
where to position the electrode. Most
EDM control circuits that I am
familiar with make decisions based
on voltage across the gap. Some
advanced circuits make decisions
based on the voltage and current as
well as other useful input. This simple
control system only uses information
gathered from the voltage.
4-3
The DC motor has two active conditions, clockwise (CW) or counter clockwise (CCW) rotation, depending on
the polarity of the electrical current applied. The third condition is passive, and the motor remains stationary.
The motor (not shown, but it connects via the J1 panel connector) receives its power from the U4 motor driver
chip SN754410. This chip receives its instructions from the U3 voltage comparator (LM339) chip.
The comparator chip is the “thinker and decision maker.” It “compares” the gap voltage to a reference voltage.
The gap voltage is reduced by R7 & R8; this voltage is stored in the C9 capacitor. Diode D6, capacitor C9,
and the R22 resistor form something of a peak detector circuit. The signal from this peak detector circuit is fed
to pins (9&10) on the comparator chip via R24. This is the signal that is compared to two other voltages. One
of the voltages that the gap signal is compared to is generated at the junction of resistors R10 & R11 and is
supplied to pin 8. The other voltage is generated by the panel-mounted (and operator-adjusted) Gap Voltage
Pot. Two of the (four) op-amps in the chip are configured into a “window comparator” circuit, which is
described below.
How Does a Window Comparator Work?
The simplest non-technical explanation I can give is this: Two separate reference
voltages are fed to the comparators via pins 8 and 11, as described previously. A
third voltage, the gap voltage, is input to the comparators via the tied -together pins 9
& 10. If the gap voltage is above, or below, the reference-generated voltages, the
comparator outputs a signal telling the DC servo motor to reposition the electrode to
make the needed corrections. If the gap voltage is “in the window,” i.e., between
these two voltages, then the comparator does nothing and the DC servo motor
remains stationary.
For example, if Pin 8 is set at 8 volts (reference Low or VLo) and pin 11 is adjusted for 10 volts by the Gap
Voltage pot.? (reference: High or VHi) and the signal from the gap to pins 9 & 10 is 9 volts, then nothing
happens; it is “in the window.” If the gap signal (at pins 9 & 10) rises above 10 volts, then the comparator
outputs a signal to the motor to advance the electrode (down). It will continue to advance the electrode until
the voltage across the gap is 10 volts or less. Should the voltage across the gap drop below the 9 volts (such as
in the case of a gap short) the second comparator outputs a signal to the motor driver to retract the electrode. It
will continue to retract the electrode until the voltage across the gap is 9 volts or larger. That is how a window
comparator works; I could go into a more technical explanation but I think this should be adequate.
Conditions at Startup
Upon startup, when the electrode is above the workpiece, an open circuit exists, and the gap voltage signal on
pins 9-10 is higher than VHi or VLo given this condition; the comparator outputs a signal to drive the
electrode down. The motor continues to drive the electrode down until an arc is struck between the electrode
and the workpiece. Soon after the arc is struck, the voltage on pins 9-10 drops below the VHi and falls “in the
window.” After that, the control circuit acts as described above to keep the arc voltage “in the window.”
Short-Circuit Condition
When a short occurs, i.e., voltage across the gap falls below VLo (usually to or very close to zero) the
comparator outputs a signal to retract the electrode The motor continues to retract the electrode until the short
is broken and arc is established between the electrode and the work piece. More often than not, when the arc is
started again it is above VHi, causing the electrode to be position down until the arc is in the window.
4-4
What to Do When the Current Stops Flowing
The success of the control circuit is primarily in the “hands” of one simple component capacitor (C9). It took
me years to discover this!!! Without this component strategically placed, and sized, the circuit will not work
correctly. Without C9 and a couple of simple passive components, this is what happens (assuming a normal
stable burn is in process): When the Hexfets switch OFF, the voltage at pin 9&10 of the window comparator
goes above the Gap Voltage pot. reference voltage (VHi) setting. This OFF time voltage looks to the control
circuit like an open-circuit condition (such as at startup). Remember when the voltage on pins 9&10 goes
above the reference voltage; the control circuit signals to drive the electrode down!!!
So during every OFF time the motor moves the electrode down!! This happens even though it was not needed
because during the ON time things were operating perfectly fine. When the timer U2 turns the Hexfets back
ON, a short-circuit condition may exist because the electrode was unnecessarily moved during the OFF
portion of the cycle. This correctly results in the motor being signaled to retract the electrode during the ON
time. When the next OFF time starts, the control circuit thinks there is an open-circuit condition and it drives
the electrode down. Quickly, an oscillation sets up between shorting and an open circuit. As a result of this
instability, little to no cutting takes place and a lot of heat is generated in the large power resistors. The
question arises: How to solve this problem?
With my limited electronic knowledge, the OFF-time problem plagued me for years; though I tried many
ideas I was not able to successfully solve the problem. I wanted a simple very rugged design that did not
require the expense and complexity of a computer, custom software, etc. My solution was to keep it simple so
I attempted to solve the problem with chips such as op-amps, timers, multiplexer and logic chips. What I did
not realize was that my solution was too high tech!! Again SIMPLE is HARD!
The Peak Detector
What I was really looking for was a peak detector
circuit of sorts, which is composed of diode D6,
capacitor C9 and resistor R22 (see schematic on
page 4-1). To have a better understanding of the
following description, read Chapter 14.
The “magic” of the control system is based around
capacitor C9. To understand how the control
system works (see the o-scope wave form at left).
To correctly understand this description, it is
important to understand that the square-shaped waves are the actual voltages across the gap; this is not the
square signal being generated by timer U2. When the square wave forms are high, that means the voltage
between the electrode and workpiece is high (open circuit); no current is flowing in the gap.
When the voltage drops to the “valleys,” then and only then is current flowing in the gap. The curvy line that
runs through the middle is the voltage seen on capacitor C9. This voltage is a reflection of the gap condition
and is applied to pins 9&10 of U3, the window comparator.
Starting on the left side: As the square wave has dropped into the first valley, this is an indication that current
is flowing in the gap, which is indicated by the little “squiggly” line on the bottom of the valley. During this
time, the charge on C9 decreases, as indicated by the sloped line running through the middle of the square
4-5
waves. There is a short OFF time indicated by a rather narrow vertical square wave; note that the voltage on
C9 goes up during this OFF time period. During the next three cycles there is good current flow with short
OFF times, which is what is wanted. During these cycles, the voltage charge on C9 is a similar curve to what
we saw previously, which is good.
The next time sequence is quite extended, as can be seen with the large square wave. For whatever reason,
when the Hexfets were turned ON the gap was not in a condition to ionize (become conductive), the voltage
across the gap remained high, and no current flowed (no metal was removed); this is not good. During this
time, the voltage on C9 rose to near an open-circuit voltage. Fortunately, after missing the pulse the gap
became conductive for the next few cycles. Note that in the last two time sequences the OFF time was
extended for some reason (gap-related). For whatever reason, it took an increased time for the gap to ionize,
which reduced the time current was able to flow in the gap.
In this example we do not see a short-circuit condition. Should a short occur, the voltage on C9 would drop to
near zero, and thus be “below the window,” causing the comparator chip to send a retract signal to the motor
and pulling the electrode up to break the short.
In a nutshell, if the gap is not conductive the voltage on C9 goes high, and if the gap is conductive the voltage
on C9 drops; if there is a short, the voltage on C9 drops to zero or near zero.
The disadvantage of a simple circuit such as this is that, as the ON and OFF times start to approach each other
in length, the control system starts to lose control and may become sluggish under certain conditions. With a
simple control circuit it is nearly impossible to have a longer OFF time than ON time. Fortunately, for
virtually all situations we commonly encounter, the ON-time/OFF-time ratio is such that this simple circuit
performs adequately.
For years (the late 1950’s thru much of the 1970’s), the loss of control during short ON times and long OFF
times was a problem that was constantly being addressed by EDM designers. For the mid-1960’s and early
1970’s the most expensive commercial EDM machines used control circuits only slightly more complex than
this one. In the 1950’s, for sure, and during a large part of the 1960’s, the controls systems were less complex
than what you have built.
Manual Up/Down and Dither Circuit
The purpose of the dither circuit is to retract the
electrode for a predetermined length of time to
improve the flushing condition especially when
using jet flushing (see page 16-2). With an
understanding of how a window comparator
circuit works, the understanding of the dither
circuit should be simple. The dither control
provides a means of applying a signal to pins 9
& 10, that is, above or below the reference
voltages.
Let’s consider the manual up/down (SW7)
switch first. It is an Single Pole Double Throw
SPDT switch with one side via TB7 tied to 12+,
and the other side via TB6 is tied to the ground. When SW7 is toggled to the servo Down position, it applies
+12V to pins 9&10 on the comparator; this forces it above VHi, which causes the comparator to output a down
4-6
signal. The motor will run down as long as the switch is toggled to this position. When SW7 is toggled to the
servo Up position, it ties pins 9&10 to the ground, causing the comparator to think a short exists. The servo
retracts as long as you hold the switch.
To perform this operation automatically during a burn, it is necessary to introduce two chips. One is the second
555 timer chip U5, and the second is a solid state relay U6.
SW8 is the on/off switch for the dither circuit; it supplies 12VDC via TB 9&10 to the U5 chip. The output of
U5 closes the solid-state relay U6, which connect pins 9&10 to the ground, causing the comparator to think a
short exists. The electrode is retracted the length for which the duration (R37) is set.
Once the solid-state relay releases, the control circuit takes over and advances the electrode until an arc is
reestablished; things remain fine until the dither circuit upsets the apple cart again. This timer circuit is very
similar to the circuit used to switch ON and OFF the Hexfets, except it is much slower. The duration of the
electrode retractions is controlled by R37. The interval between the retractions is controlled by R36.
The dither circuit will most likely find its greatest use when doing low-power and or finish burns, with
graphite electrodes in particular. By monetarily retracing the electrode it permits swarf to be flushed out of the
gap between the electrode and the workpiece.
Now that you understand the theory of the power supply operation it is no longer a “black box.” It is this
knowledge that will make you a much better operator and you will even be able to trouble shoot the power
supply in case of failure.
4-7
Chapter 5
Building the Power Supply
Soldering Tips
Before getting started, let me share a few words on making good solder connections. For some this
may seem like a useless tidbit but, believe me, it is NOT. How would I know?? I have been there.
Failure to make good solder connections is one of the frequently cited reasons for non-working
boards and circuits. Trying to track down a poor solder connection will drive you crazy!! I took my
time soldering and thought I had good solder connections, but when it came time to test the
protoboard the master clock (timer) on the board would not operate!!! To make a long story short, it
had a bad solder connection.
Looking at the board under a magnifying glass, it was not discernible that there was a cold solder
joint. All things checked out correctly (most of the time) and appeared to be in order, but there was
no output from the chip. Hint: The statement it “works most of the time” is frequently an indication
that there is a solder problem. Suspecting this was the problem, all joints related to this chip were
reheated, and the problem went away and has not reoccurred.
I am not going to go into an in-depth discussion of how to properly solder, as there are many good
sites on the web as well as videos on YouTube (How did we ever live without YouTube?) that do the
subject more justice than I could ever hope. Take a few minutes and view a few of these sites if you
feel a little rusty. I will cover the basics and, with a little practice, you should be able to make good
solid electrical solder connections.
You can spend lots of money on a fancy soldering (station) iron but a
simple pencil -type 25-35-watt iron (see photo at right) with a conical
shaped 1/16” (.062 or 1.6mm) tip works well and is all that is really needed.
It is very helpful if you have a circuit board holder (but it is not absolutely
necessary). I use one like that is shown in the photo at left. It uses two alligatortype clips to hold the board and permit easy swiveling from front to back; it also
has a nice magnifying glass to check the solder connections. It was less than $5
from the local Harbor Freight store
(http://www.harborfreight.com).
If you don’t use a board holder, it is helpful to have a folded shop towel between the PCB and the
table; this helps stabilize the board during the soldering process.
Make sure your soldering iron tip is nice and shiny; if it is not, wipe it against a moist sponge or cloth
to clean, then apply a rosin core solder to the tip and wipe it on the moist surface again. This will
spread a thin coat of solder over the soldering tip; this process is called “tinning.” If the tip will not
tin, you may have to resort to using a wire brush to remove the oxidization buildup on the tip. In
extreme cases, you may find it necessary to file the tip down to copper and proceed to tin the tip. The
tip must show a shiny silver surface.
5-1
To ensure good solder joints the soldering iron tip, the PCB pads, and the component leads need to be
free of oxide. Oxide on the board and leads may be removed using a Scotchbrite-type pad.
For electrical wiring use only ROSIN CORE solder. It comes in leaded and lead-free versions. The
lead-free one is usually about 97% tin and 3% silver. The leaded solder is about 60% tin and 40%
lead. I prefer to use the thin .02 wire solder but the thicker gauge solder works well also.
Insert the components through the board and bend the leads slightly outward to hold them in place.
Don’t put more than a few components in the board at a time. When soldering touch the solder iron
tip to the pad and component terminal follow by the solder wire. The PCB pad and the lead must be
hot enough to melt the solder in order to get a good solder connection. Inspect each joint, preferably
under a magnifying glass, to ensure a good solder connection. Once you are satisfied that the joint is
good, clip the excess lead with a small pair of wire cutters and proceed to the next connection.
The green solder mask on the PCB greatly reduces the chances of solder bridges (from one soldered
connection to another) accidentally forming, but they are always a possibility. Double-check that the
solder connections are good and that there are no solder bridges. Use “solder wick” (also known as
“desoldering braid”) to remove excess solder and/or bridges.
Construction Notes
There is no “magic sequence” when building the power supply, though I think it is wise to first select
the cabinet and lay it out so all the hardware fits with sufficient air flow. Once the physical layout is
finalized, the next sequence for me would be to wire the high-voltage circuit and then wire the
control circuit board. If the determination is to not hand-wire the control board, but instead to use the
available PCB see page III in the front of the book, then all that’s required is to “stuff” the
components on the board and wire it to the remaining power supply components.
When wiring the power supply and the PCB, note the polarity-sensitive components, capacitors,
diodes, etc.; they must connect to the correct polarity, or otherwise they WILL fail and usually very
quickly. The polarity-sensitive components’ positive lead is always the longer lead.
All resistors are 1/4 watt, unless otherwise noted. The edge-finder buzzer P1 must be the specified
buzzer, or one of the same physical size and lead spacing, in order to fit on the board.
Note: At the beginning of each task, be it electronic or mechanical, a [ ] box will be found. Once the
operation is performed place an [X] in the box. This will assist in making sure that all required tasks
are completed.
When purchasing electronic components, I have not been able to find one source for all the needed
parts at what I consider a reasonable price. I have sourced parts from several suppliers.
Mouser Electronics (http://www.mouser.com, phone number: 1-800-346-6873) has all the needed
parts but the price difference in the three power transformers alone makes it worthwhile to purchase
from a second supplier. Of course, if you prefer the highest quality (highest priced) transformers, then
Mouser is basically one-stop shopping as they have virtually all the remaining parts.
5-2
The transformers and panel switches were sourced from Marlin P. Jones & Assoc. Inc. Their prices
on these two items are significantly lower than those of anybody else: WWW.mpja.com, phone
number: 1-800-652-6733.
Many of the small components may be sourced from Jameco as they have very good prices:
Jameco Electronics (http://www.jameco.com), phone number: 1-800-831-4242
Power-Supply Housing
The power resistors and transformers generate lots of heat; therefore, the power supply should only
be housed in a metal enclosure (plastic enclosures are not acceptable). It will be a snug fit, but you
can squeeze all the electronics into a box that is 12” x 12” x 5”. A larger and deeper box is preferred,
making the building much easier; the enclosure that I am using is 17 x 14 1/2 x 5 1/4”. Looking
through catalogs, it becomes obvious that a purchased metal enclosure of the minimal suggested size
can easily be over $100. I think that is a bit steep for a metal box, but that is for you to decide.
There are a couple of options: one is to buck up, pay the money and buy a nice commercial cabinet.
Another possible option is to purchase a NEMA (Nation Electrical Manufacturers Association) steel
enclosure (designed to house electrical switches, fuses, etc.). The NEMA 1, with no knockouts, is
usually the least expensive: a 12” x 12” x 6” box painted gray with a hinged cover and a coin-slot
latch in 16-gauge steel; one sells locally for about $62 (2011). NEMA-type boxes are very sturdy and
should serve well as an enclosure, though I have not actually used one for this purpose. Other choices
are spend a little time getting to know a sheet-metal brake and shear a little better and fold up a
cabinet; the final option is to use an old computer case. The latter is what I opted for.
I suspect most have or can locate an old computer for free or for a few dollars. Check yard sales, visit
secondhand stores such as the Salvation Army, Deseret Industries, and Goodwill, etc. Ask friends if
they have an old computer that they would be interested in getting rid of; that is how I located mine.
The older desktop computers, such as the one shown in photo at
right, I think are preferable to the tower-case computers so
common today. There are some desktop computers still being
built today, but most that I am familiar with appear to be a little
on the thin (height) side for this project. The height required is
determined by the height of the transformers and fan used.
If a tower case is used it has some advantage. By placing the transformers on the bottom and mount
the power resistors on the side of the case, where the computer motherboard was mounted, and
locating the fan at the top it provides a natural flow for cooling. Another option (especially if you
have the tower case with sides that may be removed independent of each other) is to lay a tower case
it on its side, similar to a desktop computer.
5-3
One of the problems of using a tower case,
especially the newer ones, is that many are quite
flimsy and it’s questionable as to how well they will
support the weight of the transformers. The
transformers alone may weigh as much as 15
pounds or more. If you use a wimpy case, it may be
necessary to install stiffeners inside to carry the
weight of the transformers (see the section below).
In my opinion, many of the older computers (286/386/486) have cabinets that are quite robust. The
old computer (386) that I used has a case built like a tank. If you use a computer, gut the case. The
photo above shows the case after everything was removed. It may be necessary to do some cutting of
the case to get everything to fit in an orderly fashion.
Case Stiffeners
As mentioned earlier, many of the computer cases are not strong enough to adequately
support the transformer’s weight. In this situation, lay support strips to
spread the weight over a larger area of the case. I have used shelfmounting rails for this purpose in some prototype supplies. The rails are
available from many home improvement centers. Their intended purpose
is to mount shelving to walls, as is pictured at right. A different version of
the same type of product is shown in the photo at left. They are a rigid
“C” shaped channel with slotted holes that readily lend themselves to being bolted to the
computer case, as well as having the transformers bolted to the rails. Another possibility for stiffeners
is the use of DIN rails (also known as top hat rails), which are used to mount relays and circuit
breakers, as well as other electrical control equipment. The DIN rails work fine, though my
preference is the shelf rails. Using these rails permits the weight of the transformers to spread out
over a distance that is more acceptable for the thin metal computer cases. I would use 10-32 screws
with flat washers for mounting the heavy transformers to the rails and the case.
Front Panel & Power Supply Layout
In most instances, it’s
necessary to remove the
computer’s front face
and replace it with a flat
piece of sheet metal (see
photo at right). A
removable front makes it
very easy to layout and
drill the required holes
for mounting the
controls. The layout of
the front panel will be based on the enclosure chosen; thus it is almost impossible to give a layout
scheme. I suggest grouping similar controls together in a manner that is functional and yet pleasing to
5-4
the eye. Locate the power switches near the power resistors; the switches and controls that interface
with the PCB are located near the PCB. This keeps most of the wiring clean and to a minimum.
Miniature toggle switches will work in this application but ensure that they are rated for at least 6
amps. Jameco Electronics has very good prices for “house brand” miniature switches. I used, and
recommend, full-size 15A toggle switches; Marlin P. Jones has very good prices on them.
A volt meter and an amp meter are pictured on the power supply, yet the schematic only shows a volt
meter. Neither the volt nor the amp meter is necessary for operations, though they are helpful. The
amp meter was purchased for use during the testing phase of the power supply and I decided to install
it in the finished version, mainly because two meters looks more impressive in photos. The amp
meter is 0-30 amps, and is available from Marlin P. Jones for a very reasonable price. They sell a
matched set of amp and volt meters; what is shown in the photos is not a matched set. I recently
noticed that Marlin P. Jones has a 0-15amp meter, which would be much better in this application.
To take full advantage of the design, the following controls are needed:
(1.) AC ON/OFF switch; this may be mounted in the back or on the side of the case,
(2.) Polarity switch,
(3.) Power/Edge finder switch,
(4.) Five gap power switches,
(5.) Servo ON/OFF switch
(6.) Servo Manuel up/down switch
(7.) Servo indicator light
(8.) Servo ON/OFF switch
(9.) Servo plug
(10.) Gap power plug
(11.) Dither ON/OFF switch
(12.) Dither interval pot.
(13.) Dither duration pot.
(14.) Gap ON time control switch/pot.
(15.) Gap OFF time pot.
(16.) Volt meter.
With the advent of readily available, low-cost label makers such as DYMO, a respectable-looking,
labeled case is possible. Of course, there are computer graphic programs available to produce a very
professional-looking front panel, or it could be engraved with a computer controlled (CNC) mill.
The photo at left is a top view of my
power supply. Give careful thought to
the interior layout of the power supply
to ensure proper heat removal.
Consider drawing air through the case,
i.e., negative pressure in the enclosure
as opposed to “blowing” air into the
case for cooling. Position intake vents
(holes) into the case so that the
incoming air is forced to flow over and
around the heat-generating components
before being drawn out of the box via
the fan. The power resistors and
transformers generate the vast majority
of the heat but make sure that the PCB
gets good air circulation.
5-5
Power-Supply Layout
Ventilation is one area where you do not want to cut corners: purchase a good-quality, high-volume
AC fan. I would recommend at least an 85CFM fan. The air inlet area should be equivalent, in square
inches, to the swept area of the exhaust fan. Holes may be drilled or cut at various locations in the
chassis to provide the needed ventilation.
Ensure that the air cannot easily take shortcuts from intake vents to the exhaust fan. This may mean
that a partition/s is required to control air flow (in the photo on page 5-5 to the left side of the
transformers, note the partition which forces air across the Hexfets). If ventilation holes are in the
bottom of the case, the feet on the case need to be high enough to permit good inflow to the cabinet.
If there is overheating, perhaps more vent holes are needed or higher volume fan/s are required.
Major Components and Mounting
[ ] Drill the necessary holes for mounting the main AC switch. I was lucky
enough and used the existing AC computer switch (see photo at left).
[ ] Make the necessary provision for bringing the AC line into the case
either by passing the AC wiring through a wire grommet or using an AC
receptacle, such as is found on most computers. Fortunately, I was able to
use the existing AC computer plug; this computer case was made to convert to an EDM power
supply.
[ ] This is VERY important: Tie the green ground wired in the AC cord to the chassis, and remove
any paint where the ground wire attaches to get a good electrical contact with the case.
[ ] Mount the F1 fuse holder
[ ] Determine the cooling fan location, and cut an opening in the case large enough for full air flow. I
used the slots where the computer’s plugin boards mounted; this made for simple mounting. Drill
mounting holes and mount the fan.
Capacitor Note: This is a case where more (microfarads) is better. These
large filter capacitors can get expensive. The filter capacitor I am using is
the large one shown in the photo at right. It was salvaged from a piece of
surplus equipment which is maybe 15 years old. I have not been able to
locate this exact capacitor for sale from electronic suppliers; I suspect it is
discontinued. This particular capacitor has screw terminals and is rated for
2300uf @ 200VDC and 85C. Such a capacitor sells for about $28 (2011).
The smaller capacitor is rated for 2200uf @ 200VDC and 85C, and the load life is 2000 hours (I was
unable to find a load life for the larger cap). The load life is the estimated number of hours the
capacitor will operate at a maximum temperature of 85C (185F) before a failure might be expected.
Capacitors such as these sell for about $8. I am not sure what the difference is other than 100uf and
5-6
physical size. I have not used the small caps but, if I were buy caps, I would be tempted to purchase
two (or three) of the smaller caps and connect them in parallel for 4400uf (or more) and still be under
the price of the single big capacitor. The smaller cap is designed for PCB and are “snap in”; all that
means is connections to the cap. are made via soldering instead of using nice screw terminals.
No matter what cap is used, you should purchase a mounting ring to hold it securely in the cabinet.
The Mouser Electronics part number for a cap similar to the large one in the photo is
75-36DX202F200BC2A $22.01 (2011).
The smaller cap. is 661-ESMH201VNN222MA $9.16 (2011)
[ ] Drill the holes to install the filter capacitor(s) mounting ring(s) near the D20 rectifier.
[ ] Secure the C20 capacitor in the mounting ring
Transformer Note: By far, the single most expensive (and heavy) items in the power supply are the
transformer. In this design, I am using three (3) 10-amp 24VAC transformers. When rectified, the
output voltage is about 110VDC. I have found the higher voltage works better than the 75VDC from
using two transformers, though using two transformers should work.
The best price by far for transformers is from
Marlin P. Jones & Assoc. Inc. Marlin’s part
number is 7846-TR. I paid $18.95 (plus
shipping) each in 2011. These transformers are
made in China, and the quality of wiring and
wrapping is not the best, but so far I have not
had any problems with any of their
transformers. Similarly rated transformers are
more than twice that amount.
A high-quality transformer with similar ratings
(that I have used successfully) is the Stancor
P-8619, which typically sells for at least $75
each (2011) through Mouser and other major electronics suppliers. Jameco Electronics sells a
24VAC, 12-amp “house brand” transformer for $65.95 (2011); their part number is 221411
[ ] Drill the necessary holes and mount the power transformers using 10-32 screws and washers. It
may be necessary to mount the transformers on stiffener rails as previously suggested. The photo
above shows the transformer and large filter capacitor section of the power supply.
Wiring the High Voltage Section
Note: When wiring the high-voltage power section (see the schematic on page 5-8 & 5-11), use at
least number 16 gauge wiring. Where possible, make all high-voltage/current connections with
crimped connectors. The crimpers, connectors and wire are available from most automotive parts
stores. Make sure that connections are tight crimps and that the wires are clear of the power resistors
to avoid melting the plastic wire coverings.
5-7
[ ] Wire the AC switch, fuse holder, fan and transformer’s primaries as per the schematic below left;
do not wire the transformer’s secondaries at this time. The schematic may appear a little confusing
but basically the primary wires to all the transformers are connected in parallel. Wire nuts are useful
when tying this many wires together.
[ ] At a location convenient to the control board, mount the
low- voltage transformer (T4).
[ ] In close proximity to the power transformers and filter
capacitor, drill a hole to mount the D20 power rectifier. Sand
any paint from the case in the mounting area.
[ ] Apply heat-sink grease to the back of the D20 rectifier and secure it in place with a screw and a
nut.
[ ] Repeat the same process for mounting D21. This bridge should be located near the power output
plug, through which power goes to the gap.
Safety Note: When wiring the transformers, make sure that secondary wires not being used have
wire nuts (or electrical tape) covering the bare ends. You DO NOT want short circuits to occur
between transformer leads or with the chassis. Pay particular attention that the transformer’s
secondary outputs are connected to the AC terminals of rectifier D20. Ensure that Cap. C20, which
is polarity sensitive, is correctly connected to the + & - marked terminals of rectifier D20.
[ ] Wire the secondaries of transformers T1-T3 together; they are connected in series. When
connecting the secondaries, make sure that the transformers are not “bucking” each other. To check
for bucking, wire only the secondaries of two transformers together (in series), then check across the
two remaining leads (with a volt meter); there should be about a 50VAC reading. If it is only about
25VAC the secondary side need to be swapped on one of the transformers. When the third
transformer is correctly connected, a voltage reading of about 76VAC should be observed.
[ ] From the Negative (-) terminal of C20, attach a wire to the enclosure chassis (i.e., ground), and
remove any paint where the wire attaches to the case. It is a very good idea to tie it to the same
location that the green wire from the AC line connects.
[ ] Once the high-voltage supply is wired as per the schematic, connect to the AC supply and measure
the voltage across C20’s terminals; the readings should be about 110VDC.
Power-Resistor Sizing and Mounting
I am sure I’ll take some “heat” from those in electrical quarters for the way I size the power resistors.
After reading through this section, make a decision on how many and what wattage ratings of
resistors to use. The resistors used are all one size: 25 ohm with a 50 watt rating each. These are
rather small and reasonable in cost, at less than $4 each (2011).The resistors are sized for “normal
working currents,” NOT for worst-case shorted conditions. This keeps the number, size and cost of
5-8
the resistors down. I haven’t had any failures with the resistors thus far. If this machine were to be
operated commercially day in and day out, then I would consider increasing the power resistor
wattage rating. I expect these machines to be used mostly for hobby or small shop work where the
demands are not 8-12+ hours per day.
One of the objects in EDM work is to never have the electrode and the workpiece ever come in
physical contact (i.e., short-circuit). We are not in a perfect EDM world and, on occasion, shorts do
occur. The servo system detects and quickly retracts the electrode. Short-circuit conditions seldom
exist for more than one second at maximum. Yes, there are exceptions to this but they are few and far
between. Even when shorts occur, the resistor(s) is seeing pulsed current. When the OFF time is
shortest, the duty cycle is about 90%; usually the duty cycle will be less. What all this means is that
the resistors are not seeing continuous DC current.
Below are typical numbers from real burns.
Electrode is a ½ x 4”
Workpiece Steel
ON time 200us OFF time 20us
Resistance
25
12.5
6.25
5
3.5
3.1
Voltage drop
30
32.5
29
29
30
30
Current
1.2
2.6
4.6
5.8
8.5
9.6
Wattage
36
87
135
168
257
290
From this data, the maximum wattage being dissipated across the power resistor bank in a normal
hard burn is 290 watts spread over 8 resistors which are rated for 50 watts each. The eight combined
have the capability of safely dissipating 400 watts. Yes, when a short occurs the brief load on the
resistors may greatly exceed 400 watts, but it is for a very brief period. The resistors are forcibly
cooled and are not seeing a pure DC load because they are pulsed. Under a sustained burn, the
average resistor surface temperature is about 275-300 degrees F. The photo above shows the
enclosure around the resistor. This ensures maximum air flow over the resistors. The cooling tunnel is
sized such that it fits around the outside of the cooling fan
Long story short, I have yet to experience a resistor or transformer failure with this system. Power
resistors are tough little critters and can take quite a bit of abuse. If you are concerned, size the power
resistor for worst-case shorted conditions. The formula for calculating the wattage is P = I x E
(P [watts] I [current] E [volts]). The resistors pictured above and on page 5-10 are the Vishay HL
series silicone resistors. Jameco & Marlin do not sell high- power resistors; Mouser is a source.
Mouser Electronics part numbers for both are given below.
Vishay HL series silicone resistors
71-HL50-06Z25 (this is the 50-watt version; $3.54 each)
71-HL100-06Z25 (this is the 100-watt version; $6.50 each)
5-9
Ohmite
588-L50J-25 (this is the 50-watt version; $7.79 each)
588-L100J-25 (this is the 100-watt version; $9.83 each)
The commercially available power resistor mounting hardware is usually for
mounting single resistors horizontally. With so many resistors being used, that
takes up lots of space. For a few pennies, you can make custom vertical mounting
hardware quick and easy.
[ ] Manufacture 4 as per the drawing at
left. Stock should be at least .062 thick.
[ ] Cut lengths of 8-32” thread rod (the
length of the power resistors plus 1/2”).
[ ] Determine the location for the two
banks of four power resistors in the case
near the exhaust fan.
[ ] Mark and drill the resistor mounting holes for #6-32
screws.
[ ] Mount the resistors between the mounting hardware as
per the photo at right. This is a photo partway through the
wiring process; two resistors are missing
[ ] Mount the power resistors.
[ ] Design and fabricate a cooling tunnel
similar to what is shown in the photo on
page 5-9 and the photo at left. The tunnel
may be made from sheet steel or
aluminum (roof flashing also works
great).
The photo at left shows another view of
the power supply. The cooling tunnel is
to the left and the PCB is located in the
center of the picture; the current
controlling 15-amp power switches are
to the right.
5-10
Finish the High-Voltage Wiring
The high-voltage schematic below shows the gap, which is external to the actual power supply. It is
shown here for clarification as to how the circuits are wired.
[ ] Finish wiring the HV power supply as per the schematic; use the previous spec. wire size.
The R38 bleeder resistor is a 22K, 2-watt Mouser (594-5083NW22K00J).
CB1 is a 10-amp Circuit Breaker Mouser (655-W57-XB7A4A10-10).
A panel-mounted AC power entry module is used for PP1& 2; see the photo at left.
There is a good chance of recycling one from the computer’s power supply. Power
is supplied to the gap via this module. These modules are inexpensive, rugged, and
can handle the power
demands.
A standard 115VAC computer power cord is
used for the power transfer. The female end of
the plug inserts into the AC power entry
module. The male end is removed and either
alligator clips or appropriate ends to match
your servo ram are attached to the cord.
The photos at right show views of the back
side of the front panel. The photo above
shows a close-up of the power switches and
associated wiring. Keep the wiring neat and
use wire ties to group them together.
The bottom photo shows a full view of the
front panel’s back side, with transformers
located on the left and the PCB positioned in
front of the cooling tunnel.
5-11
PCB Location & Heat Sink Selection
[ ] Determine where the control board is located in the
cabinet. In most cases, assuming that proper cooling of
the PCB-mounted Hexfets is assured; mount the board
in an area close to the front panel, with the long side of
the board parallel with the front panel. Make sure that
there is enough room between the front panel and the
Terminal Blocks TBs to make the necessary wiring
connections to the board. See the photo at left.
[ ] Mark and drill mounting holes for #4-40 screws. The
board should be mounted on four 1/4’”- 3/8” long standoffs. DO NOT permit the bottom of the board
to come in contact with any metal surface. Perhaps the standoffs used for the PC mother board could
be recycled.
The purchased PCB provides for mounting six Hexfets with heat sinks on the PCB. If you not using
the purchased PCB and are wiring a control board from scratch, it may be simpler to use a single
Hexfet such as a IRFP250N (TO-247 footprint) and mount it to an external heat sink. If you do, make
sure the heat sink surface area is at least 20 square inches (on one side). The Heat sink MUST be
insulated from the case. If not, the Hexfet(s) MUST be insulated. Use thermal grease and a thermally
conductive insulator such as Mouser #532-4180.
Wiring the PCB
A description of how to wire the purchased PCB is not given in this text. Individuals who purchase
the PCB (see page III front of book) will receive detailed instructions for wiring and testing the PCB.
The control and power circuit development is an ongoing process, so the PCB purchased may differ
from the schematic and photos of the board found in this publication.
For those not using the PCB, a complete schematic of the PCB is found on page 4-1. Follow
established wiring and solder practices when wiring the control and pulse circuit.
EDS Issues
Though I am not covering the assembly of the PCB, it’s necessary to mention
the Electro Static Discharge (ESD) issues in particular as they relate to the
Hexfets. All of the electronic chips should be handled with care, but the
Hexfets (or power mosfets) are particularly sensitive to ESD. For those not
familiar, know that a static charge that you cannot see or feel will destroy
these solid-state switches, so here are a few suggestions for handling them.
For starters, use a soldering iron that has a ground plug on the AC line. When
soldering the Hexfets, make sure that the PCB, the Hexfet, the soldering iron and YOU are all at the
5-12
same electrical potential. It is a simple matter of working on an electrically conductive surface, such
as a sheet of aluminum foil. Set the soldering iron on the foil and place everything on the foil that you
are using; this automatically keeps everything at the same electrical potential. Should you leave and
come back, make physical contact with the foil before touching the Hexfet.
The critical terminal on the Hexfet is the Gate terminal. With the Hexfet laying metal side down, the
Gate terminal is the first one on the left. Once the Hexfets are soldered in the circuit with the 10-ohm
resistor in the Gate leg, most of the issues with static go away, though it is a wise to touch the PCB
prior to positioning the header strip in the Terminal Blocks (TBs). If you follow these guidelines,
destroying the Hexfets with ESD should be a non-issue.
When mounting the heat sink to the Hexfet, apply thermal heat sink grease to the Hexfet and then
attach the heat sinks with a 6-32 screw.
After mounting each Hexfet, it is a VERY good idea to check and make sure that it is still good.
Why is a Hexfet so sensitive? In the simplest terms, there is a VERY thin insulator (dielectric)
between the Gate and other internal parts of the Hexfet (i.e., capacitor); if this insulator is ruptured,
the Hexfet is BAD. Testing is to see if the insulator is intact; hence, if there is a low resistance
reading between the Gate and the Drain or Source terminal, the Hexfet is bad.
There are more involved test procedures than what I suggest here; if you are interested, check online
for detailed instructions. I use the following procedure and it has not let me down yet.
When making this test, it is important to keep in contact with the
aluminum foil work area. Lay the old static-free bag that the Hexfets were
shipped in on the aluminum foil, then position the PCB on the bag. The
reason for this is to ensure that you are not getting a conductive path
through the aluminum foil, thereby giving an error to the reading.
Next lay the test meter on the aluminum foil and touch a probe to the
aluminum foil, and then touch the positive probe to the Drain (middle)
terminal, and then touch the negative probe to the Gate terminal. The
digital meter reads open circuit (OL), or a very high (meg. ohm)
resistance. Reverse the probes, and the same open circuit or VERY highresistance reading should be repeated; this indicates the Hexfet is good.
PCB to Panel Connections
At this time, the front panel should be completed with all components
mounted as per the instructions starting on page 5-4. It is a simple process
of making the necessary connections from the Terminal Blocks (TB1-26)
on the PCB to the front panel the see schematic left for instructions below.
Using differently colored wires makes wiring and future troubleshooting
much easier. Unfortunately it appears virtually impossible to buy multiconductor (25 conductor is perfect), multicolored, stranded wire
5-13
(preferred) in short lengths (only about 12” are needed). If you know an IT guy or someone who
wires intercom systems, you may be able to talk them out of a short length of multicolored wire. I
came up with a 1-foot piece from the electronics guy at work; lucky me!
Note: Mark the components on the front panel (the lead pencil works fine) for identification during
the wiring process. Be consistent when wiring the components, and have all the switches turning on
the circuits in the same direction, i.e., up is ON, and down is OFF. Pots turned clockwise increase;
when turned counterclockwise they decrease.
By using a length of .200 header (see photo below), all wires connecting to the PCB are soldered to
the header and the header plugs into the TB on the PCB. When completed, the PCB may be
disconnected and reconnected without the concern of the circuit being accidently getting mis-wired.
[ ] Cut one length each from the .200 header (5mm).
A length of 16 pins
A length of 6 pins
Two lengths of 2 pins
[ ] Insert the 16-pin header into TB 1-16 and tighten each pin.
[ ] Insert the 6-pin header into TB 17-22 and tighten each pin.
The connecting wires should be long enough to reach their respective components on the front panel or, in
some cases, other components in the case.
[ ] When wiring TB1 &2 to the J1 panel jack, make sure that the common ground terminal on the jack is NOT
used. R31 is a 100-ohm rheostat; this provides a very simple speed control for the DC servo motor. Diodes
D16 and SW2 (when closed) permit a fast retraction of the servo motor but when current is reversed,
depending on the setting of R31, the motor speed may be slowed. This is extremely useful in assisting in
damping out oscillations that might occur under some burn conditions. Switch SW2 is optional; by opening it
the advance and retract speeds are controlled by R31.
LED D15 is a bicolor LED; by observing its color you instantly know which way the servo is being driven.
This is most useful in a noisy environment. R30 limits the current flow in D15 to keep it from burning out.
[ ] TB3, 4&5 are connections to the 50K Gap Voltage Pot.
[ ] TB6, 7&8 is the manual Up and Down control for the servo. Note: SW7 is a SPDT switch but it MUST
HAVE three distinct positions of the handle: (ON) OFF (ON). This switch may be spring-loaded so that after
pushing the switch handle up (or down), it always returns to the middle position when released
[ ] TB9&10 turn on and off the dither circuit, which is controlled via pots. R36&37. There are two wires
which connect to TB9; one goes to SW8, and the other to R37.
[ ] TB11&12 connect to R34 (a 5K pot.), which controls the OFF time of the power flowing through the gap.
[ ] TB13&14 connects to the resistance represented by the R35 pot. in the schematic on page 5-13. This
resistor may be a series of fixed (stepped) resistances which are switched in (and out) with a rotary switch. The
value of the resistances control the length of ON time that current flows in the gap. I highly encourage reading
page 4-3 prior to wiring TB13&14.
5-14
[ ] TB9, 15 &16 are tied to two 100K pots., which are wired in series with a tap at the common junction.
These pots control the duration and interval of the dither circuit. The dither is used when burns are unstable
which frequently results in poor flushing.
The settings of these two pots. determine:
(1.) How frequently the electrode is retracted from the burn which is controlled by R36 (i.e., the interval).
(2.) How long the retraction last, controlled by R37 (i.e., duration) for future reference mark which pot. is
"duration” and which is “interval.” The previously wired switch SW8 turns on and off the dither circuit.
[ ] TB17& 18 connects to the 12VAC transformer T4. This supplies power to the control board circuit.
[ ] TB19 connects to (DPDT) SW15 switch. This line provides a ground (via the electrode and workpiece) in
the edge finder circuit. DO not get the connection to SW15 confused with the wiring from TB22 which is also
connected to SW15.
[ ] TB20 goes to the (-) negative side of the panel meter.
[ ] TB21 goes to the (+) positive side of the panel meter.
Should you use a Radio Shack volt meter on the panel and want to change the face of the meter to correspond
more accurately the meter reading, this web site is very helpful.
http://www.sci-experiments.com/RadioShackMeter/RadioShackMeter.html
[ ] TB22 is the high-voltage (approximately 110VDC) input to the control circuit. This wire MUST connect
between the power-resistor bank and SW15. If it connects between D20 and the resistor bank, the control
circuit will not work.
[ ] TB23&24: This circuit needs to be wired with at least 16-gauge wire. As much as 10 amps flow through
these two TBs, back to the – side of D20.
[ ] TB25-26: This circuit needs to be wired with at least 16-gauge wire. As much as 10 amps flow through
these two TBs from the + side of D21.
5-15
Chapter 6
The Servo
This servo head requires machined parts, though I have attempted to keep that to a minimum and yet
design a head for doing precision EDM work. Should you not have the skills or facilities for
performing the machine work, consider having the work done by a high school or technical school
shop class (assuming that you can find that kind of class existing in our service-based society).
Another possibility is to have the work farmed out to individuals via the online machine tool user
groups. I am sure there are individuals that would be glad to do the work for a reasonable fee.
In most “small” EDM machines the ram, or slide (the terms are generally used interchangeably in
EDM lingo), is usually an electro/mechanical device consist of an electric motor of some type which
rotates a lead screw. This rotary motion is converted into a linear movement to position the electrode.
You can have the best electronic control system, but if the mechanical system that physically
positions the electrode is sloppy or binds, it is all for naught.
In small benchtop-type EDM machines, the forces generated by the EDM
cutting process are nowhere near those found in conventional machine
tools; thus, the mechanical parts do not need to be nearly as massive.
What is needed is a ram and servo system that is highly responsive. The
ram must have minimal frictional drag, yet it must be stiff enough to
maintain the required rigidity for precision EDM work. The ram described in this text uses lowfriction ball slides, not unlike what is used in many commercial machines.
The foundation of the head is built around a piece of 1/8”x 2”x2”x2” angle
attached to a mounting arm (see photo, above left). The motor mounts on the
horizontal member of the angle; the mounting arm attaches to the vertical
member.
The photo at right shows (unfinished head) the motor
mounted on top of the angle, along with the up/down
limit switches. The Indexer chuck is mounted to a thin
profile ball slide.
The Indexer chuck is what really sets the mechanical
part of the EDM apart from any small commercial or
homebuilt EDM that I have ever seen.
As a result of this compact design, the total travel of the ram is 1”. For the
vast majority of the EDM work, this travel is more than enough. When the
short travel head is coupled with the back slide design (see page 12-7),
the 1” travel become a moot point; in fact, in my opinion, it actually works
to an advantage. The photo at left shows the finished head with the depthsetting DRO.
6-1
DC Gear Motor vs. Stepper
From my experience, this appears to be a very sensitive subject for some individuals, and each
quarter can justify their position so I suppose it comes down to personal preference. It is my opinion
that a stepper motor greatly increases the complexity of the design and I see no major advantage (in
most cases), unless the EDM is using a micro-processor. It is also my opinion that most “simple”
(i.e., non-CNC) EDM’s, even a pulse machine, have no need for a micro-processor. There are
thousands of commercial EDMs operating today, around the world, that use a simple analog control
system very similar to and, in many cases, not as sophisticated as what is found in this machine.
As a result of the design of this machine, as well as that of others I have designed, I must come down
on the side of a DC gearmotor. It is a very simple, rugged system, which is also low-cost. For those
who would like, the analog signal maybe fed into the required chips to generate the need signals to
operate a stepper motor.
Servo Motor Selection for the Ram
I use the same (Pittman) DC gearmotor (see photo at right); this is the design that I
used in The EDM How-To Book. The particular motor I used was designed to
operate on 19.1 VDC with a gear reduction of 60.5:1.
New Pittman motors sell for over $100. There is nothing special about Pittman
motors; it just happens to be the one I purchased. Check the online auction sites as
well as electronic, robotic and surplus outlets; you should be able to pick up a
suitable motor for $15-$25 (2011).There are lots of manufacturers of these small
gearmotors; The names you will see are: Barber-Colman, Von Weise, Merkle,
Korff, and Molon, as well as others.
In searching for a motor, make sure the DC motor is a multi-poled motor; the more poles the motor
has, the better. My experience, as well others’, has shown that motors used in battery-powered
screwdrivers and small radio-controlled servos don’t perform well in this application. This is true in
particular when burning with low power and small electrodes (which is a worse-case situation). I
suspect the main reason for the poor performance is because of the low number of poles found in
most of these motors. Some of these types of motors have only two or three poles.
Specification on a motor similar to mine are as follows:
MFG: PITTMAN
P/N: GM8712F191
COIL: 19.1VDC/12ohm
82RPM @ 19.1V, .08A NL,1.5A Stall
51RPM @ 12V, .07A NL, 1A Stall
.187” dia. X .625” steel shaft. Three 4-40 threaded mounting holes.
L: 2-3/4"(Body) DIA: 1-3/8" WT: .4
6-2
The Lead Screw
Before making the lead screw, read the following two sections on removing shaft endplay and
attaching the leadscrew to the gearmotor shaft. I used a 3/8-24” lead screw in the design, but other
thread pitches may be used (though I have never use a pitch coarser than 20TPI).
[ ] The diameter and length of the gearmotor output shaft needs to be known. Drill and ream a hole
(slightly deeper than the length of the output shaft) for a sliding fit on the gearmotor output shaft. The
gearmotor I am using has a 3/16” diameter shaft.
[ ] It is critical that the lead screw be fabricated accurately to get the
smoothest operation. The most accurate method is to us a lathe and
single point the threads on a piece of cold roll steel. If threading on
a lathe, cut the leadscrew a little long to provide material to hold while single-pointing the threads.
For the most accuracy, hold the stock in a collet. To avoid threading on a lathe, an acceptable
alternative is to purchase a length of all-thread rod and cut to length.
Removing Shaft Endplay
Before fabricating the lead screw, it is a good idea to determine how the output motor shaft end play
will be removed, and how the leadscrew is attached to the gearbox shaft.
The endplay problem is dealt with using two different spring options. One is to
use a small coil spring between the bottom of the gear box and the end of the
leadscrew (see the photo at right. The other method is us to a “wavy washer”
between the gear box and lead screw to take up the endplay (photo below). No
matter which is used I suggest placing a nylon washer (I’ve on occasions cut
the washer from a plastic milk container) to act as a bearing surface between
the gearbox and the tensioning device. A small roller
thrust bearing in place of the nylon washer is a better, though more
expensive, choice. The wavy and nylon washer, as well as small coiled
springs, are usually sold in most hardware stores’ specialty fastener section.
[ ] The nylon washer goes on first (see the photo at left, and a wavy washer
(or spring) is placed between the motor housing and lead screw.
Attaching the Leadscrew to a Gearmotor Shaft
Pin Method
Two methods of attaching are suggested; both have been used with long-term success. One of the
most common methods of securing two rotating shafts is with a setscrew. With our short coupled
application the problem is it tends to pull the leadscrew slightly off the centerline, causing binding as
the leadscrew nut approaches the motor. To get around this problem, drill a hole completely through
6-3
the lead screw and the gearmotor shaft and insert a pin through both. The following describes how to
correctly position the holes to ensure the output shaft endplay is removed.
[ ] Place the nylon washer, or roller thrust washer, on the gearbox output shaft. Next, place the wavy
washer or coiled spring over the output shaft then slide the lead screw in place.
[ ] Apply a force, partly squeezing (hold the entire assembly between the jaws of a vice or “C” clamp
to facilitate taking the measurement) the wavy washer, or the coils spring. In this position, take the
measurements to know where to drill the hole in the leadscrew and the gearbox output shaft.
[ ] Use standard machine shop procedures to find the exact center of the lead screw.
[ ] Insert the gearmotor shaft into the lead screw to the depth determined by
previous measurements (some may prefer to drill the hole through the lead
screw and the gearbox output shaft separately). Drill a small hole through the
leadscrew and gearmotor shaft; I drilled a 1/16” hole
[ ] Install the nylon washer and spring device. Slip the lead screw over the
gearmotor output shaft and align the holes and insert the required pin through the leadscrew and
gearbox shaft. I used a 1/16” aluminum welding rod and bend the end to keep it from coming out.
Glue Method
If built correctly, the torsional loads on the lead screw are not great, so a two-part epoxy glue, such as
JB Weld works well for attaching the leadscrew to the motor output shaft. The glue method also does
not require the measuring and setup required to drill the cross hole for the pin method. The
disadvantage is that, if it ever needs to be disassembled, it is a difficult process.
The following steps are recommended for gluing the lead screw and motor shaft:
[ ] Make sure all surfaces are cleaned with acetone or another similar degreaser.
[ ] Place the nylon and wavy washer, or coiled spring, on the shaft.
[ ] Put a two-part epoxy glue (I do not recommend the 5-minute two-part epoxy for this application)
in the lead screw hole and cover the motor shaft with glue as well. Insert the shaft into the lead screw
with a twisting motion to distribute the epoxy evenly.
[ ] Place the motor and lead screw assembly between the jaws of a vice (or “C” clamp) and squeeze
until the spring device (wavy/coiled) is partially collapsed.
[ ] Leave the motor in the clamping device and let the glue fully cure (this usually takes 12 hours,
read the glue instructions).
6-4
The Slide
A more precision and MUCH more expensive commercial slide
such as the Parker slide (or similar) (show in the photo at left) may
be used instead of the drawer slides described in this section. I will
not cover the details of using these slides in this text.
The ram slide described below is based on a “center under drawer
slide,” intended to be used as a drawer slide in kitchen cabinets. I
purchased them from a home center (Lowe’s) store. Purchase one slide, such as a Knape & Vogt
KV1129PZC16 rated for 35 pounds. If you are unable to purchase the specified slide, the dimensions
of the drawings in this section may need to be altered. If possible, before purchasing, extend the slide
to its maximum length to check for rigidity. If there is play, reject it, and select another one.
These slides have four parts; I assigned names to the four
items and will use these names in the text:
1. Outside (wider) Traveling Slide,
2. Inside (narrow) Stationary Slide,
3. Ball Bearings,
4. Ball Retainer Cage.
The specified drawer slide has 12 ball bearings, but only four bearings are needed to build the slide.
A modified slide is shown in the photo at right.
[ ] Start by completely dissembling the purchased slide
Cut the following lengths from the two slides:
[ ] One Length, 2.750” long of the wider-width traveling slide
[ ] One Lengths, 2.750” of the stationary slide
[ ] The ball retainer cage should be cut so that there
will be a minimum of four balls spaced at least
1 1/8” apart.
[ ] With a file (or milling machine) square the ends
and deburr all the pieces.
[ ] Drill two mounting holes as per drawing in the
large traveling slide piece.
[ ] Drill two holes as per the drawing in the stationary
slide piece.
6-5
Slide Assembly
Slide assembly can be a little tricky until you get the hang of it. I suggest you practice assembling and
disassembling the slides prior to the final assembly.
[ ] First, remove all dirt and metal chips from the bearings and surfaces.
[ ] Mount the stationary (small) slide with the three holes in it to the
2”x2” angle and the mounting arm using two 6-32” socket head cap
screws (see top photo on page 6-9).
[ ] Start by holding two balls in the ball retainer cage, and insert them into the (wider) traveling slide.
[ ] Push the stationary slide forward into the ball retainer cage. The cage will move forward until it is
time to install the last two balls.
[ ] With the balls in position, slide the traveling slide further into the stationary slide.
[ ] When the slides are assembled and lubed with a light oil or grease, the movement should be silky
smooth; if not, clean and reassemble.
[ ] If final assembly is not ready, secure the slides with a piece of tape to keep them from accidentally
coming apart.
Final Slide Adjustment
Once everything is finished and you are ready to make final
adjustments of the head, the correct slide- and ball-retaining
cage position need to be set.
[ ] The final slide adjustment is correct when the ballretaining cage, the traveling slide and the moveable slide is
fully extended and in the position shown in the photo right.
Make sure that there is no interference between the ballretaining cage and any of the screws mounting the small
slide or the larger slide. If there is an interference problem,
either file the heads of the screws or bend the ball-retaining
cage to permit clearance.
In operation the slide must be rigid yet exhibit the least
friction possible.
6-6
Mounting Arm
Fabricate from a 1.250”-diameter (or larger) solid aluminum or steel stock as per the drawings. The
length of the long arm is not specified and should be made as per your application.
Suggestion: The combined length of the two pieces should be such that the center line of the
electrode reaches 3/4 of the way across the work tank. Using the suggested work tank
(see page 11-1), the total length of the arm should be approximately 6 3/4” long.
The mounting arm permits adjustment left and right and up and down. With this capability it is
possible to square the electrode with the work piece in the two planes using the squaring tool
depicted on page 7-8.
This particular design permits operations of the head at various angles, including completely
horizontal (which is really advantageous in certain EDMing applications, namely horizontal
EDMing).
It is important to hold the specified +/- .005” tolerance on
the location of the drill & taped 1/4-20”- hole in the
drawing at left. This is true in particular if you are building
this head to go on a Sherline Mill (see page 9-2).
If the arm is used on the box slide setup (see page 12-2), it
needs to match the Mounting stud shown below.
Mounting Stud
The mounting stud mates with the previously described mounting arm. The mounting
stud (photo at right) is basically a copy of what Sherline uses on their lathes and
milling machines. This quick, easy, and very flexible mounting method may be used to
mount the servo ram on any number of setups from a Bridgeport type mill, to a drill
press as well as other machines.
6-7
All that’s required is to make a fixture to hold the mounting stud to the desired machine. For those
not familiar with this system, the mounting stud is machined so that when the 1/4-20” cone-pointed
setscrew (in the mounting arm) screws into the 45-degree groove (on the mounting stud), it pulls the
mounting arm tightly against the slide’s front face. By slightly loosening the set screw, it permits the
mounting arm (and head) to be rotated to any position.
Using the squaring tool (see page 7-8) and the flexibility of the mounting arm, it’s possible to
completely true the EDM servo head in relationship to the work plate in a few seconds!
[ ] General dimensions are shown in the next two drawings. Start by
machining a partly finished stud as shown in the first drawing at right.
The .688” dimension (second drawing) from the front face of the slide
should be verified before machining.
[ ] This verification process is accomplished by actually placing the
mounting stud in the counter-bored 1/2” hold in the center of the slide
(see page 12-2). Position the short length of the mounting arm (previous
page) over the mounting stud and hold it tight against the slide’s front
face, all the while making sure that the mounting stud is pushed all the
way into the slide counterbore on the back side.
[ ] In this position “lightly” tighten (see note below) a 1/4-20” conepointed set screw against the mounting stud, leaving a mark on the stud
indicating the center line of the set screw.
Note: Do not tighten the set screw to tight; it will raise a
burr on the mounting stud, making it difficult to remove
the short mounting arm. Do not cut the 45-degree angle
(drawing above) on this center line location; instead
move approximately .015”-.025” to the outboard (left)
end of the mounting stud and cut the 45-degree angle.
When assembled, and the setscrew is tightened, this
offset will force the mounting arm tightly against the face
of the slide.
Construction of the Head Bracket
[ ] Cut from a piece of 1/8”x 2”
angle (steel or aluminum) a piece
that, when finished machining, is
2”x2”2” (see drawing at left).
[ ] Start by drilling the three holes
on the part center line as shown in
the front view. All three are
6-8
clearance-drilled. Two are clearance for 6-32” screws and the other one is clearance-drilled and
counter-sunk for a flat head 8-32” screw. See the photo on the bottom of the previous page.
[ ] DO NOT machine the top side of the motor-mount bracket at this time.
[ ] Attach the motor-mount bracket to the mounting arm using the 8-32”
counter sunk hole (see photo at right).
[ ] The smaller stationary slide and the motor bracket are attached to the end of
the mounting arm as per the photo at right (the 8-32” countersunk screw is
covered by the stationary slide).
The following operation cannot be completed until the Indexer Chuck (described in the next section)
is complete
[ ] Machine a 60-degree (sharp-pointed) taper on the end of a 1” piece of lead
screw of the same thread pitch (3/8-24”) used in the indexer chuck.
[ ] Thread the lead screw into the indexer chuck nut with the sharp point as
per the photo at left.
[ ] Mount the indexer chuck head to the traveling slide (see photo at left).
[ ] Assemble the slides as described on page 6-6. Once assembled, you have a sliding transfer punch
for locating the exact centerline of the lead screw as it travels on the moveable slide.
[ ] Slide the indexer chuck forward till the sharp-tapered lead screw end contacts the bottom side of
the motor mount bracket; then gently tap the indexer chuck, leaving an indentation on the head
bracket.
[ ] With this location, determine the center punch and proceed to perform the necessary machining
operations to prepare the top of the head bracket to accept the DC gearmotor (of your choice) as well
as other required machine work as per the previous drawing.
Head machining notes: The locations and size of the holes required will, of course, depend on the
motor selected, hence the vague dimensions shown in the drawing. The Pittman motor I used had a
1/2” boss surrounding the output shaft so I drilled and reamed a hole centered on the transferred
location from the lead screw.
The two holes with the note “as required for switches” need to be located so that the hole in the back
does not interfere with the ball slide and the front switch hole needs to be aligned so that it contacts
the indexer chuck when in the full UP position.
The SW wire hole (.187” diameter) between the two large switch holes is necessary to permit wires
to pass through the base to switches on the front panel. The two D&T 6-32” holes are for mounting
the head cover.
6-9
Front Cover
It is not required that a full-length front panel such as I illustrate be
on the servo head, but some kind of panel is needed to mount the
switch and two LEDs. The drawing at left shows the layout of the
panel used. Before mounting the cover, read the section on page
6-11 on wiring options.
[ ] Lay out the cover on a flat, .0625” aluminum sheet as per
drawing. Drill and deburr all holes.
Prior to folding the cover, check for possible clearance issues with
the leadscrew and switch holes on the bottom side of the mount.
To get the needed clearances (see photo below) may require some filling
or milling in areas to clear the holes. An example is when the nut that
secures the front switch in place needs clearance.
[ ] Once the nut and screw clearance issues are resolved, fold the cover as
per the photo at right.
[ ] If you do not have a sheet metal brake
a shop vise or even two pieces of angle
iron clamped together with Vice Grip
type pliers may be pressed into service
as a temporary brake. Use a plastic hammer or a rawhide mallet in
making the 90-degree bends.
[ ] Before mounting the cover you may want to install the various
head components and wire them (see page 6-11).
[ ] Mount the cover to the bottom side of the head mount with two 6-32” screws.
The Finished Head
This is a photo of the finished head showing the layout of the cover.
The LED on the left indicates the Gap condition.
The second LED over is used when setting the burn depth. The switch
on the right is used to toggle from a motor-burn configuration to a
burn-depth configuration. The switch may also be used to turn the
servo on and off. With a setup such as this, you have a visual
indication of the gap condition by the color of the LED.
6-10
Head-Wiring Options
There are several options when wiring the head. I encourage you to wire
the head with the full complement of Up and Down limit switches, as
well as an ON/OFF switch and various LED indicators. The switches,
diodes and LED shown in the schematics that follow are all mounted on
the head cover. The photo at left shows the UP/DOWN switches prior to
being wired.
These two switches are normally closed switches; when tripped they
open, shutting off the DC motor. Diodes are wired across the switches,
so that when the current is reversed through the circuit it flows around
the respective switch via the diode, thus permitting the indexer chuck to
be moved.
Make sure when wiring the diodes that the banded end is wired as per the schematic. The vertical
lines on the schematic diodes represent the band drawn on each diode.
In the photo above note the two switches; the UP limit switch is shown on the left. The top of the
indexer chuck has just tripped this switch. The Down limit/depth switch is on the right. When wiring
it, make sure that the wires do not come in contact with the slide when in the full-up position.
The Down limit switch serves two purposes; in normal operation it may be used to limit the indexer
chuck’s travel and keep it from running off the end of the lead screw. With the addition of a depth
Digital Read Out DRO, a Double Pole Double Throw DPDT switch and an LED can be set to control
the depth of a burn. If you are using this EDM for more than just a tap burner I highly recommend
you opt for the depth control feature.
In the following schematics the .01uf cap. is to help filter out spikes that could cause failure of the
driver chip. PL1 is the plug connecting the servo to the power supply.
The Bicolor UP/DN LED shown in most of the following drawings connected across the DC motor is
connected such that it changes colors depending which way the motor is running. A wiring
suggestion Green (GO) indicates that the electrode is advancing toward the workpiece. Red (Retract)
indicates that the electrode is being retracted. This is a very useful visual indicator of what is going
on with the servo, and this is especially valuable information in noisy machine shop environments
where it is difficult to hear the spark. If your colors are reversed, it is easy to correct; just reverse the
connections of the LED.
Wiring the Head to the Power Supply
[ ] Cut a 4+-foot length of two conductor-stranded wire (at least 22-gauge) to make the electrical
connection cord from the power supply to the ram head. Wire that is designed for wiring intercoms or
even light duty speaker wire works well for this application.
6-11
[ ] Solder one end of the wire to the two short terminals (non-common) of a 1/8” stereo plug. The
long (common conductor) strain-relief crimps around the wire but no electrical connection is made to
this strain-relief. Thread the stereo plug cover in place. The following drawing shows the electrical
connections on the plug. Verify with an ohm meter that you are using the correct connections.
ON/ OFF Motor Control
The absolutely simplest wiring is shown left. It is the output of the motor
driver applied to the motor. This works fine, except you have no ability to
turn the motor off at the head. The next simplest is
shown in the schematic on the right. It is the servo
motor with a series switch mounted on the head. This
gives the option to turn the motor ON/ OFF at the head.
ON, OFF and UP limit switch
A step up from the simplest (see the schematic at left) provides an upper limit
switch. The purpose of this switch is to shut the power off to the motor when
the indexer chuck is fully retracted. This virtually eliminates the possibility of a
blown motor driver chip, stripped or stressed lead screw nut threads or even
damage to the motor itself. When the polarity of the voltage sent to the head is
reversed, the 1N4001 diode bypasses the switch permitting the motor to move
off the UP limit switch. This drawing also provides for a bicolor LED, which
gives a visual indication as to the direction the electrode is moving for me this
is a very useful indicator.
ON, OFF, UP and DOWN Limit Switch
The next option is to use both an UP and DOWN limit switch but with no
depth setting capability. This requires a second 1N4001 diode by passing the
Down switch. The Down limit switch is to stop the indexer chuck from
running off the end of the leadscrew.
I do not show a provision in the text of this book of tripping the Down switch,
but it would require a fixture similar to but in place of the DRO (see page 8-1)
to trip the switch. In my opinion, if you are going to go this far you should go
ahead and add the necessary hardware for setting the burn depth with the
DRO.
6-12
ON, OFF, UP, DOWN Limit and Depth-Control Switch
The final option is full control of the UP/DOWN-limit switches as described in the previous section.
What sets this configuration apart is the addition of the DPDT switch, which directs the current in the
correct path so that the DRO might be adjusted for a specific burn depth.
The DPDT mini-toggle switch is toggled to the burn mode in
the schematic at left. This is the switch position during a
normal EDM burning operation. When toggled to Depth/
Setup position, the Dpth Ind. (D16) LED is activated. See
pages 17- 2 & 17-3 for the correct setup procedure.
[ ] When wiring this configuration to a full-length cover (as
shown in all head photos) the quarters for soldering behind
the cover can be a little cramped. I suggest that, prior to
mounting the cover on the head insert, the two LEDs and the
DPDT toggle switch should be mounted to the cover and wire
them as per the schematic.
[ ] Next, solder the 1N4001 diodes across the switches, and
ensure that they (or any wiring) do not contact the slide or
other parts in particular when the indexer chuck is in the fully
retracted position.
[ ] Once these connections are
UP/DOWN-limit switches to
cover.
made run the wiring from the
the components mounted on the head
[ ] Make the necessary
connections to the DC servo motor.
[ ] Once the electrical wiring is
bottom of the motor mount
finished, mount the cover head to the
bracket with two 6-32” screws.
[ ] With the advent of the
possible to have a very
LEDs and switches, as shown
digital label maker, nowadays it is
professional-looking head labeling of
in the photo at right.
6-13
Chapter 7
Indexing
Why is indexing so important in doing accurate EDM work? I will use a very simple example to
illustrate. Suppose you need a square blind (not a thru) hole in a harden work piece. The hole requires
close tolerances and a good finish. It is virtually impossible to perform the operation with a single
electrode. In many cases at least three electrodes are required, maybe more.
If it were a thru hole a single electrode is always a possibility, assuming the
electrode is stepped. The first step (see drawing at left) would rough out the hole,
the second step would bring it to near-finished dimensions, and the third step
would provide the finished dimension and/or the surface finish.
If the hold is a blind hole, the stepped electrode method is not possible. This
means fabricating three individual electrodes with the same geometric shape but
making each one progressively larger. To successfully achieve the desired results,
the successive electrodes must be perfectly aligned with the previous burned hole.
This can be a time-consuming and frustrating task without a means to index the electrodes. You not
only need a way to index the electrode on the machine ram, but you also need a method of indexing
them during the electrode fabrication process.
The indexing system consists of an arbor (see a commercial arbor photo at
right) to which electrodes are attached and an indexing chuck. The onmachine and off-machine index chucks need to be identical in their critical
dimensional layout, i.e., arbor hole size and indexing pin location(s). The
chucks themselves may have different physically outside dimensions.
The indexer system (it is a system) described in this book (see photo at right) is
a greatly simplified version of one design and is sold by System 3R; it is called
the “Mini”.
The arbor inserts into a block called a chuck (see the two System 3R chucks
below), which has a precision machined hole to accept the arbor. After placing
the arbor in the chuck, it is rotated until the arbor’s dowel pin contacts the
dowel pin on the indexer chuck face. Once located, the chuck is tightened, thus
holding the arbor.
The commercial tooling is expensive; the chuck below (left) sells for about
$1000 on ebay (2011). Your complete Pulse built EDM will
cost less than this single item but then I don’t anticipate you
making parts for the aerospace industry with your machine.
FYI, the one on the right was $800! The version you will be
making is much simpler yet works on the same principle.
7-1
Indexer Chuck Overview
Before proceeding with building the indexer chuck, read about indexing on the previous page.
The indexer chuck is made of plastic (a plastic called delrin preferred); it electrically isolates the
electrode from the ram and provides a means for indexing the electrode.
Prior to machining, a couple of decisions need to be made. One, determine the size arbor you wish to
us. I use a 1/2”arbor (the commercial System 3R uses a 13MM arbor); all the drawings that follow
are based on the 1/2” arbor design. It is advantageous to standardize on one size. If you deviate from
the 1/2” size you will need to calculate the dowel pin offset.
Another decision is how to apply power to the electrode. This greatly affects the design of the indexer
chuck. The simplest power-applying method is to place a large alligator clip (such as are used on
battery chargers) on the EDM power supply lead and clip this to the arbor or, in some cases, to the
actual electrode. This is a simple, effective method, but there are some applications) such as when
small electrodes are used) for which this set up is not a good idea.
An alternative is to install a metal insert inside the plastic indexer chuck and connect the insert to the
power supply lead via a screw connection. With this method, when the electrode, or the metal arbor
holding the electrode, is placed in the metal-inserted indexer chuck, the electrical connection is
automatically made. Of course, the metal-inserted indexer chuck comes at the cost of a more
sophisticated design which is harder to manufacture.
Indexer Chuck with no Metal Insert
The following instruction is for the ram with no metal insert. The lead screw nut is
an integral part of the indexer and is machined first; see photo, right, and drawing,
left.
[ ] The indexer chuck is made from a 1 1/4” by 2.875”-long
length of plastic Delrin. Cut and machine to the required
dimensions; double-check that the ends and sides are true
and square.
[ ] Position the block vertically in a milling vice, and locate the exact center (it
may be machined on a lathe with a four-jaw chuck if you prefer).
[ ] Step drill to .332 (Q drill) 1.375” deep; be careful not to drill through,
Delrin wants to grab the drill. Tap with 3/8-24” threads.
[ ] With the indexer in the milling machine vise, go ahead and machine the
“half-moon” notch .250” deep; see the drawing on page 7-3 for dimensions.
This notch serves as clearance for a switch.
7-2
The drawing at left is quite busy and might be a little
hard to figure out. Simpler versions of the drawing
are posted in the following text.
[ ] This operation is on the opposite end from the
leadscrew threads. Position the block vertically in a
milling vise, and locate the exact center (it may be
machined on a lathe with a four-jaw chuck if you
prefer). Step drill to 7/16”diameter (assuming a 1/2”
arbor), 1.375” deep, and be careful to not drill all the
way through, delrin wants to grab the drill.
[ ] Ream or bore a 1/2” hole (assuming a 1/2” arbor)
1.375” deep.
[ ] From the center of the arbor hole move (X axis)
.1875” (or a distance equal to the diameter of the
dowel pin used).This is a critical dimension, so keep
this dimension as close to the size of the chosen
dowel pin as possible to ensure accurate positioning .
[ ] Move .375” on the Y axis
[ ] Drill and ream a hole .1875” (or the size of the
precision dowel pin) to a depth such that at least the
diameter of the dowel pin (3/16” in this case)
extends above the surface when pressed in place.
[ ] Press the dowel pin in the reamed hole; do not use a spring-type roll dowel
pin in this application.
[ ] Position the indexing head so (what will be) the back side is facing up, drill
and tap the two 6-32” holes as per the drawing at left. These are the mounting
holes to attach the traveling slide (see photo at left on page 6-9 and at bottom
left on page 7-4).
[ ] If using the metal insert, do not perform the following operation until the
metal insert is pressed in the indexer (see page 7-4).
[ ] Position the indexer chuck with (what will be) the front facing up. Drill and
tap a #10-24 hole (location “A” on drawing) on the center line 2.500” from the
top. Through this hole threads either a nylon tipped or a nylon screw to apply a
frictional force against the mandrel (or electrode).
7-3
[ ] On the right side of the indexer chuck, drill and tap two 6-32” holes as per the
drawing at left. The .563” dimension is from the back side of the chuck. These
holes are used to mount the Digital Readout (see page 8-1).
[ ] *Optional* Turn the delrin piece 180 degrees on the center line 1 3/16” from
the bottom of the delrin drill and tap for 1/8” National Pipe Thread NPT. It is
through a pipe barb at this location that dielectric may be pumped through the
electrode (see section 29-1 for a more complete understanding of flushing); see
the photo on page 7-5.
Metal-Inserted Indexer Chuck
The following instructions are for preparing the indexer with a metal insert. Follow the above
instructions for making the indexer chuck except the 1/2” reamed hole is enlarged to .750” to accept a
metal insert as described below. An electrical connection via a screw is made to the insert at location
“B” (see drawing on page 7-3), eliminating the aforementioned alligator clip to the electrode.
[ ] Position the plastic block vertically in a milling vise and locate the center line. Step drill to 1 1/16”
diameter, and 1.375” deep on center. Do not drill all the way through; delrin wants to grab the drill.
[ ] Ream or bore a .750” +0 -.001”-hole 1.375 deep to accept the metal insert, as described below
The Metal Insert
[ ] Machine insert as per the drawing at right; the insert may be made of
any electrically conductive material, copper and brass being the best
choices; I have used steel and aluminum, though the latter is not the
best for wear. Use a file to form a slight taper and radius on the .752”
end. The bore (.438”) is left undersized and reamed later.
[ ] Press the metal insert into the previously machined plastic block (see
photos below). The forces are not great; I actually press the insert into
the delrin right in the mill using the mill quill.
[ ] After pressing the insert, find (verify) the center line of the indexer chuck.
[ ] From the center line of the indexer chuck (i.e., the
arbor hole) move (X axis) .1875” (or the width of the
dowel pin being used).This is a critical dimension,
so keep this dimension as close to the chosen
precision dowel pin size as possible.
[ ] Move .375” on the Y axis.
7-4
[ ] Drill and ream (the size of the precision dowel pin) a hole .1875” through the metal insert and into
the delrin to a depth such that at least the diameter of the (3/16” in this case) dowel pin will extend
above the surface of the insert when pressed in place.
[ ] Press the dowel pin in the hole; use only a precision dowel pin in this application.
[ ] Move back to the center, and verify that the center location is still correct after pressing the dowel
pin in. Next ream (or bore) a .500 hole through the metal insert (assuming a 1/2” arbor). Performing
the machining operations in this sequence with the dowel pin keeps the insert from spinning in the
plastic during the reaming operation.
[ ] Position the indexer chuck with (what will be) the front facing up. Drill and tap a #10-24 hole
through at location “A” on the drawing (on page 7-3) 2.500” from the top of the delrin. This hole
goes through the plastic and metal insert. It’s through this hole that either a nylon-tipped set screw or
a nylon screw threads, applying a force against the arbor to hold the electrode in place during the
EDM operation.
[ ] Drill and tap a #10-32 through at location “B,” as shown on the drawing (on page7-3) 1.750” from
the top of the delrin. This hole goes through the plastic and metal insert. This is the screw through
which electrical power is applied to the metal insert in the indexer chuck.
Thought not shown in the drawing, the following is a nice addition to the indexer head.
[ ] Turn the delrin piece over on the center line 1 3/16” from the bottom of the delrin drill and tap
for 1/8” NPT. The previously inserted, electrically conductive insert will intersect with this hole. A
pipe barb at this location provides a means of pumping dielectric through the electrode to achieve the
best flushing conditions during a burn, (see page 16-1 for an understanding of flushing).
[ ] Run the 1/2” reamer (by hand) back in the indexer chuck bore to clean up any burrs.
Misc. Indexer Chuck Photos
Notes: As you can see from one of the photos on page 7-4
(and shown as an option on the drawing), a second dowel pin
hole was located for different indexing options; this isn’t
necessary but while in the mill it was quick and easy.
Another rendition of the indexer chuck is show in the photo
at right. Note the drill and tap a hole for a 1/8” pipe to 1/8”
hose barb in the side of the indexer. This provides a means
for through electrode flushing. Dielectric can be introduced
into the electrode for through electrode flushing.
The photo at left shows the indexer chuck with an electrode inserted and the cross pin in the electrode
rotated such that it is against the alignment pin (correctly indexing the electrode).
7-5
Electrode Manufacturing Fixture Block
There is a need to have a “system” for manufacturing the electrode off the EDM that is compatible
with the index chuck on the servo head. An electrode manufacturing fixture block, i.e.,
manufacturing block provides a quick and simple method to make multiple electrodes (in particular if
you have CNC capabilities) that are all indexed in the same direction. The electrodes may be taken
from the manufacturing block to the EDM; it then may be removed from the EDM “sharpened” (or
replaced with a new electrode), and put back in the EDM you can be assured that the orientation of
the electrode is correct throughout this process. Once the system is set up it requires no dial
indicators, vee blocks, clamps, or aligning tools. This system is so slick, and yet simple, that it makes
the EDM process a joy to behold.
To hold the electrode during the electrode fabrication process I suppose (though I have never used it)
that a duplicate of the indexer chuck, as previously described, may be used but only if very light cuts
or engravings are made on/in the electrode material. It is my opinion that a method is needed which
grips the arbor more securely than the friction of a nylon screw against the arbor.
The manufacturing block seen in the drawing and photos below has a (pinch) clamp to securely grip
the arbor. The manufacturing block may be made of aluminum, though steel is a better choice. The
outside dimensions are not critical; I had a piece of 2” square aluminum in the scrap box so that is
what I used in making this prototype. For precision work it is critical to make sure that the block is as
square as possible. The .500”-reamed hole (assuming you are using a 1/2” arbor) needs to be as
parallel as possible with the sides; take your time and get this correct, and manufacture as per the
drawings below.
There is a horizontal cut into the block to just past the center line; the drawing indicates that it should
be .063” width. That is not critical; what is important is that it frees the upper section of the block to
pinch the arbor when the 1/4-20” socket-headed cap screw is snugged up. This cut may be made with
a band saw, or hacksaw. If you have a slitting saw large enough to make the cut, use it.
7-6
The indexing pin(s) should only be located on the stationary side of the hole. The location of the
indexing pin is critical; remember this manufacturing block needs to be a mirror image (as far as the
relationship between the arbor hole and the indicator pin) of the one on the ram. A second locating
pin may be placed as indicated in the alternate location.
In use, the manufacturing block may be stood vertically or laid on its
side in a mill vice. It is possible to place the actual electrode (assuming
that the electrode is the size of the arbor hole) with the dowel pin
through it as shown in photo at left. Rotate the electrode/arbor until the
two dowel pins contact, then tighten the 1/4-20” screw to pinch the
arbor/electrode so that it will not move during the machining process.
This photo shows a finished electrode ready to be installed on the EDM
Note: Though not described in this text, if you want to build the very
best indexing chuck for the ram on your EDM, I think the pinch design
is excellent. You just need to scale it down to fit on the ram. The photo
(right) is a pinch design I use on a different style ram that is not
described in this book; as you can see, it uses a pincher to grip the arbor.
The following two items, the Electrode Holder and the Squaring Tool
need not be made now; I placed them here as they flow with the
discussion at hand.
Electrode Holders
Electrodes may be attached to the arbor via mechanical methods (most
common and perhaps the easiest) as in the holder photo at right or they may
be glued with electrically conductive glues (see Section 18-1) as well as
soldered in the case of copper and other metallic electrodes.
There are many ways to make holders of electrodes but, no matter how they
are made, all must have a cross-precision dowel pin located on the center line
of the arbor. The simplest version is to use the electrode
material itself for the electrode and arbor, as shown in the
photo above left. If the electrode itself is not
used as an arbor, then an electrode holder is
necessary. The relationship of the holder (the
lower part that actually holds the electrode)
(see photo above) and the dowel pin in the
arbor (the upper part that goes in the indexer
chuck) should be consistent so that when
using multiple electrodes the alignment is
maintained from one electrode to another. In
other words, make all the “like kind” of
holders as identical as possible; the key to accurately indexing the electrodes starts
with the accuracy of the holders.
7-7
The electrode holder in the drawings and the photo on the previous page is designed primarily for
holding rectangular electrodes, though it is not limited to rectangular electrodes. The clamping device
may be as simple as two 10-32” screws, or it may be some type of gripping or pinching collar/bar.
The holders may be made of any material as long as it conducts electricity.
The arbor is pressed or threaded into the holder. If pressed, it should be pressed to a square shoulder.
If it is thread, use a jam nut to secure it. The advantage of the threaded type is the electrode holder
may be rotated to different angles, but then that may also be its downfall.
Of course, the holders may be machined from solid stock to the required size and configuration. I
have found locating the dowel pin in the arbor as the last step in the manufacturing process to be the
simplest and most accurate method. Machine the holder and arbor, and press them together and while
in the press (vise) I tack weld together. Using the milling vise, grasp the rectangular holder and then
proceed in locating, drilling and reaming the dowel pin in the round arbor.
The actual fixturing on the ends of the arbor is up to your imagination. A few ideas for these
electrodes holding fixtures are shown below; they could also include drill chucks, collets, flat plates,
reamed holes, vee blocks, etc.
The Squaring Tool
While not necessary for operation of
the EDM, the squaring tool assists in
the setup of the EDM head. The
object is to create a tool that has a
true 90-degree angle. Insert it in the
EDM head-indexing chuck, bring the
EDM ram down and adjust the head in two planes (fore and aft and left to right). This will require
rotating the tool 90 degrees to ensure that the head is square with the plate in the work tank or to the
part to be EDM’ed.
Manufacture the tool as per the drawing. The round .500” arbor maybe pressed or threaded into the
foot. I prefer pressing and silver soldering it in place; if pressed, it should be pressed to a square
shoulder on the arbor. Place the assembly in a lathe and facing off the foot. If you have access to a
surface grinder and can grind the tool, all the better, but
lacking a grinder the lathe will
work fine.
The only critical aspect is the
squareness of the arbor to the
foot. Of course, the arbor should
be the same diameter as the size
arbor used on your indexing
chuck.
7-8
Chapter 8
Depth Control
Note: It is best if this construction operation is performed after
the EDM is finished; you may use the EDM to cut various parts
of the hardened digital caliper beam and jaw.
To control the depth of burn, I modified a small digital caliper
(hereafter referred to as a Digital Read Out, i.e., DRO) to work
in conjunction with the lower limit micro switch. The photo at
left shows the DRO in its operating location mounted on the side
of the indexer chuck. In operation, the DRO rides up and down
on the indexer. When the remaining jaw of the caliper contacts
the down switch it stops the rams down feed.
I purchase a 4” digital caliper on sale for about $10 from Harbor
Freight that is almost perfect for this application. The Harbor
Freight part number is 47256
(http://www.harborfreight.com).
I am sure that there are many other sources for similar calipers
that will work just as well. The finished DRO, with its mount, is
shown on the following page.
Note: Make notes as you disassemble the caliper, as it will be necessary to reassemble in reverse
order once the caliper is cut.
[ ] Peal away the backing on the caliper and remove the steel back (4 screws) and set it aside.
[ ] Remove the 6 3/4” (stationary member) beam from the caliper.
[ ] Carefully remove the electronics from the sliding member and set aside.
The caliper is hardened and quite difficult to cut with common tooling found in most hobby shops so,
naturally, I used the finished EDM to do the cutting and saved my saw blades, knuckles, and
language much wear and tear. Several parts of the caliper need to be cut away, as the photos show.
The electrode used for the job was a (.062 x 1”) piece of
copper sheet. A quick, low-cost copper electrode may be a
length of copper water tubing smashed flat in a vise. The
EDM is used to literally burned through the caliper beam in
“chop saw” fashion; life is good with an EDM .
[ ] Cut the beam (at a minimum) the length of the lead screw plus the length of the electronic head; in
my case this was about 4 3/4” long, including the front jaw (see photo above).
8-1
[ ] Repeat the EDM-cutting process to remove the other parts of the caliper to make it look like the
modified caliper shown in the photos in this section.
[ ] Once EDM cuts are made, sand to form a smooth surface on the caliper parts.
[ ] With care, reassemble the digital head, and the remaining parts
of the caliper so the finished version looks like what is shown on
the previous page ( minus the mounting hardware).
A mounting bracket is required to mount the DRO to the right
side of the indexer chuck. The bracket may be from a piece of
steel or aluminum angle or it bent from flat stock.
[ ] Manufacture the bracket as per the drawing at left. All holes
should be 9/64” (.140).
The “See text” note shows a possible cutout should a
clearance space be needed to open the DRO to install the
battery. Disregard the third hole in the photo at right. The
“As required see text” note is to screw access holes to
assist in assembly of the DRO once glued to the bracket
(see photo below).
[ ] Degrease the bracket, as well as the sliding portion of the caliper, using a solvent such as acetone.
[ ] Mix a two-part epoxy such as JB Weld
(http://jbweld.net/index.php) and glue the back part of
the caliper to the mounting bracket; make sure the two
screw holes are lined up with the screws and that the glue
does not cover the screw heads. Access is needed to
these screws should you need to disassemble the DRO in
the future (see photo at right).
[ ] Attach the DRO to the indexer chuck with two
6-32” screws as per the photo at the top of the previous page.
Instructions for using the depth Micrometer are found on pages 17-2 & 17-3.
8-2
Chapter 9
Finish the Job
Now that the power supply and head are built you are anxious to actually start EDMing, right??
HOLD ON; it has been my experience that, for most home builders, the storage of the dielectric and
the configuration of the work tank and filtering system usually takes a backseat to the power supply
and the ram. Oftentimes, this is the last item built and many times it is done with a halfhearted effort
or it is never built!! Then the builder contacts me to explain that he can’t figure out why his
sophisticated electronic box does not produce the finely detailed work that he had hoped for.
It is not uncommon that I hear of, and/or see; builders throw a vise in a plastic dishpan with some oil
and start burning. The result is a dielectric so filled with burn partials and graphite that it looks like
thick syrup. The experience is so frustrating that they become discouraged with the EDM process and
think it must all be smoke and mirrors. The problem is that only half the EDM was completed!
Building a Dedicated, or Add-on, EDM
At this point, decide if you want to build a dedicated, stand-alone EDM, which is what I focus on in
this book, or if you want to build an EDM that may be adapted to an existing machine such as a drill
press or a mill (similar to what was suggested in The EDM How To Book).
If the decision is to go the “adapted” route you will need to design your own device for attaching the
ram’s mounting arm to the particular machine it is being adapted to. If mounting to a mill or drill
press, the arm may be attached using the existing drill chuck or
a collet or perhaps a clamp-type device that attaches around
the quill. The latter, in my opinion, would be the most rigid
and accurate as many mills and drill presses do not have a
locking device to ensure that the spindle remains completely
stationary.
Most conventional EDMs position the work tank on an XY
movement similar to the below-mentioned tables, with the
servo head (ram) held stationary above it. This is the design I
pursue in more detail in this text (see photo at left). In a
dedicated EDM there is usually a storage tank for holding the
dielectric when not in use. The dielectric is pumped from the
storage tank into a second tank called a work tank. The work
tank has movement in the X and Y axis and has a means of
rigidly securing the workpiece during the EDM operation.
Overview of the Work Tank and XY Movement
In the design I am proposing, the storage tank is a structural member of the machine. The XY table is
mounted above the storage tank and the work tank is attached to the XY table.
Inside the work tank is a metal plate with drilled and tapped holes to provide a means of clamping the
workpiece or work-holding device (vise) in the tank. The lower storage tank contains a pump and
filtering system. In operation, the dielectric is pumped from the storage tank (below) into the work
9-1
tank, It flows through the work tank and returns to the storage tank below by gravity. The servo head
is held over the work tank by an overreach arm mounted on a vertical back slide. The ram maybe
raised and lowered (manually or with a lead screw) to assist in placing work in and removing work
from the tank.
The Sherline Conversion
A (very nice) compact version of this system is shown in the
photo (right) using a Sherline lathe with a milling column
attached. This is a photo of the system as it debuted in 2006 at
a CNC workshop, along with an early prototype version of
the pulse supply described in this book.
In this photo, the assembly sits atop a rigid fiberglass storage
tank which contains the filtering system and pump. The small
dielectric work tank is attached to the XY movement (cross
slide) on the lathe. The storage tank holds 3 gallons of
dielectric; the work tank holds about a quart. Dielectric is
pumped from the storage up to
the work tank (via the flexible
black bilge hose on the right) it
flow through the work tank
and returns by gravity. The plastic work tank is a modified electrical
PVC box. If only small EDM work such as jewelry or other small
parts making is the rule, this is a very nice compact design that is easy
to put together.
Originally I designed this small EDM system as a “show and tell” to
take on the road and exhibit. I have fallen in love with this design and
find that I use it often. No additional details of its construction are
given in this text. The photo left shows the new designed head
mounted on the Sherline set up.
XY Tables and Selection
In a dedicated EDM there is a means for accurately positioning the workpiece under the electrode; to
do this the ability to move the work tank in the X and Y axes is necessary. A simple, off-the-shelf X
Y movement is available from most Industrial suppliers. These tables are called mill/drill tables
(sometimes they are called
compound slide tables). All I
have found are quite rigid,
operate relatively smoothly
and have a respectable travel
in both the X &Y axis. They are a good choice for a homebuilt EDM. Depending on size and table
travel, such tables cost $100 and up (plus shipping). The XY table above (right) was purchase from
Enco (http://www.use-enco.com). It is model #201-2536 and weighs 63 pounds.
Grizzly Industrial (http://grizzly.com) offers several tables, including one that appears to be the
same as the Enco table. The Grizzly models are G8750, G5757 and H7979 (which is shown in photo
above left).
9-2
Chapter 10
Dielectric & Filter Selection
For years I used kerosene (known as paraffin in some countries) or lamp oil as a dielectric; it is an
acceptable dielectric and, in some parts of the world, I understand it is still in use commercially. The
advantage is that it is usually readily available from an oil/fuel supplier as well as hardware stores
and some mass retailers in the U.S. It is relatively low in cost, though the price in recent years has
gone up sharply. In my opinion, one of the big disadvantage of kerosene is it is basically a solvent. If
hardware is not properly selected it will attack items such as pump seals, insulations on pump wires,
electrode leads, flushing lines, and many glues (if used in the system). Not only is it hard on
hardware, but it is also very hard on the hands (strips most of the oil off the skin) there is also a loss
of dielectric through evaporation, particularly in warm environments and frequently there is a strong
odor associated with its use.
An alternative dielectric to kerosene that I have used with good success is WD-40, but it is not
inexpensive, costing nearly as much as some commercial EDM fluid. It usually can be purchased in
the U.S. from home improvement stores in one-gallon cans, and it is not harsh on hands or hardware
and there is virtually no unpleasant smell. The single biggest disadvantage (other than cost) is that it
turns black quickly and the swarf appears not to settle out of it readily, and when it does settle out the
remaining WD-40 is darker in color.
Commercial dielectrics do not have the disadvantages to hardware and persons previously mentioned
for kerosene; I would encourage builders to purchase a commercial dielectric if they can afford it
(though I am not sure a person can justify the cost of commercial EDM dielectric if the use is only
occasional). At the present time, McMaster-Carr (http://www.mcmaster.com) sells an EDM
dielectric from Synthetic Lubricants called Sinksyn-1 in 1- and 5-gallon containers for about the
same price ($18.25 [2011]) per gallon as WD-40. The name is a little deceiving, as it is 100%
petroleum-based (not a synthetic) and it has all the advantages of being a true EDM dielectric without
the disadvantages of kerosene and WD-40. Crystal Brite is my favorite dielectric and it costs slightly
more than Sinksyn-1. MSC http://www1.mscdirect.com sell Crystal Brite and other dielectrics.
I have used several commercial dielectrics; they include:
Rustlick’s EDM-30, 100% petroleum based
Hangsterfer’s Crystal Brite, 50% petroleum and 50% synthetic
Hirschmann’s IonoPlus, 100% synthetic
Synthetic Lubricants Sinksyn-1, 100% petroleum-based
The disadvantage of the commercial dielectrics is that it is usually not available locally and the cost is
about twice (or more) that of kerosene. To my knowledge, all these dielectrics may be purchased in
quantities as small as 1- to 5- gallon lots, which usually are more than enough for our applications.
When purchasing commercial fluids, I recommend getting the dielectric designed for fine and/or
finishing work. Commercial dielectric fluid generally costs between $18-30+ per gallon (2011).
10-1
How Flammable is the Dielectric
This is a subject that keeps coming up and I want to make a few comments about my experiences
with some of the commercial dielectric fluids (kerosene excluded). From my experience, the
dielectrics commonly used are not highly flammable, especially when kept at or near body
temperature. In trying to determine how flammable they are I conducted a non-scientific test and this
is what I discovered; you can do your own test and draw your own conclusion.
For the experiment, the dielectric was Rustlick EDM 30 which, to my knowledge, is 100%
hydrocarbon, i.e., it is an oil-based dielectric, unlike some of the synthetic or partial synthetic
dielectrics. For the test the dielectric was about room temperature (70 degrees F) as it entered the gap.
The first test was to do a non-submerged burn; in other words, the only dielectric present at the gap
workpiece interface was what was being pumped through the electrode. The arc was not submerged
in a bath of dielectric. During this test, very small flames would flash and then extinguish themselves
immediately; I never had a continuous combustion. I do not recommend burning under these
conditions (the arc should be under at least 1” of dielectric)
I must make it clear I have never had a fire with any of the EDMs I have built. The most powerful
homebuilt machine I have operated had a maximum average current capability of 15amps.
Another experiment was conducted by placing a small amount of EDM 30 in a metal container and
heating it with a propane torch all the while monitoring its temperature. From time to time during the
test, I would pass the flame of the torch above the surface of the liquid. From room temperature to
250 degrees F I was unable to get the dielectric to ignite. At a little above 250 degrees F, as the torch
passed over the dielectric surface, a flame would ignite but extinguish as soon as the torch was
removed. As the fluid approached about 300 F, a flame would ignite as the torch passed over the
surface and it would continue to burn (weakly) after the torch was removed. It did not violently burst
into flames, nor did it burn aggressively.
From my primitive, non-scientific test and over nearly 25 years of experience with these small
homebuilt machines, I do not view them as being a serious fire hazard. That is not to say that it can’t
happen or want happen.
NOTE: When operating the EDM, have a Class B fire extinguisher ready at hand
and know how to use it to put out oil fires.
Dielectric & Materials Compatibility
Pump housing wiring, as well as seals and other rubber/plastic parts used in the EDM system that
come in contact with the dielectric, must be compatible with the dielectric used. There are times you
may have materials of unknown composition, i.e. flush tubing/hose plastic valves, glues, etc.; the
only way I know to determine what to expect is to place a sample of the material in the dielectric and
over time check to see how it is affected. Some materials change their physical characteristic in as
short a period of time as overnight; others deteriorate much more slowly. Some materials take weeks
or months to be affected, and others are not affected at all after years of exposure.
10-2
This testing is not so easily done with a pump, so try and obtain the specifications on materials used
in the pump before purchasing. From my experience, most pumps that are designed to pump fuel oil
(diesel) and solvents appear to be compatible with “most “dielectrics, but this is no guarantee.
When cleaning the work and storage tanks and connecting hardware, pay particular attention to the
cleaning solutions used. Chemicals such as trichloride, a member of the “chlorinated hydrocarbons
“family” of cleaners, can be death on some dielectrics, especially some synthetic dielectrics. The end
result is the formation of hydrochloric acid in the dielectric. If cleaning (other than wiping with a
towel), it is necessary use the manufacture-recommended cleanser.
Below is a list supplied by several manufactures of dielectric and their compatibility with various
materials.
Hangsterfer’s Crystal Brite:
Excellent: Viton, Nitrile (Buna-N), Polyurethane
Good: Neoprene, Hapalon, Sorbothane
Fair: Silicone,
Not recommended: Natural Rubber, Butyl, EPDM & SBR
Rustlick EDM 30, EDM 250 & EDM 500 :
Good: Viton, Nitrile, Urethane, HDPE, Neoprene
Not Recommended:EPDM, Silicone, Neoprene (EDM30 only)
Hirschmann Iono-Plus : Manufacturer failed to respond
Filters
If you want a neat and clean operation, I recommend keeping the filter and as much of the associated
plumbing inside the storage or work tank as possible. EDM is, by its very nature, a messy
proposition. Following this design idea makes installation of new filter(s) a clean and painless
process, as any spills are contained in the tank. If you have ever changed the oil in your car, you
know what I am talking about. If a non-submergible pump is used, consider mounting it above either
the storage tank (if it has an open top) or above the work tank, where any leaks are collected in the
tank.
Automotive Spin on Filters
The filters I have used the most in the homebuilt EDMs are the automotive spinon filters, designed to filter oil in automotive engines (see photo at right). I have
used these in the past as they were easy to adapt to our application (filters and
mounts), are available in most places worldwide, inexpensive, and do an
acceptable (not great) job of filtering the dielectric The price of the filters has
certainly increased in recent years, and the filtering capacity of these” oil filters”
is not as good as I would like. The filtering medium in the ones I have cut apart is
paper. Most of these only filter down to 20 microns, though I know they are
available in finer micron filtering ranges for an increased price.
10-3
Automotive Filter Mounts/Filters
Remote-location oil filter mounts (filter mount heads) are used in the automotive industry for
remotely locating oil filters in the engine compartment. Single and dual oil filter mounts are
available. These remote oil filters mounts may be purchased from many auto parts stores, as well as
online.
When purchasing a spin-on oil filter mount, be sure and note which
type of oil filter they use. Differently designed mounts accept
different filters; this is usually determined by the thread pitch of the
pipe nipple in the mount and/or the diameter of the sealing surface.
Make sure that the purchase filter is not too long to be installed in
your application. If the specified storage tank is used, a short filter is
required. The filter I purchased was the FRAM TG-16.
Jegs High Performance. (http://www.jegs.com) is one online source I have used for mounts. Their
part number for the single mount is 969-1413. The part number for the duel mount (as shown in the
above photo) is 259-15707
Diesel Filters
Diesel fuel filters are also spin-on filters and look almost identical to the automotive oil
filters, though they filter a much finer micron range. They are usually larger filters than
are the oil filters and cost 2-3 times the price. One of the main reasons for filtering the
diesel is to remove water, and some have a drain on the bottom of the filter for that
purpose (see photo at right). If you purchase these filters, make sure you purchase the
ones specifically designed to remove particulates. I have not used these filters in an
EDM application but know of people who have and they indicate good results. If using
the diesel filter, consider using a pre-filter (discussed below) to extend the life of the diesel filter. I
am not sure if the automotive oil filter mounts will work with the diesel filter; I suspect you will need
to buy the mount designed specifically for them.
Water Filters
Another filter idea suggested by a reader that I think perhaps has some merit is the use of
a whole house water filter (shown in the photo on the left). These usually use a clear
plastic housing and a pleated paper filter element. A quick check at a local retailer
revealed that 5-10 micron paper filter is commonly used in home water filter systems.
Assuming these paper filters will work to filter the dielectric, this is as good, if not better
than, most of the spin-on automotive oil filters and is more in the desirable micron range
we need for EDM work. The filtering element is physically longer than the oil filters I am
familiar with but that does not necessary mean that there is more surface area exposed the
filtered fluid. The price for the filters is very reasonable and you frequently find them on
sale, which would help reduce the price even more.
Filter Bags/Prefilter
If using a gravity flow filtering system, or a low pressure (less than 9 PSI) centrifugal submergible
pump, I recommend not using spin-on cartridge-type filters; instead use filter bags or a filter sock.
10-4
Even if you use the spin-on filters with a higher-pressure pump, it is a good idea to use the filter bags
as a pre-filter, especially if you have the more expensive diesel fuel filters.
The filter bags come in various “trade sizes;” for our application, I suggest using a polyester felt bag
with a “trade size” 1 which is a 7” diameter bag 17” long. This is the pre-filter used in the tank on
page 13-7. The bags are available in a wide range of micron ratings, starting with 1 micron and up.
For our use, 5- and 10- (and perhaps 25-) micron bags work fine. The bags are sold by many
industrial suppliers.
One online company is Filterbag.com (http://www.filterbag.com); they have a wide selection,
including bags large enough to go in a five-gallon bucket (see the following section). They, like most
suppliers, want to sell in box lots of 25 to 50 bags and, for some bags, this can
be quite an
investment.
One suppler, McMaster Carr (http://www.mcmaster.com), sells in
quantities as small as one bag, and their price is very reasonable for such
small quantities.
Below are McMaster Carr numbers for their Polyester felt bag trade size #1:
1-micron bag: #9299T31
5-micron bag: #9299T32
10-micron bag: #9299T33
25-micron bag: # 9299T34
5-Gallon Bucket Bags
For those building a simple gravity flow filter system, the previously mentioned Filterbag.com
company sells a nylon, mesh-type filter bag designed to fit in a standard 5-gallon bucket (see picture
at right). These bags are ideal for a low-volume gravity-feed process. They are
available in several different micron ratings. They may be used as a pre-filter to
finer micron “spin-on filters”, or they may be used as a stand-alone filter.
Unfortunately, there are two big drawbacks to using these 5- gallon (bucket)
bags
1. The suggested company wants to sell them by the box of 50 filters. They will
sell in quantities as small as 25 but there is a handling fee.
2. The real bad news, in my opinion, is that they are expensive! The 10-micron,
5-gallon bag part number is NMO105GP, and it sells for $44/bag.
The 5-micron bag is NMO105GP, and it sells for $46/bag.
As of the printing of this book, I have not been able to find another source for the 5-gallon bags but
who knows what will be available; do an online search and maybe someone else will be selling such
a filter and sell them in small quantities. If you find a good source, please email me the information.
A possible alternative is to use the much smaller #1 bag (see “Filter Bags” in the previous section).
This would require building a framework to hold the size bag you select inside the 5-gallon storage
container. I will leave it up to the individual builder to design their own frame work. Keep the bag top
below the bucket top and centered in the container being used to avoid over flowing the bucket.
10-5
Chapter 11
Tank & Pump Selection
The dielectric storage tank may be any open-topped-type tank that is compatible with the dielectric
being used; I chose fiberglass tanks for both the storage tank and work tank. The particular tanks are
manufactured by the Molded Fiber Glass Tray Company and are sold through various outlets in the
US. Mine was purchased from the US Plastic Corp. (http://www.usplastic.com ).
The storage and work tank are as listed below. You can scale up or down by selecting different tanks.
The storage tank (US Plastic part number #51081) is approximately 21”L x 15 5/8”W x 6’ D.
The work tank (US Plastic part number #51083) is approximately 17 7/8”L x11 1/4”W x 6’ D.
You may think these tanks are a little expensive and I agree, but you must understand the tanks are
quite rigid so it is possible to use them as a structural member of the EDM. It helps keeps the overall
cost down and greatly simplifies construction. These particular tanks can take lots of abuse; they look
very professional and they have proven to work well with all EDM fluids including kerosene, the
harshest dielectric I have worked with.
If you use the storage tank as a structural member, purchase one with a wide footprint (length and
width) that will store the need quantity of EDM fluid yet keep the machine mass as low as possible.
In applications where a rigid storage tank is not used, it requires a framework (around and/or over the
tank) to carry the weight of the XY table and the work tank. I am not going to go into detail as to how
to build such a structure but I will make a few suggestions. Metal, in my opinion, is by far the best
choice for this framework but it has become increasingly expensive and would be very heavy (which
in many ways is an advantage). A framework of square steel tubing would provide a rigid structure
that is reasonable in cost and not excessively heavy. Another option would be to use wood and build
a “box” around and over the storage tank. If you do use wood, make sure to seal it well to minimize
moisture absorption
When using the rigid fiberglass tanks, there are at least two options for supporting the XY table and
work tank above the storage tank. One is to create a solid deck over the entire tank and attach the
table to the decking. The other option is to construct a frame using square steel tubing welded
together and bolt the table to this frame. Both work well, though I prefer the steel tubing frame.
Other Storage/Work Tank Options
In this section are some suggested simpler designs for the storage and work tanks; these may be more
applicable to your needs and budget than the previous described tanks and XY movement. Most of
these are concept ideas without any drawing and construction steps. Some of the ideas I have built,
and some have been built by others. Builders ask if it is possible to use only one tank to serve a dual
roll as both a storage and a work tank. The answer is yes, but I would discourage such a system
except for only occasional EDM work. If you are not careful it can be messy and waste dielectric, and
usually involves getting your hands in the dielectric. Avoid such a design unless you are on a very
tight budget or have a real space problem.
11-1
I think that the simplest storage system is one where the dielectric is stored when not in use in a
separate, covered container such as a five-gallon bucket. The work tank could have a flexible drain
tube on the bottom or side of the tank. When the EDM job is finished, or when the dielectric level
needs to be lowered to facilitate working in the tank, simply drain the tank into the storage container.
In such a design, the free end of the drain tube must be secured above the highest dielectric level in
the tank when the tank is filled. To fill the tank, dielectric is manually poured from the storage
container into the work tank. It’s a bit messy, but it is low-cost and certainly is a vast improvement
over the previous idea
To advance the idea a little further and keep it simple, you could include the EDM head mounted on a
vertical post (with an overreach arm) attached to the work tank plate. The plate with the post and
head are removable from the tank, i.e., it just sits in the tank but is not attached to the tank. The
location of the workpiece would be “eyeballed” using either a vise or clamps to hold it in place under
the electrode. There is no provision for moving the workpiece with any kind of precision XY
movement. By having the plate, post, and servo head assembly removable, it greatly facilitates setting
up the workpiece (under the electrode).
I suggest lifting the assembly above the tank and placing it on some kind of rack
during the workpiece setup phase, so that any dielectric would drip back into the
tank. Such a design could be incorporated into the idea first presented in this
section; at least this would keep your hands out of the dielectric during setup. I
actually have used such a setup to test various ideas in designing homebuilt EDM
units. The fine example shown (left) was built by Brian, one of the readers of the
first book. I thought Brian did a great job in executing this design concept.
Another alternative design that I would suspect (though I have not built) would
work well and could be easily adapted to Brian’s design is to mount a small XY
table/vise inside the tank, on the removable work plate. In this case the XY table/vise would be
completely submerged in the dielectric. This gives some means of accurately positioning the
workpiece under the electrode during setup. Know that the dielectric can get dirty, especially if
graphite electrodes are used; in such a design the XY table/vice could become “gunked up” rather
quickly with the black swarf from the burn unless care is taken after each burn to clean it.
The next design idea (see photo at right) incorporates a pump
and filter system; this was one of the setups I put together for
demonstrating the EDM at shows. I needed a quick set- up that
was compact and simple.
It uses two long ball bearings (under) drawer slides attached to
the top of a work tank, forming the X axis. The Y axis is
composed of two similar slides attached at a right angle to the
X- axis slides; these are the two closely spaced slides in the
picture running crossways on the tank. The Z axis slide attaches
11-2
to the Y slides (with a piece of angle aluminum) at 90 degrees (vertical and horizontal). These slides
may or may not have a lead screw for positioning them; the ones shown do not have a lead screw but
can be locked in place.
A similar design would leave the work tank stationary and move the head above it in all three axes
(X, Y, and Z); the slides would not be attached to the tank, but instead they attach to a frame above
the tank. Another option is to move the tank in only one axis, say the X direction, and move the head
only in the Y axis (or vice versa); this would simplify the framework. In some ways I like this idea,
but I have not built it.
In designs where there is a pump and filtering system in the
work tank, the pump could be used to pump the fluid into a
storage tank above the work tank, see the photo left. A
covered, vented, five-gallon plastic bucket works well as a
storage tank. Cut a hole in the storage container and place a
Thru-hull (bilge) fitting (or connector) in it, see photo on
page 13-1. Thru-hull fittings come in straight and 90-degree
configurations. They may be purchased at your local
boat/marine supplier, or do an online search for “thru-hull
fitting.” I would use at least a 3/4” or larger-size fitting and bilge hose. The plastic bilge drain hoses
that mate to the thru-hull fittings are very flexible and work well in this application.
An inexpensive yet effective valve to control flow from the
storage tank (to the “Y” valve) I have used is the plasticball type used on garden hoses (see photo at the left). The
“Y” garden valves (see photo at right) permits flow from
storage to the work tank. Then it may be adjusted for circulation only
in the work tank. Hardware stores and home-supply centers sell
adapters to adapt the garden-hose threads to standard pipe threads, or
hose barbs.
One very important item that is missing in simple systems is a means of circulating and filtering the
dielectric. If the EDM burn goes on for more than a few minutes, it definitely needs to be filtered (see
page 10-3&4).
Pump Selection
At this point I would like to address the pump and dielectric filter designs, as this may affect the
options you choose to exercise in building the remaining part of your EDM. If the EDM operations
last for more than a few minutes, it is necessary to supply a clean flow of dielectric fluid to the gap
area. This greatly assists in promoting an efficient, stable burn under most conditions. I cannot overstress the importance of having a good dielectric filtering and flushing system. A good flushing
system will make a poor EDM generator look good; a poor flushing system will make a good EDM
generator look bad. The dissatisfaction with many homebuilt EDM machines, I think, comes
from a poorly designed, or nonexistent, dielectric cleaning and flushing system. I have said it
before, and I will say it again: With a pulse machine flushing is, MUCH more important than with an
RC machine. If you do not have good flushing with a pulse machine, you will not be happy with its
performance.
11-3
In the past, I have used a brushless 120VAC submergible pump to circulate fluid through the gap and
filters. Such a system is very quiet and vibration-free and, for a small compact EDM, I am convinced
this is still the best choice in most cases. Of course, there are exceptions and you may find that other
options are more suitable for your application; some of those options are explored below.
The Pump Dilemma
In designing this EDM, I came up against an unexpected problem of how to pump (and filter) the
dielectric. In The EDM How-To Book, I detailed the use of a 115VAC submergible pump and
canister spin-on (oil) filters (all located in the work tank in that design) to take care of the dielectric
cleaning and circulation. The design was based on a (single source unfortunately) ready available,
reasonably priced, submergible pump, with the ability to deliver fluid at approximately 9 psi (pounds
per square inch). This particular submergible centrifugal pump was critical to the success of the
filtering system. Most small fractional horsepower centrifugal pumps are not capable of delivering
nearly 9 psi. The importance of using a pump capable of developing higher pressures is to extend the
useful life of the filters, i.e., more pressure to force the dielectric through the filter(s) as they start to
fill up with debris.
About two years after writing that book, the specified pump was discontinued and no suitable
replacement was to be found! Recently a different manufacturer started building a pump with
specifications closely matching those of the original pump. At the present time, I have only a couple
of sources (U.S. only) for this pump. These new pumps are manufactured in China and, sorry to say,
the fit and finish is nowhere near the quality of the early pumps. I hope that the mechanicals of the
new pumps are good. I have one of the previous pumps still running strong after nearly 15 years of
service.
When selecting a pump, submergible and/or non-submergible, select one that is made of materials
compatible with the dielectric you use; see the section on Dielectric & Materials Compatibility on
page 10-2. Choose a pump that is capable of delivering at least 7-8 psi. or greater. If the pump is
positioned above the dielectric tank it MUST have the capability of generating suction to lift the
dielectric from the tank (submerged pumps require a flooded inlet).
Submergible Pump
The submergible pump which I recommend is a Dayton model 2HNN9 (see photo
at right). It is a 1/20hp 15, 60Hz pump with overload protection. Maximum head
pressure is 18ft. The pump has a Viton seal and the case material is an acetal
copolymer. The discharge fitting is a 1/2” male NPT. There are a few sources that
I know of for purchase, but I think they all source it from Grainger Industrial.
11-4
Grainger Industrial :
http://tinyurl.com/6b76ee
As far as I know Grainger only sells to companies. If you do not have a company, but know someone
who owns a business, perhaps you could have them pick the pump up for you.
Bizrate
http://tinyurl.com/6gsoy54
Wholesale Industrial Supply
http://tinyurl.com/5n26z4
Drillspot
http://tinyurl.com/4gc2rup
Other submergible pump options are solvent pumps. The Little Giant (1-YS, see photo above), has
been around for years. Solvent pumps are an excellent choice, as they are hardened against solvents
but the head pressure for most of these pumps is about half that of the 2HNN9. The head pressure on
the 1-YS is 7.4 feet. If you want to use a spin-on type filter using the lower head pressure-solvent
pumps, you definitely will want to use a prefilter such as the filter bags (see page 10-4). The other
option is to use only the filter bags as a gravity flow filtering system. . The 1-YS is available at some
of the suppliers above, as well as on line (enter 1-YS in a web search engine).
Non-Submergible Pump Options
Should you have to use a non-submergible pump, there are a few things you should know. Most of
the small non-submergible pumps sold are not rated explosion-proof. Most that I am familiar with use
brushes, which is a source of sparks. If the dielectric is hot enough to evaporate, i.e., generate fumes;
then there is always the possibility of the motor sparks causing a fire. It is my uneducated opinion
that most of the fires in commercial EDMs come from the dielectric getting too hot. I do not
recommend operating your machine with the dielectric much above body temperature. If you choose
to use a non-submergible pump to be completely safe, the pump should be explosion-proof. It is my
opinion that this is taking it to the extreme and not really necessary. An explosion-proof pump
REALLY increases the price and greatly decreases the selection of pumps. I am aware of a small
commercial EDM (10amps) was sold in the mid-1980s by a large EDM manufacturer that used a 12VDC external bilge (marine) pump to circulate the dielectric! I am not recommending the use of
such a pump, though I have used them and never had a problem; maybe I was lucky.
In good conscience, I can only recommend using a pump that is USCG, ISO
8846 compliant for diesel fuel transfer. Such a pump is a Flojet #04105501G
(see photo at left). It is a quad-chamber, diaphragm pump. It is self-priming to
8 feet, with a maximum flow of 3.3 GPM, and a maximum pressure of 25psi. It
operates on 12VDC and is fused for 7amps. This pump is rated for intermitting
duty, which usually means less than an hour of continuous duty. The pump sells for about $100
(2011). Grainger Industrial (http://tinyurl.com/6py8b7) handles this pump as well as several online
pump suppliers
11-5
Northern Tool (http://tinyurl.com/6e6ldu) sells a pump, built by Tuthill (see
photo at left). It is a Model# FR1612, operates on 12 Volts, DC and delivers 10
GPM; it too is designed for intermitting duty. I do not know if this pump has the
official IOS rating; their advertisement says that it is for “diesel fuel transfer.”
The two pumps suggested are able to easily create discharge pressures far in
excess of 9 psi. By going this route, suctions flushing (page16-4) is achievable
which was/is not possible with the submergible pump. An added benefit is that
the filter life will be much longer than what can be expected from most submergible pumps.
All the less expensive diesel fuel transfer pumps I can find are designed to operate 12VDC; I wish I
could find a version of these that operated on 115vac. Most have lower flow rates than I would like
(though they are acceptable). The 12VDC means purchasing a transformer and rectifier and building
a power supply for them, or using an auto battery just to run the dielectric pump. Perhaps an old car
battery that does not hold a charge well enough for automotive use would work for this application.
These motors don’t really care if the battery can only deliver 11 volts instead of 12. The battery could
be kept on a trickle charger or charged on a regular charger from time to time
11-6
Chapter 12
The Back Slide
The servo ram is attached to an easily constructed box slide that permits the head to be rapidly moved
from one position to
another. The slide may be
equipped with a lead
screw for raising and
lowering the head or, as in
this design, the leadscrew
may be omitted.
Without the lead screw, it
permits very rapid
positioning with the
benefit of being able to
completely remove the
head from the machine.
The ability to be able to
quickly move, or even
remove, the servo head
from the machine and,
then reposition it in
exactly the same location
has a tremendous
advantage when there is a need to load or unload a (large) workpiece. It is also nice if you want to
inspect the EDM surface in the middle of a burn, and then return to the exact position.
I use a box slide design in this project because it is simple to manufacture, low in cost, and works
extremely well for this purpose. A dovetail ways and slide would be excellent, but for most it is not
as easy to fabricate. If care is taken during the machining process a box slide, properly adjusted, can
approach the accuracy of a
dovetail slide; this is particularly
true in this application.
[ ] Machine the box slide from
steel or aluminum as per the
drawing above. The 1/2" X 3/4"
thru and counterbored hole in the center is where the mounting stud mounts
(see page 6-7).
[ ] From a piece of CRS machine the slide clamps (2) as per the drawing on
the right.
12-1
[ ] Machine the 1/4” square key stock gib as per the drawing on the top of page 12-1. The gib is
shown on the extreme right side of the drawing.
[ ] Completely deburr all machined parts.
[ ] After machining the mounting stud, the photo at left shows what it should
look like installed in the slide (the bottom side in the photo at right).
[ ] The two 10-32 SHCS (Socket-Headed Cap Screw)
on the end of the slide (with the jam nuts) are adjusted
to take all side play out of the slides movement.
If there’s too much play fore and aft (front to back), this will require removing
the clamps and milling the needed amount to adjust for the right sliding fit. If
binding occurs between the CRS ways and the clamp pieces, cut a little deeper
or shim the clamps above the ways.
The center SHCS screw is used to lock the slide in place along the ways.
The 6-32 screw (second from top) that threads into the 1/4” gib is left loose; its
purpose is to keep the gib from falling out when the slide is removed from the
machine.
To assist in getting the slide to easily slip on to the CRS ways (after all
adjustments are made), file a small taper on the CRS end.
Frame Work Overview
Before building the frame, read the section on the Storage and Work
tanks. In the proposed design, a frame work is necessary to hold the
XY table, the work tank and the EDM head. The frame is constructed
using 1” and 1 1/2” square steel tubing (14- or 16-gauge).
Two options are offered: one requires minimal welding, and the
second option is a completely welded frame work. If you are not
familiar with welding thin wall steel tubing know that it wants to twist
and warp unless you securely clamp and properly tack the weld prior
to welding solid. If welding, clamp the tubes to a flat metal surface to
ensure they are true and parallel. Picture frame clamps are a low-cost
way to ensuring squareness of the alignment, but even using the frame
clamps still tack-welds everything before welding solid.
Back Slide Rail
Start constructing by fabricating the back slide rail. This vertical rail
has a box slide arrangement that permits the head to be raised and
lowered (see photo on pages 12-4 & 12-6).
[ ] Fabricate the back slide rail from a 17” length (or longer) of 1 1/2” square tubing as per drawing.
12-2
Two holes are drilled (1/4” and 7/16”) through the side (on the bottom section of the tube) .750 and
4.250 from the bottom. When assembled, the back rail pivots around the upper 1/4 bolt. The 7/16”
bottom hole provides clearance for tilted fore and aft to correct for any slight misalignment of the
frame.
CRS WAYS
The box slide (see page 12-6) rides on WAYS formed by using a piece of CRS. CRS (unlike hot
rolled steel), is held to a few thousand tolerance and has no scale on it. It’s a readily available
material purchased from online steel suppliers or, better yet,
from a local machine shop. Cold roll steel requires minimal
machining and can deliver the accuracy needed in this
application. Hot rolled steel, i.e., structural steel, will not
work for this application.
[ ] Machine the WAYS as per the drawing (right); note that
the seven holes along the edge are 3/16” reamed holes.
Tubular Frame Option
Note: There is an aluminum
product called 80/20 which
could be used to fabricate the
frame if you feel
uncomfortable with welding.
Basically 80/20 is an industrial
erector set; it comes at the
expense of less rigidity and
increased cost:
http://www.8020.net
The company often has
“scratch and dents” for sale
through, on online auction sites.
The suggested fiberglass
storage tank has a recess step
approximately 1/2” deep and
3/8” wide running the length of
the tank top. I use this recess to
mount the tubular frame. The
frame is all welded construction
using 1” square steel tubing cut
14 7/8” long to fit snugly into
this recess, see photo following
page.
[ ] There are two spacers (see photo above)(the front one is 1” sq. and the back one 1.5” sq.); both
were cut 5 1/2” long so that the mounting holes (slots) in the XY table used would line up on the
12-3
center line of the two long tubes. Cut as per your XY table to ensure the mounting holes line up on
the center line of the tubes.
[ ] Manufacture the tubular frame as per the pictures and drawing in this chapter. The two bolt holes
in the 5” uprights is 1/4” in diameter.
[ ] Place the back slide rail, with a note card on each side, between the two 5” uprights and bolt all
three pieces together and then weld in place on the 1.500 base tube. After welding, remove the note
cards. The cards provide a little clearance between the uprights and the back slide rail; this space is
needed for adjustment later.
In the finished machine the work tank sits atop the XY table (which bolts to the tubular frame
horizontal member).
Position the frame on the work tank (see photo above left) ;the photo above right shows the XY table
in position on the storage tank.
Wood Deck Option
Note: You might want to review
the section on Plumbing the
Tanks on page 13-3 before
proceeding with this section. Also
note that the tubular frame option,
as described previously, may be
mounted on top of the wood deck
as opposed to the design described
previously. There is also the option of mounting the tubular frame
as previously described and using a wood deck with the cutout to support the filter bag.
If you choose a solid decking material, I am sure metal would be the best (most expensive). Wood is
not necessarily the best choice but it will work; I had a piece of 5/8” plywood in the shop just waiting
to be used. You may also have a scrap piece that is suitable. If not, a quick trip to the home center and
you are in business for a reasonable cash investment. This section assumes the use of plywood for the
decking material; the minimum thickness should be 5/8”. The dimensions in this section are a little
vague, as I am not sure what size tank you will be using.
12-4
[ ] Cut the plywood top to the outside (top) dimension of the work tank (21 x 15 5/8” in this case).
The drawing on page 12-4 will give an idea as to how I positioned the various holes in the deck. If
you use a #1 filter bag, the filter bag hole should be
7” in diameter or slightly larger. Locate the XY
table mounting holes, as per your table’s
requirements. The dielectric hose hole may also be
on the opposite side of the decking, i.e., position the
dielectric hose hole as needed to feed fluid to the
work tank.
[ ] Cut four 1/2” square wooden pieces at least 6”
long and attach to the bottom of this plywood deck.
These strips act as “flanges” extending down into
the tank. When making these wooden flanges,
ensure they are a (very) snug fit to keep the XY
table and work tank from moving around.
[ ] Locate the XY table atop the deck and mark the
location for the mounting holes for the table.
The filter system may be spin on type or bag or a combination of both. There is a potential interface
with the wood deck and the spin-on filter mounts if the filter mount is too high.
If a bag is used, a large hole needs to be cut in the deck to insert the bag through. See page 10-3 for
filter options.
The vertical back slide rail (which carries the head) mounts to the wood deck behind the XY table.
[ ] Fab. from 1 1/2” sq. tubing for it as per the drawings below.
[ ] Place the back slide rail, with a note card on each side, between the two 5” uprights and bolt all
three pieces together and then weld in place on the 1.500 base tube. After welding, remove the note
cards. The cards provide a little clearance for alignment.
[ ] Bolt the base to the wooden deck
[ ] When finished, the wood should be treated with a good coat of urethane or other sealer to seal
against moisture.
12-5
Final Assembly of the Slide and WAYS
[ ] Mount the complete assembly on the base between the two
5” uprights using two 1/4-20 x 5 inches-long bolts.
[ ] Mount the CRS WAYS using 1/4-20 flat head bolts (see
photo at left).
[ ] Position the box slide on the lightly oiled Ways and adjust
the gib screws for a smooth-sliding fit and zero play left and
right .
[ ] There are seven dowel pins holes on the WAYS. By placing
a precision dowel pin (.1875) in a selected hole (see the 4th hole
from the bottom) it permits the slide to go only that far on the
back slide WAYS. The slide may be secured in place, with the
middle Socket Head Cap Screw SHCS lightly tightened.
In operation, the head slides down the ways until it contacts the precision dowel pin. At that position,
the middle gib screw is tightened, locking it in place. By loosening the locking gib and screw, the
slide (and servo head) can be quickly raised and clamped at any location along the WAYS for parts
inspection or removal. The head is designed such that the dowel pin spacing is 1” and the ram travel
is slightly over one inch (so the ram travel over laps the pin spacing on the WAYS).
Note: The head can be removed from the back slide ways and then repositioned and you can rest
assured that the electrode is in the exact position (assuming the servo motor was not activated during
the removal)!!! With the indexing chuck system, it is possible to remove the electrode while the head
is off the back slide, and then reinstall the electrode and then place the head back on the ways and
know that it is properly located!!! This is a very flexible, low-cost, easy-to-build system with a high
degree of repeatability.
12-6
Chapter 13
Tank Assembly
Dielectric is pumped from the storage tank on the bottom to the work tank above
and then drains by gravity back to the storage tank. The fluid exits the work tank
via a 1 1/8” thru-hull connector (see the photo at right and also the photo at the
bottom of this page). These connectors may be purchased from a local marine
supplier or purchased from numerous retailers online; one such retailer is:
http://www.jehmco.com/html/bulkhead_fittings.html
[ ] The drain is located in the offset that is the back right hand corner of the work tank. To locate the
position of the drain hole, position the thru-hole connector inside the tank and mark the location.
Don’t get it too close to the sides; there is a radius in the tank and the connector will not seal if it is
too close to the curve. The approximate location of the hole is 1 1/4” from
the back side and 1 5/8” from the offset side. See the photo left.
[ ] Drill a hole for the thru-hull connector. If you have a hole saw, it
makes life really simple; otherwise, drill with a large bit and use a rattail
file to file to the required size (see photo below of a finished drain).
[ ] Install the thru-hull connector.
Work Tank Support and Mounting Feet
In some cases (depending on the selected tank and XY table), the work tank may overhang, or be
longer than the XY table by a distance that is more than desirable. If such is the case, the work tank
needs to be supported on the bottom. Materials that may be considered are wood (including particle
board, as well as Medium- [and High-] Density Fiberboard [MDF & HDF]), as well as myriad
plastics, steel, and aluminum. Metal is the best choice but, in most cases, is the most expensive and or
most difficult to work with given the size required.
I suggest using 5/8 to 3/4”-thick MDF. This material is dimensionally quite accurate and stable, and
easy to work using woodworking tools. It is readily available in most home improvement centers.
Home Depot and Lowe’s, in my area, sell MDF in 2’X4’ and 4’X4’ sizes, which is more than enough
for our needs. I do not recommend using MDF for spanning spaces more than a few inches
unsupported. Over a period of time, the MDF has a tendency to sag if not properly supported; hence,
I did not suggest its use to cover the storage tank. In the application of the work tank, the amount of
unsupported area is usually minimal and the MDF or HDF is more than
adequate. Attach the support board to the table with “T” nuts and
flathead (countersunk) screws.
Note: In some XY table work tank , combinations such as those when
using the specified (US Plastic # 51081) fiberglass tank on the Enco
(201-2536) XY table, there may be no need for the support/mounting
board described above. The work tank and the XY table are almost
made for each other. All that’s needed is to glue “Feet” to the work
tank end( see photo at left on the following page).
13-1
[ ] Cut the MDF board the width of the work tank; the
length should be the tank length plus 3”.
[ ] Drill and countersink holes in the board to accept flathead mounting screws, which are used to attach it to the
table. The hole spacing should match the “T” slots on
the purchased XY table
Note: It’s necessary to cut a clearance hole in the MDF
board for a work tank drain tube .
[ ] The feet are cut from a 1’ x 1/8” piece of angle iron; cut two lengths as per your work tank
requirements. It may be necessary to alter the end of one of the feet to match the end of the tank that
has offsets (see photo above).
[ ] Drill two holes in each foot to match the “T” slots on your table (or holes drilled in the board).
[ ] If used, secure the MDF board to table with the T-nuts and flathead screws.
[ ] If the board is used, position the work tank on top of the board and center it, leaving equal
overhang distance on each end. The increased length of the board (over the tank length) will provide
an area to bolt or screw the feet on the tank to the board.
[ ] Mark the area on the tank ends where the feet are to be glued, rough the marked area with
sandpaper to remove the shiny fiberglass finish, and then degrease with a degreasing solvent..
[ ] Degrease the angle iron feet.
[ ] Mix a two part-epoxy, such as JB Weld (commonly
found in hardware stores and auto parts houses in the
U.S., http://jbweld.net/index.php and apply it to both
the tank and angle iron feet. Position the feet at the
desired location on the tank ends and permit it to fully
cure.
For gluing in the correct location, I actually positioned
the tank and feet on the XY table and used “C” clamps to temporarily hold the feet in position until
the glue cured. I placed masking tape on the table to keep the glue from sticking to the table at the
glue location. Note: If you are using the suggested fiberglass tanks, they have slightly tapered sides.
This leaves a gap near the table that needs to be filled with glue; I used a tooth pick to work the glue
(completely) into this gap. Additional glue may be added once the first “gluing” is cured.
[ ] After all the gaps are filled and the glue is completely cured, securely attach the tank to the table
or the support board.
Work Plate
[ ] Decide on the size work plate you want; this, in part, is
determined by the size of the work tank. The work plate may be
steel or aluminum, with a 3/8’ minimum thickness. If steel is
the choice, a piece of Cold Rolled will ensure a good flat
surface without machine work. If you are using the 51081 tank,
a 6”x12 piece would work. I used a 7 1/2 x 12” piece of
aluminum because I happened to have a piece available.
13-2
[ ] I drilled and tapped a series of 1/2-13 holes (on 2” centers) in the work plate. I chose the 1/2-13
size, as it was compatible with existing tooling in the shop. Choose a size that is compatible with
your tooling. The holes provide a means of securing a vise or other work-holding devices in the tank.
[ ] Completely deburr all holes and edges, particularly on the surface of the plate that will contact the
tank’s bottom. When glued in place, the plate must be as parallel to the tank bottom as possible in
order to accomplish accurate EDM work.
[] Mark where the plate is to be positioned in the tank.
[ ] Rough the bottom of the tank and the work plate with coarse
sandpaper (see photo at right).
[ ] Make sure all traces of dust are removed, and then wipe both the
tank bottom and the plate with a solvent such as acetone, to remove
any oil/grease.
[ ] Place the fiberglass tank on the XY table in the position it will be when actually doing an EDM
burn. Do not bolt the feet tightly in place at this time.
I used the two-part epoxy JB Weld; this epoxy has a good resistance to the hydrocarbons found in the
commercially available dielectric fluids. If you use kerosene as a dielectric, it will work but the life
expectancy of the glue is significantly reduced (see chapter 10 for more information on dielectric
selection). No matter what glue is used, make sure it is
resistant to solvents. The photo (left) is the plate with the glue
mixed and ready to be placed in the work tank
[ ] Apply the mixed-epoxy glue on the backside of the work
plate. I ran two parallel beads of glue and actually mixed the
glue on the plate. Leave a glue-free space around each
threaded hole in the plate (as seen in the photo at left); this will
minimize the amount of glue that “squishes” into the holes
when the plate is set in the tank.
[ ] Position the plate in the work tank and work it around the tank to uniformly smear the glue
between the tank and the plate.
[ ] Once in position, use a “Q” tip to remove any glue in the threaded holes.
[ ] Apply a heavy weight (at least 50 pounds) to the plate to ensure it sits flat on the tank bottom.
Full-bond strength of this particular glue occurs in 24 hours.
Plumbing the Tanks
The description that follows assumes you are using a submergible pump; you may or may not be
using spin-on filters. If you are not using spin-on filters, ignore the instructions pertaining to them.
My experience indicates that PolyVinyl Chloride PVC is not attacked by any of the dielectrics,
including kerosene. All of the plastic parts described in this section, except for the bilge hose, were
purchased at the local Lowe’s home center http://www.lowes.com
13-3
[ ] Do a mockup of the plumbing in the storage
tank. Position the dual- (or single-) spin-on
filters and the pump along the left-hand wall of
the storage tank. Using the suggested fiberglass
tank, there is just enough room to install both on
one side of the tank, if you set the pump at an
angle in the corner (see photo at right).
The filter mount should have the words “IN”
and “OUT” labeled on them. The inlet on my
filter was on the side closest to the tank wall,
which was very convenient in plumbing with
this layout. The outlet of the pump must feed
into the inlet on the filter mount. I used
common white 1/2” PVC tubing, which is available at virtually all home center/hardware stores. I did
not glue the PVC, even in the finished project. The pressure developed by the submergible pump is
not great enough to create a leak if assembled tightly together (if it were to leak, the leak would be
captured in the tanks with this design). By not gluing, it is easy to take apart and replumb if you
change designs in the future.
[ ] Thread two plugs into the filter mount outlets that are not
used in plumbing the tank; this effectively seals them off.
[ ] Once the mock-up is finished, mark and drill a hole(s) to
attach the filter mount to the tank wall. Keep the mounting hole
as high as possible without letting the top of the filter mount
rise higher than the top of the tank. This is particularly
important if you have a solid deck over the storage tank; the
high mounting of the filters could interfere with the decking. If
the tubular frame is used, the filter mount being a little high
will not cause a problem. The filter mount needs to be high
enough to permit the easy removal of filters.
[ ] There is not a lot of space between the filters and the tank bottom. It may be necessary to shorten
the length of the nipples that the filters thread on to in order to facilitate easy filter removal.
[ ] Thread the fitting into the filter mounts (usually 1/2” male NPT adapter) with a 1/2” PVC slip fit
on the opposite end. The female version of the adapter threads on the pump.
[ ] Cut the required lengths of PVC pipe to make the connections from the pump to the filters.
[ ] The discharge line of the filters should just clear the top of the pump and make a 90-degree turn
and comes about 1/2 to 3/4 of the way down the tank’s back side, where it end with a 1/2” male PVC
adapter fitting (see photo below) just under the tubular frame member (see photo above).
The flexible line that goes from the storage tank to the work tank is a 2-foot length
of 3/4” (slinky type) bilge hose. The bilge hose’s inside diameter is such that it will
fit tightly over the 1/2” NPT (male) threads (see photo at left). Twist (i.e., thread)
the 1/2” male adapter fitting in to the bilge hose. The connection is tight enough
that no clamp is needed once they are together.
13-4
Note: I found one type of bilge hose (white in color) that was attacked by the dielectric and started to
leak. All of the black bilge hose used has not been affected by any dielectric, including kerosene.
[ ] The opposite end of the bilge hose needs a 1/2” male PVC
adapter screwed into as well, i.e., the bilge hose has a PVC
adapter in each end.
The flexible bilge hose comes across the top right-hand side of
the work tank and connects with a length of PVC across the
back of the work tank see photo right.
[ ] The PVC is held in place with a
wire that goes just to the left of the
1/2” coupling. Drill two 1/16” holes
in the tank; the wire loops around
the PVC and attaches to the tank
(see photo at left).
[ ] On the left hand side of the work tank a 90-degree 1/2” slip
fit to a 1/2” male NPT fitting is placed on the PVC pipe. See
the photo below right. This end of the PVC is also secured to the tank in a manner similar to the
previous mounting method.
Dielectric Flow Control
The EDM process requires a flow of dielectric for flushing, and a flow of dielectric to circulate
continuously throughout the tank for filtering. This requires a splitter with the ability to be able to
accurately control both flows of fluid. Ball valves are the best valve for controlling flow but they can
get expensive. A low-cost solution is an inexpensive garden hose “Y” splitter.
An adapter is needed to go from the 1/2 NPT thread to the garden hose tread. All
hardware stores and home centers sell such adapters, a Watts A-668 GH4 hose
adapter is the one I used (see photo at left).
[ ] Assemble the splitter assembly using the PVC
elbow, the garden hose adapter, and the “Y” splitter as
shown in the photo at right.
The valve that faces down controls the high-volume
flow through the work tank. To reduce splashing of
fluid when first filling the work tank, I picked up a 5/83/4” female hose repair end. This end is threaded on
the “Y” splitter and a length of nylon webbing (from a
re-sharpened end mill) slides over the end. The horizontal leg of the “Y” valve has a 1/4” faucet
adapter sold by Mister Landscaper. It is described as a 1/4”’ Faucet Adapter 3/4”’FHTx1/4” Barbed
Outlet http://www.misterlandscaper.com.
Most home centers sell these, or similar items; they are used to
reduce a water faucet down to 1/4” for drip irrigation systems. If
you need to split the dielectric flow and/or need additional control
over the flush flow, consider the little 1/4” ball valves and “Ts”
shown in these photos; they are also available from Mister Landscaper.
13-5
Unfortunately, the 1/4” hose (or at least the hose I bought) is attacked by the dielectric; it becomes
hard in less than 12 hours, and the “Ts” and miniature flow valves have not shown deterioration
problems. An acceptable flush hose is an auto vacuum or windshield-washer hose which may be
purchased from any auto part house. This hose is also attacked by most dielectrics, but the
deterioration process takes weeks and it becomes very soft, not stiff. Only about 12-24” of this hose
is needed, the parts stores usually have dropped from rolls lying on the floor that are too short for
most retail sales; I have had them just give me the drops as I need such a small quantity of hose.
There are other flush-line and control options available, as suggested below.
Other Valve Options
Other options for controlling flow also exist in the garden center. There is the single shut-off valve
and the multi-valves built into one manifold (see photos at left). Also available nowadays are myriad
of quick disconnect hose ends to consider, as well as thread on caps. Sometimes the
plastic/rubber O rings used are not compatible and you may
need to replace them with compatible O rings. The ones in
the photos (left) are plastic and manufactured by the
Gilmour Company. There are many other manufacturers of
such valves, some made in brass and some in other metals.
Browse the garden area and let your imagination run wild.
There are several versions of the garden-hose-to-pipe-thread adapters available; they come in either
male or female, or a combination of both.
Flexible Flushing Line
In addition to the garden hose Y
valve, you may want to consider
adding another flushing line as
shown in the photo at right. This
type of flush line made by Locline http://www.loc-line.com.
It is particularly useful when
doing jet flushing (see page
16-2). You can bend the Loc-line line to most any location you want and it will stay in that position.
There are other manufacturers of similar products; they are sold under the names Snap Flow, SnapLoc, etc. These products are available with various valves, nozzles, connectors, etc. I like the product
in most applications, but the valves tend to stick and are hard to adjust for precise dielectric flow so I
don’t use them often where the flush flow needs to be accurately controlled. See page 16-2 for
information on connecting to a flexible rubber flush line.
[ ] If you elect to use a Loc-Line-type product purchase at least one male 1/8” NPT valve, a round
1/8” nozzle, and a foot of the Loc-Line. Many industrial suppliers sell kits with these basic items.
Manhattan Supply Company (MSC) http://www1.mscdirect.com is who I purchased a kit from.
[ ] In the PVC line, which runs along the back side of the work tank, place PVC couplings where you
want to install the Loc-Line valves (see photo above right). The coupling increases the thickness of
13-6
the PVC line in that area, giving a little more material to thread into. Drill and tap these PVC
couplings for 1/8” NPT and thread in place.
Variable-Speed Motor
Valves are not the only way to control flow of the dielectric. A variable
speed pump is very convenient as well. A low-cost way of controlling
the submerged pump is with a light dimmer switch. These switches are
available at home supply centers and hardware stores. Wire as per the
instructions; only replace the light with the pump in the storage tank.
Mount the dimmer switch in a plastic handy box used for home light
switches; the light switch covers finish off the controller nicely. Mount
the motor speed control to the tank using 6-32 screws (see photo at left).
Work Tank Dielectric Level Control
As the photo on page 13-6 (bottom right) shows, the dielectric level control is determined by a length
of pipe inserted into the opening of the thru-hull connector. The length of the pipe determines the
level of dielectric in the work tank. Several tubes may be cut to different lengths to provide for
various levels of dielectric need in the work tank.
Pre-filter (Bag Filter)
To reduce spin on filter load, a gravity flow pre-filter is
beneficial; see the bag filters on page 10-4. Overflow from the
work tank flows by gravity into the filter bag, which is
suspended in the storage tank via a 3/16” rod running across
the storage tank (see photo at left).
[ ] Drill holes as required along the storage tank-recessed lip to
support the pre-filter. The #1 filter bags are 7” in diameter.
[ ] Thread the support rod through the bag handles.
The filter bag wants to collapse and lie flat on the tank bottom.
Therefore, it’s necessary to insert a bag expander inside. An expander is made by rolling an 8” length
of (needlepoint) plastic canvas from Hobby Lobby into a 5” diameter.
Before, rolling I weaved a 1/16” aluminum welding rod through the mesh
to add some rigidity once rolled.
[ ] Cut a piece of piece of plastic canvas 8”x16 and weave 1/16’ welding
rod through the middle.
[ ] Roll to form a 5” diameter tube, and turn the end of welding rod inside
the roll to keep it from snagging the filter bag. Place the plastic canvas
inside the filter bag and place the bag in the tank as per the above photo.
13-7
Chapter 14
The Spark
What Happens in the Gap?
To see what is electrically happening in the gap, an oscilloscope is needed; the drawings below are
taken from actual waveforms observed on an oscilloscope. Gap waveforms show what is happening,
and the order in which they happen. The 555 timer on the control board controls the ON/OFF time
sequence; this pulse train is shown across the bottom of the drawing below. An oscilloscope
waveform and descriptions of how the control circuit works are found on page 4-5.
The top section of the drawing shows the voltage waveform
across the gap. When the timer turns (the Hexfets) ON, the
voltage goes from zero to the open circuit voltage which, for our
machine, is about 110VDC. After a short ionization period, the
voltage drops to the working voltage which is normally 2050VDC.
At this point, current starts to flow through the gap, which is
represented by the current (middle) waveform. Current
continues to flow until the timer turns OFF. All of the cutting
(work) is accomplished during the ON time of the cycle. The
amount of metal removed is directly proportional to the amount
of current that flows through the gap during the ON time. The
more frequent and the longer the current flows, the more
workpiece material is removed.
After a short OFF time in which the gap has a chance to deionize (becomes an insulator), the timer
turns the Hexfets back ON, unfortunately there is delayed ionization this time; it started to ionize but
then reverted back to the open-circuit voltage. Finally ionization was successful, but most of the ON
time has passed. This is evident in the current section of the drawing; it shows a brief period of
current flow, resulting in only a small amount of
metal being removed. The timer turns OFF for the
specified time period, then it turns back ON. This
time gap conditions were such that no current
flowed during this ON time, though it approached
ionization about halfway through the ON time
cycle. The photo at left is of an actual
oscilloscope waveform recorded during a burn.
Across the bottom is a series of “square” waves.”
This is the wave output from the (555) timer.
Follow the square waves (left to right). When the
line drops from a high (near the number 1) to a
14-1
low (near the number 2) it means the GATE (which is the first pin on the Hexfet when looked at from
left to right) of the Hexfet was turned OFF.
When the signal rises the Hexfet turns ON. This is the actual electrical ON/OFF signal being applied
to the gates of the Hexfets. In this example, the Hexfets are turned ON for about a 50% duty cycle,
i.e., the ON and OFF signal are almost equal in length, unlike the previous drawing which is
asymmetrical. The upper “jagged” looking wave form above the square wave is the voltage signal
seen across the gap at the same exact time, corresponding to the timer square wave. This
may look confusing but understand that when the Hexfet gates turn ON (and if the
conditions in the gap are good), current starts to flow through the gap; thus, the voltage
across the gap drops. It is what we want to happen but it does now always happen.
See photo at left, starting on the left side with the gate signal at the first full ON time, as
soon as the timer turns on the gap voltage drops meaning that the gap ionized and started
conducting (current is flowing in the gap ). When the timer turns OFF, the
gap voltage went high with a little spike and then flattened out.
See photo at right: As soon as the gate turns back ON, we have a good example
of the gap ionizing and then conducting the voltage drops sharply and at the very bottom a
little choppy line slopes to the right, indicating the burn taking place. The gates then are
turned OFF. At the next ON time, the gap hesitates (see photo at right); it takes a while to
ionize the gap and then conducts (the narrow vertical drop), but there is not much time
remaining in the ON time sequence then and the gates are turned OFF. This is
similar to the 2nd pulse in our previous drawing.
In the next ON cycle (see photo at left), there is no flow of current in the gap. We know that
because the voltage stays high during that time, even though the Hexfets were turned ON. The
gap is not in a condition to ionize; no current flowed and no metal removal.
Then there is a good run of five almost picture-perfect examples of the gap
ionizing and conducting (as seen in the photo at right); this is what you want to see all
the time.
Next there is another example of a few missed pulses (see photo at left).
As you move to the right, there are examples of when the Hexfets were turned
ON and the gap started to ionize (the voltage dropped partway) but then
stopped, in a fashion similar to the 3rd pulse in the previous drawing. Looking
at these wave forms it is very revealing as to what is happening in the gap. It can really give
you a real understanding of the EDM process.
If most of the electrode wear takes place in the 1st microsecond (see “Electrode” 15-1) and
in at least one of the time periods shown above (in the oscilloscope pictures), the gap was conductive
for only about 10us then the wear for that time period was quite high. In other ON times the gap was
conducted for the full time period, so for that ON time cycle, the wear would be significantly
reduced.
14-2
In high-dollar commercial EDMs with much more complex control systems than those used here,
there are ways to greatly reduce the ionization problem and achieve less than 1% wear on the
electrodes. During the design phase of this machine, it became necessary to make a decision on how
complex to make the control circuit. I opted to design a simple, yet robust, control circuit as possible,
hence the reason for the current design.
In the current design, there is a free-running oscillator (555) that turns the Hexfet switches ON; if
conditions in the gap are good, the channel ionizes right away for that pulse you have very low wear
(maybe 1%). If the conditions in the gap are poor, it may take a long time (if at all), before the gap
ionizes. Given these conditions, the electrode wear for that particular pulse is significantly increased.
There are occasions, as you have seen, when the Hexfets turns ON and no ionization takes place. In
this condition true zero wear is present, but also no metal was removed from the workpiece. Even
though the desired goal of 1% or less electrode wear was not achieved (all the time) with this
particular design pulse machine. it is a quantum leap over an RC (Resistive Capacitive) machines.
There are hundreds, maybe even thousands, of commercial EDMs worldwide in operation today that
use power supplies and control circuits that are no more advanced than what you find in this machine.
When you compare the dollars invested in this machine’s power supply versus those invested in the
commercial machines, it is astounding.
Duty Cycle
The duty cycle is a percentage of the ON time relative to the total cycle time. The duty cycle is
calculated by dividing the ON time by the total cycle time (ON time + OFF time). The result is
multiplied by 100 to obtain the duty cycle, which is stated in the percentage.
ON-time
Duty cycle (%) = ------------ X 100
Total Cycle Time
The frequency is the number of cycle (ON time + OFF time) produced in one second. Frequency (in
kilohertz) is calculated by dividing 1000 by the total cycle time (ON time + OFF time) in
microseconds. When the ON time is short, only small amounts of metal are removed per discharge;
thus, the finish is finer but the electrode wear is increased. The higher the frequency, the finer the
workpiece finish. Conversely low frequencies are used for roughing burns. This gives quick metal
removal and leaves a rough finish and very low electrode wear, as has been discussed.
14-3
Frequency of the Spark
The graph at left shows the frequency in
kilohertz along the left side and the Total Cycle
time in microseconds across the bottom. As can
be seen, the frequency of the longest ON and
OFF times of our EDM is quite low.
When operating the EDM in certain ON/OFF
time combinations, expect to hear sounds other
than the “frying bacon” that you may be
accustomed to if you built the RC machine
described in The EDM How-To Book.
Most humans’ hearing range is from a couple of
hundred Hz to around 18-20KHZ. A large part
of the EDM’s operating range is in the human
audio range. This is the reason that frequently,
when operating the EDM, you will hear the
operating frequency tone as you manipulate the
ON and OFF time controls.
The EDM process may also be heard on the AM
radio band by tuning to a non-used portion of the band. I have listened to the EDM process in an
adjoining room and was able to determine if the EDM needed my attention by the sound on the radio.
Some advanced EDM control systems use this RF emission to precisely determine what needs to be
corrected for the best burn.
Metal Removal Rates
The rate at which metal is removed is determined by several factors but, generally speaking, long ON
time and short OFF time results in the highest removal rate, i.e., more current is flowing. Testing
from POCO , a manufacturer of graphite electrode materials, indicates that, frequently, once the ON
time goes beyond 200us the metal removal rate may actually decrease! This is not true of all
electrode workpiece combinations, but it is generally true of the graphite and steel combination.
Other factors which affect removal rates are flushing (gap contamination), the type and polarity of
electrode material, and the conductivity of the workpiece. By using negative polarity, metal removal
rates may be as much as twice that of positive polarity. To accomplish this higher metal removal rate,
the electrode will be in the 15-20% (and greater) wear range regardless of the control settings. As
mentioned previously, materials such as nickel, chrome, molybdenum, and carbide tend to slow down
metal removal rates.
14-4
DC Arcing
What is the difference between a spark and a DC arc? When the EDM system is operating as
designed, sparking occurs at the point of least resistance and current flows. When a (destructive) DC
arc is established, current continues to flow at only ONE location instead of various locations on the
workpiece. Oftentimes this is the result of poor flushing, especially when using graphite electrodes in
particular. A low-resistance debris bridge is built between the electrode and the workpiece and
current flows only at this location each time the power supply switches ON.
In some cases, it continues to flow even when the power supply is switched OFF. This is possible
because there is stored energy (capacitive and inductive) in the leads, running from the power supply
to the workpiece. When the power supply switches OFF, a small current continues to flow. If the
Hexfet is switched back on before this energy is dissipated, then the gap never really deionizes;
hence, there is an electrical bridge built at that location, and thus the next flow of current remains in
the same location. If this continues, cycle after cycle, a DC arc is established. A DC arc generates the
maximum amount of thermal energy, quickly heating and breaking down the dielectric; if permitted
to go unchecked, the dielectric will heat up and significantly increase the possibility of a fire.
This condition is a difficult one for the control system to detect and correct. Even more advanced
control systems than are used in this machine have problems detecting and correcting a DC arc
condition. To the control system, the input it is receiving during a DC arc appears very similar to the
signal it receives during good sparking conditions. Under certain DC arcing conditions, the control
system senses a problem and tries to take corrective action, which results in erratic servo operation.
As an EDM operator, you will notice an auditable change in the sound of the arc from the normal
operation. Instead of the usual bluish sparking, you will usually see a yellowish glow in the gap area.
So what should you do should a DC arc occur? If it is a high-power burn and the arc condition is
completely out of control, turn the power supply OFF and retract the electrode. If it is in the
beginning stages immediately increase the OFF time; this permits the gap to fully deionize and many
times will take care of the problem. Other corrective actions that may be taken are:
1. Change or alter the flushing.
2. Switch to the dither mode (i.e., the electrode is retracted a preset interval for a given time).
This permits the flushing to “wash away” debris.
3. Increase or decrease the dielectric flow to the gap. The condition may be caused by too much
flow or too little flow through the gap. Two to three PSI of flush pressure is a good rule of
thumb.
By retracting the electrode and draining the dielectric tank, you will see a buildup of graphite and/or
dark scorched or burned area on the workpiece. If it is within the workpiece tolerance, it may be
saved by removing the graphite buildup or machining the burned area away. Make sure there is no
residue left on the workpiece to encourage a debris bridge to form once you start the burn again. The
electrode may require machining to bring it back in an acceptable dimension. The lesson to be
learned in most cases is the area that was damaged was not getting proper flushing. Maybe another
flush hole is needed to increase the flushing of the EDM cavity.
14-5
Chapter 15
Electrodes
When EDMing, unfortunately there is almost always wear, or erosion, that takes place on the
electrode as the cutting actions takes place. This is not unique to the EDM process; when cutting with
conventional machine tools, the tool becomes “dull” or wears. Through research, it has been learned
that, under certain conditions, the wear can be greatly reduced. In the EDM industry, this is referred
to as “zero wear” or, more correctly for us, the “lower wear” mode. With very sophisticated, stateof-the-art EDM equipment it is near 1% and less; with the relatively simple pulse supply described in
this book, we can reduce electrode wear from what is experienced with an RC-type power supply, but
it is not possible to reach ”zero wear.”
Research shows that more than 90% of the electrode wear occurs within the first one or two
microseconds after the ionization channel is established. The ON/OFF times used for most EDM
work are usually expressed in microseconds (1 micro second = 1 millionth of a second).
Oscilloscope wave forms across the gap may be seen in Chapter 14-1. The amount of wear during the
wear period increases in relation to the applied amperage. If the major wear period is 1 microsecond
(one millionth of a second) and the ON time is 10 microseconds, the wear is 10%. By increasing the
ON time to 100 microseconds, the wear is reduced to 1%.
Factors Contributing to Reduced Electrode Wear
(1.) A long ON time, though once the ON times pass 200us metal removal may decrease.
(2.) A low flow of dielectric to the gap. Defining “low flow” is difficult without some way to
measure it. It is flow that is just enough to keep a good stable burn and no more.
(3.) The electrode material must be either copper or graphite.
(4.) The workpiece must be steel.
(5.) The electrode must be connected to the positive side of the power supply. This may be verified
by using a volt/ohm meter. Connect the red lead to the electrode and the black lead to the workpiece.
With an analog meter, the needle kicks “upscale” or to the right indicating the electrode is of positive
polarity. If it deflects to the left, the electrode is of negative polarity. With most digital meters, if the
polarity is positive; there will be no +/- indicators indicated on the display.
When operating in the low-wear zone, you might notice that the graphite electrode takes on a light
grayish color and the copper electrodes will have a blackish appearance; this is good. Under some
conditions, the material in the gap is “plated” back on the electrode, meaning it is possible to have the
electrode actually “grow” which is not a good condition. This plating condition is a combination of
electrode material, workpiece material and the dielectric that is used.
When you find a machine setting that gives good, low-wear results, note the setting, including
electrode material, workpiece material, dielectric used, ON time setting, power setting, etc. Keep a
log book and refer back to it when needed. Commercial manufactures of EDMs spend hundreds, even
thousands, of hours and lots of $$$ collecting this data for users for their high-dollar machines. I have
not collected and compiled much hard data on my machine at this time.
15-1
The shortest ON time possible (the way I have the power supply configured) is limited to about 25us.
When the ON and OFF times become shorter, it becomes more and more difficult for a simple
control circuit, such as we are working with, to control the electrode for a stable burn. I should also
stress that with very short ON times flushing becomes critical in getting good results; this is
especially true when using graphite electrodes.
To illustrate the wear that may be expected on commercial machine, the following is a compilation of
data taken from various charts developed for commercial EDMs and found in literature published by
the EDM Today magazine http://www.edmtodaymagazine.com
6 amps, using a positive graphite electrode, steel workpiece:
50 micro seconds on 1% wear
25 micro seconds on 2% wear
12 micro seconds on 6% wear
4 micro seconds on 18% wear
As you can see, with the shorter ON times (which we cannot achieve), electrode wear goes up.
3 amps, using a positive + graphite electrode, into a steel workpiece:
25 micro seconds on 1% wear
6 micro second on
12% wear
With a negative polarity, minimum electrode wear is 15% and up (as high as 30% wear), with an
average being about 20% but the metal removal rate is double that for a positive electrode.
Electrode Polarity
Depending upon the application, EDM machines use both positive and negative polarity electrodes.
Whenever polarity is mentioned, it is ALWAYS the polarity of the electrode that is referred to.
Polarity affects speed, finish, wear, and stability. In virtually all cases, positive polarity will machine
slower than negative polarity. Despite this disadvantage, when using a pulse machine, positive
polarity allows you to have more control over the electrode wear. As previously noted, positive
polarity is required to achieve low electrode wear. When a significant amount of time has been
invested in the manufacturing of the electrode, a bit of instability and reduced cutting speed can be
tolerated, to conserve the electrode.
Some electrode workpiece polarity combinations work better than others, though there are some
materials that WILL NOT work with a positive electrode. When machining carbide, it must be done
using negative polarity and very short ON times. Copper requires a positive polarity electrode. Below
are a few workpiece and electrode polarity combinations:
Positive Electrode --- Graphite to steel, low wear, general use
Negative Electrode ---Graphite to steel, fast cut, 20% wear
Negative Electrode -- Graphite to copper
Positive Electrode --- Copper to steel
Negative Electrode ---Copper tungsten to carbide
15-2
Volumetric Wear
Volumetric wear is the volume of electrode material lost in relation to the material removed from the
workpiece; it is a combination of end wear and corner wear (described below). Frequently these are
the “magic numbers” we often seen quoted by manufactures when they say “low wear” or “no wear,”
or whatever the phrase or number is that they are excited about in order to sell a machine.
End Wear
End wear is the amount of electrode material lost from the end of the electrode (i.e., you start with an
electrode that is 3” long when measured. After completing the burn it is 2.997 long; .003 was lost in
end wear). This is often expressed as a percentage.
Corner Wear
Corner wear is a little more difficult to understand; imagine a 1/2” square electrode
(90-degree external corners) burning through a 3/4” steel plate. In a perfect world, the
electrode would be unaffected; all the edges of the electrode would be nice and square.
The electrodes live a hard life; if examined after completing the burn notice that the
end of the electrode is (hopefully) only slightly tapered on the corners. Then the
remaining portion of the electrode appears nice and square and unaffected by the burn.
The area that has the slight taper on the end, and on the corners in particular, is what is referred to as
corner wear. The wear on the corners is almost always greater than what is lost in end wear. The
sharper the corner, the greater the corner wear.
Electrode Wear Test
To give you an idea of the kind of wear you might experience with this
machine, I did a few short test burns using graphite and copper
electrodes which were .125 thick with into a .500 wide piece of tool
steel with the following results (see the copper electrode test piece in the
photo at left). All burns were for 5 minutes using the shortest OFF time
possible to ensure a stable burn. The ON times were 400, 200, 100, 50
& 25us. The current was held at 6 amps of average current (peak current is higher) and the electrode
was of positive polarity; the dielectric, Rustlick EDM 30. I don’t have a means for measuring surface
roughness or corner wear; thus, the following statements are subjective. Both the copper and graphite
produced similar finishes on the steel for all ON times, with the exception of the 25us burns. The
copper, as well as graphite, finishes for the 25us ON time gave the appearance of being much
smoother than any of the other time burns.
With the copper electrode, the maximum metal removal occurred at 200us and with zero end wear;
the corners remained square and appeared to have little if any wear. The second most metal removed
(only slightly less) with same wear as noted for 200us. The third most metal was removed at 50us,
with slight rounding of edges and corners and .001 end wear. The third most metal removed was at
15-3
400us with no end wear. The least metal was removed at 25us and the electrode end wear was
significant (.009, with pronounced edge and corner wear), but the surface finish was the smoothest.
With graphite, the maximum metal removal occurred at 100us, with zero end wear and the corners
remained square. The second most metal removal took place at 200us and 50us (identical amounts
removed); 200us had zero end wear and no discernible edge and corner wear, at 50us there was
slight rounding of the corners and .0015 end wear.
The next most metal removed was at 25us., electrode end wear was .004, with some rounding of the
corners and edges but not nearly as obvious as with the copper electrode.
The least metal removal was at 400us, with no end edge or corner wear.
Overcut
Overcut is the gap distance between the electrode and the workpiece. Because the EDM process
always produces a cavity larger than the electrode, the electrode must be slightly smaller than the
specified cavity dimension. This size difference is called overcut. An overcut from a sharp corner of
an electrode produces a corner radius instead of a sharp corner in the workpiece; this is because all
the spark lengths are the same for a given voltage. In most cases, I don’t think the home builder will
find this a serious problem, but in an industry where tolerances may be tight, it can become a
problem.
Overcut can vary from a few ten thousandths to a few mils, depending on the dielectric condition
(mainly determined by the filtering system), spark ON and OFF times, power level (discharge
energy), electrode and workpiece material and polarity, etc. As the gap voltage or power level
increases, so will the overcut. The magic number to know as an operator is the amount of the overcut.
It is beyond the scope of this book to determine these numbers for your applications. The simplest
method is to do a test burn using the same type dielectric, workpiece, voltage, current, ON and OFF
times, etc., as will be used to machine the finish part. Keep logs of these setting, as well as the
overcut data for your machine. Soon you will have accumulated valuable information unique to your
machine for the type of work you most frequently do. Measure the overcut and manufacture your
electrodes based on the data collected. Overcut is fairly predictable and, given the usual
homebuilders’ tolerances, I don’t see the variations as a serious problem, assuming the dielectric
conductivity (i.e. filtering) does not change a large amount.
Recast Layer
In the EDM process, material is removed from the workpiece by thermal action. This results in a
layer of melted and solidified material on all cavity surfaces, known as recast. The recast is brittle and
forms because the metal is heated to the melting point and then is quickly quenched; cracks may
develop in this recast surface. It has been found that the recast layer can produce stress risers from
micro cracks on the work surface; these micro cracks could result in part failure in critical
applications. The recast surface may be in excess of 65 Rockwell (Rc) hardness. The recast is usually
between .00001” and .002” thickness. Higher energy-per-pulse (i.e., coarse burns) will generally
leave a thicker recast layer. If machining of the EDM surface is required by conventional methods, be
aware of the hardened recast surface.
15-4
Electrode Materials
The only physical property that an electrode must have is that it must be electrically conductive, but
there are other desirable qualities to make good electrodes. Electrodes should have a high melting
point (the higher, the better), be ready available and reasonable in cost and, in our case, be machined
using common machining tools and practices.
There is no “perfect” electrode material. The most commonly used electrode materials are graphite
and copper. Metallic electrodes, mainly copper and brass, were the first electrode materials used. In
about 1960, graphite was introduced as an electrode material. In the U.S., the primary electrode
material is graphite; in Europe and Asia, I understand copper, is preferred. The five materials
commonly used in the industry for electrodes, in order of melting points, are (lowest to highest): zinc,
brass, copper, graphite and tungsten.
The EDM described in this book burns nicely with most metallic and graphite EDM electrode
materials. When graphite is used, it should be of the type manufactured specifically for EDM
applications. Carbon welding rods, carbon rods from dry cell batteries as well as motor brush
material, and other forms of carbon and graphite are poor choices; their use will be very
disappointing at best. My recommendation is stay away from any form of graphite not specifically
designed for EDM work. All types of copper that I have used burn VERY well with this machine.
The more pure forms of copper appear to have better resistance to wear.
Metallic Electrodes in General
The main advantage of metallic electrodes is that they have high electrical conductivity and are
mechanically very strong. This is important when EDMing with poor flushing conditions on parts
that require sharp corners or lots of details. The disadvantage of most metal electrodes is that they
have slow cutting speeds and many are metals that are difficult to machine.
Brass
Brass was one of the first EDM electrode materials used; it’s readily available, inexpensive, and easy
to machine. Today it is not frequently used in commercially ram type (sinker) applications, mainly
because of its high wear rate, though there is an applications for brass electrodes in “hole poppers”
(special EDMs for burning small holes). In the hole-popper applications where the high wear rate is
not a serious issue, the low cost of brass, coupled with its stiffness, makes it a desirable material.
Though brass is not use widely used in commercial applications, that is not to say it would not be a
suitable electrode material in the home shop where production usually is not the order of the day. If
low cost and ease of machining is important (and high electrode wear can be tolerated), this may be a
good electrode choice for your application.
By using low power and long ON times, good results may be produced using brass electrodes. A
brass Chuck E. Cheese token was used as the electrode material to produce the image on a piece of
steel as shown on page 1-2. Brass, as well as copper, is more resistant to DC arcing than graphite.
15-5
Copper
When pulse-type EDMs replaced the RC machines, it was learned that, at a certain power supply
setting, very low electrode wear was experienced using copper; this was especially true of pure
copper (Electrolytic copper).
Copper is physically quite strong and is known for working well in poor flush applications because it
resists DC arcing, which can quickly destroy an electrode and workpiece. Another advantage of
copper is it can produce very fine surface finishes (at a reasonable electrode cost). Copper is a very
clean electrode material to work with (when compared to graphite) and perhaps, in many instances, it
is the reason in Europe and Japan that it is preferred over graphite. I really enjoy EDMing with
copper on this machine. Using copper creates less swarf in the dielectric, thus reducing the filtering
load.
With all these advantages, there are also disadvantages; anyone who has machined pure copper
knows that it is gummy and leaves lots of burrs!! These require removal, which can be timeconsuming and a real pain to deal with. To help improve the machining of copper, tellurium is added.
With the addition of 2% tellurium, most, if not all, the problems with machining copper (and if you
have ever machined it you know what I am talking about) go away. But as with most things, there is
a tradeoff; copper burns only about half as fast as graphite, and the addition of tellurium causes
increased electrode wear and a reduced burn rate. As with brass, copper is quite stiff and relatively
low-cost, making it a good candidate in hole-popper applications as well.
I really enjoy burning with copper using a pulse machine, especially when doing detailed work that
requires good surface finishes. As previously mentioned, the resistance to DC arcing is a huge
advantage and the swarf is not nearly as “heavy” as with graphite. Frequently, in work such as token
and jewelry-making, existing copper tokens and jewelry provide ready-made electrodes
(see page 18-1).
Copper/Silver Tungsten
If there is a need to machine carbide, the electrode material of choice, without a doubt, is copper
tungsten; as its name indicates, it is a mixture of copper and tungsten. It has high electrical
conductivity and extremely good wear resistance; thus, it holds up very well where sharp corners are
required. People tell me it is readily machined but I cannot address that issue from experience. I have
only used it as an electrode material to do through holes in parts using round tubular electrodes.
Also in the tungsten family of electrodes is silver tungsten. I have never used this material but I
understand from several sources that it has very low wear and is capable of producing extremely fine
details. It is quite expensive and availability is limited, so it is not an electrode you will find very
often in most EDM shops.
If you have a requirement for doing really small holes, consider tungsten carbide. This material is
quite brittle, so handle with care. The smallest tungsten carbide electrodes I have used were .009 in
diameter; they come in even smaller sizes.
15-6
Graphite Electrodes
Graphite is the preferred electrode material for 85-90% of all sinker EDMs used in the U.S. As a
general rule, graphite machines twice as fast (and with less wear) as copper. With certain grades of
graphite workpiece surface finishes approach that of copper. Graphite has very high melting point
(about 6062F [3350C]), which is one of the qualities which makes it a great candidate for electrode
material. Actually, “melting” is a misnomer: it does not melt; it sublimes (goes from a solid to a gas,
similarly to what happens to dry ice or carbon dioxide). The temperature at which graphite sublimes
is close to the temperature at which tungsten melts.
Graphite is not nearly as strong, hard, or stiff as metallic electrodes. Graphite is made from a powder
and binders pressed together under a thermal process. It is available in different densities which are
determined by the powder particle size used in the manufacturing process. Some powder grain sizes
used for electrode materials are as small as one micron! The human hair (.003) is roughly 100
microns in diameter. The coarse grade of graphite electrode material is composed of graphite
particles about 100 microns in size: the smaller the particle, the higher the density of the electrode.
Electrodes composed of the smallest graphite particle size give the finest detail, lowest electrode wear
and finest surface finish and, of course, they cost the most to purchase.
Perhaps a little understanding as to how the particle size affects graphite electrode performance
would be helpful. When burning, realize that not only is the workpiece being attacked by the sparking
conditions in the gap, but the electrode is also under attack by the sparks. This attacking often breaks
out particles from the graphite electrode (metallic electrodes do not have particles to break away). If
the electrode is composed of large-size particles then, as the particles are removed, the definition of
the electrode (shape) deteriorates rapidly. If the electrode is made up of small-sized particles, then the
electrode will not lose definition as quickly.
There is a relationship between particle size and the level of detail that is attainable using graphite,
i.e., small particle size has a direct relationship on electrode strength, corner wear, and surface finish.
Basically, all graphite raw material is sold based on particle size; of course, the small particle size
comes at an increased cost.
Below is a generalized idea of available grades of graphite electrode materials; the name may differ
depending on the manufacturer. The important thing to remember when buying graphite is the
particle size, not the name.
Sub-micron graphite has particles of less than one micron in size; this is the most expensive. It is
used where extremely fine details are desired and or where very sharp corners are needed.
Premium usually has particle size in the 1- 5-micron range. This is the most commonly used graphite
for high-precision work, such as small molds and dies.
15-7
High-performance graphite particles are in the 10- to 20-micron range. This grade of graphite is
commonly used for large molds and dies. Of course, the cost is going down as the particles get
bigger.
General-purpose graphite has particles of 20 microns and larger. This is a good grade to start with
for the home shop, but it does generate lots of swarf in the dielectric.
There are intermediate grades of graphite, but it is beyond the scope of this text to go into the
specifics. Most manufacturers/companies selling graphite color-code the stock for easy identification.
If you are using graphite as an electrode material, in my opinion, it is a good idea to get markers or
paints of these colors and mark your graphite once the piece is cut.
Graphite has several disadvantages. From an electrode fabrication standpoint it is a very “dirty”
electrode material to work with; just handling it will ensure dirty, blackish hands at best. When
machining graphite an electrically conductive dust is generated, if not controlled; it gets in and on
everything and everyone. This electrically conductive dust is especially hazardous for CNC
equipment and computers. When burning with a graphite electrode and the flushing deteriorates,
graphite will form the dreaded “DC arc” much faster than any of the metal electrodes. Using graphite
significantly increases the workload for the dielectric filtering system.
Historical note: One of the first (if not the first) EDM grade graphites was Gentrode 10, made by the
General Electric Co. This early generation of EDM graphite had a “Grain” pattern resulting from the
manufacturing process, which pressed the graphite in only one direction. Later manufacturing
processes eliminated the “grain” and today we have isotropic (non-directional) graphite.
.
Copper Graphite
As the name indicates, it’s a mixture of graphite and copper. This mixture gives higher electrical
conductivity and physically it is stronger than plain graphite. It machines well and has many of the
attributes of copper in the burning process. Copper graphite is a good choice for machining hightemperature metals like titanium and Inconel. It should be noted that copper graphite does not resist
wear on its corners during a burn as well as the same particle-size straight graphite.
Graphite Machining Tips
Graphite has good machining ability and does not require deburring the way that copper does. It may
be machined with common cutting tools found in small shops. High-speed steel tools can be used, but
expect high cutter wear; carbide is the preferred cutting tool. Keep spindle speeds up; graphite loves
the speed.
Most grades are brittle and machining sharp corners should be done with care. To reduce chipping,
climb the mill and reduce feed rate when about to exit a cut. When turning, roughing feed rates are
.015, and finishing should be around .005. Cut depths of .015-.020 are better than light .005 cuts,
except for when finishing, of course. Use positive rake tools with a nose radius of at least 1/64”.
15-8
As mentioned, the real problem with graphite is that it is “dirty” and “messy” to machine; it does not
make chips, it makes dust. Fortunately, in most home/ hobby shops this is not a serious problem as
we are not making huge electrodes or machining them around the clock. Many machinists like to
machine wet to reduce the dust problem, but you end up with a black, abrasive, slurry mess. Often I
use WD-40 when cutting; it helps to keep the dust down, but it leaves a pasty residue that may be
(frequently) wiped up with a shop towel. If machined dry, use a respirator and a well-filtered vacuum
to collect the dust. Remember that graphite dust is electrically conductive; do not permit the dust to
enter electronics cabinets; i.e., computers, power supplies, etc., as strange things start to happen.
Suppliers of EDM Graphite
The largest manufacturer and supplier of commercial-grade EDM graphite in the U.S. is POCO. A
good general-purpose grade of POCO graphite to start with is sold under the name POCO EDM-2
and EDM-1.
POCO has developed a huge wealth of information in graph form about metal removal rates,
electrode wear, etc. for their particular electrode materials. You will not be able to replicate
completely their results with this machine but you will be close; do test burns to confirm.
The Poco website is WWW.Poco.com.
Other suppliers of EDM consumables mostly sell POCO graphite, but some sell less expensive EDMgrade graphite. An industrial supplier that carries some EDM electrode material and EDM- related
items is Manhattan Supply Company or MSC: 1-800-645-7270,
http://www1.mscdirect.com/cgi/nnsrhm.
McMaster-Carr sells some EDM consumables, including various grades of graphite on the most
common shapes: http://www.mcmaster.com,
phone: (404) 629-6500.
For the home shop, a good source for graphite and metal electrode materials in small quantities are
the online auction sites. They often sell short pieces that are no longer useful for commercial EDM
operations, but usually are more than what is required for a home-shop application. Some retailers
sell “drops” as well.
15-9
Chapter 16
Flushing
Of all the functions performed by the cutting fluid, perhaps the removal of particles, or swarf, is the
most important. If the swarf is not properly flushed from the gap, the gap becomes overly conductive.
Shorting conditions occur more frequently, i.e., the servo ram becomes unstable, oscillating up and
down, and efficiency drops sharply, surface finish is poor etc.. Such conditions frequently lead to DC
arcing (see page 14-5), which will destroy the workpiece as well as the electrode. When doing
finishing work at a low power setting using graphite electrodes, DC arcing is a real possibility.
Under such conditions, usually the gap distance is very short, which restricts the removal of the
swarf; thus, the conditions are such that DC arcing may easily occur. There are three rules for good
EDM work
1. Flush 2. Flush 3. Flush
This is so important I will state it again. There are three rules for good EDM work: 1. Flush 2.
Flush 3. Flush. This is especially true when using a Pulse power supply in the finishing mode
described above. For those of you who have built and operated an RC-type power supply, if you think
flushing was important for good operation, it is nothing compared to the importance of flushing with
a pulse machine under low power settings in particular. A pulse machine will not work if you do not
have good flushing!!!! It is a fact of life; accept it. Should there a means of measuring through
electrode flush pressure, a good pressure for most flushing is 2-3 PSI.
For those who are RC-machine users, I venture to say that most of your flushing was surface or jet
flushing. It WILL not work with a pulse machine, particularly when using very short ON/OFF times,
unless your burn is only a few thousands deep. Good flushing is key to good machining and good
surface finishes. Poor flushing will result in erratic metal removal, rough surface finishes, and
increased machining time as the servo will “hunt” (i.e., oscillate back and forth) and not be stable.
The design of the indexing chuck on the servo head is
such that a flexible flush line may run from the “Y”
valve (seen in the picture at left) to the 1/8” hose barb
on the indexing chuck. If a hole(s) is drilled through
the electrode, then you have a quick and easy way to
attach the flushing line to the electrode. The picture left
has two flush-line setups.
EDM Flush Fittings
In some cases, you may not want to (or can’t) drill a hole completely
through the length of the electrode. Or you may be running multiple
electrodes attached to a single arbor. In these cases, multiple flush
lines may be run from the main flush line to the electrode(s). In such
a scenario flush fitting, or flush adapters, are necessary to attach flush
16-1
tube(s) to the electrode(s). This is accomplished by drilling a small hole in the side of the electrode,
which intersects with the flush hole(s). Then a flush fitting, with a long, tapered nozzle, (is pressed
into the side hole in the electrode and a flexible flush line is connected
to the fitting (see photo at right and at the bottom of this page).
The other end of the flush line either slides over a 1/8” plastic Loc Line
nozzle, as shown below, or is connected to the garden hose via a 1/4”
reducer (as shown on the previous page). At one time, the little tapered
plastic flush fittings were sold by EDM suppliers, but in the past few
years they have been increasingly difficult to obtain. I only know of one remaining source for these
fittings and they want to sell in lots of about 100, which is more than a
person needs for a home shop. I am considering purchasing a couple of
lots and reselling to interest builder. Email me for availability, sizes, and
price.
It is possible to modify plastic vacuum line connectors sold in
automotive supply stores to serve this function. File or machine a taper
on these connectors to fit into the small hole drilled in the electrode. The
other option is to make your own flush fitting from metal or plastic.
Jet Flushing
The most simple and least effective (though most often attempted by
homebuilders) flushing method is jet flushing. If a shallow through,
or blind hole, is burned in the workpiece, jet flushing may be used. It
is flushed by forcing dielectric from a nozzle into the area around the
electrode workpiece interface; see the drawing at left. It’s
advantageous to extract the electrode to permit clearing of dirty
dielectric fluid; this improves stability, speed and finish. The Pulse
EDM has a dither mode, which assists in this operation. The frequency and length of the dither may
be independently adjusted for best results. Jet flushing is an application where the Loc Line system
works very well, as it holds the nozzle in place: no fuss, no muss.
Through-electrode Pressure Flushing
One of the most common efficient flushing methods is through electrode
flushing. In pressure flushing, the pump forces clean dielectric fluid through
a hole in the electrode. The drawing and photo suggest a method that
facilitates this technique. The hose used to carry dielectric fluid to the
electrode from the flow control valve should
be flexible to avoid deflecting the electrode.
I often use rubber automotive vacuum line
tubing. It is low in cost, readily available,
and flexible; the problem is that it tends to degrade after continuous
exposure to many of the dielectric fluids. For real small flexible lines
I have used plastic gasoline lines that are designed for use on string trimmers (Weed Eaters).
16-2
The pressure flushing method is easy to accomplish and supplies clean fluid to the arc gap. The
correct pressure and flow of dielectric through the gap is critical for burn stability. Too high a
pressure or too much volume can cause an unstable burn (see page 13-5 for control valve info).
Through-electrode pressure flushing does have its drawbacks. One is that it causes tapers to develop
in the burned hole. As the dielectric leaves the arc gap area, it is swarf-laden. The dielectric and swarf
are forced between the electrode and the workpiece (in the overcut area). This creates sparking along
the side of the electrode as it exits, and will create slight tapers as shown in the drawing on the
previous page (exaggerated).
Through-electrode flushing leaves a core of metal extending into the
flush hole, in the electrode, as shown in the drawing at left. The core
restricts dielectric flow. With deep holes, it’s necessary to extract the
electrode and break the metal
core off. One way to reduce the
core problem is to drill flush
holes in the bottom of the electrode at an angle to the main
feedhole. This is show in the drawing at right. By doing this, the
cores are mostly burned away as the cavity is formed.
Sealed Flush
The best conditions for pressure flushing are when a sealed flushing condition exists. Sealed flushing
is when the whole face (end) of the electrode is in a closed burn cavity. This condition forces the
dielectric to flow completely across the face of the electrode as it removes swarf. When burning a
through hole, as the electrode starts to break through the sealed flush, condition is lost, the burn
becomes unstable, and the dielectric quickly escapes through the hole. The solution is to mount the
workpiece on top of a sacrificial metal block. As the electrode burns through the workpiece, the
sacrificial metal block maintains the sealed flush condition. This method will result in the best
surface finish and the lowest and most stable burn time.
Pot Flushing
Another means of pressure flushing is pot flushing (as seen in the
drawing at left and the photo below). A pot or plenum is provided
below the workpiece. Dielectric fluid is forced into the pot and
through the gap via a hole drilled in the
workpiece. This method is especially
convenient in mold-making, where drilled
ejector and/or core pin holes are already in
the workpiece. Pot flushing is excellent to use with long or thin electrodes
that are difficult to drill flush holes in. The photo at right is a plenum that I
made for a small specialty job; the offset to the right is used to clamp the
plenum in place on the work plate.
16-3
Suction Flushing
If flushing is accomplished by suction, as shown in the drawing
(left), the tapered electrode and workpiece problem is virtually
removed. This is because clean dielectric is drawn through the
overcut area and the swarf-laden dielectric exits the gap through
the hole in the electrode and is returned directly to the filtering
system. The hole in the electrode should be at least .062 in
diameter or larger.
Although this is the preferred flushing method, it is more difficult, since it requires a means of
drawing (sucking) the fluid through the electrode. If there are holes in the work piece already, this
method of flushing may work well with the above plenum (which is a similar flush method). If the
suggested submergible pump is used, this flushing method is not possible. If you are using a nonsubmerged pump that has the ability to create “suction,” then this is a viable flushing method.
Trepanning
Should there be a large area of metal to remove, and it is a through-hole, time can be saved if you use
a trepanning electrode (see drawing at left). Trepanning is a “holed-out” or tubular electrode used in
through-hole machining to remove a large amount of material from the solid workpiece. A core is
formed during the cut, and drops out when the electrode penetrates the workpiece. Trepanning is
always the fastest method for creating large through-holes using
the EDM process.
Remember that a good flushing system will make a poor EDM
generator look good, and a poor flushing system will make a good
EDM power supply look bad. Flush, Flush, Flush.
Also remember that a good flush pressure for most applications is between 2-3 psi. As a reference,
the recommended pump is Dayton Model 2HNN9 (see page 11-4), which has a maximum head
pressure of a little over 7 psi.
16-4
Chapter 17
Your First Burn
Now that your EDM is complete, it is time to actually make a burn; complete instructions will be
given for accurately setting up the machine for any burn. This includes finding the workpiece edge
and positioning the electrode over the workpiece and determining the burn depth. For this first burn
you may, or may not, want to go through all of these steps. For example, you may not care if the
electrode is precisely centered in a particular location for the first burn.
In the course of operating the EDM, you may find combinations of ON and OFF times that will cause
the servo ram to be sluggish. This will be most noticeable when “long” OFF times are used in
conjunction with the 25 and 50us ON-time settings. More often than not, this is not a serious issue, as
short OFF times usually will work fine with short ON times. The correct procedure is to set the
desired ON time and adjust the OFF time for the shortest OFF time that will give you a stable burn.
You should also know that it is not uncommon to discover the setting of the Gap Voltage control pot.
It has a narrower operating window in the 50 & in particular the 25 us ON time setting.
For the first burn, I suggest using a graphite electrode with negative polarity in conjunction with a
high-power (at least three of the power switches toggled ON) roughing cut with ON times longer than
100 microseconds. In most cases, the higher the power settings, the easier it is to achieve a stable
burn; thus, at lower power settings and short ON times, the cut will be the most unstable. Should you
want to start with a copper electrode, you must use positive polarity.
[ ] A good choice for the first burn would be a graphite electrode at least 1/2” in diameter.
[ ] The workpiece should be steel and secure in the work tank.
[ ] Set up the machine so the electrode is negative (positive for copper); remember polarity is always
in reference to the electrode.
[ ] This first burn will not be very deep; jet flushing should suffice (see page 16-2). Should the burn
be deeper than a few thousands- through electrode flushing is recommended.
[ ] Make sure that the electrode and workpiece are dry (free of dielectric) and clean.
[ ] Turn the power supply to the edge-finding mode.
[ ] Jog the electrode so that the tip is extended below the top of the workpiece; turn the Servo switch
OFF.
[ ] Bring the electrode physically close to the workpiece using the X axis.
[ ] Move the X axis slowly until you hear the high-pitch- sound of the edge finder buzzer. The edgefinding circuit is indicating that the edge of the workpiece has been located, so stop moving the table.
17-1
[ ] Toggle the Servo switch ON and raise the electrode above the workpiece using the manual
up/down switch, or by manually moving the entire head up the back slide, temporally locking it in
place on the ways.
[ ] Move the X-axis half the diameter of the electrode (.250), which puts the center line of the
electrode on the edge of the workpiece. Now move the desired distance over the workpiece.
[ ] Repeat the same process for the Y-axis.
[ ] Turn the power supply ON, using the manual up/down switch; position the ram to its middle
position on the lead screw (if it is not already in that position).
[ ] For locating the electrode on the top of the piece, there are a couple of ways to achieve this. Which
method is used is up to you.
[ ] Release the back slide and slowly let it down until the electrode contacts the workpiece and lock it
there.
The second method is a little trickier but is the method to use if you are burning to a specified depth
and or want the ability to retract the electrode (or the whole head on the back slide) and reposition it
and not lose position.
[ ] With the ram in the mid- to full-up (or near full-up) position, release the back slide and slowly let
it down until it comes in contact with a pre-positioned dowel pin in the back slide WAYS. This dowel
pin should be in the lowest hole on the back slide WAYS that will not permit the electrode to come
in contact with the workpiece.
[ ] With the full weight of the slide against the dowel pin; lock the slide in position on the WAYS.
[ ] Toggle to the edge-finder mode on the power supply
[ ] Turn the power supply on. With the manually UP/DOWN switch toggle to the down position.
I suggest using the servo speed control knob and turn it to the slowest speed as the electrode
approaches the workpiece; you do not want to crash the electrode into the workpiece. Keep your
hand at the ready on the servo ON/OFF switch. As soon as the edge-finder buzzer sounds toggle the
Servo ON/OFF, switch OFF. The edge-finder buzzer will continue to sound as long as the power
supply is on.
[ ] Turn the AC power to the power supply off.
The following sequence may be a little confusing when you read through it, but what is happening
will become clearer if you follow the steps word for word.
[ ] With the electrode in physical contact with the top of the part and the machine in the edge- finder
mode (the edge-finder buzzer will be sounding) perform the following operation to set the depth of
the burn. Flip the switch on the servo head to setup.
17-2
[ ] Move the beam (the sliding portion) of the digital readout down until the LED on the head comes
on (this opens the lower limit micro switch).
[ ] Zero (0.000) the digital depth indicator.
[ ] Slide the beam back (i.e., up) the desired burn depth (as indicated in the digital readout window).
Lock the beam in place with the thumbscrew.
[ ] Hold the manual servo switch to the UP (retract) position, and then turn the servo switch on. This
will move the ram away from the work piece a few thousands.
[ ] Turn the power supply AC switch off and let the (panel-mounted) Manual Up/Down servo switch
return to the middle (off) position.
[ ] Toggle the Power/Edge Finder and switch to “Power.”
[ ] Toggle the Burn/Setup switch on the head to “Burn.”
So what have we just done? Let’s assume that the desired burn depth is .125 deep. The electrode will
be advanced by the servo motor during the EDM burn; the preset distance (the distance you manually
move the beam up on) and then the lower limit micro switch trips. This opens the switch, stopping
power from reaching the servo motor. The power supply will continue sending power pulses and the
burn will continue until the gap becomes too great for the sparks to jump; at that point all EDMing
stops. This is called “sparking out.” The burn depth should be very close to the preset .125. After a
few burns (keep a log), you will develop a feel for any correction you need to make to get the desired
burn depth. It is also helpful to record the ON and OFF settings of the power supply when doing the
burn.
The nice thing about the quick-remove back slide with the dowel pin is that the power supply may be
turned OFF during a burn; the servo head moved up the side (and/or was removed from the slide),
making it easy to inspect the burned cavity and/or take a depth measurement. Once inspected, the
back slide is returned to the dowel pin location, locked in place, and the burn resumed; this is
possible without losing any position accuracy!! I love it; this is a GREAT system if I must say so. As
you do more and more EDM work, you will really appreciate this feature.
[ ] Position the nozzle-directing dielectric at the workpiece/electrode interface (assuming that jet
flushing is used). Jet flushing is an application where the Loc Line-type flushing line really shines;
the nozzle(s) stay where you position them.
[ ] Fill the tank with dielectric. Check that the drain valve is closed; if you have a drain valve tube,
adjusted it to a height to ensure that the workpiece is submerged by at least 1”.
[ ] For this next operation, make sure that the servo switch is turned off. You don’t want the servo to
advance until the power supply is fully adjusted.
17-3
[ ] Set the power supply for a high-power setting, with at least three power switches switched ON.
[ ] Adjust the ON time to 100 (or longer) microseconds.
[ ] Adjust the OFF time pot. to mid-position.
[ ] Turn the Servo Gap Voltage pot. to middle position.
[ ] Position the servo speed pot. to the middle position (1/2 speed).
[ ] Turn the flushing pump on and adjust the flushing flow for a steady, but not forceful, flow.
[ ] Turn the servo switch on; this will start the ram advancing.
If the ram does not advance, rotate the Gap Voltage pot. and/or the Servo Speed control until the
electrode starts to advance toward the workpiece.
Within a few seconds, you should hear a sound somewhat like frying bacon. For those that have built
and used the RC power supply depicted in The EDM How-To Book, do not expect the same fryingbacon sound that you are used to.
[ ] Adjust the OFF time for a shorter off time if possible. If you are unsure as to which direction to
rotate the pot., listen to the sound of the burn; most likely you may hear a tone change in the burn.
Shorting the OFF time will result in the tone becoming higher in pitch and vice versa.
Another indication that the OFF time is becoming shorter is that the panel volt meter will show a
decrease in voltage. If an amp meter is used, the current will increase.
Make the OFF time as short as possible; should the servo start to operate erratically, increase the OFF
time until it stabilizes. A short OFF time will provide the fastest burn.
The Gap-Voltage control may be adjusted to stabilize the burn as well but ONLY after attempting to
stabilize with the OFF time control.
The following describes how to use the dither circuit.
When burning with low power and, in particular, when using short ON times (finishing operations),
keeping the gap properly flush is CRITICAL; this is true especially when using graphite electrodes.
The dither circuit causes the electrode to be retracted at a predetermined rate and length of time. By
retracting the electrode slightly, the trapped particles are able to be flushed from the gap more
readily. Dithering may be used at any time that less-than-desirable flushing conditions exist.
[ ] Toggle the dither switch ON.
[ ] Adjust the duration (length of time that the electrode is retracted) and interval (the frequency with
which the electrode is retracted) to a timing sequence that fits the operation. This timing will be
17-4
learned through practice. For starters, adjust the duration for about as short a time as is permitted.
Adjust the interval for the cycle to repeat about every 5 seconds. When the dither circuit is actuated,
the Ram Indicator LED will switch colors every time the electrode is retracted. When the dither
circuit is not needed, toggle the Dither switch OFF.
[ ] Let the EDMing proceed until you can’t wait to see the result.
[ ] Turn the machine off, retract the head, and drain the dielectric tank.
[ ] Remove the workpiece and see the results of many hours of work!!!
At this point you are set for lots of fun and interesting experimentation.
Try different workpiece materials; use different types of electrodes and different size electrodes. Play
around with all the different types of flushing. Keep a log of every burn and note:
1.
2.
3.
4.
5.
6.
7.
8.
9.
Power level
ON and OFF time setting,
Type of electrode material
Polarity of the electrode
Type of workpiece material
Flushing method & number of flush holes
Dielectric used; if you have a way of indicating flush pressure, this is useful to note.
Measure the overcut.
In most cases, the electrode corner wear will be a subjective evaluation, but note it as best you
can in the log. It is easy to measure the length of the electrode before and after the burn.
There is a lot to experimenting and learning to become proficient. This is a metal-cutting method that
is not commonly understood; soon you will be the expert EDM person in your circle of machinist and
electronic friends. As the word gets around that you have an EDM, you may find that you have more
work than you want, but then maybe you are looking for a niche machining market; EDM is one of
those markets.
You will find out that this machine has the capability of doing more than just burning out broken taps
and bolts. Put a “real copper penny” (prior to 1982) and burn the impression into a piece of tool steel.
I am in no way advocating going into manufacturing money; it is illegal!!! It is fascinating to see
Lincoln’s head appear in a piece of steel. The image will need to be buffed out using jewelers’ rouge
to see the amazing detail possible. I’ve used a brass Chuck E. Cheese token, and copper earrings (see
back cover), as well as many other brass and copper object and burned these impressions into tool
steel!! Of course, I also use it for machining operations such as splines in gears, blind keyways,
square holes, blind hex holes, squaring of milled pockets, and of course removal of broken drills and
taps.
Have fun; you are going to enjoy this machine!
17-5
Chapter 18
Jewelry Manufacturing and EDM Tips
Jewelry Manufacturing
Another aspect of EDM work is using the EDM process for adding lettering and original designs to
existing or partly finished jewelry. For the craftsman, designs can be made in copper and/or graphite
electrodes and then burned into hardened tool steel, thereby forming a die.
Frequently, the work done for jewelry requires only shallow burns; when making such burns for dies,
it is critical that the electrode be completely parallel with the workpiece. The piece in the photo
shown on the back cover was burned about .015 deep in the steel; thus, squareness with the
workpiece was very important so that one end or side is not deeper than the other.
The electrode needs to be supported especially in the thin areas and at the tips. Heat generated in the
electrode during the burn (metallic electrodes in particular) causes it to bend and warp; thus, the ends
or tips will be deeper than the center parts if not properly supported.
A significant timesaving (as well as better) surface finish is accomplished by horizontally EDMing
(see page 18-3) Attach arbors via solder or glued, as described below.
For both roughing and finished burns, consider using the dither setting to assist in keeping the gap
clear of swarf, particularly if a graphite electrode is used. Start the roughing burn @ 100us or 200us
ON time and at about 6 amps (or less). An intermediate (if used) roughing burn should be at 3 amps
at 50us ON time, and the finished burn should be 25us ON time (l amp or less).
Using Tokens and Coins as an Electrode
The idea of using existing jewelry, findings, tokens, or even coins for an electrode is an exciting
prospect in jewelry-making using an EDM. Don’t misunderstand; I am in no way suggesting that
you copy any kind of currency. It is illegal to make money of any kind!!!
The following photos and suggestions are to show the capabilities of this
machine when it comes to jewelry-making. I am by NO means a jeweler;
this machine, in the hands of a jeweler, can produce much finer work than
what is illustrated here. On page 1-2 of this book is a photo of an
impression made using a Chuck E. Cheese token and on page 18-2 is work
using a copper earring.
The following photos outline how the Canadian penny was burned into a
steel piece. First, the coin was soldered to an arbor (see photo at left). A
mill was used to hold the arbor square with the penny. A piece of EDM
graphite was readily available to insulate it from the mill vise during soldering. Make sure that the
electrode is completely square with the workpiece prior to the burn. The electrode polarity needs to
be positive for this burn. The” roughing” burn was 10 minutes @ 200us, ON time, with 3 amps of
power. The finish burn was 10 minutes @ 25us, ON time, at 1 amp.
18-1
After the burn, the electrode shows very little
wear; the fine line of “dots” is still clearly visible
near the outside rim of the penny (see photo at
left).
At the top, 12-o’clock position, darkness is
evident. This is an area that was not receiving
proper flushing; fortunately, it did not result in a
DC arc (typical of copper electrodes). Should it
have arced, it would have destroyed the electrode and die. The photo
above right shows the die after the finish burn. The die was buffed for
about 10 minutes and the photo right below shows its appearance. More
precision buffing would yield a much better die. From start to finish, the
total time was less than 40 minutes!!
Once you have a die, soft metals such as brass, aluminum, copper, zinc,
etc. may be pressed into any die in order to make reproductions, but not
COINS; it is illegal to manufacture money!! After all, you will go
broke; most coins are worth less than their face value
Using a Copper Earring as an Electrode
The photo at left (as well as on the back cover of the book) is an
example of using a commercially available copper earring as an
electrode. The 1/2” steel arbor was soldered to the back of the
earring. The reverse image is shown burned in the rectangular
tool steel block. The feather image in front of the block is an
impression in copper pulled off of
the EDM (die) impression.
The complete process, from
beginning to end, including the
impression in copper, took less than
30 minutes. The roughing was at 200us ON time, which took 6 minutes.
The finish at 25us, ON time, took 12 minutes.
The photo at right shows the same copper earing “nested” in its burned
pocket. There is a second image of a feather also burned in the vertical
face for the die block and the feather impression in copper is lying horizontal in front.
18-2
Pressed Arbor
Many times, it possible to use small, lightweight thin electrodes to do the
operation (see photo at left). The graphite electrode has a reamed hole in the
center and the brass arbor is pressed into the hole; (a round piece of graphite
may be used in place of the brass). The brass piece has a hole through the
center for flushing. The electrode itself need not be thick to cut a pocket much
deeper than its thickness. This design also minimizes hole taper. This 1/8”
thick electrode was used to burn the gear shape 1/4” deep into a hardened workpiece, for a plastic
injection-molding project.
Glued Arbor
Another technique for holding electrodes is to glue them to an arbor (rather than pressing or
soldering). Use superglue for the adhesive and mix graphite dust with the glue. The graphite dust
provides a conductive path for the current to flow through. The glue bond needs to be electrically
conductive with less than 1 ohm of resistance; check with ohm meter after gluing. If you end up with
a couple of fingers glued together or are glued to the electrode, use acetone to free yourself.
Glued-Together Electrodes
Sometimes it is faster and more cost-effective to make electrodes in parts and then glue them
together. Use the superglue and graphite dust combination as described above for glued electrodes.
Upside Down and Sideways
In some applications (see “Jewelry Manufacturing,” above), it is beneficial to mount the electrode
on the work plate and attach the workpiece to the ram (assuming that the workpiece is small). This
arrangement greatly simplifies the flushing process, as the swarf naturally falls out of the gap.
Perhaps an even better method is horizontal EDMing. To do this, turn the head horizontally (i.e.,
parallel with the surface of the dielectric in the work tank), with a leg extending off the end of the
ram and down into the dielectric. The electrode (or workpiece) is attached to this leg. The workpiece
is standing vertically in the work tank on a right angle plate. The flush should be from the top,
causing the swarf to fall or be forced out of the gap. This is also a good configuration for machining
where surface finish is critical, an example would be making a mold for tokens or medallions. Using
either of these methods greatly reduces the chance of arcing, which is always a possibility (especially
with graphite electrode).
Carbide Insert Tool Holder
Another very useful application of the EDM process is burning the pockets needed to hold carbide
tool inserts. Often carbide inserts may be found on internet auction sites for very reasonable prices.
Many times these are discontinued inserts, so the inserts are sold cheaply. The problem is that holders
for these inserts are often difficult to locate and expensive. With an EDM, this becomes a moot point;
in fact, it has the potential for being a money-making side line.
18-3
Attach the actual carbide insert to the end of an arbor via
the described glue method; it may also be brazed. Use the
actual insert as the electrode material, as seen in the photo
at left. This will give a perfect match to the inserts. It
takes only a few minutes to burn an insert pocket.
All that is needed to finish the holder is to machine a
finger to hold the insert in place. Of course, it is possible to make a graphite or copper electrode of
the required shape and burn the pocket. The graphite and or copper electrode will give a slightly
better surface finish, with sharper corners, than will the carbide electrode.
Multiple Electrode Burns
It is possible to burn multiple holes at the same time with an EDM. Let’s assume that five 3/16” holes
are needed in a part. Instead of burning each hole individually, make an arbor with five electrodes
attached. The electrodes may be machined from one piece of electrode material, or individual
electrodes can be glued, pressed or mechanically attached to an arbor. It will take five times longer to
burn than with a single electrode. However, you will save set-up time and minimize the chance of
losing position, by accomplishing the task with one set up.
During the burn, if any of the electrodes short, all will be retracted, as they are all attached to the
same arbor. Remember, the generator only generates one spark at a time; it will spark at whichever
electrode’s gap ionizes first. In the end, things average out to equal burn depths, assuming good
flushing.
Broken Screw and Bolt Removal Techniques
To remove broken screws, bolts, studs, etc. (in many cases), it is not necessary to burn the complete
broken item out. I have, on occasion, made a six-sided hex electrode and then burned a hex shape in
the broken part. Once deep enough, insert a standard hex key and remove the part. Screwdriver slots
may also be burned into broken parts to assist in removal. When removing a broken tap (especially
for larger taps), make a trepanning-type electrode and just cut through the flutes in the area where the
chips normally collect.
18-4
Chapter 19
Glossary
Ammeter - An instrument for measuring electrical current in amps.
Amperage - The amount of average current measured while cutting is taking place.
Ampere - The unit of electric current produced by 1 volt across 1 ohm of resistance.
Anode - The positive terminal of a cell. Often applied incorrectly to the EDM electrode.
Arc - A continuous flow of electrical current (also known as D.C. arcing) between the electrode and
workpiece. It is an undesirable condition and is normally recognized by a yellow flash.
Arc suppressor - A circuit in the control section of an EDM that reduces the possibility of arcing.
Automatic depth finder – An electrical device that stops the EDM ram slightly above the workpiece
and maintains that position.
Average current - The average value of all the minimum and maximum peaks of amps flowing in
the gap.
Burning - A slang term for the EDM process.
Capacitor - An electrical device that stores a charge. It is used in some power supplies (mostly older
designs). It is also used occasionally in newer supplies to intensify the spark with increased electrode
wear.
Carbon - A naturally occurring material used to make graphite. Many times it’s “used” to describe
the electrode material made of graphite.
Cathode - The negative terminal of a cell. Often applied incorrectly to the workpiece.
Circuit - A continuous path for the flow of electric current.
Condenser - An outdated term used to describe a capacitor.
Conductor - A material which will carry electric current.
Contamination - The accumulation of debris in the dielectric fluid, which results in a decrease in the
fluid's dielectric strength.
Coolant - See dielectric.
Copper graphite - A graphite electrode material infiltrated with copper.
Copper tungsten - A porous tungsten material infiltrated with copper.
Core - The slug that remains after EDMing with an electrode that has a flush hole in it.
Corner wear - The wear which occurs on the corners of an electrode (the highest wear)
Crater - The small cavities left on the surface of the workpiece by the EDM sparks.
Cubic inches per hour - The unit of measure used to describe the rate of metal removal.
Cut - To machine with the EDM process.
Cutting rate - Same as “machining rate.”
DC arcing - See Arc.
DC/Direct Current - Constant polarity current, as opposed to Alternating Current (AC), which
changes polarity from negative to positive in cycles.
Deionization – A return of the dielectric to a nonconductive state. Failure to reach this state is a
major cause of DC arcing.
Depth of crater - The distance from the peaks to valleys on the workpiece.
Depth-to-diameter ratio - The ratio of the depth of a blind hole compared to the diameter of the
electrode used to make the hole.
Diameter (dia) - The straight-line distance through the center of a round object.
Diametral sparking distance - The difference between the electrode dimension and the dimension
of the hole produced.
19-1
Dielectric fluid - A light oil which insulates the spark gap between the electrode and the workpiece
until a high voltage ionizes the spark gap and causes it to become an electrical conductor. This
ionization permits an electrical current to flow through the dielectric to the workpiece. The dielectric
also serves to cool the work and to flush away the particles removed from the workpiece.
Dielectric strength - The voltage at which the insulating qualities of a material break down.
Discharge - The EDM spark.
Discharge channel - The conductive path formed by the ionized dielectric and vapor between the
electrode and workpiece.
Dither - A vibrator motion of the electrode used to improve cutting stability.
Down feed - A control circuit to advance or retract the electrode.
Duty cycle - The percentage of the on-time relative to the sum of the on-time and off-time setting for
a particular cut.
Edge-finder - An electrically activated device to aid in the accurate location of the workpiece with
respect to the electrode.
EDM - Electrical Discharge Machine or Electrical Discharge Machining. A metal removal process
using a series of electric sparks to erode material from a workpiece.
EDG - Electrical Discharge Grinding.
Electrical resistively - The resistance of the flow of electricity measured in ohms.
Electrode - A tool used in the EDM process most commonly made of copper or graphite.
Electrode growth - A plating action occurring at certain low wear settings, which causes workpiece
material to build up on the electrode, causing it to increase in size.
Endwear - A reduction in the length of an EDM electrode during EDMing.
Eroding - Material removal by the EDM process.
Farad - The unit of electrical capacitance. One farad = a potential of one volt when charged by one
coulomb.
Filtering - The removing of debris from the dielectric fluid before being pumped back to the work
tank, or through flushing holes in the electrode or workpiece.
Finish - the surface texture produced by EDMing.
Finish cut - The final cut made with EDM on the workpiece. The finer the finish, the longer it will
take for the finish cut.
Flashpoint - The temperature at which any flammable material will burst into flames.
Flush pot - A multipurpose, box- like fixture, which is clamped to the machine worktable. It is
primarily used when EDMing through-holes. The workpiece is clamped to the box top (over a
predrilled hole or opening) that will permit an electrode to pass through the workpiece without
interference from the tank's top plate. When connected to the dielectric system, it can be used for
either suction or pressure flushing.
Flushing - Making dielectric flow through the gap to remove the debris caused by machining with
EDM.
Flush hole - A hole through the workpiece or electrode used to introduce dielectric fluid into the gap
for flushing away debris.
Frequency - The number of cycles (ON/OFF completed per unit of time, usually expressed in Hertz.
Gap - The distance between the electrode and workpiece. Also referred to as “spark gap.”
Gap voltage - This can be measured as two different values during one complete cycle. The voltage
that can be read across the electrode/workpiece gap before the spark current begins to flow is called
the “open gap voltage.” The voltage reads across the gap during the spark is the working gap voltage.
Generator - An old term for an EDM power supply.
19-2
Graphite - The most commonly used material for EDM electrodes in the U.S. It has very high heat
resistance and transfers electric current very efficiently. In Europe, copper is the most common
electrode material
Head - The part of and EDM which houses the ram or quill.
Heat-Affected Zone (HAZ) - Also called the "recast layer.” The depth of heat penetration alters the
parent material’s metallurgical structure due to the EDM process. The depth ranges from .0002 to
.008, depending on the material and energy per pulse.
Hertz (Hz) - The international term for one complete electrical wave cycle. In EDM, it is the unit of
frequency.
Hunting - An erratic bouncing movement of the ram during a cut caused by poor flushing conditions.
Injection flushing - An external flushing method (also known as jet flushing).
Initiation voltage - Same as open-gap voltage.
Insulator – A material (such as rubber) which blocks the flow of electric current.
Ionization - Generally accepted as a phenomenon by which the dielectric between two points on the
electrode and workpiece become electrically conductive.
Ionization voltage - The voltage at which current flow begins across the gap.
Ionized path - The path of electrically conductive dielectric molecules (between the electrode and
workpiece) through which the spark current will flow.
Lateral flushing - Same as surface, splash, or jet flushing.
Low wear - Setting for EDM machining which produces a very low degree of wear on the electrode.
In some cases, less than 1% wear is possible.
Machining rate - Same as “metal removal rate.” The rate at which material is removed.
Metal removal rate (MRR) - The rate at which material is removed from the workpiece by EDM, in
cubic inches/hour.
Microfarad (also uf) - One-millionth of a farad.
Micron micrometer - A unit of length equal to one-millionth of a meter.
Micro-inch - One-millionth of an inch (.000001).
Microprocessor - A computer-on-a-chip. Found in all advanced EDM systems. The microprocessors
provide many control functions.
Microsecond - One-millionth of a second (.000001).
Millisecond (ME) - One thousandth of a second (.001 sec.).
Miss - A pulse that does not produce metal removal.
Modular construction - In EDM, a type of power supply where entire circuits are integrated on
boards or modules. This makes possible rapid servicing and replacing.
Monitor - In EDM work, any mechanical or electrical device that is used to indicate various
operating conditions.
Multiple electrode - The simultaneous use of two or more electrodes to produce cavities in
workpiece(s).
Multiple-lead power supply - One power supply with multiple independent power leads.
Nanosecond - One millionth of a second (.000001 sec.).
Non-directional finish – A finish having no specific direction to its surface pattern. An EDM finish
is non-directional.
Normal polarity - Negative polarity to the electrode.
OFF-time - The time between sparks, measured in microseconds.
Ohm - A unit of electrical resistance. Current flowing in a conductor of one ampere produced by a
potential of one volt.
19-3
Oil-through chuck - A sealed holder for tubular electrode through which dielectric fluid can be
pumped or drawn.
ON time - The duration time of the EDM spark measured in microseconds.
Open circuit - An electrical circuit that is not complete.
Open-gap voltage - The voltage between the electrode/workpiece before current begins to flow.
Output - The voltage and current of an EDM power supply.
Overcut - An EDM cavity is always larger than the electrode used to machine it. The difference is
called the overcut. Diametral overcut is most often used.
Pause - Same as off time.
Peak current - The maximum current available from each pulse from the power supply.
Percent electrode wear - The volume of electrode worn away as compared to the volume of
workpiece worn away.
Platen - A large flat mounting surface affixed to the end of the ram of an EDM machine.
Polarity - In EDM, the designation of positive or negative electrical potential to the electrode.
Potential - The difference in voltage between two points of an electrical circuit.
Power supply - The part of the EDM system that supplies the voltage and current that causes the
sparks or discharges between the electrode and the workpiece.
Pressure flush - The forcing of dielectric through flush holes in the workpiece or in the electrode.
Pulse - The discharge of a quantity of electrical energy having preset voltage and amperage and
expended over a preset time.
Pulse duration - Same as “ON-time.”
Quench - The rapid cooling of the EDM’ed surface by the dielectric fluid.
Quill - See “ram”.
Ram - The moving member of an EDM on which the electrode or electrode holder is mounted.
RC (relaxation) circuit - An outmoded EDM power supply circuit that uses capacitors to store the
charge that produces the spark at the gap.
Recast layer - A layer created by molten metal solidifying on the work surface.
Rectifier - An electrical device that converts alternating current to direct current.
Relaxation circuit - See “RC circuit.”
Reverse burning - The technique of mounting the electrode on the machine table and the workpiece
on the ram.
Reverse polarity - A term used to indicate positive polarity to the electrode.
RMS (Root Mean Square) - An obsolete term used in surface finish measurement.
Rotating spindle - A device used on EDMs used to rotate the electrode to achieve more uniform
wear and to improve flushing conditions.
Roughing - The mode of EDM that removes the most material in the shortest time.
Secondary discharge - A discharge that occurs as conductive particles are carried up the side of the
electrode by the dielectric fluid.
Servo-mechanism - The device that drives and controls the movement of the ram.
Side wear - The wear along the sidewalls of the electrode.
Silicone - A dielectric fluid for special situations consisting mostly of the chemical polymer silicone.
Solid-state power supply - A power supply that uses transistors in the circuitry.
Spark - An electrical discharge of very short duration between two conductors.
Spark erosion - Another name for EDM. Used primarily outside of the U.S.
Spark gap - The distance between the electrode and the workpiece when discharges are occurring.
Spark intensity - The amount of energy in the spark.
19-4
Split electrode - Multiple electrodes on a single machine electrically insulated from each other.
Square wave - A term for an electrical wave shape generated by a solid-state power supply.
Stability - The steadiness of an EDM cut.
Stepped electrode - An electrode constructed in such a manner as to allow the roughing and
finishing of a through-hole cavity in a single set-up.
Stroke - The distance the ram travels under servo control.
Suction flushing - Using a vacuum to draw the contaminated dielectric away from the gap.
Surface finish - The relative roughness or smoothness of a machined surface.
Surface integrity - The quality of the machined surface and subsurface.
Surface flushing - The use of nozzles to direct jets of dielectric at the cutting area to flush away
debris.
Surface roughness - Surface irregularities on a machined surface.
Taper - The dimensional difference between the entrance and exit opening of a through-hole cavity.
Through-hole flushing - The use of a pre-drilled hole in the workpiece to inject dielectric fluid up
toward the gap by injection flushing or down from the gap by suction flushing.
Timer - A control unit that controls spark ON and OFF time.
Transistor - An electronic component used as a switch to turn electrical current on and off.
Trepanning electrode - A hollowed out or tubular electrode which is used in through-hole
machining to remove a large amount of material from the solid to avoid pre-machining by
conventional means.
Unstable - Erratic or intermittent EDMing.
Vacuum flushing - see “suction flushing.”
Vacuum -Tube power supply - An old power supply design which used vacuum tubes to switch the
machine pulses on and off.
Vibrator - An accessory used to move the electrode back and forth rapidly primarily for improving
flushing.
Viscosity - The tendency of a fluid to resist flow.
Voltmeter - An instrument that measures voltage.
Volumetric Wear - The total wear of the electrode expressed in cubic inches.
Waveform - A geometric display of the output of a power supply as seen on an oscilloscope.
Wear - The erosion of the electrode during the EDM process.
Wear ratio - The volume of electrode worn away as compared to the volume of workpiece material
removed by EDM.
Wire EDM, WEDM - An EDM process where the electrode is a continuously moving conductive
wire that moves in preset patterns around he workpiece.
Working gap voltage - See “gap voltage.”
Workpiece - Any part on which EDM is being used to cut holes or cavities.
19-5
Chapter 20
Reference Materials
_____________________________________________________________
Most of this book is from my own personal experience in operating EDM equipment, as well as
designing and testing various design ideas. In developing design and in writing this book, the follow
publication and papers were used without this material this book would not be possible.
EDM Technical Manual
Poco Graphite, Inc.
1601 South State St.
Decatur, TX 76234
Complete EDM Handbook
Carl Simmer M.E.
EDM Today
EDM Publications Inc.
230 W. Parkways
Pompton Plains, NJ 07444
Electrical Discharge Machining
Elman C. Jameson
System 3R USA
231 W. Parkway
Pompton Plains, NJ 07444
Vega Enterprises Inc.
Electrical Discharge Machining
EDM Experiment Manual
Decatur, IL
Metals Handbook, 8th Edition
American Society for Metals
Electrical Discharge Machining (EDM)
Metals Park, Ohio 44073
Machine Tools
Volume 1
Manfred Weck
International Conference on Manufacturing Technology
Present and Future of EDM in Japan
By Hisao Kurafuji
University of Tokyo
20-1
All part number are Mouser Electronics unless otherwise noted.
Qty.
Item
Description
Mouser number
(1)
(2)
(2)
(8)
C1
1000uf
667-EEU-FR1V102B
C2,15
.1
581-TAP104K035SCS
C3,5
33
667-EEU-FR1V330
C4,6,7,8,10
12,14, 16
.01
81-RDER71E103K0K1C3B
(1)
C9
47pf
140-50S2-470J-RC
(1)
C20
2200uf +
75-36DX202F200BC2A
Optional, use 1 (or 2) of the following capacitors see page 5-6
(1)(2) C20&20A
2200uf+
661-ESMH201VNN222MA
Note: Select and purchase a capacitor mounting ring specific to size capacitor/s purchased
(1)
CB1
35-2110
655-W57-XB7A4A10-10 Circuit breaker
(1)
(6)
(2)
(3)
(3)
(2)
D1,
D2-7
D8, 9
D15,16,17
D18,19,22
D20, 21
KBP04G
1N914
1N4742A
Bi color LED
1N4001
KBPC2502
512-KBP04M
610-1N914
512-1N4742A
604WP937EGW
512-1N4001
905-KBPC2502T/W
(1)
F1
Fuse holder
(1)
Fuse
10 amp
441-R328-GRX
576-0314010.HXP
(1)
Fan1
Fan
OA4715-12TB
(1)
J1
3.5MM jack
161-MJ355W-EX
(1)
PL1
3.5mm plug
568-NYS231B
Mouser’s meters are very high quality and expensive, average price about $70. I suggest using another
supplier. Listed below is a meter from Marlin P. Jones http://www.mpja.com/ that is fine.
(1)
Meter DC
15 VDC
17671 ME
(1)
P1
Buzzer
539-PK-12N40PQ
(1)
Power Plug
BP1&2
592-20ESRM (power plug to electrode)
(1)
(1)
(6)
Q1
Q2
Q3-8
PN2907
PN2222
IRF 640
512-PN2907ABU
512-PN2222ABU
844-IRF9640PBF
(7)
R1,2,3,15,20
21,25,26,30
R4, 11
R5-6
R7
R8
R10,24,27,28
R12
R13
R16-21
R22
R23
1K
3.3K
33K
68K
5.1K
10K
100K
100K (pot)
10ohm
1 meg
510
279-CBT25J1K0
291-3.3K-RC
660-CFS1/4CT52R333J
660-CF1/4C683J
291-5.1K-RC
660-CFS1/4C103J
660-CF1/4C104J
72-T70YB-100K
291-10-RC
71-CCF071M00GKE36
71-CCF07510RJKE36
(2)
(2)
(1)
(1)
(4)
(1)
(1)
(6)
(1)
(1)
20-2
Qty.
Item
Description
Mouser number
(1)
R31
100 Pot.
785-53C2100
(1)
R33
50K Pot.
785-53C250K
You may find a 10 turn pot for R33 (gap voltage pot.) advantageous especially with short OFF times and low
power. A suitable 10 turn pot. would be Mouser 6523590S1503l
(2)
R34,37
5K pot.
785-53C25K
If you elect NOT to use the recommended bank of stepped resistor for R35 then purchase the following pot.
(1)
R35 Pot.
50K pot
785-53C250K
If stepped fix resistors are chosen (in place of R35pot.) it is your responsibility to choose resistor’s values
based desired accuracy see page 4-3.
A suitable rotary switch for selecting the stepped resistor is listed below
(1)
Rotary switch for ON time selections part number 105-SR2611F-17-18NS
(1)
R36
1meg pot.
78553C21MEG
(1)
R38
22K 2W
594-5083NW22K00J
(8)
R40-47
25ohm 50W
71-HL50-06Z25
NOTE: all switches except for SW3,4&5 may be purchased from Marlin P. Jones at considerably less than the
Mouser price. The Marlin P. Jones switches are 15A. imported from China, see page 5-2
(3)
SW1,2,8
SPST 3A
103-0132-EVX
Marlin#
16082 SW
(2)
SW3,4
SPST
103-1016-EVX (momentary contact for servo head)
(1)
SW5
DPDT
611-7201-001 (miniature toggle switch for servo head)
(1)
SW7
15A SPDT
103-13293-EVX (ON-OFF-ON) Marlin#
16082 SW
(6)
SW9-14
15A SPST
103-0010-EVX
Marlin#
16082 SW
(2)
SW15,16
15A DPDT
103-0013-EVX
Marlin#
16087 SW
Transformers T 1,2,3, are the single most expensive items in the power supply Marlin P. Jones Co. has the
least expensive transformers by far, see page 5-7 . The Marlin part number is below after the Mouser number.
(3)
T1,2,3
10 amp
553-F401U
Marlin#
7846-TR
T4 is a control transformer; I suggest ordering transformer T4 from the supplier of the larger transformers.
(1)
T4
2amp
553-F218X
Marlin#
7839 TR
(1)
U1
LM7812CT
512-LM7812CT
(2)
U2,5
555
512-LM555CN
(1)
U3
LM339N
863-LM339NG
(1)
U4
SN754410NE 595-SN754410NEE4
(1)
U6
LCA110
849-LCA110
(6)
Knobs
5164-1500
or builder may choose knob of their choice
(1)
6 Pin Dip
575-191306
(2)
8 Pin DIP
571-26412601
(1)
14 Pin Dip
855-D2814-42
(1)
16 Pin Dip
855-D2816-42
(1)
heat sink U4 532-580200B00 clip on
(1)
heat sink U1 532-577202B00
(13)
TB1-26
Term. Block 845-MBE152
(1?)
Straight header 5mm pitch (.200) Mouser does not have what’s needed but there are options.
Mouser 534-8727 is the correct pitch, but 2 pins short of what is needed. If wires are connected to TB 23-26
then this strip is perfect. Of course you may purchase a second strip for the missing 2 pins.
Another option is to use Mouser 649-69190-236HLF (order quantity 2) cut to length and remove every other
pin. Remove pins by heating with a solder iron, and pulling them out with pliers.
20-3
The following is a parts list for non-electronic items
XY Table
http://www.use-enco.com Enco model #201-2536 see page 9-2
http://grizzly.com Grizzly models are G8750, G5757 and H7979 see page 9-2
Dielectric tanks
http://www.usplastic.com
US Plastic part number #51081 see page 11-1
US Plastic part number #51083 see page 11-1
Thru Hull connectors
http://www.jehmco.com/html/bulkhead_fittings.html see page 13-1
Misc. 1/2” PVC and fittings home improvement centers see page 13-3
Pump
http://www.grainger.com Dayton model 2HNN9 see page 11-4
Speed control see page 13-7
Filters/mounts
http://www.jegs.com Jegs High Performance.
Single mount 969-1413see page 10-4
Dual mount259-15707 see page 10-4
Filter Bags see page 10-5
Flow control see page 13-5,6
http://www.loc-line.com Loc-line also available from industrial suppliers see page 13-6
Dielectric see page 10-3
Rustlick’s EDM-30, 100% petroleum based
Hangsterfer’s Crystal Brite, 50% petroleum and 50% synthetic
Hirschmann’s IonoPlus, 100% synthetic
Synthetic Lubricants Sinksyn-1, 100% petroleum-based
Head
Drawer Slide Knape & Vogt KV1129PZC16 see page 6-5
DC gearmotor Pittman P/N: GM8712F191 or similar see page 6-2
Multi conductor wire see page -11
Miscellaneous items the following (as well as many other non-electronic items) may be purchased locally or
from on line sources. McMaster Carr is very fast shipping and is almost one stop shopping for such items.
http://www.mcmaster.com
Aluminum plate, all-thread rod, bilge hose, bolts & screws, Cold Roll Steel (CRS), Delren, Loc-line pins, JB
weld, key stock, MDF, light dimmer switch, precision dowel steel tubing, vice, etc.
WWW.homebuiltedmmachines.com has an available PCB for building the EDM described
in this book. Visit the above web site for availability and cost
20-4
Tool and Manufacturing Engineer’s Handbook
Fourth Edition, Vol. 1: Machining
EDM
SME Dearborn, Michigan
Radio Shack’s
Engineer’s Mini-Notebooks
Forrest M. MimsIII
Internal Rectifiers Power Mosfet
Hexfet Databook 1985
IRF El Segundo Ca 90245
Dan Poynter’s
Self-Publishing Manual
Santa Barbara Ca.
I would like to credit the EDM operators at the Puget Sound Navy Shipyards in Bremerton, WA.
Unfortunately I have long forgotten most of their names, but they spent many hours educating me
about EDM and loaning different electrode materials to experiment with.
Also the Puget Sound Technical Library (on base), is where I spent many hours reading and
photocopying every technical publication (prior to WWW) that I could get my hands on. If my
command (NRMTF) only knew where I went when I was “borrowing and returning tools”!!
Thanks, guys!!
And, lastly, but far from least, My Heavenly Father for the guidance, inspiration and determination to
see this project through to this point. Without Him none of this would have been possible.
20-5