Motion Control Architecture
Mini-Review
Contributors: Ed Wetherell, Kevin Tsubota, Jason Chin
11 Mar 2010
Schedule
The agenda for the review is as follows (times are HST):
• 8:00 AM: Welcome and introductions
• 8:10 AM: Presentation
• 9:15 AM: Break
• 9:30 AM: Review Comments/ Open discussion
• 10:30 AM: Review committee closed session
• 11:00 AM: Review committee feedback to team
2
Agenda
•
•
•
•
Review Committee Charter
Scope of the Review
Requirements Compliance
Motion Control
– Types
– Locations
– Architectures
•
•
•
•
Design Status
Review Committee Comments
Summary of Concerns
Plans for PDR
3
Review Committee Charter
• Reviewers:
Don Gavel (UCO, chair)
Alex Delacroix (CalTech)
Tomas Krasuski (WMKO)
• Are the requirements understood?
• Does the proposed architecture satisfy the requirements?
• Is the architecture
– Complete?
– Technically feasible?
– Cost effective?
• Is the architecture sufficiently mature that it can be
developed to the PDR level by the 2nd Qtr of 2010?
4
Scope of the Review
• Motion control electronics architecture
– Control of actuated devices used throughout the various
NGAO subsystems
• Not in the scope of this review:
– Overall architecture of the NGAO control system or the
top-level design of the NGAO control system
– Motion control required for real-time wavefront correction
(DMs, T/T) under control of the RTC
– Software controls
– Effort Estimates
– Budget
– Schedule
5
Requirements Compliance
• Locating all of the relevant requirements has proven
challenging
– Every subsystem was reviewed for requirements relating to
motion control
• PD Requirements Review (Phase II) week of 1March,
continued into week of 7March
– Impact on motion controls has yet to be assessed
• More effort required to
– Verify all requirements have been identified
– Determine compliance
– Address deficiencies in requirements and compliance
• No areas of non-compliance have been identified
6
Motion Control Types (1)
•
Shutters
–
–
–
–
•
Simple in/out devices with very loose positional requirements
Actuators other than motors (e.g., solenoid, pneumatic, etc.) may be considered
Switches or hard stops may be used to define the positions, encoders not required
Knowledge of actual position when moving, although desirable, is not required
Low precision, non-tracking
–
–
–
•
A dichroic or fold, for example, that is either in the beam or out of the beam
Moved during configuration, not during an observation
Position with encoder
Medium precision, non-tracking
–
–
–
•
Higher precision, still primarily single axis devices
Moved during configuration, not during an observation
Likely combine this category with Type 1 devices
High precision, non-tracking
–
–
aligning a lenslet or focusing a unit
moved during configuration or acquisition, not during an observation
7
Motion Control Types (2)
•
Tracking
– position calculated from and synchronized to external information (telescope az/el, etc)
– servo loops closed during an observation, command rates of 25mS to 100s of seconds
– generally more stringent requirements on servo loop performance
• want smooth motion, small following error, minimal overshoot
– various levels of precision required
– ADC, rotators
•
Extremely high precision tracking and non-tracking
–
–
–
–
•
coordinated motion with other DOF(s)
may be constantly moving during an observation, update rates of 1Hz or faster
generally requires a high precision actuator, not a servo motor
examples include steering mirrors and tip/tilt stages
Pickoff arms - coordinated high precision non-tracking
–
–
–
–
–
most demanding DOF
position calculated from and synchronized to external information (telescope az/el, etc)
coordinated motion with other DOF(s)
spatial position constraints to avoid collision
mechanical design may require the device to servo in position
8
Location Overview
BGS
(f/15 module)
[16 DOF]
AO Bench
(Cooled)
[38 DOF]
Laser
Switchyard
(El Ring)
[13 DOF]
Off AO Bench
(Not Cooled)
[16 DOF]
Current NGAO System Total: 54 AO + 29 Laser = 83 DOF
Original Estimate: 150+ DOF
Existing K2AO system: 29 AO + 22 Laser = 51 DOF
9
LGS WFS [13]
Low Order WFS [9]
LOWFS unit focus (z)
NGS WFS [9]
NGS WFS FSMs (θ,Φ)
Laser System [29]
Star imager pickoff (x)
Laser Asterism rotator (θ)
[2] [1]
Type 0 - Shutter
Type 1 - Low precision, non-tracking
Type 2 - Med precision non-tracking
Type 3 - High precision non-tracking
Type 4 - Tracking
Type 5 - Extreme High precision
Type 6 - Pickoff arms
[6]
Type 5 [18]
[6] [1] [2] [1]
Type 4 [20]
[2] [1] [1]
Type 3 [8]
[2]
Type 2 [16]
[1]
Type 1 [5]
[1]
Type 0 [4]
Color Codes: Blue - Cooled AO bench, Green - Off-bench AO device,
Brown - Laser enclosure and Gold - Telescope secondary.
LTO Polarization sensor (x)
[3]
LTO Cover (x)
Beam Expander Focus (x)
Laser Asterism Tip/Tilt (θ,Φ)
[3] [3]
Patrolling LGS steering (x,y)
[1]
Alignment beamsplitter (x)
[1]
Patrolling LGS beam splitter (x)
[2]
Polarization 1/2 waveplate (θ)
[4]
Pointing/steering mirrors (θ,Φ)
[6]
Polarization 1/4 waveplate (θ)
[1]
Science Instrument ADC assy (x)
[1]
Science Instrument ADC (θ,θ)
IR ADC [3]
NGS WFS Assy Focus (z)
[3]
NGS WFS Camera Focus (z)
[2]
NGS WFS Lenslet (x,y)
[1]
NGS WFS Dichroic (x)
[1] [1]
Instrument Select (x)
[6]
HODM slow tip/tilt (θ,Φ)
[1]
High-Order Relay [3]
[6]
LOWFS pickoff (θ,Φ)
[1]
LGS WFS Assy focus (z)
LGS WFS PnS slow Tip/Tilt (x,y)
LGS WFS Pickoff (θ,Φ)
[1] [4] [1] [3]
Acquisition Camera Focus (z)
Acquisition Camera Fold (y)
IF Dichroic (y)
Input Image Rotator (θ)
Turbulence Generator (x,θ,θ)
Source Positioning (θ)
Astrometric Source (grid) (x,y,θ,x)
Flat Field / Star Simulator select (x)
Cal/Sim Injection Fold (x)
Hatch cover (x)
Low-Order Relay [17]
Device Summary by Location
Type 6 [12]
10
Motion Control Architecture (1)
•
Centralized
–
–
–
–
All components are rack-mounted in a single location
Individual cables flow from the rack to each DOF
Primary approach taken throughout the observatory
Pros
•
•
•
•
Familiarity
Straight forward heat / power management
Single starting point for troubleshooting
Potential for reuse of SW and/or HW
– Cons
• Cabling – lots of it to a single location
• Longer cables may exclude use of low cost PWM amplifiers
• Scalability due to space constraints
– Proposed use
• AO Electronics Vault
–
Most AO devices
• Telescope Secondary or Laser Service Enclosure
–
–
Laser Beam Generation System devices
Laser Switchyard devices if equipment in LSE
11
Motion Control Architecture (2)
•
Distributed
– Equipment located in close proximity to actuator
– Several options with varying amounts of distributed equipment
• Distributed amplifier, central controller
• Distributed controller and amplifier
• Smart motor: controller and amplifier integrated into motor
– Pros
•
•
•
•
•
Significant reduction in cabling effort
Very scalable
Possible improvement in servo bandwidth
Short cables allow use of lower cost PWM drives for some axes
Allows partial system reset which may reduce the recovery time
–
all stages may not require homing
– Cons
• Distributed thermal loads
• Troubleshooting requires knowledge of physical layout with multiple device locations
• Integration with E-stop system
– Propose use of smart motors for low/moderate precision non-tracking devices
12
Design Status (1)
•
Controllers
– Use of programmable, multi-axis controllers with Ethernet
• Delta-Tau PowerPMAC a possibility
– Use of coordinate systems for multi-axis stages
– Prefer distributed control system to communicate via engineering units (mm, field
position), not encoder counts
• Translation at the controller level
•
Actuators
–
–
–
–
–
–
•
must operate in specified environment: -15°C in cold box, -10°C on secondary
DC motors (brush or brushless)
Drive or load position encoding, precision dependant
Precision actuators (piezo t/t stages, linear piezo, voice coil, etc)
Smart motors
Normally closed (open when triggered) end of travel switches
Crates / processors
– Separate embedded processors (VxWorks) likely not required
– Higher level server to handle communications and monitoring functions
13
Design Status (2)
•
Observatory E-stop interface
– Part of the motion control requirements
– Will likely remove power to all motion amplifiers
– Preferably preserve encoder/limit switch power to easy recovery
•
Cables will need to remain flexible at operating temp
– Combine multiple axes on single cable where possible
– -15°C in cold box, -10°C on secondary
– Attention needed on pickoff arms and devices mounted to an in/out or focus stage to
provide adequate range of travel
– Motors supplied with an integral cable could be a challenge
•
Instrument mechanisms
–
–
–
–
Not part of NGAO motion control design
Imager+IFS (DAVINCI) has its own motion controls
Interferometer DSM supported by IF ancillary electronics rack and IF control system
Interferometer will have devices on AO bench
• Devices will be controlled by NGAO to eliminate any impact on non-IF AO observing
• Details not known
14
Proposed Layout of AO Controls on
Left Nasmyth
Interferometer
Electronics
Rack
w/
Motion
Controls
Supervisory/Control
Electronics
Electronics Vault
Motion Contollers
Servo Amplifiers
Power supplies
(39 DOF)
Ethernet Switch
El Cable
Wrap
Ethernet
E-STOP
Protocol
conversion
Power
Supply
Mains Power
E-Stop
Interface
Power
Supply
Time Sync
Analog
I/O
Glycol
120 VAC
Power
Controllers
(6 DOF)
Az Cable
Wrap
Mains Power
Emergency
Stop
“Smart” Servo
Devices
Input Hatch (1)
IF Output Hatch (2)
Time Sync
Glycol
Calibration/Simiulation
Servo Motion Devices
Injection Fold (1)
Source Select (1)
Astrometric Source (4)
Source Positioning (1)
Turbulence Generator (3)
HODM
Slow T/T (2)
ADC Servo Devices
ADC Rotator (2)
LGS WFS
DAVINCI
AO COLD BENCH
LO Relay
“Smart” Servo Devices
Interferometer Dichroic (1)
Acquisition Fold (1)
Acquisition Focus (1)
“Smart” Servo Device
ADC In/Out (1)
Instrument select (1)
NGS WFS
“Smart” Servo Devices
Dichroic (1)
LOWFS
Servo Motion Devices
TT/TTFA Pickoff arms (6)
TT/TTFA Focus (3)
NGS WFS
Servo Devices
FSM (4)
Lenslet (2)
Focus (2)
non-Servo Devices
PnS Pupil Pointing (6)
Servo Motion Devices
LGS WFS Pickoff arms (6)
LGS WFS Focus (1)
Imager+IFS
Interferometer Devices
LO Relay Servo Devices
Input Image Rotator (1)
Motion Devices
Servo Motion Devices
IF DSM
AO CLEAN
ROOM
Ethernet
non-Servo Devices
15
Proposed Layout of AO Controls on
Left Nasmyth
•
•
Multi-axis servo controllers located in e-vault
Smart motors used for intermittent, low or moderate precision devices
–
–
–
–
–
•
32 servo devices on cold bench
–
–
•
anticipate ~15 W (total, 0.5 per device) of power for encoders alone
some DOF in the Cal/Sim unit may have dedicated controllers
13 devices in AO Room
–
•
6 of these motors inside the cold box, 3 outside
Continuous power dissipation of ~0.9 W each
Stages must accept NEMA frame motor
Requires power supply and terminal server port
estimate ~$400/axis savings (procurements) over centralized approach
7 servo motors, 6 piezo (or equivalent)
Cabling
–
Anticipate around 16 cables between cold box and e-vault
•
•
•
Interface with facility Emergency Stop
–
–
•
Assumes 2 DOF per servo motor cable
Cold box will have bulkhead connectors; environmental seal but not hermetic
Prevent motion to protect personnel and equipment
Keep encoders/limits powered if possible to speed recovery
Volume ~ 22U (half 7 foot rack), similar to existing AO system
16
Proposed Layout of Laser devices w/
controllers in LSE
Beam Generation
Servo Motion Devices
PnS beamsplitter (2)
Star Imager Pickoff (1)
Asterism Rotator (1)
Beam Expander Focus (1)
Polarization Sensor (2)
Launch Telescope Cover (1)
Piezo Device
Asterism Tip/Tilt (2)
Protocol
conversion
Ethernet
Ethernet
E-STOP
Mains Power
Time Sync
6x HV Drive
Linear Piezo
Controllers (6)
Power
Supply
1x Strain
Gauge
4x Servo
Piezo Controllers
(8 DOF)
Analog
I/O
120 VAC
Power
Glycol
Heat
Exchanger
1x HV Drive
Motion Contollers
Servo Amplifiers
Power Supplies
(15 DOF)
SYD Piezo Device
Pointing/Steering (6)
Power
Supply
SYD Servo Motion Devices
Polarization Waveplates (6)
Alignment beamsplitter (1)
(Elevation Ring)
6x Encoder
Laser Enclosures
El Cable
Wrap
(Telescope Secondary)
Linear Piezo Devices
PnS Steering (6)
17
Proposed Layout of Laser devices w/
controllers in LSE
• Motion controllers for laser switchyard and beam generation
system located in LSE
• Pro:
– Limited requirement on Elevation cable wrap
– Space, power and glycol are available in LSE
• Con: cables between secondary and LSE
• Cabling
– Anticipate 18 cables between LSE and secondary
• 12 of the 18 for the linear piezo devices, may be able to combine into
fewer cables
– Only infrastructure cables (power, Ethernet, time sync, e-stop)
between LSE and e-vault
• Volume: 15U of 19” rack (~0.12 m^3)
18
El Cable
Wrap
Power
Supply
Motion Contollers
Servo Amplifiers
Power Supplies
(10 DOF)
Ethernet
Piezo Controllers
(2 DOF)
Analog
I/O
Servo Motion Devices
PnS beamsplitter (2)
Star Imager Pickoff (1)
Asterism Rotator (1)
Beam Expander Focus (1)
Polarization Sensor (2)
Launch Telescope Cover (1)
Piezo Device
Asterism Tip/Tilt (2)
Beam Generation
120 VAC
Power
Linear Piezo Devices
PnS Steering (6)
(Telescope Secondary)
Heat
Exchanger
Linear Piezo
Controllers (6)
(Elevation Ring)
Protocol
conversion
Laser Enclosures
Alternate Layout of Laser devices w/
controllers on Secondary
Power
Supply
Glycol
Mains Power
Time Sync
Ethernet
E-Stop
HV Drive (3)
Strain Gauge (3)
Servo (4)
SYD Piezo Device
Pointing/Steering (6)
SYD Servo Motion Devices
Polarization Waveplates (6)
Alignment beamsplitter (1)
19
Alternate Layout of Laser Devices w/
controllers on Secondary
• Controllers for BGS devices located on secondary, Switchyard
controlled from e-vault
• Pro: Minimal cabling to secondary
• Con:
– Space and mass on secondary
– Will require (custom?) cooled enclosure for electronics
– LSE cables in elevation wrap
• Cabling
– Requires infrastructure cables (power, Ethernet, time sync, e-stop)
between secondary and e-vault
– 10 motion cables between LSE and e-vault (el wrap)
• Volume: 10U of 19” rack (0.08 m^3)
20
Alternate Layout of Laser devices w/
all controllers in E-vault
•
•
Controllers for all laser devices located in E-vault (not illustrated)
Pro:
– Heat/power management
•
Con:
– Some equipment still likely in LSE or secondary
• Would require cooled enclosure
– 40m cables from secondary to e-vault, through elevation wrap
• Devices must meet spec
• Voltage drop across cable must be managed
• Differential drivers required for encoders
–
Actuators with single-ended encoders would require design of custom driver board
– LSE cables in elevation wrap
•
Cabling
– 18 cables between secondary and e-vault
• may be able to reduce this to 12
– 10 cables between LSE and e-vault
•
Volume: 15U of 19” rack (0.12 m^3) in e-vault
21
BREAK
22
Review Committee Comments(1)
• Mahalo to everyone for your detailed review
of this material
• Revision to documents forthcoming
23
Review Committee Comments(2)
• Need definition of terms
– Reviewers noted inconsistency of naming
– Project needs to publish preferred names and definitions
• Need agreement on required controls (tracking) and precision
of devices
– Perhaps some differing ideas about definition of tracking
• Servo loops closed during observation
• Recycling of existing equipment (OBS)
– Not likely despite apparent compatibility
– Much is obsolete or would require upgrade
• Concern about smart motors inside cold enclosure
– Not a significant heat source, compared to traditional motors
– Some in-house testing is required to verify manufacturer claims
– Changing the design to conventional servo motors straight forward
24
Review Committee Comments(3)
• Comment regarding controls split between AO and IF in current system
– NGAO will control all devices on AO bench
– NGAO will control all hatches
• Eliminate problems in existing distribution of control
• NGAO will need the output hatches closed during observing (and presumably
daytime prep/stabilization)
– NGAO is required for IF observing, the converse is not true. IF should not be
required for NGAO to work
• Is there any allocation for expansion?
– Not in the strict sense of x% free channels
• Need guidance from the project
– Adding smart motors to the ring is easy
• Probably a limit based on communication bus speed
– Depending on choice of controller, expansion would only be limited by
available rack space
25
Review Committee Comments(4)
• Type and amount of diagnostics was questioned
– This needs work.
– At present, no additional hardware is anticipated
• Heat analysis of Cold Enclosure
– Estimate of steady state load provided
• ~20W of encoder
• Active limit switches have negligible contribution
– Active state much harder to predict
• Need payload information from subsystems designers to estimate required motor
power
• Need duty cycle information for devices with intermittent motion
– Work with Mechanical engineers to estimate thermal constants of enclosure
– Need analysis of how thermal gradients , bench ‘seeing’, local hot spots, etc.
impact performance
26
Review Committee Comments(4)
• Reviewer suggests a survey of existing motion systems for ‘likes’ and
‘dislikes’
– Worthwhile
– Some of this is already included given the experience of the team on both AO
an IF
• Reviewer comment on missing reference on pg13 of KAON 715
– Typo, should be KAON 643 section 7.6 (not 6.7), will be corrected
• Reviewer responded to concerns about probe arm limit switches
–
–
–
–
Updated design that uses load cells to determine direction of travel
This helps recover from a limit condition
Concern about interface to motion control system
Still a concern about homing these stages
• Reviewer concern about flow-down requirements listed in KAON 715
– These are flow-down, not functional requirements
– Not aware of a decision to manage this type of requirement in Contour
27
Reviewer Feedback
• Any questions?
• Detailed responses to individual comments
are (or will be) posted on the TWiki
28
Summary of Concerns (1)
• Better collaboration between subsystem
design teams
• Need agreement on required controls and
precision of devices
– Need completion of Master Device List with all
relevant information
• Need better understanding of pickoff arm
controls (homing)
29
Summary of Concerns (2)
• Nearly every subsystem requires motion control
– Insufficient detail on some subsystem designs
• Need to understand goals for DD and I&T
– Which team is responsible for what
• How much duplicate equipment is required in California
• What level of performance validation is performed by subsystem
designers
• Understand cabling requirements
– Clean enclosure
• work out baseline for connectors/cables
– Telescope cable wraps
30
Plans for PDR
• Work with Systems Engineering to get a complete
approved set of requirements
• Work with subsystem designers to complete the
Motion Control design
• Decide on location of laser control electronics
• Provide estimates for power/volume/mass
• Maintain KAON 682 (Master Device List)
• Complete KAON 715 (Preliminary Motion Control
Design)
• Identify risks and mitigation plans
• Budget and Schedule
31
Review Comittee Session
• Given the short time to the PDR, we request
an informal report via email, rather than a
formal write-up
– this will decrease the turn around time and limit
additional effort required by the reviewers
• If the reviewers prefer a formal write-up,
please provide this within a week
32