Proceedings of the ASME/BATH 2014 Symposium on Fluid Power & Motion Control
FPMC2014
September 10-12, Bath, United Kingdom
FPMC2014-7846
The Impact of Peak-And-Hold and Reverse Current Driving Strategies on the Dynamic Performance of
Commercial Cartridge Valves
Farid Breidi
Purdue University
West Lafayette, IN, USA
Tyler Helmus
Purdue University
West Lafayette, IN, USA
Michael Holland
Purdue University
West Lafayette, IN, USA
John Lumkes
Purdue University
West Lafayette, IN, USA
ABSTRACT
INTRODUCTION
High speed valves have an important role in many existing
fluid power systems and are an enabler for many proposed
digital hydraulic systems. One method commonly used to
improve the dynamic performance of on-off valves involves
modifying the electrical input signal to the solenoids to reduce
the inductive lag and eddy current decay. This research
examined two commercially available direct actuated and
pilot-stage actuated cartridge poppet valves and the role of
peak-and-hold voltage and reverse current input profiles on
opening and closing switching times. A test stand was built to
characterize the performance of these valves. The valves were
placed between two high frequency pressure transducers and
the pressure differential across the valves was recorded,
allowing the calculation of transition and delay time. The peak
and reverse voltage duration was tested over a range of zero to
ten milliseconds and an optimum response was found at a
peak duration of six to eight milliseconds. Peak voltages
ranged from 50 to 55 volts, followed by a holding voltage of
12 volts. Reverse current profiles were used to turn off the
valves with a maximum peak current of three amps. The
reverse current was used to increase the decay rate of eddy
currents thus improving the turning off performance of the
valves. Commercial valves that had a range of 33 to 55
millisecond turn-on response without input signal
modification; these same valves had response times reduced
to a range of seven to nine milliseconds after applying the
peak and hold method. The turn-off time was reduced from
130 milliseconds to a range of 16 to 50 milliseconds after
adding reverse current inputs. This improvement in valve
performance can lead to siginificant energy savings due to
reduction of transition losses and can widen the useful
application of the valves.
In the United States, conventional hydraulic systems are
inefficient, averaging 22% in overall efficiency, with losses
totaling more than all renewable energy sources combined [1].
Valve throttling losses on an excavator can account for as
much as 43% of the total energy consumed in the system [2].
Research into more efficient system topologies is being
pursued and includes displacement controlled actuation,
hydraulic transformers, and independent metering valves [36]. The need for more efficient fluid power systems is
paramount, yet research is limited by the performance of the
components available.
On/off valves are common components in many hydraulic
systems. As these systems are developed and their functions
become more complex in the pursuit of efficiency, the need
arises for valves with increased performance. The biggest
problem with commercially available on/off valves is slow
and varying response time [7]. These types of hydraulic
system applications may include camless engines, active
vibration control, single valve virtually variable displacement
pump/motors, multi-valve virtually variable displacement
pump/motors, high speed selection control, hydraulic
transformers, improved actuator control, and pulse width
modulation (PWM) control. Specifically, high performance
valves are essential in the development of four-quadrant
digital pump/motors. The efficiency of a digital pump/motor is
directly related to the reliable, rapid response of the valves
being used. If the repeatability of the valves varies by even 2
ms, the throttling losses experienced can be significant [8,9].
Commercially available valves are relatively slow when
precise actuation is needed and thus much research has been
done to improve valve performance. Such research includes
1
Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 11/25/2014 Terms of Use: http://asme.org/terms
Copyright © 2014 by ASME
piezoelectric actuators, single solenoid actuators, dual
solenoid actuators, use of ferrite, soft, and hard magnetic
components, linear and rotary stepper motors, voice coil
actuators, shape memory alloys, dual impacting moving
masses, torque motors, and self spinning rotary actuation [10].
a
solenoid
circuit
previously
open.
1
Current (unitless)
This paper highlights results from the testing of the peak-andhold and reverse current strategies on two types of Sun
hydraulic solenoid valves. In order to perform these tests, a
test circuit was constructed and multiple valves of each type
were tested using identical control and data acquisition
procedures.
0.5
0
-0.5
BACKGROUND
To improve the opening of these normally closed valves, the
peak and hold strategy was implemented. This involves
sending high initial voltage and current to overcome
inductance and eddy current lag while generating high flux
levels across the air gap. After these effects have been reduced
by the peak voltage, a holding current is applied to the
solenoid to keep the armature in place without expending
undue energy to resistive heating. It should be noted that the
transition time of the valve is not effected by a peak duration
that is longer than optimal.
-1
-0.5
0
0.5
Time (unitless)
1
1.5
Figure 2: Reverse current applied current versus time
Batdorff developed a theoretical equation (Eq. 1) for the decay
of the magnetic flux density (B) when a reverse current is
applied [10]. While normal decay is realized proportional to
the zero applied magnetic field and time, adding the reverse
pulse adds the multiplier of one plus the magnitude of the
reversed pulse, greatly reducing the decay time realized in the
solenoid.
Figure 1 shows the peak and hold strategy implementation in a
solenoid circuit previously closed.
B( z , t ) 
1
 (1  mr )  z, t  t  0     z, t  t  tr 
Initial Steady State
2.5
Reversed Pulse
Eq.1
Zero Applied Magnetic Field
Current (unitless)
2
Within this equation, z is the distance into the plate, tr is the
normalized duration of the reversed pulse and mr is the
relative normalized magnitude of the reversed pulse. δ is the
unit step function which is 1 when the first term is greater than
or equal to the second and is 0 otherwise. β is the the diffusion
of magnetic flux from one side into a plate in response to a
step magnetic field intensity change. Figure 3 shows the
theoretical normalized comparison of magnetic flux effusion
and reverse current effusion.
1.5
1
0.5
0
-0.5
0
0.5
Time (unitless)
1
An important factor that must be taken into account when
using the reverse current method is that the magnetic field can
be reestablished by the reverse current resulting in an increase
or less than optimal transition time if it is left on for longer
than necessary. The proper length of the applied reverse
current will first quickly wind down the forward current and
then counteract the lingering eddy currents and residual
magnetism. This results in a critical pulse duration for optimal
transition time of the valve beyond which the reverse current
hinders valve closing transition time. This point and the
reestablishment of the magnetic flux can be seen graphically
1.5
Figure 1: Peak and Hold applied current vs. normalized time
The reverse current turn off method is more complex than the
previous method. Valve closing lag is mainly due to lingering
current in the solenoid and residual magnetism which counter
the closing force of the spring. The reverse current method
decays these residual effects more quickly than a flyback
diode [10]. Figure 2 shows reverse current implementation in
2
Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 11/25/2014 Terms of Use: http://asme.org/terms
Copyright © 2014 by ASME
can be controlled by manipulating the switches. These states
are shown in Table 1. Forward current and voltage can be
achieved by closing switches 1 and 4 and opening switches 2
and 3, while reverse current and voltage could be achieved by
closing switches 2 and 3 and opening switches 1 and 4. Off
states can be achieved by opening all switches, opening
switches 1 and 2 and closing switches 3 and 4, and closing
switches 1 and 2 and opening switched 3 and 4.
in Figure 3. This critical pulse duration is a function of
forward current, supply voltage, and material properties. This
is dissimilar to the peak and hold strategy where the penalties
for a longer than necessary peak duration include loss of
efficiency and potential overheating of the coil due to
increased power drop over the coil when the solenoid is fully
turned on, but does not include an increase in the transistion
time.
Normalized Total Magnetic Flux ()
1
Table 1: H-bridge states
ON
Normal Effusion
0.5
Reverse Pulse Effusion
Zero Net Magnetic Flux
0
Critical
Pulse
Duration
-0.5
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Normalized Time (t/MDT)
0.8
0.9
Off
Switch
Forward
Reverse
Case 1
Case 2
Case 3
1
Closed
Open
Open
Open
Closed
2
Open
Closed
Open
Open
Closed
3
Open
Closed
Open
Closed
Open
4
Closed
Open
Open
Closed
Open
Holland describes implementing a valve power electronic
circuit is shown in Figure 5 [11]. A LMD18200 H-bridge was
used for peak and hold and reverse current strategies; this Hbridge featured built-in logic and current sense output, as well
as 3A and 55V capabilities which were enough to apply the
peak and hold and reverse current strategy for a 12V coil.
High speed optocouplers were used to protect the control
system from the high peak voltages. Optocouplers invert the
input signal, so a 74LS04 hex converter was used to invert the
signal back.
1
Figure 3: Comparison of Dimensionless Magnetic Flux Effusion [10]
ELECTRIC CIRCUIT
An H-bridge circuit, shown in Figure 4, was implemented to
power the valves with either the peak and hold turn-on
strategy or reverse current turn-off strategy. The peak and
hold turn-on strategy can be implemented with simpler
circuits, but a full h-bridge is nessary for reverse current turnoff strategy.
Figure 4: H-bridge circuit
Figure 5: Valve power electronic circuit, Holland (2013)
The H-bridge consists of four solid state switches
(MOSFETs), so the polarity of voltage and direction of current
Pulse width modulation (PWM) and direction (DIR) pins were
used to achieve the peak, hold, reverse current, and off states.
3
Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 11/25/2014 Terms of Use: http://asme.org/terms
Copyright © 2014 by ASME
As shown in Table 2, a high signal at the PWM pin was used
to enable the H-bridge, a high signal was used for the peak
strategy for the duration of the peak required. A low DIR
sends the signal in the forward direction, while a high DIR
signal would achieve a reverse current. A modulated PWM
was used to hold at a constant voltage. An off state was
achieved by having a low DIR and PWM signal.
The differential pressure across the tested valve was measured
during opening and closing. As shown in Figure 8, the delay in
the valve opening was estimated by the duration from the
signal trigger to a drop of 10% of the difference between the
initial and final values of the differential pressures. The
duration of the drop between 10% to 90% in pressure
difference was used to estimate the transition time. Similarly,
the valve delay time during closing was estimated by the
duration from the trigger to a 10% rise in the difference
between the initial and final pressure values, while the
duration of the rise from 10% to 90% represented the
transition time. The 10% and 90% points were chosen to be at
the first instance the pressure reaches these values as opposed
to subsequent instances in an overshoot-settling condition as
this gave the most consistent results and permits test results to
be summarized in an automated fashion.
Table 2: Truth table for H-bridge circuit
State
Peak
Hold
Reverse Current
Off
Direction
Low
Low
High
Low
PWM
High
Modulated
High
Low
HYDRAULIC CIRCUIT
The circuit used to evaluate the response time of the valves is
shown in Figure 6. The valve was placed between two 2000
Hz pressure transducers measuring the pressures at ports 1 and
2. The fixed displacement pump was capable of providing
around 31 l/min at 124 bar. The pressure relief valve was used
to set the operating pressure and the needle valve was used to
limit the flow across the tested valve.
Figure 6: Valve response test circuit
Figure 8: Time estimation during turn-on response
Forward flow was defined as from port 1 to port 2 and reverse
flow was from port 2 to port 1 as labeled in Figure 7.
EXPERIMENT SETUP
Two normally closed cartridge valves were selected for the
experiment; a Sun Hydraulics DTDA-XCN valve with a 770212 12V coil and a modified Sun Hydraulics DTDA-XCN
valve with a 760-212 12V coil which uses a solenoid tube and
coil that are used in Sun Hydraulics DAAA valves. The
ratings of the coils used in both valves are shown in
Table 3. These valves were selected because of the possibility
of implementing one of them in a digital pump/motor
assembly. The experiments were conducted at a differential
pressure of 52 bars, 28 l/min flow, 12V holding voltage and
55V peaking voltage.
Figure 7: Port and Valve Diagram
4
Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 11/25/2014 Terms of Use: http://asme.org/terms
Copyright © 2014 by ASME
Table 3: Coil ratings
Coil
Rating
Supply Voltage (V)
Power Consumption
(W)
Maximum Coil
Temperature (oC)
Connector
770-212
12
760-212
12
22
12
105
105
ISO/DIN
43650A, Form A
ISO/DIN
43650A, Form A
National Instruments hardware was used for testing the valves.
A PXI-1031 chassis with a Field Programmable Gate Array
(FPGA) card were used. The peak and hold turn-on and
reverse current turn-off strategies were programmed in FPGA.
FPGA was also used to read the pressures from the pressure
transducers. The sensor calibrations were done in
Matlab/Simulink and then compiled into NI Veristand which
was used to interface the FPGA and Matlab/Simulink and
provide the user with a control interface.. The Veristand panel
allowed the user to specify the peak and reverse voltage
duration for both turn-on and turn-off response. Holding
current was controlled with the duty cycle of the PWM signal.
Figure 9: Modified DTDA-XCN turn-on response at 10 milliseconds
voltage peak
RESULTS
Figure 9 and Figure 10 show the turn-on and turn-off response
for the modified Sun Hydraulics DTDA-XCN valve under
peak voltages. The signal was sent at time zero, a delay in
valve opening and closing was recorded to be 5.9 ms and 19.9
ms respectively. The transition time for opening was estimated
to be 2.2 ms with a 10 ms voltage peak duration, while the
transition time of closing was estimated by 4.8 ms at 5 ms
voltage peak.
Figure 10: Modified DTDA-XCN turn-off response at 5 milliseconds
voltage reverse peak
Both valves were tested in forward and reverse flow, while
opening and closing under peak voltages ranging from zero
milliseconds to ten milliseconds peak with an increment of 1
ms peak. The experiment was repeated three times under the
same conditions, the averages for the calculated response
times are presented below.
5
Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 11/25/2014 Terms of Use: http://asme.org/terms
Copyright © 2014 by ASME
Figure 11 and Figure 12 represent the turn-on response for the
DTDA-XCN and the modified DTDA-XCN valves
respectively. The t10 value represents the time the valve needs
to drop or increase by 10% of the pressure difference, and is a
measure of the the valve delay time, while ttrans represents the
time the valve needs to drop or increase from 10% to 90%,
which is a measure of the transition time of the valve. Both the
modified and unmodified valves had the fastest forward
response at a peak duration higher than five milliseconds,
while the fastest reverse response was found at a peak higher
than seven milliseconds. The delay and transition times for
forward flow were brought down from 16.8 ms and 10.0 ms to
around 3.8 ms and 2.4 ms respectively, for the DTDA-XCN
valve, while it was brought down from 16.8 ms and 37.7 ms to
around 5.3 ms and 2.4 ms for the modified DTDA-XCN valve.
As for the reverse flow, the delay and transition times were
brought down from 15.4 ms and 25.4 ms to around 4.5 ms and
2.5 ms respectively, for the DTDA-XCN valve, while it was
brought down from 21.2 ms and 12.1 ms to around 3.9 ms and
2.6 ms for the modified DTDA-XCN valve.
Figure 12: Modified DTDA-XCN turn-on response
For both the forward and reverse flow there was a
considerable improvement in both the delay and transition
times under higher peaking voltage. No more improvement in
the delay time is noticed when the peak voltage duration is
larger than the delay time, which is trivial because the excess
peaking would occur after the transition phase has ended, so
no further improvement in delay time could occur. However,
the transition time reaches its optimum value when the peak
duration is equal to the summation of both the delay and the
transition time, where the excess peak would be acting for
holding and not for improving the transition since the
transition phase would have already ended.
The turn-off response times for the DTDA-XCN valve and the
modified DTDA-XCN valve are shown in Figure 13 and
Figure 14 respectively. Notice that the delay for the DTDAXCN in both forward and reverse flow drops from 152.6 ms
and 93.6 ms to reach its minimum with a delay of 21.6 ms and
15.6 ms at a peak duration of six milliseconds. The modified
DTDA-XCN valve shows a similar behavior with a drop in
delay in both directions from 76.3 ms and 127.9 ms to a
minimum of 18.0 ms and 42.9 ms at a peak duration of five
milliseconds. However, as the peak duration further increases,
the delay time increases. This is because the excess duration of
the reverse current would start to regenerate the magnetic
field, thus slowing down the valve when closing as anticipated
in Eq.1. The transition time was not improved during turningoff for both valves in both flow directions; this is because the
valve closing is based on the spring stiffness.
Figure 11: DTDA-XCN turn-on response
6
Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 11/25/2014 Terms of Use: http://asme.org/terms
Copyright © 2014 by ASME
strategies, but stayed relatively constant during closing
(reverse current turn-off strategy) because it is dependent on
the stiffness of the spring.
Additional electrical energy consumption would be needed to
implement these strategies, this energy would be dissipated in
the coil and armature leading to heat generation. This heat
generation can be minimized by precise timing of the peak and
hold and reverse current strategies determined by testing of
valve performance for different signal durations. Though
additional energy is needed, no physical or mechanical
alterations to the valves are required to achieve these results.
Many applications would benefit from this strategy, especially
those in which additional energy consumption and heat
generation wouldn’t be a problem. This strategy could be used
in many digital hydraulic systems where valve timing and
accuracy is a need such as digital pump/motors.
ACKNOWLEDGMENTS
Figure 13: DTDA-XCN Turn-off response
We would like to thank Sun Hydraulics for their help
customizing and donating a new set of valves for our needs.
This research was supported by the Center for Compact and
Efficient Fluid Power, a National Science Foundation
Engineering Research Center funded under cooperative
agreement number EEC-0540834, and by the National Fluid
Power Foundation.
REFERENCES
[1] Love, Lonnie. (March 5, 2014). Energy Impact of Fluid
Power [Powerpoint Slides]. Presented at the 2014 IFPE
Technical Conference, Las Vegas, NV.
[2] Love, Lonnie. (August 17, 2009). Fluid power research:
A fundamental concern for U.S. energy policy
[PowerPoint slides]. Presented at the National Fluid
Power Association Industry and Economic Outlook
Conference, Wheeling, IL.
[3] Williamson, C. and Ivantysynova, M. (2007). The Effect
of Pump Efficiency on Displacement-Controlled Actuator
Systems. Proceedings of the 10th Scandinavian
International Conference on Fluid Power, Tampere,
Finland.
[4] Bishop, E. D. (2010). Digital hydraulic transformer –
Efficiency of Natural Design. 7th International Fluid
Power Conference, Aachen, Germany.
[5] Heikkila, M., Tammisto, J., Huova, M., Huhtala, K.,
&Linjama, M. (2010). Experimental evaluation of pistontype digital pump-motor-transformer with two
independent outlets. Proceedings of the Bath/ASME
Symposium on Fluid Power and Motion Control.
[6] Lumkes, J. and Andruch, J. 2011. Hydraulic Circuit for
Reconfigurable
and
Efficient
Fluid
Power
Systems,Proceedings of the 12th Scandinavian
Figure 14: Modified DTDA-XCN turn-off response
CONCLUSION
This work experimentally examined the effect of peak and
hold and reverse current strategies on the turn-on and turn-off
response of two Sun Hydraulic valves. Experimental results
show a decrease of more than 80% in turn-on response time
and more than 64% decrease in turn-off response in both
valves, for both flow directions when when using peak (55V)
and hold (12V) and reverse current strategies compared to
steady 12V input. The delay time was reduced in both opening
and closing phases for both flow directions. The transition
time for opening was improved under peak and hold voltage
7
Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 11/25/2014 Terms of Use: http://asme.org/terms
Copyright © 2014 by ASME
International
Conference
on
Fluid
Power,
SICFP (1096):16pg.
[7] Mikkola, Jaakko, Ville Ahola, Timo Lauttamus, Markku
Luomaranta, Matti Linjama, and Matti Vilenius.
"Improving Characteristics of On/off Solenoid
Valves." The
Tenth
Scandinavian
International
Conference on Fluid Power (2007).
[8] Merrill, K. J., 2012, “MODELING AND ANALYSIS OF
ACTIVE VALVE CONTROL OF A DIGITAL PUMPMOTOR,” Ph.D. thesis, Purdue University, West
Lafayette, In.
[9] Merrill, K. J., Breidi, F., & Lumkes, J. (2013).
“Simulation Based Design and Optimization of Digital
Pump/Motors”. Proceedings of the ASME/BATH 2013
Symposium on Fluid Power & Motion Control.
[10] Batdorff, M. A., 2010. “Transient Analysis of
Electromagnets with Emphasis on Solid Components,
Eddy Currents, and Driving Circuitry”. PhD Dissertation,
Purdue University, West Lafayette, In.
[11] Holland, M. A., 2012, “DESIGN OF DIGITAL
PUMP/MOTORS
AND
EXPERIMENTAL
VALIDATION OF OPERATING STRATEGIES,” Ph.D.
thesis, Purdue University, West Lafayette, In.
8
Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 11/25/2014 Terms of Use: http://asme.org/terms
Copyright © 2014 by ASME