0
\$\begingroup\$

I'm currently trying to regulate the voltage on an inductive load. The idea is to use an opamp to divide the actual voltage across the coil down, compare it to a set voltage and drive a FET so that the desired voltage is reached.

This is my current design. The first try was without the push-pull buffer but adding it (because the OP can't deliver much current) didn't really improve things. I know that I'm actually measuring the voltage across the transistor, not the coil, but since they're in series, it should have the same effect.

Current through the coil will be between 1 and 1.5A, the coils DC Resistance is 10.5 Ohms.

schematic

simulate this circuit – Schematic created using CircuitLab

The circuit works in simulation but is heavily oscillating in reality. When I turn the potentiometer, the "duty cycle" of the oscillation changes, but it only ever stops oscillating if the control voltage is either 0 or 5V.

It looks like this on the scope, where the pink trace is the control voltage on the - input of the OP, yellow is the divided down voltage across the FET on the + input of the OP and blue is the output of the OP:

enter image description here

I would be very greatful if someone could point me in the right direction. I don't know, what else I could try.

Thanks in advance

Edit: I kind of improved the situation by leaving out the push-pull stage and conencting the opamp output via a ~200kOhm series resistor to the FET gate. The circuit looks like this now: The first stage calculated the differential voltage across the coil and divides it by approximately 7 so it can fit in the 5V rails of the opamps.

schematic

simulate this circuit

It does what it's supposed to, now, at least somewhat reasonably. Any hints on how to reduce the ringing on the step response would be much appreciated though. (Blue trace is a step on the control voltage, yellow is the response) enter image description here

\$\endgroup\$
  • \$\begingroup\$ if you try to keep a DC voltage, say +7,5V staying stable at the drain of the M1, your load must have substantial DC resistance to prevent some enormous DC current which draws the PSU on its knees.. How many ohms you have hidden in L1 and what is the intended voltage range at the drain of M1? How many amps you expect as Id for M1? In addition there are the common control system stability problems. We can sort them out after seeing the solution is possible in terms of volts, ohms and amps. \$\endgroup\$ – user287001 Dec 31 '17 at 15:08
  • \$\begingroup\$ Hi @user287001, L1 has 10.5 Ohms DC Resistance, the current range is between 0 and 1.5A, so with an operating point of 1A, there will be around 9.5V left at the Drain of the FET. I calculated the power dissipation curve and mounted the FET on an appropriate heat sink. I'll add this information to my original post, thanks for pointing it out. (I also partially fixed the problem by leaving out the push-pull buffer and adding 200kOhm in series between the op output and the FET Gate, Ill put a screenshot in my original post, maybe you can help interpreting it and further improving it) \$\endgroup\$ – Felix S Dec 31 '17 at 15:19
  • \$\begingroup\$ Replace the the load with a fixed 10 Ohm resistor to see, if your control circuit rings. Be sure that you have both GND clips connected in your scope probes (=no long GND routes, which act as resonant coils). Have you some hidden capacitance, too in the load? That would make the circuit resonant. If you inductance 220mH is written right, the resonance frequency of the inductor probably is much lower than the ringing frequency assuming even 500pF self capacitance. \$\endgroup\$ – user287001 Dec 31 '17 at 17:14
  • 1
    \$\begingroup\$ Try reducing the gain of OA4 to something like 100. \$\endgroup\$ – Spehro Pefhany Dec 31 '17 at 18:10

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.