3
\$\begingroup\$

I know this topic has been discussed a lot, but there seems to be varying opinions about the solutions..

The problem is we are trying to make a current source with an op-amp driving a FET as the load. Our goal is to have a fairly high bandwidth for the regulating loop(couple of 100kHz?), but the output of the op-amp is oscillating.

First iteration

Using a ADA4807 and a ST P36nE06 mosfet. The values for the series(R2) and parallel(R3) resistors were chosen from the op-amps datasheet about driving capacitive load.

There is a 100nF bypass cap very close to the supply pins, and a 2uF cap also.

We have tried a few things which includes:

  1. Upping the series resistor to 50ohm or 100ohm. Did not fix it.

  2. Adding 1nF cap from output to -vin of the op-amp, together with a 10k or 100k resistor (miller integrator). Did not help.

  3. Changing the 10mOhm resistor to 100mOhm. Did not help either.

The prototype is done on verroboard. I don't know if that could be an issue?

I would think this op-amp would be able to drive any capacitive load, but it doesn't seem to work for us. What are we doing wrong?

\$\endgroup\$
1
  • \$\begingroup\$ Comments are not for extended discussion; this conversation has been moved to chat. \$\endgroup\$
    – Dave Tweed
    Jul 17, 2018 at 20:04

1 Answer 1

Reset to default
2
\$\begingroup\$

The op-amp's capacitive load drive stated in the data sheet is 15pF giving rise to an overshoot of 30%. Yes, you can improve this by using a series resistor to drive the capacitor (The MOSFET is about 2 nF) but you are still driving capacitance within the closed loop feedback network and that makes adding a resistor pointless - you just shift the phase further and make it oscillate at a lower frequency.

In a closed loop situation like this you have to look at the op-amp's open-loop gain and recognize that it is never ideal and might have a phase margin that is quite poor at high frequencies. A poor phase margin means that negative feedback is close to becoming positive feedback and an extra bit of RC filtering (due to the gate-source capacitance) can easily tip the balance from stability to oscillation.

You need to put a capacitor from op-amp output to the inverting input and put a resistor (try 1 kohm) from that input (pin 4) to the feedback point on the current sense resistor. You may be lucky and get the bandwidth you want but there is no guarantee.

\$\endgroup\$
4
  • \$\begingroup\$ Hmm alright, it makes sense then why the series resistor didn't change anything then.. I believe the input capacitance of the mosfet is 2-3nF actually \$\endgroup\$
    – Linkyyy
    Jul 16, 2018 at 18:06
  • \$\begingroup\$ Step voltage incremental prop delay is 100ps/pF load and output impedance rises with slew rate. Yet the FET Ciss is very nonlinear. I would insert an Emitter Follower before the gate with Re=180 Ohms and keep all jumpers as short as possible. And use 10V step input reduced to 100mV with ~500:5 Ohm divider. \$\endgroup\$
    – Tony Stewart EE75
    Jul 16, 2018 at 18:53
  • \$\begingroup\$ @Linkyyy yes it is more like 2 nF - I have corrected my answer and added more explanation as to why it becomes an oscillator. \$\endgroup\$
    – Andy aka
    Jul 16, 2018 at 21:08
  • \$\begingroup\$ @Linkyyy if we are done here, please take note of this: What should I do when someone answers my question. If you are still confused about something then leave a comment to request further clarification. \$\endgroup\$
    – Andy aka
    Oct 5 at 8:20

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.