Project: High power constant current programmable load

[GK]

Hi. Here is an alternative (loop gain bode plot) sim. Using your values I get a phase margin of only 40 degrees with a 1mR source resistor, which explains the huge transient overshoot.

I'd be looking for either an alternative op-amp with a lower GBWP, or at perhaps implementing a discrete (bipolar transistor) driver stage with a significantly lower than 60 ohm Z-out between the OP07 and the mosfet gate, to push the "2nd" pole that is degrading your phase margin higher up in frequency.

[GK]

Oh never mind; I just read through the previous page of posts.

[GK]

OK, just some playing around before I go to bed. I have taken your "original" circuit and have configured the op-amp as a unity gain inverting stage instead of a unity gain voltage follower. The feedback resistors are R4 (Rf) and R3 (Ri). However R5 has been added to increase the noise gain to 3. This has the effect of lowering the unity loop gain frequency of the OP07 by itself from ~900 kHz to ~300 kHz. Phase lead compensation is then added in the form a a zero by C3, to effectively cancel out the pole formed by the 60 output resistance of the OP07 and the 28nF of MOSFET input capacitance. As you can see, this results in a simpler circuit and almost ideal loop gain plot with over 80 degrees of phase margin and a vastly improved transient response. The only downside is that bipolar op-amp supplies will be required and the input control voltage need to be negative.