From: Winfield Hill
Subject: Re: Current sinking circuit instability
Date: 17 Jan 2003 07:09:06 -0800
Organization: Rowland Institute
X-Newsreader: Direct Read News 4.20
> I have a simple current sinking circuit designed to test DC power supplies
> up to 5V, 10A. Not being much of an ASCII artist, the circuit is described
> as follows:
> An LM351 opamp output is connected to the Gate of a BUZ71 (50V,14A) MOSFET,
> the Source of the FET is fedback to the -ve input of the opamp and a
> potential divider with POT provides an input to the +ve input to the opamp.
> There is also 0.1 ohm power resistor connected between the FET Source and
> ground (so a 200mV change on the opamp input sinks 200mA).
Oops, stop right there! You're going to have trouble with loop stability
or out-right oscillation because the FET's very high gate capacitance is
loading down the opamp's output at high frequencies. My recommended
solution is to add a resistor to isolate the sense resistor from the
opamp's -ve input (15k), plus a capacitor from -ve input to the opamp's
output (2.2nF). This new feedback path removes the MOSFET from loop at
frequencies above 5 kHz (usually the cutoff frequency isn't important).
BTW, you can add a resistor from the opamp's output to the FET gate if
you want, but if that's the only change you make, the loop-stability
situation will be even worse!