# Electronic load, Problem turning MOSFET on

I am trying to make an adjustable dummy load to test a 12V supply, I’m having trouble with the op-amp feedback circuit. Attached is a simplified schematic. I have tried lots of things and stripped it down to this.

Basically when +In of the op-amp is 0V, the output of the op-amp goes to the negative rail. In this case the negative rail is 0V, but when powered with a +/-5V supply, the output goes to -5V. In this state the MOSFET is fully off

When I add just a tiny bit of positive voltage to the +In of the op-amp, The output shoots straight into positive saturation and goes as close to the positive rail as it can get (about 3.7V) and stays there no matter how much voltage I give to the +In. At this point the MOSFET is on, and allows approximately 250mA to flow through though.

Any ideas on what this could be?

Things I have tried: - removing R6 and Q1 and connecting R8 to -In. At this stage I get a normal voltage follower. - Changing the FET - Raising the resistance of R6

On a side note, there is one way that I can get the load to be variable, and that is by removing the heat sink from the MOSFET. I have only done this for short periods of time.

• Is the circuit oscillating? Jul 10, 2019 at 21:43
• To what value did you raise R6? R4 and R5 are sized to deliver about 450mV to the $V_+$ terminal of the amp; R6 should be sized so that at maximum design current its voltage is about the same. Then you should check for oscillation -- usually when I do a circuit like this my op-amp to gate resistance is about 1k-ohm, and (for an LM358) I'll have about a 470pF cap from the amp output to it's $V_-$ terminal for compensation. Jul 10, 2019 at 21:53
• "...the load to be variable [by] removing the heat sink from the MOSFET" - page 9 of the datasheet says the tab is connected to the Drain. Is the heat sink shorting out the MOSFET? Jul 10, 2019 at 22:08

What you have built is an oscillator.

There is nothing wrong with the LM324, it has been updated through the years, but I do agree with @laptop2D you could make a better choice of devices. You don't need a low offset opamp in this application.

The problem you have is that the FET you chose has a huge input capacitance (>5000pf). With the 180 Ohm series resistor you guarantee that the circuit will oscillate around your setpoint.

You can try simply removing the 180 Ohm, which allows that LM358 to current limit into the gate, that might just work though it isn't pretty. However to fix the problem you really need a power driver stage for the FET gate, or chose a FET with much lower Gate capacitance.

The circuit has another problem though, the VGS(th) range for the IR1404 is 2-4V. You may have lucked out and got one that is closer to 2V ….but you will run out of drive voltage and not be able to use one that comes in at close to 4V.

You should at least double your supply to 10V (perhaps you could use the 12V supply you are testing to generate this).

To charge your gate capacitance at a reasonable slew rate I'd suggest you need to allow for up to +/-500mA drive. You could do this by using an NPN/PNP pair on the output of your opamp.

Perhaps something like this:

simulate this circuit – Schematic created using CircuitLab

• Thank you! Raising the voltage to 12V solved my problem. 5V supply wasn't quite enough to turn on the FET. As for stability, this was solved by adding 10N in series with 1k from the output of the op-amp to -In. it works great now! Jul 20, 2019 at 13:25

The LM324 is old. Without too much analysis...

It could be that the voltage offset (which for the LM324 would be high at 3mV) is causing problems with the feedback loop as there would be unertanty when the voltage across R6 goes lower than roughly 3mV (the input offset voltage will vary from amp to amp and temperature). Another problem could be the common mode range. Anyway this is an old opamp, my suggestion would be to find a new rail to rail opamp with a low input voltage offset.

Yes. The loop gain is to high, so the circuit will oscillate even if you replace the opamp with better one. You would need at least a low pass filter on feedback to damp the high frequencies. Since it is not oscillating yet, the problem could be a poor opamp as laptop said.

• Yes this was part of the problem. Placing 10N in series with 1K from the output to -In of the op-amp helped a lot with stability. thanks! Jul 20, 2019 at 13:23
• Jul 20, 2019 at 15:35