0
\$\begingroup\$

So I have an interesting request that I've searched the internet for and have not been able to find, or been able to come up with in my head.

I have a circuit with a feedback loop that is meant to control a MOSFET. The feedback loop has a difference amplifier that outputs the difference between the current output and the desired output. Based on this, the voltage to the gate of the MOSFET should change.

In other words if the output of the system is currently 11V, and the desired output is 12V, the difference amplifier has an output of 1V, and this 1V output should somehow signal that the MOSFET gate voltage needs to increase. Similarly, if the current output is 13V, the difference amplifier would have an output of -1V, and should signal that the MOSFET gate voltage needs to decrease.

I initially tried hooking the feedback output from the difference amplifier directly, but quickly realized that wouldn't work.

Rather, I need a circuit where a + voltage causes a constant voltage increment until the feedback reaches 0, or a - voltage causes a constant voltage decrement until the feedback reaches 0. So the feedback should be a sort of specification of the rate of change to the base voltage.

In a nutshell, this is what I'm referring to:

enter image description here

My initial guess is that this might have to include a non-inverting amplifier, or even just a comparator comparing the feedback signal to ground, and charging of a capacitor that powers the base voltage to the MOSFET.

Edit

Came up with something, looks like this works fairly well. Still open to better suggestions though.

enter image description here

\$\endgroup\$
  • 1
    \$\begingroup\$ This just sounds like an integrator to me, so clearly I am missing something. If you take your output, subtract your setpoint, and feed that into an opamp integrator doesn't that behave as you describe? \$\endgroup\$ – Matt Dec 19 '13 at 17:45
  • \$\begingroup\$ @MattAnderson I think you're absolutely right, this didn't occur to me at all. You must obviously command a better understanding of math and circuits than I do. Thanks! \$\endgroup\$ – krb686 Dec 19 '13 at 18:16
  • \$\begingroup\$ Why do you need a transistor at all? How much current do you need to supply to the output? \$\endgroup\$ – Joe Hass Dec 19 '13 at 19:13
  • \$\begingroup\$ @JoeHass I didn't really give a lot of detail about the entire circuit. I'm not using it drive anything, I'm using it as a variable resistor :) \$\endgroup\$ – krb686 Dec 19 '13 at 19:42
1
\$\begingroup\$

Sounds like all you need is an integrator.

https://en.wikipedia.org/wiki/Operational_amplifier_applications#Inverting_integrator

| improve this answer | |
\$\endgroup\$
1
\$\begingroup\$

You need a closed loop control circuit such as one where the opamp's feedback comes from the output. In effect the mosfet (or external transistor) becomes regarded as part of the opamp. The opamp will then deliver whatever control voltage is needed to the transistor in order to make the output equal the demand.

On your circuit the opamp has local feedback but if this were removed and replaced with overall feedback you will get what I believe you are requiring. It's a little trickier though because circuits like this can be unstable but there are many examples of a MOSFET or BJT being used this way. Typical of this is the ubiquitous 7805 voltage regulator. It's demand is fixed internally and it uses overall feedback and internal high gain amplification to regulate the output voltage to 5 volts.

Here's a great example of an op-amp driving a bunch of power transistors but taking its feedback directly after the transistors i.e. the op-amp is "measuring" the output and correcting the anomalies output by driving the transistors harder or lighter as required: -

enter image description here

Here's a simpler circuit where the op-amp is "measuring" the output current and ensuring that it remains constant over a large load range: -

enter image description here

Note that the feedback point comes not from the op-amp's output but from the resistor above the transistor.

| improve this answer | |
\$\endgroup\$

Your Answer

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

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