e2g-based PSU

180

02:02:56

I made this "op-amp e2g" about a year ago, intending to use it for a power supply. It's not a realistic op-amp, but it behaves in a realistic way and it realistically has to be powered; that's the point of it. So now I've built a voltage regulator around it. Its output is the green node.

The 'op-amp' in the corner outputs the error voltage times 1000 and clamped to 3V to the blue node. 

The rest of the circuitry should be quite realistic. I'm not sure why I chose 2V, but it works for the LEDs, one of which draws 2A. That's close to the PSU's limit. If you want to change the voltage, change the two 5kR resistors below the 10mm wide N-channel MOSFET.

If you want to limit inrush current with an inductor, 1mH should do the job.

Original op-amp e2g:
http://everycircuit.com/circuit/4998086501072896

What did the op-amp designer say when he found it had a bad frequency response?
"This does not Bode well."

published 5 years ago

PrathikP

5 years ago

Nice circuit @eekee. Compact full bridge rectifier, I'll remember the layout. One thing, you have too much output capacitance. 5mF will dump a huge amount of damaging energy into you circuit in the event of an overcurrent. It is always good to minimize output capacitance. In your case, it's not really needed but you can always whack a 10uF at the output. More that sufficient. And instead of the op amp, just stick a voltmeter between the + of the 2V voltage source and the + of your output to show the deviation from the expected output voltage.

eekee

5 years ago

Thanks @PrathikP! I forgot about the output cap. If I make it 10uF now, I get a convergence error. Don't know why, yet. What do you mean by 'overcurrent'? Something presenting too little resistance to the output? If so, I agree. I put the op-amp in so I could see where the deviations occurred relative to the other traces.

eekee

5 years ago

With output cap reduced to 10uF, it starts up all right if the heavy load (blue LED) is not connected. It can be switched in after start-up.

eekee

5 years ago

...With 10uF, it sometimes has a nasty fluctuation when the heavy load is switched in. I remember now, I was going to try to fix that. I left it for a few weeks, and it got forgotten. (I make too many notes; can't read them all.)

PrathikP

5 years ago

Overcurrent/short circuit is a fault that occurs when a load draws a current higher than expected, usually due to a fault in circuit design and rarely due to component failure. When such an even occurs, the output of the PSU powering the circuit is expected to fall to zero (or a low voltage) and the current limit to kick in ASAP without sending too much overcurrent through the faulty circuit, in order to prevent damage to the circuit components. Having a low output capacitance means that the output voltage can fall very quickly and the PSU can go into CC very quickly, dumping an overcurrent into the circuit for a very low amount of time, thus saving circuit components. This is why it is important to minimize PSU output capacitance.

PrathikP

5 years ago

As for CSF (Cannot find solution sim error), reduce the sim speed to 1ms/s and removing the op amp helps. Op amps are notorious for causing CSF in this app.

ViolationMad

5 years ago

Adding series caps to the opamp inputs also helps. Hurz and I have been trying to figure out why but I'm not sure anymore. Anyway it helps tho.

eekee

5 years ago

Thanks, guys! I'll try some things when I can. Series caps might perhaps reduce EC's idea of the gate capacitance, making the FETs respond faster and thus the whole thing would be less prone to delay-based oscillation.

eekee

5 years ago

It's better if the blue led only draws 1A, but has some problems on switching it; it can go into bad oscillations. A 1nF choke (simulating long lead) doesn't help.

eekee

5 years ago

series gate capacintace does improve things at high speed, but isn't all good. simple test: https://everycircuit.com/circuit/5619100666298368

add comment in editor