# Op amp current regulator

This is probably a marginal mistake.

This circuit, a current regulator:

simulate this circuit – Schematic created using CircuitLab

Works fine on simulation.

However, it does not once built. I built it twice, once on a PCB once on a breadboard, using different components, and both behave the same.

The OA2 output is either at High or Low, depending on the V1 but is not regulating anything.

I also observe some kind of schmitt trigger behavior, current is 0 and V1 is 0, then slowly increase the voltage, it suddenly switch to the max output. Then I have to reduce V1 to almost 0 to stop conducting.

There is no oscillation or anything like this. it's either fully on / off instead of regulating.

At this point I do not believe it's a mistake on the build.

• Opamp are single rail supplied at 10V.
• V1 is 0-1V to control 0-1A.
• I tried with BA2902 and OPA227
• It looks unnecessarily complicated. The node above $R_7$ should replace where $R_2$ goes. Meaning that $R_2+R_5+R_6+R_3+R_4+OA_1$ = unnecessary. - This is a comment, not an answer, because I'd like to hear why you've gone with this design, which I might have missed. Commented Sep 27, 2018 at 3:02
• What op amps? What supply voltage? Commented Sep 27, 2018 at 3:07
• It is actually a current bypass, in some situation, R7 might not be directly tied to the ground, so OA1 is there as a differential comparator. @HarrySvensson Commented Sep 27, 2018 at 3:07
• Perusing the OPA227 datasheet reveals that the common mode range does not include v- or v+. It looks like you are using a single supply but your op-amps are not rated for rail-to-rail inputs. The inputs must be at least 2 volts from the rails for it to work. Commented Sep 27, 2018 at 3:18
• My first impression of this circuit is to worry about oscillation and the need for a gate drive that is higher than your rails. (I know you say there is no oscillation in the practical build.) But, if you set $R_1=10\:\Omega$ then a suitably dissipating NPN BJT would eliminate much of those worries. I'm curious if you might try one in your practical circuit to see how it performs. If it works better, that tells you something. Any chance you could attempt it? (Obviously, keep your load's required voltage drop well within range of your supply, when testing.)
– jonk
Commented Sep 27, 2018 at 18:26