Hey guys,
I'm doing a differantial op amp circuit followed by a comparator. It will eventually serve me in a DCDC converter.
The idea is to fake the output of the convert ( from 0V to 12V) and generate a PWM from that comparing the error with a saw tooth signal.

My problem is the following : the circuit only seems to work within some boundaries.
When my error is at 12V, I have no PWM generated. The PWM only starts close to 9V. When the error keeps getting smaller the PWM eventually stop being generated too ( around 2-3 V ) .
Does anyone happen to know why I am seeing this behavior ? I would obviously expect the PWM to work through all the voltages.

Thanks a lot guys.

 

I suspect the allowed input common-mode voltage range was exceeded. Especially since LT1028 is not a comparator but an opamp. You need to use an input rail-to-rail comparator.

 

I made your circuit with parts closer together.
Tomorrow I will see the wrong biasing.

 

The datasheet for the LT1028 opamps you are using say that their inputs do not work if they are within 2v from a supply voltage. Your circuit has pulse inputs idling at the 0V negative supply.
Fix the circuit by adding a negative supply of a few volts for those opamps.

EDIT: Also your inputs go as high as the +12V supply so increase your positive supply 3V or attenuate the input signals to 9V.

 

Audioguru again has what is probably the best solution. If you are stuck and can't change the power supply or op amp, you can try this:


In your case you can probably make all the resistors the same value, say 10k and everything should work fine except the output will be half the voltage it was.

 

@ericafoster Did you get any good ideas reading this thread?

 

Every PWM circuit has a defined control voltage input range, and outside that range no satisfactory function is claimed. Not all PWM circuits will allow zero to 100% control, nor is it always required. That may also be part of the problem.