comparator "terminology" and use

Hi all, before i go blundering in and wasting time I'd like to clear a few small details up. I'm thinking of a circuit that will output a signal to drive a mosfet when a voltage in one point "exeeds" voltage at another point, this needs to take place with just a few millivolts ideally. so does the input offset voltage get in the way of pefection. for example if I have a LM393 with a 2 mV offset, does this mean that there will be an error of 2 mV in "reading" the two voltages, so potentially the ouput could flip 2 mV before the deadline is met ? my idea is to control the direction of current flow, so really I don't want it flowing the wrong way even with a sloght difference although I think practically it won't be an issue

The short answer is yes, the voltage offset is the limiting factor.

Operating the comparator in the linear region can also happen when the input voltages are similar and can cause oscillation.

Hero999 said:

Operating the comparator in the linear region can also happen when the input voltages are similar and can cause oscillation.

Click to expand...

I had a suspicion that was the case, I think practical experimentation will be the only answer.

You can null out the offset.
You can purposely introduce an offset to guarantee that one input >= the other at switching.
You can induce hysteresis which is purposeful offset which depends on which input was larger before.

I've never tried nulling a comparator before.

You could trim the reference by +/-2mV but how can you measure voltages that accurately anyway?

The chances are the tolerance of resistors in the potential divider and the voltage reference are introducing a greater error than the comparator.

Manually nulling out an op amp will have some offset drift over temperature. If you truely need zero offset, and good temperature offset stability, you should go with a zero offset amp. These null themselves out automatically in the background. You can get better then 20 uV offset performance.

RCinFLA said:

Manually nulling out an op amp will have some offset drift over temperature. If you truely need zero offset, and good temperature offset stability, you should go with a zero offset amp. These null themselves out automatically in the background. You can get better then 20 uV offset performance.

Click to expand...

the 1-2 mV may not be an issue, and yes if need be I'd just get a better comparator