Testbench for a simple uncompensated opamp as a comparator

Hi guys,

Can someone help me to mount a comparator testbench? I have seen several approachs and I have tried two of them and they produced different results.

For example, using a vpulse directly between the two inputs with the addition of a vdc supply to bias the input transistor.

Another testbench that I have seen is one where we put a vpulse in one of the inputs and on the other input a reference voltage. In this case, this testbench produces a different fall time when compared to the above testbench.

I am confused.

Appreciated.

This probably involves the underdrive/overdrive of the inputs.
Inspect the common mode, pre- and post-transition difference
voltages. Comparator transient performance specs are often
quite specific about how "wound up" the front end is initially
and how much past null the input transition goes, which
affects the degree to which the front end is steered and
thus how much internal current imbalance is there to swing
the active, parasitic and in this case compensation
capacitances. There will be a great difference in the rise/fall
and delay behavior with variations in overdrive, when that
overdrive is in the few-mV (anything below 50mV probably)
range.

Apples:apples input voltage waveforms is where to start.
And then, your third apple is some relevant comp part's
test figures for Tr, Tf, Tplh, Tphl.

Thanks dick, for your reply.

How can I test the common mode range? How can I set up my testbench to do that in a comparator?

Regarding the underdrive/overdrive of the input: from what I have understood I have to vary the input of the comparator from a negative voltage up to a positive one and the back to the negative voltage. Is that it?

Silly question: What is the most suitable way to do the transient analysis, using the positive/negative voltage at the input of the comparator or to just put the square wave varying from 0V to VDD in one input and then have a reference in the other input?

- - - Updated - - -

For example, when I am using a vpulse (say -100mV to +100mV) connected directly to the differential input of the comparator (with a vdc to bias the differential pair) the comparator will change its output each time the vpulse crosses the offset?

- - - Updated - - -

And by the way, how can I test the comparator sensitivity?

Large signal transient analysis will look better but be
a poor indicator of in-application performance with
small input difference. "Test it like you fly it" is the word.
What does the intended application tell you?

Check input differential mode range on op amps. Many
comparators are designed to tolerate full supply range
independently on +. - but many op amps put cross-
clamps on the inputs to prevent them jacking apart
and damaging the diff pair (BVebo on bipolars). Not
all op amps are suited to this role, even if you are OK
with SLLOOOWWWW....

- - - Updated - - -

For common mode range I would just do a parametric
analysis with the difference fed by one pair of sources,
each 1/2 Vdifferential (VDM) with the midpoint stacked
on a DC source that is VCM. Loop VCM for all tests and
where you break out of compliance, you've found it (or
more properly, the last passing point). You could fail for
prop delay, for AVOL, for simple functionality, for bias
current etc. Common mode range is the "it's all good"
range.

Thanks for your suggestions dick. I will came back to you later or tomorrow to see if I managed to put the ICMR working.