comparator with differential input(s)

[zappedagain]

Most 'differential' comparators are really singled ended devices; they compare the single-ended input-A against the single-ended input-B. Does anybody know of a comparator IC that has true differential inputs?

My application is comparing the instantaneous amplitude of a differential signal (A+ minus A-) against a threshold. While I can use an amp to do a differential to single-ended conversion and then compare it with a standard comparator, I'm space constrained so I'm trying to reduce my IC count. A 1 uS response time is adequate.

 

[nbucska]

I don't think you will find one. I have never seen
any op.amp or comparator with more than two analog inputs.

What is your application? Wouldn't be a way to design around it? What are the signal level etc. specs?

Transformer?




----------------------------
Please read FAQ240-1032
My WEB: <

http://geocities.com/nbucska/>

 

[zappedagain]

I have a DC coupled signal with bandwidth up to about 200 KHz that's going into an A/D with a 2.5V range. So far my best bet looks like an AD8130 Differential Receiver Amp. It is much faster than I need but it will do the job. I was hoping to not have another SOIC-8 on the board.

 

[itsmoked]

Hey zapped; Have you looked at the SOT23-5 amps/comparators? They are small and often easy to jam into places. You might be able to put 2 of them into the same area you can't cram an SOIC-8 because of the layout efficiency they present. Course if you're psychotic you could use some of the uMax10 or flip chips. You could fit 5 or 6 into an SOIC-8 footprint.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[nbucska]

The common mode rejection of modern comparators is fairly good -- perhaps you don't have to add anything or just
a LP filter ( 2 R + 1 C ) to reduce HF noise.



----------------------------
Please read FAQ240-1032
My WEB: <

http://geocities.com/nbucska/>

 

[zappedagain]

nb - I want to compare my amplitude of the differential signal against the threshold. Using a single ended comparator means I need to first convert my differential signal's amplitude to a single ended signal. There's that extra part again!

it - My days of avoiding those teeny weeny parts may be over... I successfully touched up teh pins on a DFN-8 package today with 0.010" wide leads, so there might be hope!

 

[IRstuff]

I'm not sure I follow. Most double-rail comparators will correctly compare a negative signal against a positive signal:

http://www.analog.com/en/subCat/0,2879,759%5F776%5F%5F352%5F6%5F352de,00.html#352

Not sure if they're compatible with your space constraints, but something like:

http://www.analog.com/en/prod/0,,759_776_CMP402,00.html

seems to specified to do the "true" differential comparison.

TTFN

Eng-Tips Policies FAQ731-376

 

[nbucska]

Can you tell us more about your signals and operation?



----------------------------
Please read FAQ240-1032
My WEB: <

http://geocities.com/nbucska/>

 

[ScottyUK]

Combined comparator and op-amp in one SOIC-8 package? Depending on how good a diff amp you need you may be able to build your own with a handful of precision resistors.

http://www.maxim-ic.com/quick_view2.cfm/qv_pk/1838

----------------------------------

Sometimes I only open my mouth to swap feet...

 

[itsmoked]

He was trying to avoid a single soic8 I think

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[ScottyUK]

I thought he was trying to avoid another SOIC?

I was hoping to not have another SOIC-8 on the board.

No doubt we will find out when the OP comes back!


----------------------------------

Sometimes I only open my mouth to swap feet...

 

[zappedagain]

Thanks Scotty, I'll give these Maxim parts some thought.

nbuksa -

I have a differential receiver channel and I'm looking into adding gain control. My signals are DC-coupled pulses that have about 200 KHz of interesting bandwidth (down to -70 dBc). AC-coupling is not an option because the pulses can also stretch into the millihertz range.

At this point I'm keeping the AGC simple with bimodal control (high gain (typical) or low gain (avoid clipping) so I want to compare the amplitude of my differential signal (it can have a lot of common mode) against a threshold to decide which gain setting to use. I need to ignore the common mode voltage in my decision. About 20 dB of gain adjustment should be adequate for the application (did I mention I'm chasing some loose specifications?!?).

Presently I'm thinking about an AD8130 to convert it to single ended and then I can compare with it with a simple comparator.