3

For cost reasons, I want to use an opamp as a comparator. So far in terms of making sure the opamp is suitable, I have:

  • Making sure the output can slew fast enough
  • Making sure it can handle large differential inputs
  • Making sure it can go rail to rail
  • Ensuring that the current draw at maximum swing is acceptable

Are there any other crucial points?

BeB00
  • 5,303
  • 2
  • 18
  • 34
  • Lock up will be a consideration – Scott Seidman Apr 05 '17 at 00:56
  • 2
    Related: [What is the difference between “opamp” and “comparator”?](http://electronics.stackexchange.com/q/90657/11683). Also, I'm curious about the "cost reasons" -- comparators are at least as cheap as opamps. – Dave Tweed Apr 05 '17 at 02:10
  • 3
    @DaveTweed, "cost reasons" makes sense if you've already designed in a quad op-amp (for example) but are only using 3 of the amps. – The Photon Apr 05 '17 at 02:28
  • 1
    Using an op amp as a comparator is [not recommended](http://electronics.stackexchange.com/a/194324/51760). And as @DaveTweed notes, you're unlikely to be in a situation in which an op amp is actually cheaper than an comparator -- unless you have a spare op amp in a dual or quad package and that op amp just happens to meet your requirements as a comparator. A comparator is more likely to have better performance, too -- it will have lower quiescent current, faster response time, etc. – Null Apr 05 '17 at 02:44
  • Rail-Rail Op Amps are severely current limit on output thus slew rate depends on reactive loads or R< 10k. Rail to rail input is common but below ground is not except certain comparators, so depends on requirements. Vio, Iio may be important as well as PSRR vs f. Comparators with pullup R are Rail to rail but asymmetric slew rate but much faster if active Vol/Iol is more important. If this is for 100k pcs price differential is pennies. It comes down to all the specs and cost /volume criteria, so no precise answer is possible. – Tony Stewart EE75 Apr 05 '17 at 02:45
  • @DaveTweed and Tony: Yup, for the cost i was mainly talking about using a spare opamp already on the board, compared to the price of both the extra component and having an extra BOM line, as well as the extra space required. – BeB00 Apr 05 '17 at 03:57
  • Just to add, the question is general, but in this specific situation the comparison is for a PIR sensor, so the response time isnt an issue. Unfortunately I cant use the opamp I originally wanted (TLV8802) because of its high current draw when the output is at full swing. It's surprising the price difference between opamps and comparators for TI, the TLV3492 is over triple the price. – BeB00 Apr 05 '17 at 04:14

3 Answers3

1

This application note answers exactly your question and covers a few less obvious issues associated with incompatibilities that may arise due to the different optimization and expected usage of op-amps and comparators. Give it a read:

https://www.analog.com/en/analog-dialogue/articles/amplifiers-as-comparators.html

DKNguyen
  • 54,733
  • 4
  • 67
  • 153
0

Yeah you can absolutely use operational amplifiers as comparator, however if you want to consider the frequency or time factors, then you should consider CMOS rather than opamps.

Mark Jhomel Olimon
  • 11
  • 1
  • 1
  • 4
0

I have:

it is a lot more application specific than opamp vs. comparator specific.

Making sure it can handle large differential inputs

handling small differentials accurately is a lot more important than handling large differentials accurately.

Making sure it can go rail to rail

output R2R is not as important than input R2R. input R2R is not as important as the ability to swing to ground (for inputs).

I would also put current drive up there as well.

dannyf
  • 4,272
  • 1
  • 7
  • 9
  • 1
    The issue with large differential inputs is that some op-amps might consume excess power and overheat in this condition. – The Photon Apr 05 '17 at 01:12
  • Yup, even when used as a comparator, you need to avoid violating absolute maxima – Scott Seidman Apr 05 '17 at 02:33
  • With the large differentials, it was more about the fact that some opamps actually can't deal with large differentials, for exampe the OPA228. If you feed it a differential voltage over ~1.4V, it will start sinking current, and if you dont put any current limiting it will release its smoke. It's a bit sneaky because it's not explicitly mentioned in the absoulte maxima on the datasheet, you have to go to page 17 where it starts talking about diodes. – BeB00 Apr 05 '17 at 04:06