Issue in driving opamp as comparator to drive mosfet

kubeek

Joined Sep 20, 2005
5,794
But didnt datashhet says "Wide Bandwidth (Unity Gain): 1 MHz"?
1. If 1Mhz is not freq to which it can operate, then what parameter define what max freq opamp will work fine?
2. Otherwise why 1Mhz is mentioned?
That is the gain bandwidth product figure. That means that at 1MHz you can get only gain of 1 i.e. a follower and not more however you may try, at 500kHz you can get gain of 2, 200kHz gain of 5 etc. This figure also is taken at a very small signal amplitude, so that the output waveform is not afffected by the slew rate as slew rate is important for large amplitude signals.
You want a comparator, so you want the gain to be in the hundreds or thousands for it to be useful. That would mean limit of useful operation somewhere about 1kHz-10kHz.
 

MrAl

Joined Jun 17, 2014
11,389
That is the gain bandwidth product figure. That means that at 1MHz you can get only gain of 1 i.e. a follower and not more however you may try, at 500kHz you can get gain of 2, 200kHz gain of 5 etc. This figure also is taken at a very small signal amplitude, so that the output waveform is not afffected by the slew rate as slew rate is important for large amplitude signals.
You want a comparator, so you want the gain to be in the hundreds or thousands for it to be useful. That would mean limit of useful operation somewhere about 1kHz-10kHz.
Comparators also have limits too though.
 

kubeek

Joined Sep 20, 2005
5,794
Comparators also have limits too though.
Yes they have, but you just reminded of a thing that I forgot to mention. Comparators usually have defined how long it takes them to change state according how different is the voltage between the inputs, say for 5mV difference or 50mV difference.
Opamps internally go into saturation when used as a comparator, and it takes them considerable time to come back out of saturation and start changing the output, and this is not in the datasheet beacuse they are not meant to be used like that and the manufacturers dont specify these values. The resulting functionality with this in mind is even worse than what the GBW and slew rate figures alone would predict.
 

MrAl

Joined Jun 17, 2014
11,389
Yes they have, but you just reminded of a thing that I forgot to mention. Comparators usually have defined how long it takes them to change state according how different is the voltage between the inputs, say for 5mV difference or 50mV difference.
Opamps internally go into saturation when used as a comparator, and it takes them considerable time to come back out of saturation and start changing the output, and this is not in the datasheet beacuse they are not meant to be used like that and the manufacturers dont specify these values. The resulting functionality with this in mind is even worse than what the GBW and slew rate figures alone would predict.
Hello again,

Yes good point. The op amps can be strange sometimes in saturation.
I got away with it several times because the frequency was low. Perhaps 20kHz is pushing it. Some testing is in order but not sure if the spice models will properly include the saturation delay characteristics. Perhaps we can check this in a sim or else it has to be tested on a physical breadboard. Not sure if you want to try that or not, but it would be very informative.
 

crutschow

Joined Mar 14, 2008
34,281
That means that at 1MHz you can get only gain of 1 i.e. a follower and not more however you may try, at 500kHz you can get gain of 2, 200kHz gain of 5 etc.
That's true if the op amp is being used as a non-inverting follower.
For an inverting amp, those values must be cut in half.
 

MisterBill2

Joined Jan 23, 2018
18,167
It has already been explained why an op-amp would not work in the application, so now how about moving on? I suggested an LM311, but that may not be the very best choice for this application, since that is a fully mature product, been available for perhaps 20 years. Not obsolete, though, just mature. An LM339 would be cheaper but still possibly not the best choice, since it is also a mature product.
 

MrAl

Joined Jun 17, 2014
11,389
It has already been explained why an op-amp would not work in the application, so now how about moving on? I suggested an LM311, but that may not be the very best choice for this application, since that is a fully mature product, been available for perhaps 20 years. Not obsolete, though, just mature. An LM339 would be cheaper but still possibly not the best choice, since it is also a mature product.
What is it you dont like about the LM339? You want faster speed?
 

MisterBill2

Joined Jan 23, 2018
18,167
What is it you dont like about the LM339? You want faster speed?
My point being that while the LM339 is certainly an OK comparator, a lot of advances have been made in the past 20 years. So there is probably an even better choice someplace. AND, because it has an open collector output the rise time may not be fast enough.
 

MrAl

Joined Jun 17, 2014
11,389
My point being that while the LM339 is certainly an OK comparator, a lot of advances have been made in the past 20 years. So there is probably an even better choice someplace. AND, because it has an open collector output the rise time may not be fast enough.
OH yes i see what you mean. Even 40 years ago there were better comparators. One of them made by National had 20ns response time for a normal input step. Some 10 years later Analog Devices had super fast comparators 5ns i still have a few of them around here somewhere.
So these comapators were around for a long time they were just more expensive :)

I cant believe i forgot that 5ns comparator part number. I had to use that for an automatic test jig design that was mean to test some analog driver cards for large 100 amp bipolar transistors.
I think i'll look it up, but it may be discontinued by now.
It may have been the LM161 but cant remember now.

We also used the LM318 which has a floating output emitter that helps in tight coupling of the output to the next circuit stage.
 
Last edited:
Top