LM393 Input Common Mode Voltage Range question

[Anthony Fremont]

I've been looking at some op-amp material and I have a question. The
datasheet shows a Max value for the "Input Common Mode Voltage Range"
spec of Vcc-2 over the full temperature range. Ron Mancini says that
this parameter refers to the average voltage at the input pins. If I'm
using a 5V Vcc and have one input biased at 2.5V when the other input
exceeds about 3.5V aren't I exceeding the spec? IOW, it appears that
it's not kosher to drive the other input past 3.5V (3.5V + 2.5V) / 2 =
3V = Vcc-2

Am I understanding this? If so, any recommendations for a single supply
comparator with a bigger common mode input range?

 

[Fred Bloggs]

I've been looking at some op-amp material and I have a question. The
datasheet shows a Max value for the "Input Common Mode Voltage Range"
spec of Vcc-2 over the full temperature range. Ron Mancini says that
this parameter refers to the average voltage at the input pins. If I'm
using a 5V Vcc and have one input biased at 2.5V when the other input
exceeds about 3.5V aren't I exceeding the spec? IOW, it appears that
it's not kosher to drive the other input past 3.5V (3.5V + 2.5V) / 2 =
3V = Vcc-2

Am I understanding this? If so, any recommendations for a single supply
comparator with a bigger common mode input range?

If you have one input >3.5V and the other at 2.5V, the CM is immaterial
because the comparator is overdriven into its output state. We have
repeatedly told you that as long as one input remains below 3V, you can
take the other one to 36V regardless of Vcc and the comparator operation
will be the same. How difficult is this to understand? The CM
specifications apply to the analog switching performance of the
comparator at threshold, i.e. both inputs ~ equal, and not whole Volts
after the switch has taken place.

 

[Anthony Fremont]

If you have one input >3.5V and the other at 2.5V, the CM is immaterial
because the comparator is overdriven into its output state. We have

How did you determine this?

repeatedly told you that as long as one input remains below 3V, you can
take the other one to 36V regardless of Vcc and the comparator operation
will be the same. How difficult is this to understand? The CM

I guess I'm stupid Fred, since I just don't know when to listen to you
instead of taking the datasheet at face value.

specifications apply to the analog switching performance of the
comparator at threshold, i.e. both inputs ~ equal, and not whole Volts
after the switch has taken place.

Ok, I just didn't see it explained like that in the Ron Mancini material
or anywhere else for that matter. I'm not saying that I don't believe
you, I'd just like to know how to determine things like this for myself.

 

[Fred Bloggs]

How did you determine this?




I guess I'm stupid Fred, since I just don't know when to listen to you
instead of taking the datasheet at face value.




Ok, I just didn't see it explained like that in the Ron Mancini material
or anywhere else for that matter. I'm not saying that I don't believe
you, I'd just like to know how to determine things like this for myself.

Was Mancini talking about opamps or comparators? Is the LM393 an opamp
or a comparator? Keep at it, it should come to you eventually.

 

[Jim Thompson]

How did you determine this?


I guess I'm stupid Fred, since I just don't know when to listen to you
instead of taking the datasheet at face value.


Ok, I just didn't see it explained like that in the Ron Mancini material
or anywhere else for that matter. I'm not saying that I don't believe
you, I'd just like to know how to determine things like this for myself.

Fred's explanation is correct.

For a comparator only one input need be =< VCM

For an OpAmp BOTH inputs are =< VCM, only because an OpAmp is defined
as being in its linear region ONLY when both inputs are equal...
otherwise it's a comparator too.

...Jim Thompson

 

[Anthony Fremont]

"Fred Bloggs"

Was Mancini talking about opamps or comparators? Is the LM393 an opamp
or a comparator? Keep at it, it should come to you eventually.

So this means that the exact same terminology is meaningless in the
comparator world? No wonder I find this confusing. Perhaps you could
point me to something like the Mancini document but pertaining to
comparators instead so that I no longer bother you with my trivial
questions?

However if you can't put together a response without being rude and
condescending, then do us both a favor and don't bother.

 

[Anthony Fremont]

Fred's explanation is correct.

For a comparator only one input need be =< VCM

For an OpAmp BOTH inputs are =< VCM, only because an OpAmp is defined
as being in its linear region ONLY when both inputs are equal...
otherwise it's a comparator too.

Thanks for the explanation Jim, I'm just trying to make sense of these
parameters.

 

[Jim Thompson]

"Fred Bloggs"
[snip]

However if you can't put together a response without being rude and
condescending, then do us both a favor and don't bother.

You apparently haven't lurked here long enough to realize that Fred is
treating you kindly... you should see his responses when he decides
someone is a dummy ;-)

...Jim Thompson

 

[Fred Bloggs]

"Fred Bloggs"




So this means that the exact same terminology is meaningless in the
comparator world? No wonder I find this confusing. Perhaps you could
point me to something like the Mancini document but pertaining to
comparators instead so that I no longer bother you with my trivial
questions?

However if you can't put together a response without being rude and
condescending, then do us both a favor and don't bother.

The best write-up of comparator operation, theory, and use is in Walt
Jung's OP Amp Cookbook.

 

[John Fields]

I've been looking at some op-amp material and I have a question. The
datasheet shows a Max value for the "Input Common Mode Voltage Range"
spec of Vcc-2 over the full temperature range. Ron Mancini says that
this parameter refers to the average voltage at the input pins. If I'm
using a 5V Vcc and have one input biased at 2.5V when the other input
exceeds about 3.5V aren't I exceeding the spec?

---
No.

Here's how common mode works:

Let's say that you have a +5V supply and an LM393 with 2.5V on the +
input and 2.3V on the - input. The output will go open-collector,
just like it's supposed to.

Now let's say that we have 2.5V on the + input and 2.7V on the -
input. In this case the output will go to 0V. Again, just like
it's supposed to.

Next, let's say that we increase the voltage on the + input to 3.1V
and put 2.9V on the - input. The output will again go
open-collector.

Finally, let's keep the + input at 3.1V and raise the voltage on the
- input to 3.3V. Will the output go to 0V? Theres's no telling,
because we've exceeded the common mode range of the LM393 by
allowing _both_ signals to to go higher than Vcc-2V.

 

[Anthony Fremont]

"Fred Bloggs"
[snip]

However if you can't put together a response without being rude and
condescending, then do us both a favor and don't bother.

You apparently haven't lurked here long enough to realize that Fred is
treating you kindly... you should see his responses when he decides
someone is a dummy ;-)

I have. I'm firmly getting the impression that I'm on that list now.
I'm not trying to be argumentative with him, but I would like a more
technically based explanation than "We have repeatedly told you......".

 

[Anthony Fremont]

The best write-up of comparator operation, theory, and use is in Walt
Jung's OP Amp Cookbook.

OK, thanks. Now I hope that if I read something in there, and then ask
about it here, that you won't point out that it's actually an op-amp
book. ;-)

 

[John Fields]

"Fred Bloggs"
[snip]

However if you can't put together a response without being rude and
condescending, then do us both a favor and don't bother.

You apparently haven't lurked here long enough to realize that Fred is
treating you kindly... you should see his responses when he decides
someone is a dummy ;-)

I have. I'm firmly getting the impression that I'm on that list now.
I'm not trying to be argumentative with him, but I would like a more
technically based explanation than "We have repeatedly told you......".

 

[Anthony Fremont]

"John Fields"

"Anthony Fremont" you......".

Well, I thought it was a straight-forward question. I had what I
thought was an authoritative definition of the parameter and I had a
datasheet with that parameter's value. It seemed a simple matter to do
the math as I layed it out. Jim's explanation was helpful because he
elaborated that the parameter was only really important when switching
of the output was about to occur and that after switching it wasn't that
important. That was piece that I was getting from the Mancini document
or Fred's input. It seems sensible now since an op-amp is expected to
remain linear over a broad range of inputs, and the comparator is at the
opposite extreme by basically eliminating the linear region. In the
comparator what happens after one input passes the other (and the output
switches) is mostly irrelevant.

Am I getting close yet?

 

[Jim Thompson]

OK, thanks. Now I hope that if I read something in there, and then ask
about it here, that you won't point out that it's actually an op-amp
book. ;-)

Keep in mind... there are generally only two differences,
circuit-wise, between an OpAmp and a comparator: OpAmps have
compensation to make them stable under unity (usually) feedback,
comparators do not; and comparator output stages are more amenable to
connecting to logic.

...Jim Thompson

 

[Anthony Fremont]

Keep in mind... there are generally only two differences,
circuit-wise, between an OpAmp and a comparator: OpAmps have
compensation to make them stable under unity (usually) feedback,
comparators do not; and comparator output stages are more amenable to
connecting to logic.

That's the impression I was under until he made the remark about
Mancini's book being about op-amps, not comparators. I think I get it
now (more or less, less I suppose).

 

[Anthony Fremont]

The best write-up of comparator operation, theory, and use is in Walt
Jung's OP Amp Cookbook.

I see that's available from Amazon, but I also see an apparently newer
book (1st edition 2005):
"Op Amp Applications Handbook ISBN 0750678445". This is an Analog
Devices book as opposed to the Sams cookbook. I can get this book
locally, do you think it would be a decent substitute for the
"cookbook"?

 

[John Fields]

"John Fields"


Well, I thought it was a straight-forward question. I had what I
thought was an authoritative definition of the parameter and I had a
datasheet with that parameter's value. It seemed a simple matter to do
the math as I layed it out. Jim's explanation was helpful because he
elaborated that the parameter was only really important when switching
of the output was about to occur and that after switching it wasn't that
important. That was piece that I was getting from the Mancini document
or Fred's input. It seems sensible now since an op-amp is expected to
remain linear over a broad range of inputs, and the comparator is at the
opposite extreme by basically eliminating the linear region. In the
comparator what happens after one input passes the other (and the output
switches) is mostly irrelevant.

Am I getting close yet?

---
Not really...

Didn't you read my explanation?

OK, I'll try again.

If you have a comparator with a Vcc of 5V and a common mode range of
Vcc - 2V, then that means that at least one of the inputs must be at
3V or less for the output to switch properly. It doesn't matter
_which_ input, as long as it's one or both of them.

However, if _both_ inputs are above 3V, then all bets are off and
the state of the output won't depend on whether the + input is more
positive or more negative than the - input. In other words, if both
inputs are more positive than 3V, the comparator stops being a
comparator.

 

[martin griffith]

I see that's available from Amazon, but I also see an apparently newer
book (1st edition 2005):
"Op Amp Applications Handbook ISBN 0750678445". This is an Analog
Devices book as opposed to the Sams cookbook. I can get this book
locally, do you think it would be a decent substitute for the
"cookbook"?

I think this is the one that is downloadable from Analog


martin

 

[Anthony Fremont]

---
Not really...

Didn't you read my explanation?

OK, I'll try again.

If you have a comparator with a Vcc of 5V and a common mode range of
Vcc - 2V, then that means that at least one of the inputs must be at
3V or less for the output to switch properly. It doesn't matter
_which_ input, as long as it's one or both of them.

I think I understand exactly what you are saying. I'm trying to
understand why the definition of the parameter (in terms of an op-amp)
doesn't say that. IOW, why are the comparators special. If I
understand Jim, the paramater is only important when the inputs are
close to the switching point. For example, if one input is at 3V and
the other input rises to exceed 3V, the output will be predictable. If
the rising input continues to rise (meaning the mathematical average
voltage of the inputs now exceeds the input common mode voltage range
parameter), the output will remain predictable in that it wont suddenly
invert on you. However, you will need to have things back within the
specs before you can have another predictable transition of the output.
Right?