Comparator output problem

[FreshmanNewbie]

I have this comparator circuit using an op-amp. This is my actual op-amp used in the circuit. I have used TL072 as I didn't have the simulation model.



This circuit is used to trigger the output when the inverting input is less than 80% of the non-inverting input.

But, when I use this in actual circuit, when the inverting input is less than 80% of the non-inverting input, the output is oscillating, even though the circuit is designed for hysteresis.

I read this application note from TI, but I am not sure why the output is oscillating. Can someone tell me the reasons why the output might be oscillating?

In actual circuit measurement, the input at the non-inverting node, has an oscillation of 1.4V peak to peak. I believe this is the reason why the output is also oscillating.

But I am not sure on why the input is also oscillating. I have no capacitors also in the entire circuit.

Can someone tell me the reasons and how to avoid this oscillation?

 

[KlausST]

Hi,

An Opamp is an Opam, a Comparator is a Comparator. They have different features although the symbol us the same.
If you use one for the other's function it calls for problems.

trigger the output when the inverting input is less than 80% of the non-inverting input.

No. A comparator always switches when both inputs are (almost) identical.
You rather talk about Vout and Vout2.
With Vout = 32V the switching points are: 25.6V and 24.5V (about 1.15V hysteresis)

Circuit issues:
* the power supply Vout needs a bulk capacitor and a fast ceramics at every IC supply pin.
* R5 is useless in combination with a RR push pull output.
* with pp oscillation of 1.4V at noninverting input ... it is expected to oscillate ( higher than hysteresis)
* I expect the problem is, that your Vout supply is not stable I expect it to have a pp oscillation of about 3V.

You don't tell about oscillation frequency.
I expect it to be anywhere inbetween 100 Hz and some kHz.

If it's much higher then the problem is else where.

Klaus

 

[danadakk]

Sure looks like the OpAmp compensation, possibly the magnitude of the 100Ks,
and stray C could be issues.

Regards, Dana.

 

[Audioguru]

Everything that is built on a solderless breadboard oscillates, except oscillators.
If R5 is replaced by a short piece of wire then it might not oscillate.

 

[danadakk]

For what its worth, and w/o analysis I do not suggest this is a fix
until evaluated, I put a 20 pF across the R1 in my schematic, and
that stopped the oscillation.

C's from either OpAmp INV or NI inputs to ground did not help in sim.


Regards, Dana.

 

[Easy peasy]

notice the osc in the sim is right up at Vcc - way less than a milli-volt - this would not happen in the real world - or you would not see it, the 20pF across Vout to + input is a valid fix if needs be

Also - as said above ITRW ( in the real world ) you need 10uF across the comp, a decent pull up resistor on the o/p

and the - Vin cannot be floating ( you are inviting osc to leave it so )

 

[danadakk]

notice the osc in the sim is right up at Vcc - way less than a milli-volt - this would not happen in the real world - or you would not see it, the 20pF across Vout to + input is a valid fix if needs be

Also - as said above ITRW ( in the real world ) you need 10uF across the comp, a decent pull up resistor on the o/p

and the - Vin cannot be floating ( you are inviting osc to leave it so )

Osc at ~ ( Vcc /2 ) - 1......which is the bias point in the sim of the output.

Not sure why you are recommending a pull up on output, its active P-P output....?


Regards. Dana.

 

[Easy peasy]

The OP states comparator output problem, even though the TL072 is a JFET input opamp - if a comparator were to be used in the same ckt - a pull up R would be needed - always good to explain these things to posters in case they are/were un-aware ... as most comparators have no active pull up capability ...

the oscillation shown is an artifact of the sim.

 

[danadakk]

I agree Easy, I too now think its an artifact, the sim. If I expand the time window
it goes away. Also if I look at expanded transition point response, and use hi res
time increment, I do not see any osc either. I am going to inject noise into trigger
to see if hysteresis adequate. I should look at model to see how complete it is.

Regards, Dana.

 

[danadakk]

https://www.ti.com/lit/an/snoa860/snoa860.pdf?ts=1608132362524

https://www.st.com/resource/en/application_note/dm00050759-introduction-to-comparators-their-parameters-and-basic-applications-stmicroelectronics.pdf

Above might help.

Osc circuit, you can see what possible parasitic can cause -

Note this ignores input C effects, as well as stray C of R's and layout.

When you use bypass caps for power pins not all caps of same value perform
as well in terms of ESR -

Regards, Dana.