Introduce an hysteresis without PCB redesign...

[flote21]

Hello guys!

The schematic attached is a PCB design where the comparator is ON when the Vin+ > Vin-. However the Vin+ voltage oscillates +-200mV and I am wondering if it is possible to set an hysteresis avoying a PCB redesign...Any idea?

Thanks in advance.

Greetings

 

[KlausST]

Hi,

as usual: a feedback from output to noninverting input.

Your source impedance at IN+ is 866 Ohms.
to pull it by 200mV you need a current of 230uA
(Ohm´s law)

Your output voltage range is 15V.
You need a "pull" resistor of 65k to cause 230uA.
(Ohm´s law)
reduce it by the already installed 10k pullup --> 55k
("minus" calculation)

****************
But I see you use 15V as reference.. so the +/-200mV will be caused by 15V fluctuations.
So the simplest way is to use the already installed LM4050. It avoides fluctuations at IN+.

****************

Klaus

 

[FvM]

If you don't consider patching a resistor between U5.1 and U5.3 a "redesign", it's feasible.

The circuit shows arbitrary design variations, e.g. not using U6 reference. Why? Why not at least filtering the reference voltage?

 

[flote21]

Hi!
Thanks a lot for your quick replies.
But I made a mistake. The oscillations are produced in Vin- (not in Vin+ as I said previously). Therefore, the scenario changes and I need to create an hysteresis in the comparator of the 400mV (aprox.) to avoid a continuous relay switching.
Greetings

 

[KlausST]

Hi,

The optocoupler control is (quasi) analog.
Why not clear switching?

Klaus

 

[flote21]

You mean placing a rc filter at the optocoupler output for example? But I will need to do a PCB redesign then..

 

[FvM]

You didn't answer if patching a feedback resistor is acceptable.

 

[flote21]

You didn't answer if patching a feedback resistor is acceptable.

I would need to do a PCB redesign...but if there are not other choices.
Greetings

 

[Easy peasy]

a resistor from Comp o/p pin to the (+) pin will provide all the hys you need, typ choose 100x the resistance seen from the (+) pin for ~ 1% hysteresis ....

--- Updated Nov 5, 2020 ---

about 100k in this case ...

 

[KlausST]

Hi,

You didn't anser my post#5. Instead you started a new thread which obviously belong to the same circuit.

...depends on the answer to post#5:
Instead of reducing VRef, you could increas the value of R41..with similar effect.

Klaus

 

[flote21]

Hi! KlausST!
I want to avoid to do a PCB redesign. So your solution is acceptable but I want to know if there are other options...
Regarding to the other thread. Yes I started another thread to avoid mix topics. With this thread I want to focus on the hysteresis solution and with the other one, I want to discuss the voltage reference solution.
Finally when I increase the value R41, the effect on the simulation is like increasing the gain. therefore the noise is also increased...I prefer to reduce the Vref...I made some test on the lab and reducing Vref until 0.5V with the same R41(2.2K) gives me good results...But I will need anyway to think a hysteresis solution...
Greetings.