How to obtain assymetrical delay on a comparator output

[earckens]

Purpose of this design: detect light variation at low intervals (>1 second, up to minutes).
This detection circuit works independent of ambient light.

What I do need is a delay on the output, only when LDR is uncovered, ie when the voltage on the non-inverting input recovers to a high level.

The left design uses two comparators and works well. I would however like to reduce this to a design with one comparator only, if possible.

I am in a bind how to introduce a delay circuit for the voltage on pin 9 (non-inverting): voltage going up on that pin needs to be delayed.

 

[KlausST]

Hi,

what about using a rather small microcontroller? Maybe 8 pins.
All threshold levels and timinngs can be programmed (or self calibrating) and have no drift in time and temperature.

Klaus

 

[earckens]

Hi Klaus, thanks for your suggestion. But I prefer to keep it simple and cheap.
Microcontroller: extra cost (hardware and software), extra space on the pcb, extra maintenance headache, extra production cost, .. No please, at most an extra mosfet besides C and R.

EDIT: and drift is no issue, no precision needed.

 

[KlausST]

Hi

Microcontroller: extra cost (hardware and software), extra space on the pcb, extra maintenance headache, extra production cost, .. No please, at most an extra mosfet besides C and R.

I agree with the cost for software.
But all the others are benefits for the microcontroller.
It is smaller, maintenance free (even automatic calibration), less assembling parts, no Cs (besides supply bypass)

Microcontroller replaces:
* all comparators
* all pots
* all timing capacitors
* all timing/pullup resistors

3 ADC inputs for the LDRs, 2 digital outputs. A single IC (8 pin should be sufficient) solution.

In the past 25 years of developing industrial electronics I avoided the use of pots, because of cost, size, maintenance. I never used a screwdriver to find the desired operation points (time consuming, not precise) and also avoided that a customer (or other person) manipulated my adjustment.

The cost of two trimmers may be in range of a small microcontroller (ATTINY, PIC...)

But for sure it depends on the application´s requirements, the country, ther personal preference, the volume, the mounting methods and maybe many other parameters...

Klaus

 

[earckens]

No need to try and convince me.
Still looking for a passive circuit (at most with a mosfet added) for my query.

 

[FvM]

The circuit needs two comparators (or other active components with similar function) by design. One provide the light intensity threshold, one to convert the slow varying voltage at the delay capacitor to a digital signa. Ideally it should have some hysteresis, otherwise the output can be bouncing with multiple edge under circumstandes. The second comparator can be replaced by a ST CMOS gate.

The right single comparator circuit does not provide a clean digital signal and isn't suited for all kinds of load.

 

[crutschow]

The LM339 is in a quad package.
Why do you want to reduce the number of comparators?

 

[earckens]

The LM339 is in a quad package.
Why do you want to reduce the number of comparators?

In order to have 4 detectors instead of 2 on the same pcb.

--- Updated Jan 20, 2022 ---

The circuit needs two comparators (or other active components with similar function) by design. One provide the light intensity threshold, one to convert the slow varying voltage at the delay capacitor to a digital signa. Ideally it should have some hysteresis, otherwise the output can be bouncing with multiple edge under circumstandes. The second comparator can be replaced by a ST CMOS gate.

The right single comparator circuit does not provide a clean digital signal and isn't suited for all kinds of load.

That is clear (ST CMOS gate, such as in "logical gate"?)
Query shelved, that will be two detectors per comparator. Adding a CMOS gate, although a good idea, defeats my purpose to limit the number of comparators/gates/.. per detector. I think I rather stick with 2 comparators per detector and not to add a different IC.
Thank you for that clarification.

 

[BradtheRad]

Simple rising-falling detector.
1) Capacitor charge lags behind the source signal
2) creating voltage differential
3) op amp pins to supply rail to indicate rising or falling
4) unchanging signal causes output to adopt midway volt level

Simulation with signal .001 Hz (17 minute cycle).

 

[earckens]

Simple rising-falling detector.
1) Capacitor charge lags behind the source signal
2) creating voltage differential
3) op amp pins to supply rail to indicate rising or falling
4) unchanging signal causes output to adopt midway volt level

Simulation with signal .001 Hz (17 minute cycle).

View attachment 174020

Oh ok, this is with 2 comparators.
Thanks,
Erik