Electret output and op amp choice

(I realise this sort of question comes up a lot but I can’t find the answers I’m looking for with the search terms I’ve thought of. Sorry if it’s all boring and been covered 100000000 times.)

I’m hoping to detect 1kHz sound pulses (hence the LPF and HPF) and the LM393 is just what I have without going shopping.

I’m struggling to find out what sort of output I could expect from an electret mic capsule in this circuit and whether or not I’d need to amplify it before passing it to a comparator. Any thoughts on the output, or whether I’m in the right range for it’s resistor?

Is the LM393 a reasonable choice for this or would I be better using an opamp rather than a comparator? Perhaps a dual package opamp, using the first stage as an amplifier and the second (without feedback) as a comparator? Have I understood the posibilites correctly?
If an opamp would be better, I’ve read the 741 is quite noisy with a slow slew rate so may not be a good , and similar circuits seem to use LM358.
Or (if I have actually started to understand the specs) the NE5532P seems to be faster, with suitable voltage requirements, and pleanty of bandwidth.

Any thoughts much appreciated.

#ElectronicNovice!!!

As the LM393 has up to 9mV input offset voltage, your circuit may not work unless the sound is quite loud. I would therefore recommend more gain before the comparator. However, why use a comparator? Why not turn the LM393 into an analogue amplifier with about 100x gain and feed that into an analogue input of your Arduino?

The LM393 has an open collector output, and needs a pullup resistor from it's output to +5volt (e.g. 10k).

Your circuit won't work, unless the elektret outputs 5volt t/t (unlikely), because the +input is at ground level and the -input is mid-voltage.

Your circuit would sort of work if you change R4 to ~100k.

Google "opamp bandpass filter". You might find a better circuit.
Leo..

Wawa:
Your circuit would sort of work if you change R4 to ~100k.

R2 needs to go to the mid point of the potential divider and a decoupling capacitor provided from there to ground.

The LM393 has an input common-mode range that includes ground.
Leo..

Wawa:
The LM393 has an input common-mode range that includes ground.
Leo..

Looking at a datasheet, I see input voltages can go to -0.3V. So isn't the solution to scrap R4 and make R5 = 10kΩ ? (The 10kΩ will minimise offset due to input bias currents)

Where did your R4 = ~100kΩ come from?

Some voltage difference between the inputs of the comparator (hysteresis).
One input is ground referenced, so to me it makes sense to create a small voltage (~25mV) on the other input.
Leo..

Wawa:
Some voltage difference between the inputs of the comparator (hysteresis).
One input is ground referenced, so to me it makes sense to create a small voltage (~25mV) on the other input.

I think the idea is to produce a zero crossing detector, especially important as the signal from the electret microphone could easily be less than 25mV, depending on sound level of course. A small amount of hysteresis could be desirable to prevent the comparator changing state when there is no received signal apart from acoustic and/or electronic noise. Such hysteresis would be provided by a tiny amount of positive feedback on the comparator (preferably arranging that the hysteresis is symmetrical with respect to zero crossing).

Yeah, reverse the + and - inputs and run a high value resistor from the output back to the junction of R4 and R5 and +. The filter output should go into the - input. That will give you a little hysteresis without messing the filter up.

You need to put a high value resistor between the +/- inputs, couple to just one of those
pins capacitively and then it ought to work and be fairly sensitive, so long as there is some
DC return path from the inputs to somewhere.