What cap, what resistor?

[Johan_Ha]

In another thread I showed this circuit:

This is supposed to be a peak detector. The piezo creates one pulse (due to a hit on a board by a ball). The capacitor C1 gets charged and holds the voltage. Then Micro:bit reads the voltage at P0 and resets the voltage with a signal from P3. My problem is that a shock from the piezo raises the voltage only a bit. Next shock raises it again, and again and again. The cap is probably too big, the piezo obviously delivers only a very small charge, though the voltage is probably rather high, some 5 V.
So my first question is how to figure out the proper size of the cap. It should reach the max voltage with only one pulse from the piezo (or any lower voltage, if the shock on the piezo is smaller).
My second question is to figure out why there's a continuous raise of the voltage at C1, even if there's no piezo activity. Only sources are P0 and the piezo. The piezo might produce a voltage due to sound, or it just lying somewhere, where there's mechanical vibration. How could I stop leak into the cap from anywhere else, when I'm only interested in the piezo pulse?

 

[(*steve*)]

You should probably use a proper peak and hold circuit.

See here.

 

[(*steve*)]

If you have weak pull up on P0, or if it sources leakage current, you'll slowly charge the capacitor.

The circuit you have would need to use a very small value of capacitor because the piezo sensor cannot generate a significant current.

 

[AnalogKid]

I can state with perfect authority: Yes, the capacitor is too big.

ak

 

[Johan_Ha]

I put a .47 uF cap there and just using a multimeter it seemed to work. Though the cap seemed to discarge through itself and the piezo. Of course it discharged through the multimeter, too, but also while the multimeter was not attached. But that lasted several seconds and is not a problem. A bigger problem is the microcontroller charging the cap while waiting for the piezo pulse. Would a proper peak detector circuit be better isolated from the microcontroller?

 

[(*steve*)]

Would a proper peak detector circuit be better isolated from the microcontroller?

Did you watch the video?

 

[Johan_Ha]

Yes. Although I jumped from thing to thing. I just watched more thoroughly. I didn't pay attention to the precision thing, because I thought I wouldn't need it. After watching again, I'm still not sure whether the circuit with two opamps will be unaffected by my microcontroller.

 

[(*steve*)]

I'm still not sure whether the circuit with two opamps will be unaffected by my microcontroller.

That's the whole point of it

 

[Johan_Ha]

Ok, thanks very much for it.