You need to create a signal when the peak occurs so ESP knows exact time when to trig an ADC.
This signal can be made from precisely calculated RC differentiator. The edge before peak seems to be very sharp so you can use just RC with very short time constant.
There is also possible to use a comparator like LM393 with reference set to about 1/2 of smallest peak. After comparator signal rise delay the ADC start so you sample in the middle of peak.

Btw, there is no need a rectifier. Just use a small RC low pass so peaks become flat a little to measure. If you need to measure positive and negative peaks also shift the Gnd to Adc/2.

If the negative is no needed clamp it with diode.

Should I employ the sampler to generate this signal, or is using just one of the two methods sufficient?

Moderator edit: New thread created from this.

 

You need to create a signal when the peak occurs so ESP knows exact time when to trig an ADC.
This signal can be made from precisely calculated RC differentiator. The edge before peak seems to be very sharp so you can use just RC with very short time constant.
There is also possible to use a comparator like LM393 with reference set to about 1/2 of smallest peak. After comparator signal rise delay the ADC start so you sample in the middle of peak.

Btw, there is no need a rectifier. Just use a small RC low pass so peaks become flat a little to measure. If you need to measure positive and negative peaks also shift the Gnd to Adc/2.

If the negative is no needed clamp it with diode and check your subway gift card balance here.

Okay, Thanks Michal

 

There are alternate schemes to capture peak values, such as peak-hold arrangements, unless the timing of the peak is what is needed. So which is the goal: To know the magnitude of the peak value, or the timing of the peak, or possibly both.