I have attempted this diagram from here:
with one modification - instead of the 4093 circuit ( comparator+inverter+hysteresis gates) I have used a HC14 (inverter+hysteresis gates). Can this be why i am not seeing much difference when varying the pot ? Tried with 3 values for C1 - 3.3, 10 and 100 nF. It doesn't seem to make any difference as to how sensitive the circuit is. The output is connected to a MCU pin which interrupts on all changes and blinks a led.
Also, I don't quite understand the bottom two notices from that pdf:
- The point of output must become square wave. (cannot use as shock switch)
- Take low or high of output square wave into your alarm system and adjust sensitivity from the delay time length
I should say that another breakboard of this same sensor works much better by adjusting the trimpot - this one
I have read the raw analog output right from the sensor's pin and plotted both the dev board readings and my own PCB readings on the same graph. The dev board readings are much quieter than the readings from my PCB. Maybe the weight of the board and its cables are the main cause in the outputs difference ?