Hello to everybody.
I need to estimate the energy of AC signal bursts coming out from an amplified piezo detector.
The bursts have typically 100 to 500 waveperiods in a frequency band from 20kHz to 300kHz, the signal amplitude of each burst is not constant and has maximum peak-to-peak voltage ranging from 100mV to 1V. According to that data the burts energy should be below 10 nanoJoule per event.
What I need to do is estimate the accumulated energy associated with a cluster of bursts (10 to 100 events minimum, all of them within the same 10 sec time window).
The first possible idea is to rectify the AC signal and use a long time integrator to accumulate the energy. Such a solution could be made sensitive enough for the nW range, but needs an ideal rectifier (problem since I am working with single supply and with a total pwr budget of 20uA) and will be sensitive to thermal drifts, especially for the integrator.
Anybody has suggestions about smart ideas for the ideal rectifier and micropower zero-drift op amp selection?
The second possibility is to use a thermal bolometer for the AC signal (see picture): the lower NTC of the sensing half-bridge is used as a load for the AC signal and the dissipated power increases its temperature, thus increasing the unbalancement between upper and lower arm. When needed the bridge is powered from a DC source and its unbalancement measured by an instrumentation op amp zero-drift or chopper stabilized.
This circuit would be very low power for sure (it will live switched off most of its time) but can it be sensitive enough?
Does anybody have a working bolometer project I could start from?
Does anybody have other ideas?
Thanks in advance.
I need to estimate the energy of AC signal bursts coming out from an amplified piezo detector.
The bursts have typically 100 to 500 waveperiods in a frequency band from 20kHz to 300kHz, the signal amplitude of each burst is not constant and has maximum peak-to-peak voltage ranging from 100mV to 1V. According to that data the burts energy should be below 10 nanoJoule per event.
What I need to do is estimate the accumulated energy associated with a cluster of bursts (10 to 100 events minimum, all of them within the same 10 sec time window).
The first possible idea is to rectify the AC signal and use a long time integrator to accumulate the energy. Such a solution could be made sensitive enough for the nW range, but needs an ideal rectifier (problem since I am working with single supply and with a total pwr budget of 20uA) and will be sensitive to thermal drifts, especially for the integrator.
Anybody has suggestions about smart ideas for the ideal rectifier and micropower zero-drift op amp selection?
The second possibility is to use a thermal bolometer for the AC signal (see picture): the lower NTC of the sensing half-bridge is used as a load for the AC signal and the dissipated power increases its temperature, thus increasing the unbalancement between upper and lower arm. When needed the bridge is powered from a DC source and its unbalancement measured by an instrumentation op amp zero-drift or chopper stabilized.
This circuit would be very low power for sure (it will live switched off most of its time) but can it be sensitive enough?
Does anybody have a working bolometer project I could start from?
Does anybody have other ideas?
Thanks in advance.
Attachments
Last edited:
