Advice on bidirectional switch on/off switch for low voltage daq isolation from AC?

Thread Starter


Joined Feb 21, 2017
I have a pulsed signal of a few V lasting 300uS superimposed on a 220 VAC 60Hz .
The pulses are always happening when VAC is passing -3V & rising ( we are generating them accurately)
I want to extract this pulse voltage from AC carrier with a DAQ.
For accuracy reason ( the pulse shape is important for the application) high pass filters are not considered.
I am considering using some kind of switch between the AC line and the daq input.
the switch will be open most of the time as the DAQ requirement is +/-10V and closed when the AC voltage is within this range.
This means that this switch must be bidirectional with on/off capabilities.
The daq impedance is high so current is low (mA range)
My question: what switch technology could be recommended for this application?


Joined Aug 21, 2008
One approach is to use an isolation transformer to couple to the power line. On the secondary you can use a back-to-back zener diode clipper to limit the peak voltages so you don't have to switch the whole 670 VPP waveform.

Following that you can use something like an 74HC4051, 52, or 53 to blank the signal.

Alternatively you can use a transistor designed for muting applications with the emitter grounded and the collector across the back-to-back zener clipper, and drive the base to keep the signal blanked except when you want to measure it.

Muting Transistor Datasheet

Note: Neither of the schematics in the post were drawn by me -they were found on the internet.

Thread Starter


Joined Feb 21, 2017
Thank You!
The diode clipping solution seems perfect.
I did not mention it in my post but of course the insulation transformer was included in our design.
the initial idea was to use 2 mosfet in serie to get bidirectional switching but we we were not sure it was the best idea for our application
and your answer proved it!
The idea after that is to use the accuracy of the daq at full capacity and to digital treatment so no blanking should be required, I think.
Thanks You again