I have a simple application where I need to isolate a body connected instrumentation amp from recording equipment. The floating side (patient) needs to be battery powered. Coupling is to be analogue (as is the reporting equipment). I am having trouble finding a analogue coupler chip that is low power and has reasonable isolation. Hope someone may know of one - I could probably use plastic fibers if they can be driven with an analogue signal.

  • 1
    \$\begingroup\$ can you send analogue PWM through a digital coupler? \$\endgroup\$ – Jasen Jan 23 '18 at 6:17
  • \$\begingroup\$ Yes, I was thinking of that possibility - it is really low bandwidth - maybe 3dB down at 500Hz \$\endgroup\$ – skyman Jan 23 '18 at 6:24
  • \$\begingroup\$ See Wouter's answer to electronics.stackexchange.com/questions/126610/…. \$\endgroup\$ – Transistor Jan 23 '18 at 7:09
  • \$\begingroup\$ You might want to look into the AD202KN by Analog Devices. It is an analog isolator rated for 1,500 volt isolation. It is not cheap but extremely linear over a wide range. It use PWM and a coupling transformer for galvanic isolation. \$\endgroup\$ – user105652 Jan 23 '18 at 7:27

Coupling is to be analogue

I hear what you say but...

This digital opto coupler is cheap (£5) and can be easily interfaced with a serial analogue to digital converter and a serial DAC on the output: -

enter image description here

I would choose this method over an analogue coupler because it is robust. With an analogue coupler there is a tendency to lose accuracy with changing temperatures.

I wouldn't use PWM either for the same reason although, if the PWM were quite a low frequency then the temperature dependency on rise and fall time through the coupler will be barely noticeable and not generate much of a pulse width error.

If going down the serial ADC/DAC route, the easier implementation is to have three opto couplers where the DAC side generates a start convert pulse and also feeds a clock signal to the ADC.

If you definitely need an analogue coupler then you might consider this technique: -

enter image description here

It uses near-identical phototransistors as dual receivers. One receiver is used within a local feedback system at the sending end and the other is at the receiving end and uses an identical analogue section as the send phototransistor. The idea is that because both phototransistors can be regarded as identical, providing you measure the send transistor's output and adjust the diode current continually to obtain the desired analogue level, the receive phototransistor will perform the same. In other words, it should produce a decent reproduction of the locally generated analogue signal.


Not the answer you're looking for? Browse other questions tagged or ask your own question.