In a circuit that considers two different ground domains which can change dependant on the numbers of batteries in a series array (think smart battery), I'm trying to isolate a current sense diagnostic output from a high side switch.

This high side switch will output a current linearly proportional to the load; say 5ma per 1A on the output for example. Because of the different ground domain, it's desirable to isolate this using an optocoupler. Will a non high linearity isolator (like the TCMD1000) with a CTR of around 100% provide a linear current output such that the current sense can be read as a voltage on an ADC on the phototransistor side?

  • \$\begingroup\$ The linearity of an opto-isolator is not that great. Often, when used in a closed loop controller, the non-linearity does not pose a problem. What kind of accuracy are you hoping to achieve? Since you are feeding the output into an ADC, I am guessing there is a micro-controller involved. If so, you could calibrate the opto-isolator, and have a table in the micro-controller correct for non-linearity. Otherwise, you may want to purchase an opto-isolator that is designed for linearity. By the way, opto-isolators also vary greatly in terms of speed. You do not mention speed requirements. \$\endgroup\$ – Math Keeps Me Busy May 5 at 1:20
  • \$\begingroup\$ for one of a kind probably since you can adjust the circuit. Look at the tolerance, it is quite large. Also consider what happens over the years to the LED, it will get dimmer and lower the transfer. \$\endgroup\$ – Gil May 5 at 1:21
  • 2
    \$\begingroup\$ No, you need one of those fancy optocouplers that have one LED and two detectors in the same package where the second detector is run back to the output to provide linearizing feedback. \$\endgroup\$ – DKNguyen May 5 at 1:22
  • \$\begingroup\$ @MathKeepsMeBusy as far as speed, it's low; 1 sample a second would be sufficient. For the time spent on calibrating the opto isolator, I'd imagine it'd be easier to deploy a high linearity optocoupler. Thanks for your comment \$\endgroup\$ – Alex Turner May 5 at 5:22

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.