I want to use an optocoupler to linearly step down a DC voltage which willy vary from 0V to 400V. That is, I want the output to be directly proportional to the input and I want it to be galvanically isolated from the input

V1-0 to 400V, V2- 12V

Now, for this purpose I used IL300 and MCT2E earlier. The problem with IL300 was that the slope (linearity factor) varied across different samples of the IC. I can not allow that to happen. And the problem with MCT2E was that the linearity lived only up to a particular value of input ad the output tended to get saturated after that. I realize that changing Rs might have solved this problem, but the saturation couldn't be eliminated quite very properly. Now, I'm looking for some other optocoupler for this purpose. I have shortlisted LTV 817 for this purpose but I'm not sure how it is going to perform. So, I want to select atleast two more ICs for this purpose and then test all of them to select the best one.

What I require from the IC:

  1. Directly proportional output for a DC input varying from 0-400V.
  2. Galvanic Isolation between input and output.
  3. Output to vary linearly with the input, between 0 to 12V.

Can you guys please suggest me some optocouplers that will help me solve this purpose? Note: There might be devices other than optocouplers, available for this purpose, and I welcome suggestions regarding them as well, but I prefer optocouplers.


2 Answers 2


A standard way to do this is to use a second optocoupler (preferably on the same chip) to compensate the non-linearity of the first one. The example below is from Analog Devices AN-106. The principle is to put a current through the second optocoupler's primary such that it secondary gives the same result as the secondary of the first optocoupler. Then that current must be the current through the first optocoupler's primary.

enter image description here

An IL300 optocoupler is a slight variation on this theme: it has two secondaries coupled to the same primary, so the primary side can do the compensation.

enter image description here

As Dirk suggests, you could go digital at the primary side. A digital signal is much easier to get through an isolation barrier. You don't have to build this yourself, check for instance the HCPL-7510 linear isolation amplifier :

enter image description here

  • \$\begingroup\$ The IL300 is better than using two optocouplers because two LEDs are not guaranteed to age at the same rate. The IL300 uses only one LED so it does not have that problem. \$\endgroup\$ Commented Aug 26, 2014 at 9:40
  • \$\begingroup\$ Digital isolation is practical solution and can be implemented, but expecting a linear output using optocoupler is tricky. I suggest to go with HCPL7510 option!! \$\endgroup\$
    – AKR
    Commented Aug 26, 2014 at 16:33
  • \$\begingroup\$ - LOC111 Inexpensive too.find datasheet from below link ixysic.com/home/pdfs.nsf/www/LOC111.pdf/$file/LOC111.pdf \$\endgroup\$ Commented Jun 30, 2017 at 12:20

I would suggest doing it digitally using a voltage to frequency converter one side, and a counter the other. Alternatively some kind of pulse width modulation one side and simple smoothing on the other. Either way it is going to be more complex that your preferred method - maybe a couple of PICs.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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