Optoisolator draft- RFQ

I'm putting this out there for comments. Please let me know how this works for you.


by rberne
December 16, 2012

NIce model.

Because CL doesn't properly handle subcircuits/hierarchy, you can simplify the way you have tried to make the opto into a subcircuit ...

In the simplified version you have to edit the subcurcuit pin labels whereas - in yours - you have to edit the behavioural voltage source expressions making up the ideal DC isolation transformers for each subcircuit instance.

If you want to separate the net names between the test jig and the subcircuit then you could connect them together through 0V sources:

Note, although it shouldn't affect your model, be aware of:


in case any ringing or overshoots generate negative diode currents throught the junction capacitance due to lead inductances and driver transient behaviour.

by signality
December 17, 2012

Do what in the who now? are you saying that I can just put the same labels in different spots? If I would have known that... My circuits are about to get a LOT smaller.

It would be really nice if we could just write a behavioral function inside a custom chip and call it a day.

by rberne
December 17, 2012

I'm curious where you got your values for I3. I ask because I'm trying to model an AC input optocoupler SFH620A-3 http://www.vishay.com/docs/83675/sfh620a.pdf.

by cblauvelt
July 02, 2013

Vishay do not supply a spice model for this device:


If you require a model then please feel free to email me to discuss your requirements.


by signality
July 03, 2013

Well my budget is $zero so you're probably not interested in my requirements.

From what you're saying, you'd need the SPICE model to get those values and if Vishay doesn't provide one, then oh well.

by cblauvelt
July 03, 2013


The values that @rberne used in his model came from some Vishay spice model data.

However, it is possible to build behavioural models based on datasheet performance. How good they are then obviously depends very much on how good the datasheet is.

That said, many of the manufacturer's spice models for opto devices really are very basic if not to say poor.

Ultimately, a lot depends on how accurate you need your model to be.

These sorts of models are often all you need:

(and follow the links)

You may find that one of those but modified by adding a second, anti-parallel LED, D2 (but not an anti-parallel photodiode as it will provide an unwanted input current into the 2nd stage amp) and editing the expression for the Ictr1 CCCS from:


to something like:


might well be good enough.

(It probably will since the rest of CL is not yet up to providing detailed enough modelling for the rest of the surrounding circuitry.)

by signality
July 03, 2013

Post a Reply

Please sign in or create an account to comment.