3

Background:

When using LTspice to simulate such a circuit (the opto-model is called PC357NT and is part of some exhaustive lib called "ZZZ" which I downloaded a long time ago. The PC357NT seems to be a pretty standard opto with ~100% nominal CTR):

enter image description here

I can identify three regions in the behavior with respect to the voltage V1: A linearish region, in which the output transistor is saturated. And then two regions of more constant current, where the transistor becomes base-current-starved. The starvation happens at rather different current because of the very different forward and reverse \$\beta\$ of the transistor.

enter image description here

If you pay attention to the blue circle, the behavior is rather linear around zero volts. When simulating a small bipolar current on the secondary side, it passes rather symmetrically. E.g. when applying +/- 2.5 V, I get a rather symmetric current:

enter image description here

Question:

Does this behavior qualitatively agree with real optocouplers? I.e. will they pass a bipolar current when bipolar voltages are applied to their output?

I have a project where I need an isolated switch to pass/block about 100µA bipolar current (actually +/- 2.5 V at 22 kΩ impedance) and it has to work faster than ~50 µs and I already have optocouplers in the BOM.

tobalt
  • 18,646
  • 16
  • 73
  • Regarding your specific question/project, how can a series switch (optical or otherwise) block a 100 uA current (bipolar or otherwise)? You need to show a schematic to avoid ambiguity. – Andy aka Mar 09 '23 at 14:32
  • OK. If it helps I've used regular optos for switching precision resistors across certain low power loads (and with decent accuracy) so, certainly the concept of optically switching "off and on" works but, I can't be sure about bipolar stuff. – Andy aka Mar 09 '23 at 14:47
  • @tobalt please share the model number of the optocoupoler model – Voltage Spike Mar 09 '23 at 15:02
  • @tobalt The datasheet says Ic should have a minimum of 2.5mA (don't know why). I can't find the spice model, is it possible to post the model? – Voltage Spike Mar 09 '23 at 18:11

1 Answers1

2

I measured the behavior myself, for a generic SOP-4 package opto with nominal 100% CTR.

enter image description here

In this first figure, one can identify that both forward and reverse transistor current scale with the LED current as expected, and that the reverse current is indeed much smaller, but appreciable. The maximum reverse current is about a 100th of the forward current for given \$I_\text{LED}\$.

enter image description here

This second plot shows the behavior around 0 Volts. The behavior differs in that all the curves run essentially through the origin, different from the model, which indicated that the base current would exit more through the emitter, leading to some kind of "quiescent" current at \$V_\text{CE}=0\$. This doesn't seem to be the case.

The behavior looks almost like the Shockley equation, which is linear in a small region around 0. I found that the LTspice diode model .model opto D(Is={Isat} N=1.2) works well to model this behavior, i.e. an emission coefficient of 1.2 and a saturation current according to the first figure above.

So, yes Optos can be used as bipolar analog switches if the currents are very small and linearity is not important. But in general, they will rectify the signal quite a bit.

tobalt
  • 18,646
  • 16
  • 73