We may examine the linear properties of an ET1104 opto-isolator as follows.
Note that there was no reason in particular for choosing an ET1104 to examine. There just happened to be one on hand. Also, the 1.33K to the LED was used only because it happened to come to hand at the moment. It could just as well have been a 1K.
Happily too, there is absolutely no sign of parasitic oscillation in either configuration. Even putting 1.0 µF across the emitter follower's 150Ω load only resulted in some phase shifting with no observed propensity to oscillation at all.
The LED forward voltage drop (the reverse is the drop of the 1N4148) is seen as follows:
The forward voltage drop of the LED is very nearly 1V across the entire linear operating range (see below). Therefore, the LED current at 10 mA above is (7V - 1V) / 1.33K = 4.5113 --> 4.5 mA.
The current transfer ratio here is taken as 10 / 4.5 = 2.22:1.
However, changing the x-axis from 2V / Div. to 5V / Div., we extend the view the emitter follower output and examine a break point, a non-linearity which may have consequences for us:
When we examine the slopes on both sides of the break point, we find as follows:
This 19 dB variability of opto-coupler gain could translate to 19 dB variability of feedback loop gain and could possibly translate into conditional feedabck instability problems.
That is just for starters. Photo-transistor based isolators also exhibit substantial device-to-device and over-temperature variations
Posted by: George Storm | May 23, 2011 at 05:45 PM
I know that temperature is a big problem, but what about using a dual in one package with one section in the front of an op-amp and the other in the feedback with a gain of one ? ? ?
(I am a retired analog EE. I wish that I still had a lab. at home to do some contract work.) (FUN STUFF)
Posted by: Mike Muegge | May 24, 2011 at 11:13 PM
Use a dual with feedback through the second pair in closed loop
Posted by: Paul | May 25, 2011 at 04:26 AM
Hi, Mike and Paul.
I just e-mailed each of you a dual-opto circuit that does just as you describe.
Thank you.
John
Posted by: John Dunn | May 25, 2011 at 07:09 AM