This simple circuit provides better than 1% accuracy in optically coupling a DC voltage signal. To compensate for the non-linearity of the optical isolator, an identical device is connected in the feedback loop. To further enhance stability, both optical isolators are located inside the same IC. Both the input and output circuits operate from single power supplies (one for the input side and a 2nd for the output side). Prior art generally requires both positive and negative power supplies –elimination of the negative power supplies simplifies the circuit.
Optical Isolator Matrix
Optical isolator selection
There are numerous acceptable devices –see matrix. All these devices are current with DigiKey and most cost less than $1. Also, all are in the DIP-8 package so they may be easily plugged into a protoboard. All have CTR (current transfer ratio) ranging from 50% to about 500% –quite sloppy. What makes this work is that the two devices inside the same IC are likely matched in CTR to some degree.
Note that all these devices are not necessarily interchangeable because pin-outs vary –I found this out the hard way when I attempted to get the circuit working. Check the device data sheet if not using the specified device.
Voltage isolation issues
While DIP optical couplers have up to about 5kV isolation capability, this circuit cannot support above about 100V or so because one of the outputs is fed back to the input amplifier –this puts the voltage isolation between adjacent pins on the device rather than across the device. To maintain a high voltage rating, use two single devices and match the CTR to some degree.
Op amp selection
I generally use the LM358 dual op amp whenever I need a single ground sensing op amp. I do this because there is no DIP single device that will do the job (that I know of). There is the LM321 single op amp, but it is unpopular and comes only in the SO-23-5 SMD package. At any rate, the LM358 is very inexpensive.
A positive signal input causes the output of U1 to shift positive. This causes both LEDs to conduct. U3-B turns on and provides a positive going feedback signal. When both op amp inputs are equal, the output of U1 stops integrating. At the same time the output of U3A presents a positive going signal to U2. Adjusting R2 so that R1 + R2 »4.7K causes the output of U2 to come close to the input voltage signal –further adjustment of R2 will trim for the difference in CTR so that the output voltage will exactly match the input voltage. The easiest way to test performance is to connect a DVM between input and output.
Potentiometer R4 accommodates gross variations in the CTR (both devices). R4 is adjusted so that it drops about 2V at maximum signal level –this is very non-critical.
C1 is a compensation capacitor –for faster performance, this may be adjusted down –if too low, the circuit may oscillate. I did not test the response time.
Bench test short cut
To make test and evaluation easy, I used a single dual op amp and only one 12V supply. While it provides no isolation, the transfer function is identical.
I was able to obtain a linearity of about ±10mV over the input range of 0 to 5V –not too shabby a simple circuit. That amounts to about ±0.2%. I did not test it over the temperature range, but expect it to be under 1% or so. Also a greater input voltage range may be possible –especially with a higher Vcc.