The power supply I was looking at uses an optocoupler in its path for feedback to control the output voltage. This is a common and well-established design approach, but the unconditional stability of the feedback loop was not entirely certain.

I decided to measure the transfer function of the optocoupler in two ways, as seen below. I knew the device datasheet presented this feature, but I wanted to look at it first hand (Figure 1).

Figure 1 Optocoupler transfer function test settings.

The 1.33k and 150 ohm resistors just turned out to be handy, so I used them. A pair of 1k and 100 ohms would be enough just as well, but I’m just not digging at the moment.

Both test results appear to confirm the presentation of the data sheet (Figure 2).

Figure 2 Optocoupler transfer function test results.

The upper track of Figure 2 follows the upper sketch of Figure 1, the lower track follows the lower sketch. It was reassuring to note the symmetry of the results of the two tests. However, after some thought, I realized that there was some kind of cycle-enhancing mine that needed to be known.

Using the lower trace in Figure 2 and measuring the slope of the transfer function, we see the following (Figure 3):

Figure 3 Close-up study of transfer function.

The “linear” nature of the optocoupler should not be taken too literally. The device transfer function does not switch or anything like that, but the first derivative of the output relative to the input, which means the slope of the output relative to the input, varies depending on where you place the operating Q-point of the device. The variability in the device I studied was almost 19 dB.

This large gain deviation can have a madness effect on the overall transfer function of the feedback circuit, which may lead to conditional instability of a slightly stable feedback circuit.

The warning note of all this is to make sure that you have checked the true inclination of any optocoupler you have chosen for yourself and where you choose to set its Q-point.

John Dunn is an electronics consultant and a graduate of the Brooklyn Polytechnic Institute (BSEE) and New York University (MSEE).

Related content

https://www.edn.com/linear-opto-couplers-and-the-loop-gain-booby-trap/