# [Optocoupler] Crossover frequency feedback loop in function of Current transfer ratio

#### leo0001

Joined Sep 30, 2018
79
Hi,

Here is the schematic :

Here is the following assertion :

"Small Signal Effect: A reduction in CTR by 2x will cause the DC gain and crossover frequency of the for-ward converter’s voltage feedback loop to drop by 2x, assuming all other parameters are constant. Likewise, an increase in CTR by 2x, assuming no change in other parameters, will cause the DC gain and the crossover frequency of the voltage feedback loop to increase by 2x. The voltage feedback loop must be designed ensuring that at CTR(MAX) (maximum CTR of the opto-coupler), the crossover frequency of the feedback loop stays well within the Nyquist frequency of the system (= switching frequency / 2). A good rule of thumb is to design the voltage feedback loop’s crossover at about 1/10 of the switching frequency for an opto-coupler at the nominal value of CTR. "

I do not understand why the crossover frequency is function of CTR and why CTR = a*crossoverfrequency. Could please tell me the relation between the two parameters ? If you have any documentation about this, it would be great to read it.

Thank you very much and have a nice day !

#### crutschow

Joined Mar 14, 2008
28,538
You'd have to write the equations for the loop response.
Is that posted in the information you have?
Where did that schematic and info come from?

#### leo0001

Joined Sep 30, 2018
79
Thank you for your reply. The information comes from the LT8311 datasheet. You can find it on google if you want to have futher informations. Nevertheless you will not find the transfer function. I will try to obtain approximately the transfer function of Vout/Ip (Ip = Iprimary). Do you think it is a good idea to find this transfer function Vout/Ip or do you think about an other transfer function. I think it is not a simple task to find the transfer function ... But I agree with you, it seems possible that like for a RC low pass filter (1st order), the crossover frequency is dependant on RC. It could be dependant on CTR in my case.

#### crutschow

Joined Mar 14, 2008
28,538
Do you think it is a good idea to find this transfer function Vout/Ip or do you think about an other transfer function.
It is a good idea if you want to totally understand the loop dynamics, but it's not easy to do. And it's difficult to incorporate all the variables, such as input voltage, output voltage, current, temperature, etc.,
It rather depends upon the application for the circuit, as to how much analysis you need to do.

The data sheet takes a test-and-select approach, starting with some nominal values, and then tweaking them in the system to get the desired response.

A simulation of the circuit with LTspice (free from Analog Devices) can also help greatly with optimizing the design.
The simulation makes it easy to vary component values to see the effect on the circuit performance.
Of course, the circuit must always be breadboarded to verify the operation after the simulations determine (supposed) optimum values.