Feedback Isolation Doubts

[sabu31]

Hi all,

I have a doubt regarding Feedback isolation for power converter control. In SMPS, its using TL431 and Optocoupler (say PC817). The operation of optocoupler in saturation region is clear. But for feed back control the optocoupler is operation in linear region.
How do we design the parameters for operating in linear region.?

The output of the system (optocoupler) i will be giving to analog input of microcontroller.

(2) Is this approach better or directly use isolation amplifier ?

Is this circuit good for studying gain characteristics for designing digital controller

 

[FvM]

The schematic should be supplemented with frequency compensation, otherwise it would work as comparator and can't constitute a stable feedback loop.

Your feedback solution has to be selected according to bandwidth and accuracy requirements of your application. What is it?

 

[sabu31]

Thanks FvM for the Reply. What I want is to provide compensation of a DC-DC converter. using micronctroller but use TL431 and Optocoupler as an isolation.

The schematic should be supplemented with frequency compensation, otherwise it would work as comparator and can't constitute a stable feedback loop.

Your feedback solution has to be selected according to bandwidth and accuracy requirements of your application. What is it

 

[FvM]

I see, but it's not entirely possible the way you try. Suggest to analyze the specific behaviour of TL431 reference circuits and use it appropriately for your purposes. Optocouplers can't transmit measured values accurately due to gain variation and temperature drift. Thus usual TL431 circuits transmit the output of a PI error amplifier.

Viso = (Vact-Vref)*F(s)

Despite of possible offset and gain variations, Vact will always follow Vref.

Vref will however be set on the output side and can't be programmed in the controller. For a programmable setpoint you need a true isolation amplifier or something equivalent.

 

[crutschow]

U2 is upside down.
The normal operating current is from anode to cathode.

And it looks like you are trying to use the LT1009 as the typical TL431 for this application, but the LT1009 does not work the same way.
The input on the LT1009 is only for trimming the output, not adjusting the output.

Did you not look at its data sheet?

 

[Easy peasy]

you are really looking for the IL300 then .... opto

 

[sabu31]

I see, but it's not entirely possible the way you try. Suggest to analyze the specific behaviour of TL431 reference circuits and use it appropriately for your purposes. Optocouplers can't transmit measured values accurately due to gain variation and temperature drift. Thus usual TL431 circuits transmit the output of a PI error amplifier.

Viso = (Vact-Vref)*F(s)

Despite of possible offset and gain variations, Vact will always follow Vref.

Vref will however be set on the output side and can't be programmed in the controller. For a programmable setpoint you need a true isolation amplifier or something equivalent.

Thanks FvM. I am having more clarity.

--- Updated Jul 23, 2022 ---

I see, but it's not entirely possible the way you try. Suggest to analyze the specific behaviour of TL431 reference circuits and use it appropriately for your purposes. Optocouplers can't transmit measured values accurately due to gain variation and temperature drift. Thus usual TL431 circuits transmit the output of a PI error amplifier.

Viso = (Vact-Vref)*F(s)

Despite of possible offset and gain variations, Vact will always follow Vref.

Vref will however be set on the output side and can't be programmed in the controller. For a programmable setpoint you need a true isolation amplifier or something equivalent.

Thanks FvM for the reply. I though if I there was some mechanism to use conventional opto PC817. then we can avoid additional power supply required for isolation amplifier (required at both input and output side).