Hi,
I'm working with a project where I'm building an spwm inverter using 8051 (at89c2051) uC IC.
Now I'm planing to add a feedback system to regulate the output voltage. My thought was using an optocoupler to read the transformer secondary voltage and compare the output of the opto with a reference voltage using uC's internal comparator. As voltage error detected by comparator, uC will controls the pwm at MOSFET (IRFZ44) gate accordingly. Now the problem is output current of the opto (el871) is not proportional to the input current. So, I'm not comfortable using this technique (or should I use it? Please let me know).
I also searched internet and found tl471 is used along with tl871 in many feedback applications. But I'm not confirmed whether it would be doing good in my case.
Thanks in advance.

 

Your problem will be the linearity of the opto coupler. How does the output of the opto coupler relate to the transformer output voltage?
You will usually find the comparator and reference before the opto coupler, so that the output form the opto only needs to be a logic level: indicating either higher or lower.
You will also find the opto used in conjunction with the error amplifier (maybe you mistook the error amplifier for a comparator - it will usually have a capacitor between output and inverting input) The error amplifier produces an output signal that is proportional to the difference between measured signal and reference. Because the reference and measured signal are both on the same side as the opto, any non-linearity gets ignored by the rest Of the circuit.

 

You will usually find the comparator and reference before the opto coupler, so that the output form the opto only needs to be a logic level: indicating either higher or lower.

So, the comparator/error amplifier can seat on high side of transformer and send high/low signal to uC? Can I use tl431 in this case? (I'm sorry I posted the wrong ic name in the thread.)

Because the reference and measured signal are both on the same side as the opto, any non-linearity gets ignored by the rest Of the circuit.

So, with error amplifier (e.g. tl431) working with opto (el871) eliminates non linearity produced by the opto?

maybe you mistook the error amplifier for a comparator - it will usually have a capacitor between output and inverting input.

yes, I thought they are same. Thank you for pointing that out.

 

So, with error amplifier (e.g. tl431) working with opto (el871) eliminates non linearity produced by the opto?

Yes. The feedback to the TL431 means that the TL431 output current gets set to whatever current will produce the right voltage on the opto output so produce an output that matches the reference.
But, in your case, there is a catch:
If you rectify and smooth the output voltage that creates a delay.
The feedback system needs the capacitor across the TL431 to keep it stable.
Each delay can create a 90° phase shift.
Two delays makes a 180° phase shift.
A 180° phase shift creates an oscillator.

To make it work, you have to choose the delays so that it always operates in an area with a 90° phase shift, otherwise it is unstable. That might mean that the feedback takes a long time to respond to changes on the output.

It is explained here:

 

Quite often the IC manufacturer's websites will have very useful application literature to help with getting the best performance from their components. The applications information can be very useful.

 

Please let me know

OK
It looks like this is a square wave inverter. If so, how do you regulate it?
I think you can get almost the same information from the primary side of the transformer.

 

OK
It looks like this is a square wave inverter. If so, how do you regulate it?
I think you can get almost the same information from the primary side of the transformer.

Why does it look like a square wave inverter to you? Because it is using center tapped transformer? If so, is there any inconvenience? Although, I have seen H bridge MOSFET in SPWM inverter, I thought that there is no problem using center tapped transformer.

 

Why does it look like a square wave inverter to you?

I think your project will make the red output. Not the blue.
I think you have no control over the height of the red wave.

 

I think your project will make the red output. Not the blue.

View attachment 336073

No, the Final output waveshape is sinusoidal, as you can see on the oscilloscope.

 

OK. Now I see the capacitor as a load.

 

OK. Now I see the capacitor as a load.

Yes. W/o cap, its 2khz carrier frequency. With cap, its 50 hz sinusoidal.