How to modify/ design error amplifier compensation circuit (with TL431) for variable flyback converter and make it stable.

Thread Starter

polashd

Joined Jan 16, 2014
6
I bought a power supply adopter (old, used, made for some device) of 24v 4.5A (from 230vac mains). It is a fixed output psu. My intention is to make it a variable psu of 30v to <=5v (as low as possible, and continuously variable). Schematic (simplified) of its error-amp circuitry is attached below.

I tried the following methods (varying different component values)-

1) Replaced R_lower2 with a 10k Pot (Rv_lower1 is a variable resistor of 2k- adjusted to ~1k by manufacturer).
2) Replaced R_upper1&2 with a 100k Pot.
3) Replaced R_comps with a 100k Pot.
4) Removed DZ1&2 and connected the err_amp to a 12v source (common ground -Gnd_B).

Whatever I did ends up to same situation- up to 16v it goes without any problem. But when I try to lower the voltage further the output becomes very unstable.

I searched internet (including youtube, chatgpt3.5) and got the idea that the problem is with the compensation network of the error amplifier.

I also got some good documentations on the compensation network which describe about – transfer function, bandwidth, frequency response, poles & zeros and many more things. As a simple hobbyist these terms and equations are very complicated for me. I don’t understand where to start, how to proceed and steps involved.

Can anyone please help me in this? What should be the values of different components, how do I calculate step by step (so that a hobbyist like me can do it, for continuously variable psu).

Note:
a) in the schematic I added TL431, in actual circuit it’s AZ431.
b) I didn’t get symbol and .lib/ .sub for LD7575PS- so I edited symbol of another ic in ltspice.
c) as per the equation in the datasheet of PWM IC –for a 120k resistor from RT to ground the switching frequency sb around 54kHz.
d) If possible provide me a working simulation schematic (with symbols and .lib/ .sub of important components like TL431 (or AZ431), optocoupler (in the PCB it is EL817), etc).
e) Can I do it with an op-amp (without TL431, achieve even lower output voltage like <3v).
 

Attachments

Pyrex

Joined Feb 16, 2022
239
Hi,
it seems, 431' is not a problem. Problem is the topology of the PSU. If it 's a flyback PSU, it's optimized to output 24V. And a minimal current to be provided. Flyback can't work if output current falls significantly. So, usually in such a PSU a big power resistor at the output is provided to maintain the minimal current.
If it's a half -bridge topology, the situation is a bit better.
In any case, circuit diagram of the PSU is very helpful.
One more question- how the PWM controller is to be powered , if it's powered from the same transformer and output is lowered, say to 5V?
 

Thread Starter

polashd

Joined Jan 16, 2014
6
One more question- how the PWM controller is to be powered , if it's powered from the same transformer and output is lowered, say to 5V?
It is powered from the auxiliary winding of the transformer.
While testing- i use 47ohm power resister as test/ dummy load which draws 400ma to 500ma.
 

Pyrex

Joined Feb 16, 2022
239
It is powered from the auxiliary winding of the transformer.
While testing- i use 47ohm power resister as test/ dummy load which draws 400ma to 500ma.
It is powered from the winding of the same transformer, right? So, if 5V is set in 24V PSU ?
It is wise to provide a separate small power PSU for PWM controller to have a stable and constant voltage for the controller.
Anyway, flyback is not a best topology if you want to built a laboratory PSU
 

Thread Starter

polashd

Joined Jan 16, 2014
6
It is wise to provide a separate small power PSU for PWM controller to have a stable and constant voltage for the controller.
you were right (thanks a lot). Chatgpt misguided me from the beginning by saying that its a compensation loop (of error-amp) problem.
I disconnected the connection from auxiliary winding, powered the control circuit using a 12v adopter and some other changes in the secondary (mostly various resisters). the output now can be adjusted down to 2.6v.

I want to replace the 12v adopter, power the control circuit from main DC input (~305v, rectified from ~230v AC) using voltage divider-diode-capacitor- zener diode (of ~20v). the controller IC works from 10v to 30v. Can I (should I) do it.

The reason of doing this- 1) I need to use one in the secondary to power an op-amp (maybe used for current limit. voltage feedback works without any separate power source), 2) as it's a pre-made power adopter (for some device) i have space problem.
 
Top