Questions about flyback feedback loop designing

Hello!

I am designing a 75W flyack SMPS, which input voltage is 18v~36vdc, output voltage is 110vdc, and output full load current is 0.65A.
The transformer primary turns are 4Ts and the secondary turns are 40Ts.
The pwm IC is UC3842B and the feedback IC is TL431.

The output ripple voltage occur oscillation while output load is between 0.1A and 0.2A.

So, I want to anyone help me solve this problem.Thanks.

iepower, can you offer more details? The output doesn't oscillate when the output load is greater than 0.2A? If you vary VIN, how does this oscillation change?

Can you include a schematic? (More details should be known, such as switching frequency, transformer inductance, compensation components used, output capacitor, etc.)

Hi ZekeR,
About schematic please see this attachments.
The switching frequency of this SMPS is about 50KHZ. The transformer primary inductance is 10uH.

thank you for your help!!

I have to wonder if it has to do with the fact that you have two voltage feedback loops, which isn't normal in a flyback converter. Can you verify whether just one or both optocouplers are conducting when the oscillation happens?

Also what is the frequency of the oscillation? When oscillating, does the converter operate in DCM or CCM? It might be a RHP zero issue.

edit: also what's up with the two secondary windings and rectifiers in series?

there are two control loop in the flyback. one is output voltage feedback control loop and another is output over-voltage protection control loop.

I have confirmed that when the oscillations occurs,only one optocoupler control the feedback loop.

The oscillations frequency is about 1KHz and the converter operate in DCM.

If it's operating in DCM then it should just have one pole and one zero and be very easy to make stable. And your oscillation is far below the resonant frequency of your output LC. And from what I can see on your output capacitor's datasheet, it has a very high ESR, which means the converter might be operating even above the ESR zero frequency. But I see that your feedback has a type III compensation network, so you should have huge amounts of phase margin. I'm not sure, but it might be possible that having an effective phase margin of 180 degrees or more can cause instability.

But I doubt this is a case of small signal instability, but rather some kind of relaxation oscillation. Check every node in the feedback network and see if the shape of the oscillation changes at any point. It should look the same if it's a small signal oscillation. If it's not (like it clips anywhere) then it's some nonlinear phenomenon.

also one potential issue I see it that your TL431 is biased with the same current as the photodiode. Not sure if this is really an issue, but I've always seen this circuit done with separate bias resistors.

Hi mtwieg,

First of all, thank you for your supports!

I have solved this problem in the afternoon. The compensation network parameters are as follows:
R38=220K, C27=100nF, C26=1nF,C38=33nF,R32=1K,C24=10nF and R31=100Ω。

I do not have enough time to do the small-signal analysis, because the samples of this project will be delivered to customers at the end of this month.
I also want to use simulation to analyze and solve this problem, but simulation modeling is not so easy thing.

A full small signal simulation of the converter probably isn't worthwhile, because it probably wouldn't actually reveal the problem. What you should do is take a closer look at the oscillation and how it propagates through the feedback network. For each relevant node, look at the amplitude, phase, and overall shape of the oscillating waveform. That should give you good clues as to where the trouble is.