Flyback converter: Load on unregulated winding

Hello, I have basic question on flyback converter with multiple windings.

Spec:
Input : 12V
Output 1: 5V, 1A
Output 2: 15V, 1.5A

The feedback is provided on 5V winding.

While testing i loaded 10% on 5V winding and kept 15V open, the duty cycle was around ~9%. But when i loaded 15V winding(~50%), the duty cycle increased around 40%.
/?
So my question is how the duty cycle increased when loading increased on 15V(unregulated winding), since feedback in present only on 5V winding?

This is expected. The windings of the transformers should be strongly coupled, meaning their that their voltages should always be related by their turns ratio. So when you load the 15V output and its output voltage starts to drop, that will also cause the 5V output to drop. The feedback loop will counteract this, and this will also regulate the 15V output. This is called cross regulation. The quality of the cross regulation will depend on various things like winding resistance, leakage inductance, and diode drops.

mtwieg said:

This is expected. The windings of the transformers should be strongly coupled, meaning their that their voltages should always be related by their turns ratio. So when you load the 15V output and its output voltage starts to drop, that will also cause the 5V output to drop. The feedback loop will counteract this, and this will also regulate the 15V output. This is called cross regulation. The quality of the cross regulation will depend on various things like winding resistance, leakage inductance, and diode drops.

Click to expand...

There should be strong coupling between primary and secondaries but between secondaries there should be as little coupling as possible, unless it is intended to be designed that way.
I did not understand when there is a voltage drop in 15V winding after it is loaded, how it will cause 5V to drop if there is no coupling between them.

abhinand rd said:

There should be strong coupling between primary and secondaries but between secondaries there should be as little coupling as possible, unless it is intended to be designed that way.

Click to expand...

How do you suppose the secondaries would be decoupled from each other?

mtwieg said:

How do you suppose the secondaries would be decoupled from each other?

Click to expand...

if the secondaries were coupled they would induce voltage on each other windings during switch off time in flyback.

There should be strong coupling between primary and secondaries but between secondaries there should be as little coupling as possible, unless it is intended to be designed that way.

Click to expand...

Actually - for a successful flyback, you want good coupling on the secondaries for good cross regulation AND good coupling from the secondaries to the primary to avoid large turn off spike on the primary side switch - this is well know to PE designers - we have many thousand flybacks in the field ... and many hours of development ...

- - - Updated - - -

Just to complete the picture, if you have a system where the pri and secondaries are well coupled ( per Mr Abinand rd ) then, by default, the secondaries are well coupled too ...

abhinand rd said:

if the secondaries were coupled they would induce voltage on each other windings during switch off time in flyback.

Click to expand...

A more sensible way to look at it is that the flux in the core induces voltage on all of them, and the voltage on the windings determine how the core flux changes. All windings are on the same core, so they will have roughly the same coupling between all of them.

In fact in principle, it's impossible for multiple windings to be perfectly coupled to one primary without also being totally coupled to each other.

mtweig is correct, however it is possible for some secondaries to have lesser coupling (more leakage) than others - an output with high leakage will have poor regulation, high volts at no load and lower volts at full load - hence the winding layout is critical for good cross regulation.

Another way is to stack a 10V wdg on top of the 5V winding, i.e. 5V rectified + 10V rectified, to get the 15V required (assuming common 0v is OK), with weighted feedback resistors, this often works well.

Still need good coupling all round for a successful flyback design.

yes and i hope you abhinand sandwich the two secondaries between two prinmary layers to get better coupling.