Is it feasible to use one flyback controller with two transformers, yet only one feedback? The primary winding of the transformers would be connected in parallel, isolated feedback would be taken from the most loaded output, then others would be less regulated, but yet in required range.
The energy per cycle delivered to the transformer that has the biggest load is determined by the on-time of the switching transistor. During the on-time, the current ramps up (fairly linearly) thus storing energy in that transformer's primary winding. When the transistor turns off, that stored energy is released to the secondary load and therefore joules (stored) x F (switching frequency) becomes power delivered to the load.
If you have a second identical transformer with paralleled primary winding, and the load on this transformer is much lighter than the first transformer, you have a problem because the primary energy it stores will be identical to the first transformer and if this is dumped into a much lighter load every switching cycle then the load will be over-powered and have too much voltage placed on it.
In fact, the output voltage on the 2nd transformer will rise until the power dissipated in the 2nd load is about the same as the first load.
You might be able to play tunes with more primary inductance on the 2nd transformer so that it stores less energy per cycle but this will rely on the two loads being at some fixed ratio.
It doesn't sound very feasible but maybe someone has a cunning plan?
If this is about two flyback cores and windings, then the load on the (#1 core) secondary winding may mismatch the load on the (#2 core) secondary winding, leading to different voltage/time during the 'flyback' phase. So, there's no way to prevent DC currents in one core from changing its inductance, and ruining the match (even with the primaries connected together, i.e. in parallel).
It's feasible, but may drift unpredictably from 'nominal' behavior.
Matching is improved by winding multiple secondaries on the SAME core, which no longer requires multiple primary windings.
In a flyback system, the controller controls the power stored in the transformer by adjusting the duty cycle. If you have two transformers and one controller, the controller will control the power stored in only one of them, the other one will be out of control.
On the secondary side of the uncontrolled transformer there will be voltage relative to the load at the secondary side of the controlled transformer. Because duty cycle is adjusted according to the load at secondary side of the controlled transformer.
So there will be unexpected and uncontrolled voltage at the secondary side of the uncontrolled transformer.
One controller and two transformers is only possible if below things are satisfied.
1- Both transformers have same number of primary winding turns. 2- Both transformers have same number of secondary winding turns. 3- Both transformers have same primary inductance. 4- Primaries of the transformers are connected in series. 5- Secondaries of the transformers are connected in parallel. 6- Any extra secondary windings all must be connected parallel and have same number of turns. 7- Output diodes of secondaries must be separated and winding parallel connections are made after diodes.
One of the transformer will be loaded slightly less than other because of the imperfections in windings and inductance adjustments.
This approach is good for separating heat to two transformers and also increasing the heat dissipation by increasing outside surface area of the transformer.
One more benefit is input voltage is divided to two between two transformers, so primary to secondary ratio of transformer is reduced. This increases the primary to secondary coupling ratio and reduce leakage inductance. This is good for generating low voltage high current outputs (e.g., 3 volts 5 amps) from flyback systems without having to decrease primary to secondary ratio too much.
Using more than one transformer for flyback is a good idea. Better and more sophisticated systems might be developed. The one I explained must be tested first.