Regulating flyback w/ separate secondaries

[payloadde]

I have a flyback with two separate secondaries which need to be completely isolated from each other.
So far I assume that I also need separate 431 and O.C. because there is no common ground.
correct ?
But how do I mix the signals from both couplers into the controller feedback ?
parallel, series, or any other network ?

[DUG]

Pick one.

OR

Decide that the important part is to limit the maximum output of either supply or the minimum output of either supply.

[sawreyrw]

The 2 output will tend to track, but only one can be regulated. Overcurrent protection should be implemented by sensing the source current of the MOSFET.

[SemperFi]

You can post regulate the non-sensed output if they dont track well enough.

[SemperFi]

btw, how well they track is very dependant on how the transformer is wound. Some experimenting may be needed.

[RJM1]

Generally you regulate to the highest voltage output, as the fly-back voltage level will change greater there than for any lower voltage winding . The problem with this is that you will experience more voltage drop in the high current low voltage rails as their voltage will change more due to resistive losses in their windings. It's a balance between how you design a transformer and how much regulation that you need.
Or you could just regulate to the most critical rail.

[payloadde]

Originally Posted by DUG

Pick one.

OR

Decide that the important part is to limit the maximum output of either supply or the minimum output of either supply.

So, which one is which ?

[payloadde]

Originally Posted by sawreyrw

The 2 output will tend to track, but only one can be regulated. Overcurrent protection should be implemented by sensing the source current of the MOSFET.

No worries, it's a current mode controller, so the source sensing is there.

I know how to regulate both - its a compromise but it always worked well for me.
That is its not a problem when both secondaries are referenced to the same ground as in picture #1.

But my problem is how to implement this when the two secondaries are referenced to different grounds as in picture #2.

[Sch3mat1c]

You can always add a tertiary winding and regulate off that. That way, you don't have to show favoritism, instead, both windings get regulated equally poorly.

If you want to mix optos so they are jointly regulated, you need ratio matched optos. These don't exist, so you're SOL. You can make a far-too-elaborate circuit with some means of signal isolation (i.e., a circuit which copies the error voltage over the isolated boundary, like an ISO122, but without the dollars), and do the ratio mixing from there, but you'll only do as well as the tertiary winding.

Tim

[payloadde]

Originally Posted by Sch3mat1c

You can always add a tertiary winding and regulate off that. That way, you don't have to show favoritism, instead, both windings get regulated equally poorly.

If you want to mix optos so they are jointly regulated, you need ratio matched optos. These don't exist, so you're SOL. You can make a far-too-elaborate circuit with some means of signal isolation (i.e., a circuit which copies the error voltage over the isolated boundary, like an ISO122, but without the dollars), and do the ratio mixing from there, but you'll only do as well as the tertiary winding.

Tim

The tertiary could be the auxiliary used to supply the controller, no opto needed at all I guess ...

Lets assume I had matched optos ... how would I connect them ? in series or parallel ?