Paralleling Flyback Transformers

[sean6541]

To power 4 isolated MOSFET drivers in a full-bridge inverter, I am using 4 LM25184 DC-DC flyback power supply controllers paired with 750318131 transformers (dual secondary, +15V and -4V). My question is, can I connect the primary windings of the 4 transformers in parallel to a single LM25184, so long as the total current doesn't exceed the rating of the LM25184? I am assuming so, as long as the design parameters are adjusted for the lower magnetizing inductance (1/4 of a single transformer).

[NiHaoMike]

That will work fine, but you only need 3 for a full bridge since the lower two can share the same supply.

[sean6541]

Thanks for confirming that for me. Good point about combining the lower 2 power supplies.

[TheMG]

I've seen it done that way in some commercial  products. Another way to do it is with a single transformer with multiple isolated secondary windings.

[langwadt]

I've seen it done that way in some commercial  products. Another way to do it is with a single transformer with multiple isolated secondary windings.

just make sure to check the isolation rating between secondaries ..

[sean6541]

I would prefer a single transformer but I would also prefer not to hand wind transformers and I can't find any off-the-shelf transformers that would meet the requirements (3x 12-15v and 3x -4V secondaries).

[T3sl4co1l]

Yes, parallel is fine.  Beware the leakage between secondaries is about double that of an equivalent windup in a single transformer, so will have poorer cross-regulation.  You may want to do it with local regulation anyway, to avoid any channel going too high/low in voltage.  On the upside, they should all have the same load -- give or take if you want to parallel two for the low side, at least -- with a single transformer it would have double the load so run low while the high-sides run high.

And yeah, more than one winding even (per transformer), is probably a custom job.  I haven't seen any with multiple secondaries of adequate isolation rating; they may be out there, but it's so devilishly hard to search for something like this.  So rarely used, anyway, aside from this sort of application.

Don't worry about the -4V, don't treat it as a separate circuit, just do the total (i.e. 20V or so) and shunt-regulate the tap (i.e. with a 3.9 to 5.1V zener).  It only ever gets displacement current, as long as it's adequately bypassed it's fine, and the zener will hold it more or less where it needs to be.

BTW, if this is like a big industrial module application, you may find it's better to isolate the two low-sides anyway: the stray inductance between them will generate large common-mode voltages that won't be easy to handle by the supply or driver circuits.  Whereas if it's like just a quad of TO-247s or something, that's not so bad, and the inductance of that particular path is under design control, you can account for it.  Whereas with modules and bus bars, you're limited by whatever they do, which usually isn't very great.

Tim