300W Full Bridge: Current Doubler transformer..OR split sec transformer?

[cupoftea]

Hi,
We have been sent an isolated Full Bridge, 12-36vin, 32vout, 300Wout, 240kHz, Planar transformer, synchronous. We should test it up.

We compare this split secondary version (which this is) with a Current Doubler version, which we recommend instead.

There is utterly no comparison, and it should be done with Current Doubler output. We are shocked that the customer wants split-sec output.

The split sec version has some 200% more current in the transformer secondary…..compared to the current doubler version….and since the primary (in both versions) is 33Arms…..then reducing the overall current in this transformer is crucial. Not only that, but the split sec version has worse overall coupling between primary and secondary, because two separate sec’s are harder to couple to the pri than a single sec (even though they both have the same overall turns)….its blatantly obvious that its easier to “interleave wind” a transformer with one primary and one secondary….(as opposed to a single primary and a split secondary) and this applies even more with a planar transformer.

Also, the current doubler version is much more amenable to use with secondary synchronous rectifiers….you can see this by just looking at the schematics (attached)…the current doubler version offers a more overall reduced impedance pathway to current that suddenly gets it path broken by an off-going synch FET, when that FET turns off when the current in it has reversed…..(this always happens in light load).

The sec diode off state voltages, and sec side snubber losses are the same in each case.

But it is a known fact that with a Current Doubler transformer, the leakage inductance as seen from the primary can be made smaller.

Overall no brainer…Current Doubler transformer wins…wouldn’t you agree?
LTspice and pdf attached

 

[Easy peasy]

Current doubler will cause quite a bit of heat in the chokes unless carefully designed - they see the fundamental frequency ( not 2x ) and have high ripple for the same L value as the CTapped version - we normally design for high L for current doublers.

Yes the Tx is fully utilised if driven from a full bridge , and this allows a lower leakage design, and smaller more efficient Tx overall.

Our 12V 100A continuous module uses a current doubler - we have to protect the output diodes ( as schottkies ) but this unit goes 24/7 for 10+ years.

--- Updated Dec 22, 2022 ---

" But it is a known fact that with a Current Doubler transformer, the leakage inductance as seen from the primary can be made smaller. "

respectfully - this is not a known fact - at least to PE engineers. Leakage is an inherent and intrinsic part of a Tx and cannot be altered by external circuirty - excepting that external wiring inductance effectively adds to it.

 

[cupoftea]

Thankyou.......And the real problem for the split-sec transformer version of the full bridge is its totally disastrous behaviour when used with synchronous rectifiers........compare here with the Current Doubler version of the Full Bridge........as soon as the Full Bridge goes from full to no load, the split sec version goes haywire...whereas the Current Doubler Full Bridge shows superb performance...without any synchronous rectifier current limiting it shows no high overcurrents or overly high FET voltages.
Will you support us in telling our customer to change to the Current Doubler version?
The Current Doubler version was "made in heaven" for synchronous rectifiers....just see it on the attached LTspice.
LTspice and PDF attached (showing synchronous rectifier performance)

 

[Easy peasy]

In the C Tap sec, for CCM, between the power pulses the sec current flows in both sec wdgs and both diodes, the sec's and primary are essentially hard shorted by these sec currents ( which are ~ equal and opposite ).

For the current doubler, the current flows in both chokes between power pulses, and both diodes, very little current flows in the Tx sec during these periods, just the mag current on the pri side.

As far as rectifying mosfets are concerned I can't see too much difference in the stresses they are subject too.

They are both ON in the time between power pulses ( or should be ) and both are affected by the Tx leakage at turn off, there is one ON during a power pulse.

The only difference is that the Tx sec has less rms sec current for the curr doubler - always useful.

 

[cupoftea]

Thanks, i just wonder if its best to put say a dual interleaved Boost converter in the front end to give a more reasonable input voltage to the Full Bridge, and ease its transformer design.....eg, boost up to 36V, say.

 

[cupoftea]

Hi,
I have just been searching through various semico product selectors, and none do anything satisfactory for
Full Bridge
10-36vin
32vout
300wout
surge to 170v for 2 seconds (through which it must produce 300W output at 32v)
Isolated to 1500vdc

I can go for a LTC3723-1 Full Bridge, but it means some 32A of input current, and i wonder if the filter and inrush limiter and surge protector are going to drop so much voltage that it gets a bit silly...obviosuly we can buy super expensive filter comoponents and get low loss, but we dont want the cost.

Would you recomend a dual Boost at the front end to give a vin of 36V......so the current input to the full bridge transformer was reduced?.......probably making the whole thing cheaper?