SMPS Design - 2.7kW thru an RM14

[FlapJack]

In a post in this thread.

https://www.edaboard.com/threads/343413/

Easy peasy said.

"We have put 2.7kW thru an RM14 in a commercial design that required long periods at full power (250kHz)."

I would sure like to hear some of the details of how that was done.

What was the input voltage and what was the output voltage and current. What types of winding's were used. And how did you keep the transformer cool.

If anyone else has extraordinary stories like this i would like the same details.

 

[Velkarn]

put 2.7kW thru an RM14

easy if it is a dc current inductor with 10mm gap))

 

[treez]

could have possibly been a high frequency bridge transformer , and soaked in some heat transfer potting material, which was going straight to a finned heatsink, or water cooled heatsink

 

[FlapJack]

I do not see how it could be done. Hopefully Easy peasy will tell us.

 

[FvM]

If you use the Epcos Ferrite tool for a first estimation, you get 1.85 kW transmitted power with the therein assumed standard (free convection) cooling. So it seems well possible to achieve the stated power with some kind of forced cooling - if possessing the respective knowledge.

 

[Velkarn]

yes, it is very interesting! maybe high output voltage (my guess)

 

[Easy peasy]

410VDC bus, LLC H bridge, triple ins litz wire, no bobbin, 210kHz min, slight air flow, 56V 50A nominal full power out, 3C96 core, haversine currents in wire., thousands made and in service...

 

[mtwieg]

no bobbin

Ah, that goes a long way to explaining it. Probably increases your area produce by 25% or so. Any gaps in the windings at all to let airflow through? Or is it potted/varnished to help heat travel from the core through the windings?

What do you estimate the total transformer losses and core temperature at?

 

[FlapJack]

Very cleaver design, with triple insulated litz you have no margins and with no bobbin 20.8mm-18.35mm = 2.45mm extra space to fatten up your wires.

Pressman estimated about 2 KW maximum power out of a RM14 core at 200 khz, with a bobbin. So 2.7 KW around 210 khz is extraordinary

 

[treez]

I am wondering (I do however appreciate that its none of my business and don't deserve an answer) if there was a gap in the 3C96 core, since as we know LLC's typically have a decent bit of magnetising current flowing...or was there a separate inductor parallel to the RM14 core to act as the main "magnetising inductance".
If the RM14 was not gapped, then were the ferrite surfaces lapped ?
Also, if there was a centre leg gap, presumably it had to be narrower than the litz wire. Also , was the RM14 potted.

 

[Easy peasy]

In this design we did not use a gap in the main Tx, external choke in parallel with the HV winding, we did not need other than the std surface finish, dip varnished to keep the wires in place.
Falco in Mexico make very repeatable magnetics without bobbin using TIW litz.

 

[mtwieg]

Wait, are you saying you did use an external inductor in parallelwith the primary? Never heard of that before in an LLC, and I can't see why it would be necessary...

we did not need other than the std surface finish, dip varnished to keep the wires in place.
Falco in Mexico make very repeatable magnetics without bobbin using TIW litz.

Surely they must have pulled a hell of a vacuum during the varnishing to get all the air out between the litz wire strands. In my experience with litz, it has very poor thermal conductivity due to all the air trapped between windings. Maybe they used some sort of impregnated litz wire, or maybe they're just really good at winding the thing while keeping the bundle super tight..?

 

[treez]

Wait, are you saying you did use an external inductor in parallelwith the primary? Never heard of that before in an LLC, and I can't see why it would be necessary...

Mtwieg, i know you know about this and have just temporarily forgot...I am sure you remember commenting on my post a few weeks back (thanks for your excellent comments too), which concerned an LLC with inductor in parallel to primary, to bring your "m" value down whilst keeping the transformer current relatively low.

Easy Peasy used TIW litz wire, so its "coated" litz wire....don't know what that means toward the dip varnishing.

 

[FlapJack]

Here we go, this helps explain the coil construction.

"Bondable wire is a magnet wire with an additional adhesive enamel overcoat. This adhesive has a bonding feature, which is activated by heat or solvents. Once activated the adhesive bonds the turn to turn windings into a compact self supporting coil."

 

[Easy peasy]

Or, tape that has adhesive on both sides is very useful for building self supporting coils on a plain former (no sides) for TIW wire that really maximises the winding space available...! Once the cores are fitted the Tx can be dunked in varnish to stop movement of the wires, vacuum impregnation is a step better to get the varnish into the middle, but if the outside is effectively potted by the varnish - then simple immersion varnishing is OK unless a high vibration environment is to be encountered.