Isolated DC-DC Transformer with 6 Secondaries??.......

[Smokey]

I'm trying to put together a 6-cell Li-Ion battery charger with independent charge circuits for each cell.  I need 6 floating 5V voltages for this, and it would be great to generate them all from one isolated DC-DC converter. 

I would be fine if the sum of all 6 outputs was in the 1A range, so around a 5-10W converter.  Bigger is also acceptable.   Input voltage is 24-30VDC.  I don't particularly care about regulation accuracy or ripple.

It would be optimal if I could use off the shelf magnetics since I'm going to need to make a fair number of these, but I don't think enough to get into the range that a custom transformer has a good price break.

So... The question is...
Does anyone know of an off the shelf flyback transformer in the 5-10Watt+ range with 6 secondaries?

also, if I'm thinking about this wrong and there is a better way to get 6 isolated voltages, let me know!


Notes (See Below):
1) An interesting multi-secondary solution involves either the WE-Flex or Hexa-Path coupled inductors. This still only gets me to 5 secondaries though.
2) A more cost effective solution for this particular application, considering I don't need true floating voltages, looks like a charge pump solution.

[Benta]

How do you get 5...10 W?
5 V x 1 A x 6 gives me 30 W.

[bdunham7]

This may not be an optimal solution but I was looking at DC-DC converters for another project and your question reminded me of my search.  If this is a 6-cell charger where individual cells are put into the charger and then removed, not part of another device, then perhaps all the voltages don't need to be isolated?  My first thought was to use x3 +/-5V DC-DC converters and just have each pair of cells--one anode and one cathode--connected in common to the center of the DC-DC converter output.  If the other ends of the cells remain isolated, you can use pretty standard circuitry from there to charge your cells.  Not sure if that is viable, just a thought.

[Smokey]

That would be cool, but I need to deal with cells that are all welded in series already. 

[Benta]

That would be cool, but I need to deal with cells that are all welded in series already.

Isn't that normally fixed with a balancer?

[dobsonr741]

The commercial solution is to throw more MOSFETs at it, not more copper in the transformer. Take a look at Ti EMB1428Q

[doktor pyta]

https://www.we-online.com/en/components/products/WE-FLEX
?

[D Straney]

The Coilcraft Hexa-Path transformers/coupled inductors might be useful here: https://www.coilcraft.com/en-us/products/transformers/power-transformers/configurable-transformers/  Can buy these either directly from them, or through Mouser I think.  You'd get 5 secondaries out of 1 part, but ground-reference the bottom cell and you're good.  Or use 2 of them, stick 2 primaries in series, and you get 10 secondaries.

[Buriedcode]

Yeah, the search term "coupled inductor" comes in handy, as they are generally multiwinding inductors with two coils, but sometimes many more, and often aren't under the "transformers" section of distributors (usually they are split between inductors and transformers as some places dont' know where to put them).

The most windings I've seen on these is 6 - so short of one for your application.  But you could split this up into two 1:1:1:1 inductors, with the primaries of the two paralleled on a buck converter.  The regulation isn't great, but will track roughly with whichever output is fed back.  Or of course 3x 1:1:1, but then you start paralleling quite a few primaries on the buck converter raising the inductance.  5-10W is low enough that there are off-shelf parts for 4.7-100uH, plenty of range for whichever buck converter you choose.

Just saw Dr Straney's reply    sorry for pretty much suggesting the same deal

[T3sl4co1l]

^^^

You probably want some turns ratio from 24V in, maybe using 3 in series for primary, then getting 3 secondaries out, per part.  Note the primaries (between different transformers) probably shouldn't be wired in series, as they won't share flux evenly; preferably they should be paralleled.

For flyback, this is less important (depends on the inductance tolerance, and assuming DCM operation), but for forward (or flyback CCM) it's critical, especially if the loads are unequal.

If it's not a production item, you can always wind, or get a prewound, toroid (powdered iron) and loop a hunk of multiconductor cable through it, for your own chintzy multi-secondary transformer.  This might not be too bad in production anyway, actually, though you might at that point prefer something like a ribbon or flex cable, or planar transformer even, just to save the cost of soldering all those wires.

Tim

[bdunham7]

That would be cool, but I need to deal with cells that are all welded in series already.

Then I'm not sure how isolated supplies help?

[Smokey]

How do you get 5...10 W?
5 V x 1 A x 6 gives me 30 W.

Sorry.  I did the same thing in the original post.
Sum of 1A, so 1A / 6 = 0.166A for each secondary. 

5V * 0.166A * 6 = 5W

[Smokey]

That would be cool, but I need to deal with cells that are all welded in series already.

Then I'm not sure how isolated supplies help?

Each cell gets it's own floating 5V charge supply, and connected to that it's own floating charger chip.  Each cell has external tap wires at both ends.

[Smokey]

I was reading some stuff on the interwebs, and it sounds like its a really bad idea having more than one transformer in (series or parallel) on a flyback regulator.  The reasoning was that the core energy won't be the same unless the load is the same on all secondaries.  So if the second transformer has lighter load than the first, then it's coupling energy transfer won't match and you will get wacky output voltages even when the first transformer is regulating correctly.  I'm not sure if that's actually true, but it sounds plausible.  Unfortunately, I don't expect to always have equal load on all outputs.

I'm not sure if this will ever be a production thing, but I probably need around 36 of these transformers just for the thing I'm making now.  I've hand wound some E-Core transformers before.  Not sure if I want to make 36 of these by hand.

I like the ground referencing of the bottom cell idea.  If I can get 5 floating taps + 1 ground referenced one, I think I would be good.  I'll look into that.

[nctnico]

You are overthinking this. Just charge the cells in series and let balancing do the rest. IIRC the best way is to do balancing when the cells are near their maximum voltage. Look up some posts from Siwastaja on how to do balancing correctly; dealing with Li-ion cells is one of his specialties.

If you still want to create a flyback converter then put the off-the-shelve transformers in parallel. But you'll still need some feedback for the voltage for each cell which more or less circles back to simply charging the cells in series and do balancing. The latter will be way less complex and it can be done using off-the-shelve BMS chips.

[Smokey]

You are overthinking this. Just charge the cells in series and let balancing do the rest.
...

This is going into a balancer (but not a BMS).  Part of a larger system. 

(And yes, I'm aware of Siwastaja's stuff : https://github.com/siwastaja/bms/blob/master/balancing.txt )

[Smokey]

I was reading some stuff on the interwebs, and it sounds like its a really bad idea having more than one transformer in (series or parallel) on a flyback regulator. 
...

I bet forward converters don't have that problem with multiple transformers.  Maybe that's the thing to do....

[Smokey]

The commercial solution is to throw more MOSFETs at it, not more copper in the transformer. Take a look at Ti EMB1428Q

That is a crazy expensive solution for 7 cells:
https://www.digikey.com/en/products/detail/texas-instruments/EMB1428QSQ-NOPB/6572293
https://www.digikey.com/en/products/detail/texas-instruments/EMB1499QMH-NOPB/6571840

EMB1428QSQ = $22
EMB1499 (only at Rochester?) = $8
+ all the other stuff like FETs and power supplies.

The full Eval board is $2499 direct, or $3748 from Digikey!
https://www.ti.com/tool/EM1402EVM
https://www.digikey.com/en/products/detail/texas-instruments/EM1402EVM/7563635
That has 3x EMB1428QSQ and gets you 16 cells.  But.. it only charges or discharges one cell at a time.

[Weston]

The voltages you need are not truly floating, but referenced to some (relatively) static potential above ground with a low impedance return path.

Have you considered using some sort of charge pump? You could do away with the transformer entirely but you might need some sort of pre-regulation if you intending to use the transformer for regulation too.

If you needed to extend this approach to big stacks of cells in series you could feed charge pumps providing power to stacks of 6 cells with a single transformer for galvanic isolation, which would prevent having to deal with high voltages across any of the charge pump capacitors.

You could also do a resonant switched capacitor approach to use smaller capacitors and get better switching waveforms.

[dobsonr741]

EMB14xx: consider the idea of a “high current, high voltage analog switch”. You can visit each cell in the stack with the EMB and either connect it to a charge pump capacitor, or to a constant current charger.

If the EMB pushes the BOM cost, you can always try building it from scratch.

[Smokey]

The voltages you need are not truly floating, but referenced to some (relatively) static potential above ground with a low impedance return path.

Have you considered using some sort of charge pump? You could do away with the transformer entirely but you might need some sort of pre-regulation if you intending to use the transformer for regulation too.

If you needed to extend this approach to big stacks of cells in series you could feed charge pumps providing power to stacks of 6 cells with a single transformer for galvanic isolation, which would prevent having to deal with high voltages across any of the charge pump capacitors.

You could also do a resonant switched capacitor approach to use smaller capacitors and get better switching waveforms.

Now that the forum is back up...

I've been thinking about this charge pump system, and I must be missing something because I can't come up with a circuit that works for multiple cells.  Lets say I already have an isolated supply, like you were suggesting.  So I've already got a solid isolated 5V from a flyback.  I would need to take that 5V and create 5 more 5V rails level shifted up to the cell potentials.  Is this something that would need to be all discrete with a bunch of individual FETs charging caps on the isolated 5V and then disconnecting them and reconnecting them across the cells, because I'm not seeing an IC that has this function (I'm probably just missing something, sorry).  There are high side gate drivers that sort of look like this functionality, but I'm not sure that would work here (lower current) and I think that solution might end up more expensive than the 6 tap transformer option. 

Did you have something specific in mind?

[dobsonr741]

Is it Fig. 6 you are after? There are no commercial implementations, only evaluated in simulation.

[TopQuark]

Prof. Sam Ben Yaakov recently did a video about the switched capacitor balancer idea. Don't think they went as far as making a PCB and measuring things with a scope (I could be wrong, maybe they aren't allowed to show it  ), but Prof. Sam Ben Yaakov makes really solid videos / presentations about power electronics.

[T3sl4co1l]

I think the big problem with a charge pump will be, the voltage difference is very small (10s mV at end of charge/balance) and gain is only ever exactly 1 (minus loss); you don't have any way to overcome losses and deliver a controlled current between respective cells.  And cells are effectively such massive capacitors compared to anything you'd put in circuit, not that balancing requires much work on average, but it may still require enough (and MOSFETs too) that it's not practical versus other methods.  It also diffuses the charge imbalance into neighbors, rather than either dissipating excess directly (practical as we're only talking a few percentage points of lost efficiency even in a rather unhealthy pack, and an even more unhealthy one just shouldn't be used anymore), or redistributing it evenly to the remainder.

At least ESR isn't hard to come by; a 10uF X7R will still have plenty of value even in 0805 size, at 4.2V.  10uF might have to run pretty fast still (say, a MHz for <20mOhm), so probably a somewhat larger size would be chosen (100uF?) or enough used in parallel, or electrolytic (even bigger to keep ESR down), or maybe polymer (if economical).  MOSFETs aren't a problem, indeed it's almost a problem if anything that they're hardly available at low enough voltages to keep Qg nice and low.  (IIRC, selection is narrow -- but indeed available.)

Gate drive -- would you use a DC restore network and just drive them all ganged together?  Hmm, has some possibility.  Hot-plugging, or first time wiring the balancer to the pack, might be an interesting experience.  Maybe that's fine with some clamp diodes.  Maybe a precharge circuit (all taps are wired up through modest value resistors first, then shorted out).

Reliability -- would probably want to use fancier ceramics (flexterm, short resistant) if they are to be wired across the battery at all times, if this is internal to the pack that is.  Taps can be fused, and, any tap blowing needs to be sensed by the controller as a failure.  (So, don't just keep the charge pump running mindlessly such that the controller reads perfect balance across a missing tap!)

How much is actually needed -- not sure, depends on the relative impedance of the cells during charge.  I've not designed a balancer before so that's something I'd need to look into.  Maybe it doesn't need to be the mOhms I imagine it does (say for a small pack using 18650s, or the next one up).  Then again, there will always be some scale of battery pack where even mOhms is too high, and perhaps an inductive buck-boost topology would be warranted above some point (in which case, where is that point?).  Or alternative strategies applied, like disconnecting entire affected cells (total battery voltage fluctuates over time), whether for purpose of individual attention (charge/balance), or self healing (out of spec, removed permanently from battery), in which case multiple such batteries could be wired in parallel to achieve arbitrarily high total Ah or Wh, using a converter to address the variable output voltage and deliver a fixed common bus (or perhaps widely variable as the case may be, e.g. for a VFD).

Perhaps the video goes over these finer points, I don't know.

Tim

[Smokey]

I think this is the hang up. 

I'm not trying to exclusively directly transfer charge from one cell to another, like those series/parallel cap balancers.  I'm trying to generate separate cell-referenced power supplies so I can use Li-Ion battery chargers to intelligently charge each cell.  Something like this: https://www.digikey.com/en/products/detail/microchip-technology/MCP73812T-420I-OT/1626618

So I need separate cell referenced rails in the 4.5V-6V range.  But I think Weston was right, that they don't all actually need to be fully floating.  If I have an external 24V and I used a flyback to make one floating 5V rail and reference it to the bottom cell, I should be able to use that rail to generate the other for the remaining 5 cells in series with some sort of charge pump.  That was the idea.  I can see through how to do that with a multi-tap transformer (thread title...), and conceptually it makes sense that should be doable with a capacitor charge pump but it's the details of that which I'm not cool enough to see all the way through (and how to do it inexpensively and at the required current).  The closest circuit my brain can see is a high side gate driver supply generator that ends up floating on the high side FET, but I'm not sure that sort of thing can supply a continuous 150mA.