self balancing battery charger

[cadstarsucks]

First off, hiya Safe, how've ya been?

LiFePO4s are quite tolerant of "crude" charging techniques and we had a customer that asked for a power pack for their unit, but the balked when my boss suggested SLAs. "to BIG and to HEAVY" they said, so I suggested LiFePO4s.

While 12V packs charge OK the overall +/-24V system with imbalanced loads makes it a bit problematic.

While my solution is a bit...unique....it is curiously both similar and simpler than Safe's "switcher for every cell" thread that was just locked. A single current mode switcher into a flyback transformer with individual windings, current senses and a peak voltage sense. The cross regulation of the transformer balances the charging across the packs and the individual current senses (the highest current actually controls the regulator), allows high speed charging while some of the packs might be fully charged.

Of course it could be simpler if you limit the energy to what a single cell can handle but then if it is a 2hr charge it will take 8hrs to charge the four packs.

Dan

 

[fechter]

Yes, I've seen some designs that use a single primary and many separate secondaries on the same core. It appears that if you regulate the voltage of one output, the rest of them will track at the same voltage very closely. In practice, you'd be limited in how many secondaries you can wind on a single core, but you could just gang units in series. I bet you could easily do 4, possibly 6 or 8 outputs. I'm not sure how you'd want to do the current limiting. You might be able to just limit the input. It may be unnecessary to measure each cell individually for current. Somebody on the forum built one I think? I can't remember who.

Other variations include using off-the-shelf transformer cores (2 outputs) and driving several of them from a single switcher. Inputs could be series or parallel.

The "Flintsone" chargers use a single input rectifier and main capacitor that feeds a bunch of independent switchers.

 

[TylerDurden]

Somebody on the forum built one I think? I can't remember who.

IIRC, DoctorBass started out using a large toroidal.

 

[stew007]

It would be interesting to try removing the BMS from these Lifep04 packs, and charging using a modern balance charger. The EOS 1210i can charge up to 12s any Lifepo4/a123 battery at up to 180w max and the EOS 0610i duo up to 360w total, balancing cell voltages to .002v at the end of each charge. The new Graupner Duo Ultra Plus 50 can do 14s LiFepo4 at 360w, so we could connect two 7s packs in parallel on each side (inc balancing taps) so we are effectively balance charging anything up to a 28s pack, thats how I charge my lipo packs.

 

[johnrobholmes]

It would be interesting to try removing the BMS from these Lifep04 packs, and charging using a modern balance charger.

I agree. I see no reason to keep the BMS on LiFe. As long as they are used within specs they won't even need but an occasional balance. I balance my packs every 10th charge, and check the cell voltages after every few runs to be sure they are still in balance. So far so good, but I am also running 20C cells at less than 10C peak and 2-4C average. This is certainly possible on any EV, as capacity takes precedence over discharge rate since we need long runtimes.

 

[cadstarsucks]

Yes, I've seen some designs that use a single primary and many separate secondaries on the same core. It appears that if you regulate the voltage of one output, the rest of them will track at the same voltage very closely. In practice, you'd be limited in how many secondaries you can wind on a single core, but you could just gang units in series. I bet you could easily do 4, possibly 6 or 8 outputs. I'm not sure how you'd want to do the current limiting. You might be able to just limit the input. It may be unnecessary to measure each cell individually for current. Somebody on the forum built one I think? I can't remember who.

Other variations include using off-the-shelf transformer cores (2 outputs) and driving several of them from a single switcher. Inputs could be series or parallel.

The "Flintsone" chargers use a single input rectifier and main capacitor that feeds a bunch of independent switchers.

What I was planning was a simple current feedback from each 12V pack with the maximum current and the overall voltage limiting the regulator. The transformer cross regulation then balances the individual packs with the lowest one charged first. When the max charge rate brought that one up to the next most discharged then there would be two packs at max charge rate.

At low power I can put 4 packs on an off the shelf transformer. At high power levels I can do customs, but the pricing would be prohibitive.

Dan