Balance charging with DC/DC converters?

First off, apologies in advance if this is a dumb idea, but my understanding of electronics has never got much beyond the plumbing analogy of current as flow and voltage as pressure. This is a bit of a handicap for an EV tinkerer, but it does have the advantage of forcing me to try to find simple solutions.

So I’m looking for a simple charging system for a small electric car (my Mini). The LiFePO4 battery pack will be split into nominal 36V modules for safety and ease of handling.

I really like the idea of using DC/DC converters to charge cells individually, as covered by Doc and Jeremy here: http://endless-sphere.com/forums/viewtopic.php?f=14&t=2824
http://endless-sphere.com/forums/viewtopic.php?f=14&t=22881

I’ve bought a bunch of 48V to 3.3V converters, rated at 75W, and plan to trim them for 3.6V output by adding a resistor. All I need now is a power source to supply 36–72V for the DC/DC inputs. I already had a few second-hand Curtis 36V/30A lead acid chargers, which are easily trimmed for the desired cutoff voltage, so it would be great if I could use these. Unfortunately the charger is too smart to power the DC converters alone, as it doesn’t recognise them as a battery.

I was thinking I might be able to trick the charger into starting by briefly connecting it directly to the pack, when I started wondering what would happen if I connect the charger in parallel with the DC converters and battery pack and just let them run?

When the pack is at a low state of charge I guess the charger will do the bulk of charging directly, which is fine, but what’s likely to happen as the pack gets close to fully charged?

Could this work as a balancing charger or should I just forget it and buy a 48V Mean Well power supply?

dude.... used hp 3000w 48v server power supplies... $40. That's what you should be using.

Use HP power supplies to power the 48v input DC/DC's

DC/DC's setup correctly are the best way to balance charge IMHO. They can correct any amount of imbalance rapidly and safely, which bleed types can not do.

The schematic you've drawn will work but only when the inputs and outputs of the DCDC converter are isolated.

The DCDC converters you bought, is the ground of input and output the same ? If so you're scheme will not work as you'll
short out the cells.

You need to determine which type of dcdc converter you have:

http://en.wikipedia.org/wiki/Buck_converter this one has the same ground for input/output -> will not work
http://en.wikipedia.org/wiki/Flyback_converter in this one input and output do not share ground -> should work in your scheme

Also, I would leave out the connections between the charger (bottom of your drawing) and the battery pack and charge using the
DCDC converters only (what you've drawn is a version of the 'bleed' architecture). Then you can use any voltage source you
want as long as your DCDC converters can turn it into the required cell voltage.

Thanks auraslip and Luke. I've found plenty of general references here to using server power supplies, but there's not much detailed information. Do they need much modification to set up as a bulk charger? Are they generally current limited?

Lebowski:
Yes, the DC converters have fully isolated inputs and outputs (3000V). I understand that normally there would be no direct connection between the power supply and battery pack in this sort of setup. I just wondered if I could use the chargers I already have this way, and curious about what would happen towards the end of the charging process. When you describe this as being a version of a 'bleed' architecture do you mean that some of the cells are likely to overshoot 3.6V before the others catch up?

EDIT: Found a thread with a little more information on HP power sources here: http://endless-sphere.com/forums/viewtopic.php?f=14&t=29316