Lead Acid Battery Balancer.

[waitatian]

I recently found  a lead acid battery balance IC .
 Linear Technology LTC3305.
When series connected batteries are not equal, on charge  when the regulator limits the total voltage, the voltage on the best batteries drifts up, and since the total is fixed, the voltage on the rest of the batteries must go down, so the worst battery never gets fully charged.
On discharge, one battery can be flat while the system voltage is still ok.
This chip balances up to four batteries by connecting each in turn to a super capacitor or auxiliary battery until they are all balanced.
Has anyone tried this chip or maybe made a battery balancer?

[dnix71]

The obvious way to balance a lead-acid battery can't be done because the people who make batteries don't want them to last. Lithium packs have balancing wires. A regular car starting battery could have the same if they were simply made that way. Telcos use 2v cells, as do forklift batteries.

To properly balance a sar start battery all you need is a balancing post for each cell. Charge each cell separately at 2v+ and you will always have a fully charged balanced battery.

[waitatian]

A balancer should fix the problems of replacing one battery in a string and, on discharge, the worst battery in the string  either limiting the total amp/hour capacity or being overly discharged.

[Flux]

To work reliably you need to monitor every cell. By definition a battery is a string of cells and normally you can't monitor closer than the battery terminal voltage

If for instance using 6v batteries for a 24v system, a monitor will keep all the 6v units in step. Probably in the early years this will help, but at some point one cell will fail and your equalizer will force the batteries to the same block voltage with 2 cells gassing to try to get the terminal voltage when one cell is dying.

It is probably better than nothing as during the early years it will equalize things when we are only looking at minor differences, but once a cell dies you will be back to regular monitoring to spot the impending disaster.

Flux

[lifer]

I was thinking about balancing by connecting a small dummy load across the most charged battery/batteries in a series string.

I could mount such a dummy load (10-20W, relay driven) across every battery then I could build a small monitoring board to automatically do the job.

Could it be a feasible solution?

[dnix71]

Build your stack out of single cell 2V units like forklift batteries and balancing is easy. Charge one cell at a time is you have to.

[lifer]

That's a good advice but I already have 12V batteries thus I have to pay more attention to keep them well balanced.

[kitestrings]

I was thinking about balancing by connecting a small dummy load across the most charged battery/batteries in a series string.

I could mount such a dummy load (10-20W, relay driven) across every battery

Wouldn't it make more sense to limit/reduce the charge voltage to the cell(s) drifting high rather than burn-off excess?.  I follow the logic of monitoring individual cells, but why not reduce the charge to those that are out-gassing, while others aren't quite there?

I'm starting to think a single-series string may be the best configuration, invariably when they fail it seems to become a death-spiral as Flux describes.  One cell starts to fail, compounded by say a poor electrical connection in parallel strings, the other cells over compensate, boil off more and more water...pretty soon the end is near.

Catching even minor fluctuations between even 6-volt batteries would seem to be helpful in early detection.

~ks

~ks

[lifer]

why not reduce the charge to those that are out-gassing, while others aren't quite there?

That's what a small dummy load across the most charged battery (cell) is actually doing. Being a series string of batteries, the current through every single battery should be the same. The small dummy load will divert a small ratio of the charging current through itself thus the (remaining) current flowing through that most charged battery will be smaller.

It's a lot harder to inject an additional current to a random battery in a series string (the circuit will be more complicated).

Of course, I'm talking about doing the job automatically. If you want to manually balance those bateries then you could design a simple charger to increase the charging current for a particular battery, without removing it from the circuit.

[kitestrings]

That's what a small dummy load across the most charged battery (cell) is actually doing.

Not exactly.  With a "dummy load" your clipping, or burning off excess.  I'm just suggesting you monitor and limit it before it is produced.  Probably harder to do, I'd guess.

Good luck with it.  ~ks

[Simen]

Those 'dummy loads' are a common way to balance Lithium cells. They usually have a small microprocessor with some dumping resistors on each cell monitoring the cell voltage, and start dumping excess charge when the max charge voltage for the cell are reached.

[waitatian]

This chip seems to have a few advantages over dummy loads and such.
The system capacity can be increased by the capacity of the balance source if a lead-acid battery is used.
It will also apparently allow renewing a single battery in a string, and having a string of different capacity batteries.
The power it uses to run itself may be a problem but it could be less than the improved efficiency of the system.

[ruarangifarm]

do you intend to try such a thing - Battery Balancer of some sort?
Sounds a good idea.
But I got 24 cells at 2V in series. So would I need 6 of those, and could they talk to one another to get all the cells to the same level?
something to look into!

[waitatian]

I have built one and tested it.
 My battery bank is 48v of mixed age 6v,220 ah batteries with shunt regulators to limit overcharge of the better batteries.
The balancer was hooked to 3 old and 1 new batteries in the string and used an old 6v battery as the auxiliary. It was started in the afternoon and left going overnight and into the next day.
The auxiliary battery voltage was up to the worst string battery by the time the sun went down. Once power was taken from the string, the balancer kept its 4 battery voltages more equal than the other 4 batteries in the string.
The next day was sunny so there was about 5 amps of charge going in.
As they charged, the voltage on the new battery went increasingly higher than the rest until the shunt regulator had to be connected.
On charge, the balancer did not transfer enough power between each battery to keep them balanced, but it seemed to work ok on discharge. It would probably work ok with more equal batteries than mine or with lower charge/discharge rates.

[joestue]

i think i've mentioned this 3 years ago but you can make a synchronous buck converter wrap around the cells according to this schematic
http://johansense.com/bulk/bms_1.png
or
http://johansense.com/bulk/bms_1.1.png

all you have to do is wire up a transformer to feed each fet its gate drive signal. the capacitors shown need only be  on the order of 1 uf to buffer the battery if the parasitic inductance is minimal.. the impedance of this system is fairly low.. on the order of the half dozen milliohms per mosfet. inductors required are up to you to figure out based on the battery voltage and the frequency you want to run it at. i'd suggest at least 50khz.

the duty cycle percentage needs to be exactly 50% within .1% because the battery voltage balance is in exact proportion to the duty cycle but that is not a difficult task. driving the gates directly from a transformer should yield preciously symmetrical timing.