DIY Electric Car Forums banner

balancing went out of control

Tags
balancing
5K views 12 replies 10 participants last post by  JRP3 
#1 ·
Yesterday I took all my 45 CALB cells out of the car to balance them.

I drove them for ~30km/19miles so they were not fully charged, estimated at 60%.

Wired all parallel and let them rest a while
Connected a small power suply unit at 3,5V and let some current flow (~3A) for a while.
Got of the cables and measured the same voltage at every cell.
I thought, everything works fine.

Wired all together in series and connected the charger with 157,5V.

Today I measured the voltages again, after charging over night:
3,3V to 4,6V(!!!!) at a single cell ... fu§k!!!

Something went terribly wrong and I don't know what!?!
I will drain every single battery to a fixed point by hand now ... this will take a while and after that, I hope the cells are fine.

Saturday there is the "1st northeuropean e-rallye" and I signed and want to start ... fingers crossed
 
See less See more
#4 ·
Got off the cables and measured the same voltage at every cell.
What was that voltage?

How long were the cells paralleled?

How well were they paralleled? That is, with twice the usual number of copper links, or a bunch of clip leads?

My suspicion is that you didn't have them paralleled well enough (or left them paralleled long enough; one can compensate for the other) so the cells didn't actually equalise.

But equally, if the voltage of the paralleled cells was about 3.30 V (in the middle of the open circuit voltage range), then even a few millivolts of voltage difference can mean a different state of charge. You need to get the cells at either end (I prefer the upper end) where small changes in state of charge show up as large changes in terminal voltage.

Balancing cells (especially when they haven't been balanced well before) is tricky. That's why I recommend a balancing BMS. But it should be one where it is difficult for the micro to go crazy and discharge the cell to near zero. I'm involved in a BMS design with diodes in series with the bypass transistor, so that even if the BMS tried to discharge the cell below 2.0 V, the transistor would lose bias and turn off. We have yet to find out if the transistor will overheat and possibly fail shorted (defeating the purpose of the diode protection), but we'll know soon.

In case you are interested in more details:
http://www.aeva.asn.au/forums/forum_posts.asp?TID=980&PID=30281#30281
http://www.aeva.asn.au/forums/forum_posts.asp?TID=980&PID=31289#31289
 
#6 ·
Few things:
1) Cells are only balanced at the SOC you balance them at. They will differ in voltage at all other SOC. Balancing toward 60% SOC means they will be unbalanced in voltage near full charge and near full discharge if they have some variability in capacities (think of filling different size glasses all half full with water, then adding more water, or taking water out).

2) If you balance with most cells at 60% SOC, a difference of 1 or 2% C between cells will show up as only a very small difference in voltage as Coulomb said, since the V versus Ah curve is fairly flat there. Have to leave them connected in parallel for quite a while to ensure they equalize in voltage in this case - like a day or more.

3) Most people balance at one of the exponential parts of the V versus Ah curve. So either near full charge, in which case balancing must be done during charging to stay on the exponential part of the curve, so bleeding charge from high cells with shunt resistors during charging (with bms or just connecting a resistor between a cell's terminals), or charging low cells individually; or near full discharge (say 25% SOC) where you can balance rest voltage of cells by bleeding with a resistor or charging individual cells (but allow at at least an hour after charge/discharge of a cell for it to relax close to rest voltage).

Since you balanced toward middle SOC, then charged with cells in series so they all received the same amount of charge, the cells with lower capacity reached full charge sooner than those with larger capacity, so were driven to higher voltage. The same would happen, even to a greater extent, if you bottom balance. You then must track voltage on your lowest capacity cell and stop charging when it starts up the exponential part of the curve (or lower). If you top balance, you need to track SOC on your lowest capacity cell and stop driving when it gets to 20% (or higher), and 20% doesn't leave much margin for error. You could also track voltage in the latter case, but you won't get much warning before the cell suddenly drops to very low SOC due to the V versus Ah curve being fairly flat until the exponential falloff in V at low SOC. And different voltage sag at different discharge currents makes this very confusing. I think tracking SOC of the lowest capacity cell with a charge counter (Ah counter) is better for either case, and only charge to about 97% SOC of that cell or less, and discharge to 30% SOC of that cell or greater SOC. Better yet is tracking SOC and having HVC and LVC protection (minimal bms) as backup.
 
#7 ·
I wouldn't charge them while in parallel unless you can let them sit for a long time after. When you charge in parallel there will be voltage drop with every connection, so the battery closest to the charger will be the highest, and all of the cells will be progressively lower as you get further from the charger. So unless you used buss bars or battery cables to parallel them you probably did more harm than good by trying to charge them. What you need to do is put them in parallel and leave them there as long as possible, (also this should be close to one of the ends, fully charged or fully discharged) Don't do it at 60% unless you can leave them there for a week or two.

Good luck with the e-rallye!
 
#8 ·
ahhh, yes now I can see why they are out of balance.... you basically forced them out of balance by adding parallel charge in the middle of SOC. you have to balance, either top or bottom when in the 'knee' either top or bottom.... you cannot balance by voltage in that long flat middle.

if you bottom balance, you will have to be careful not to ever go near the top when charging. if you top balance, then you have to not got near the bottom when driving.....
 
#9 ·
Quick reply out of the garage:
yesterday I took all cells again out of the car, after the failed charge at the morning and a resttime of 6 hours.
Most of the cells were down to 3,36 Volts, some cells were at 3,5V.
I picked out the highest cells by switch on the charger. It builds up the voltage slowly, so I can intervene if one cell gets in critical hights.

I located the highest cell and shut off the charger at 3,5V cell voltage.
Now I made a bridge over the cell, adjusted the charging voltage down by -3,5V and turned on the charger again.
I made it four times more and then the voltage of many cells where at > 3,44V.
The chargelevel should be average ~90% now, I guess.
Should be enough to call it "top" for top-balancing.

I tryed to drain the single cells which were to high in voltage, but I had no suitable resistor, bulb or device which would do a significant load at ~3,3V.
So I paralleld them again and let them rest for two hours.
After that short period of time, I measured the current at the half of the pack at ~ 0,01mA, so I decided to add the PSU.
One connection at the left positive terminal and the negativ at the other end of the packs terminal.

Constant Voltage (CV) fixed at 3.55V and connected to the pack.
~5Amps was initial on the display, slowly fewer over time.
Four hours later (current was < 1A) I took of the PSU, disconnected the poles and measured all cells at the same voltage. 3.36V.

I hope, that paralled them enough.
I would do it over 48h but I have not time left, I have to start the rallye in 20 hours :eek:

So everything back into the car and wired together.
Finished at 23:00 Uhr I had to rest for a while and balance myself :D

Today I will connect the charger again and look at the cell voltages.
If everything works fine, I would drive around, charge the car and look at its behaviour.

reply to some questions abouve:
- no BMS
- of course all positives where connected togehter and all negatives where connected togeher to parallel
- yes, parallel them at 60% SOC was an idiotic idea :eek:

open questions:
- What should I take to drain the cells at 3,3V?
- Did anyone a paralleld charge to >40 cells?
- Which PSU supports 3,3V, adjustable in 0,01V steps and provides enough current to do something like that?
 
#10 ·
Hi !

If you want to top balance you pack once in a while buy an inexpensive RC charger, something like this
http://www.hobbyking.com/hobbyking/store/uh_viewItem.asp?idProduct=11444

That one can charge a string of 6 Lixx cells balanced with up to 10A.
You can also use it to drain one cell at the time.
It can only draing with max 5A but that will probably do.
You can set it up to drain down to a certain voltage of a single cell.

I use one of these to measure capacity of 40ah cell right now. It takes 8h per cell and I have 120 cells to measure....... Pretty stiff job.

Really low budget drainer is a 12v H1 headlight bulb or a H4 using both hi and low beam. That also gives a nice light..... but you need to check the voltage all the time. A simple 3,3ohm 5W power resistor is good to use in the end to not get the voltage to bounce that much.

Regards
/Per
 
#11 · (Edited)
I am still fiddling around with final top balance before I consider things done.... what I am finding out is that the charge curve at the top is VERY steep, and a surprisingly small amount of energy in or out affects the finish voltage quite a bit.

When I do another new full pack, I would take my initial parallel charge much higher than the conservative 3.7 I did to nail down the initial balance MUCH closer. I would (follow the manufacturers instructions) and set the initial top balance to 4.0 and let the pack sit there until there is no measurable amps.

For balancing, I am finding individual tweaks easy to over-correct. I am using my power supply set to CV=4.0v, and find that 30 seconds raises the 'final' v on a cell by about 0.01v. I am using two 25watt resistors to take voltage down, which may be too much because it seems like it only takes about 6 SECONDS to result in a -0.01v drain after next charge.

so.... if you are top balancing, I would suggest the initial parallel balance charge set right up to the rating max of the cell, and then making your measurements at the very end of charge, with the charger still plugged in when in series, as soon as the voltage has settled a little below the very end so it is stable long enough to complete recording voltages.
 

Attachments

#12 ·
Connecting the charger has frustrated me again :-(
The cells went up and some cells scattered in every voltage, so I had to switch off the charger.
Hmmm ... I took a H4 bulb (thanks for this advise) and started to drain down the batteries with the highest voltage down to 3.35V.
Get the charger back on and the same thing started again!

So I took four(!) H4 Bulbs, wired them together and got some "alligator clips"(?) to connect the bulbs quickly from one cell to another.
Max. three cells in series could light up the bulb and drain down the cells.

It was a time-consuming procedure to get every cell down to 3.34V under load. The idea behind this was to get them to the same level as the "lowest" ones.

Time was melting like snow in the desert ... so we hurried up and adjusted the charger down to 156V / 3.47Volts per cell.

Charged the pack and all cells were under the magic 3.6V max. voltage, but not really close together. So we looked for the lowest cell and wired a voltmeter.

After charging, we took a ride to check our range.
60km/37mi without dropping the lowest voltage cell under 3.0V.
That has to be enough testing and range for the rallye ... at 00:30 a clock at night :eek:

06:00 suturday morning: loaded the beetle on a trailer, went to the rallye and drove two funny days without problems or voltage problems.

We didn' t won a cup, but we were the only handmade conversion by a private person at the world greatest e-mobility rallye!!!
It was awesome! :D

The next week I would take a closer look and decide the next steps of balancing.
 

Attachments

This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top