Trouble keeping cells ballanced.

[Ussyamoto]

I've got 2 LiFePO4 packs, 1 6s and the other 4s, and they will not stay balanced at all.

I'm using an HK 200w balance charger and I notice the same cells are always finishing first causing the rest of the pack to painfully trickle charge. In each of the packs 1 finishes to 3.6 and the others are only at 3.4~ consistently, which slows charging to a painful crawl for a 1-3 hours..

What am I missing, I doubt this is normal, anyone have any advice?

Thanks in advance.

 

[Ykick]

Yeah, it sucks waiting for "balance chargers" to trickle balance. When things take too long I usually rig a simple connection to the balance plug - identify and charge each cell individually as 1S to make sure they're all pretty much top-balanced. Once that's done, if they continue to swing wildly out of balance then it's time to dig deeper into which cell(s) have lower capacity and often poor IR compared to other cells in the series string.

Once balanced, if one or more cells continue to reach HVC (high voltage cutoff) long before the others that's the one(s) with least capacity.

Further troubleshooting would be individual cell IR testing. Not as hard or complicated as it may sound. A CellLog 8S with logging function and the software to view the voltage log can offer insight about which cells are sagging more than others during use. Or, rig a known Amp test load - lights, heaters, etc., check cell voltage drop under load to identify weakest cells.

 

[dnmun]

try leaving the batteries on the charger for 10-30 hours.

 

[Ussyamoto]

Yeah, it sucks waiting for "balance chargers" to trickle balance. When things take too long I usually rig a simple connection to the balance plug - identify and charge each cell individually as 1S to make sure they're all pretty much top-balanced. Once that's done, if they continue to swing wildly out of balance then it's time to dig deeper into which cell(s) have lower capacity and often poor IR compared to other cells in the series string.

Once balanced, if one or more cells continue to reach HVC (high voltage cutoff) long before the others that's the one(s) with least capacity.

Further troubleshooting would be individual cell IR testing. Not as hard or complicated as it may sound. A CellLog 8S with logging function and the software to view the voltage log can offer insight about which cells are sagging more than others during use. Or, rig a known Amp test load - lights, heaters, etc., check cell voltage drop under load to identify weakest cells.

Thanks Ykick, that makes perfect sense. I'll probably make up a 1s charging setup like you did until I have the down time to do a discharge capacity test. My balance charger doesn't have the graph or anthing special, It's just a cheap charger to help me get by till I can afford a better one.

try leaving the batteries on the charger for 10-30 hours.

I can try that on my next day off, I use my bike as a primary vehicle so it's difficult to work on.

If it turns out that I have 2 cells of reduced capacity would it be a good idea to separate them into their own pack?

 

[Ussyamoto]

Just started a discharge test. I can only do a 5A discharge but the cell in question drops from 3.6 to 2.88 under that 5A discharge. That seems like a very steep drop.. would you say this cell is toast?
I'm going to finish the discharge test and then post the capacity.
I should mention that these are 3c/5c 10Ah cells
*Edit*
Log:
Volts/Ah/Time
2.88 - -
2.77 2 25min
2.73 3 37min
2.70 4 49min
2.65 5 60min
2.63 6 74min
2.59 7 86min
2.56 8 98min - Figures, I just realized I might be testing the wrong cell... o well I'll just finish this up and try the other one tomorrow.
2.46 9 111min
2.33 9.5 115min
2.00 9.9 122min

 

[Ykick]

If the 5A load and voltage drop’s close to accurate that would calculate an IR of about 144mOhm. Very bad for an individual cell. I generally stop using cells for bike motor power around 20mOhm and above.

However, nearly 2 hours of 5A load is pretty much 10Ah - are these Headways? I dunno your back story?

At least you have access to individual cell connections (balance connector) so you can glean a lot of useful cell health information.

Good thread about IR testing:

http://endless-sphere.com/forums/viewtopic.php?f=14&t=48739&hilit=

From that thread - here's the video that inspired me to learn IR testing:

[youtube]usRE3gLFnOc[/youtube]

 

[Ussyamoto]

Thanks for the info Ykick, you've been very helpful.

I'm going to go ahead and start logging my cells like this guy does.

I have a mix of cells right now:
4s headway pack 5c/10c new
4s GM rectangular one 3c/10c this is the tested cell
2x 6s GM rectangular one 3c/10c

I think I killed a few cells from testing my new controller. I ran it at 60A for a few second bursts just for kicks, and then ran 30a cont and 50A with 30second bursts for about a week. Also one of my 6s packs had a dead short situation for about 2-3 minutes... xD ya I've been mean to my packs.

Now I'm running a 25a setup babying what little life they might have left.

I notice my full packs drop form 70v down to 61v under ~25a load, that's 360mOhms for all the cells if I'm understanding this right. I guess this weekend I'm going through and finding the worst ones and separate them into their own pack.

Do you know if a full discharge and charge cycle would do anything helpful for these cells? 3.6v to 2.0v back to 3.6v. I hear about some people doing this to recover bad cells but I don't know exactly how it works.

 

[Ykick]

You seem to be on the right track. Other than "calibrating" how much is actually put back into your cells full CC (charge cycles) won't do much to improve poor condition Lithium cells.

That is a major difference compared to how Nickel or Lead may "come around" with a few CC. My experience with both Lifepo and Li-Ion is that they're either good or bad. Saggy ones (high IR) can be useful for low current applications (LED lights?) but at some point the sag is simply too much for motive duty.

In series strings it's very important to discover which cells suck/sag and which cells can deliver near labeled C rates. Basically, a matter of qualifying the "health" of your cells.

As you're discovering IR testing is straightforward stuff only needing a few simple tools/meters and calculator.

 

[Ussyamoto]

I am a little surprised that IR is tested on a cell fresh off the charger. To me it would seem like a better plan to test after the cell stabilizes at nominal and then testing voltage drop...

 

[Ykick]

I am a little surprised that IR is tested on a cell fresh off the charger. To me it would seem like a better plan to test after the cell stabilizes at nominal and then testing voltage drop...

SOC (state of charge) will influence IR values and I suggest you do some mid-bottom charge comparisons as well. Trouble is, it's often hard to tell exactly where mid-bottom state is therefore full charge seems to deliver the most consistent results from cell to cell, day to day.....

Basically, if you measure shit IR on fully charged cells it's only gonna get worse as SOC is consumed. However, the purpose is to compare so a common starting point is the best way to achieve a comparison.

 

[Ussyamoto]

Thanks, that clarifies things a bit.

When I'm testing these cells, should I test them immediately right after they charge and are at 3.6 or should I wait for a few minutes for them to settle down to where they sit stable, usually 3.56-.54?

 

[Ykick]

I give em a few minutes to stabilize and then apply roughly 10 count load to capture sag/drop. The more you do it the more you'll recognize patterns and trends.

 

[Ussyamoto]

Thanks, you've been a great help. I'll do this tomorrow and post my results. I'm going to keep using these cells regardless until they kick the bucket because I can't afford to replace a lot of them at this time. but maybe I'll be able to find their new c rating xD

 

[Ussyamoto]

Ok, so I tested 2 of my 4s packs.

At first I was getting upset because they were all in the 180mohm range and I couldn't understand why my packs were still working. So I hooked up my multimeter in parallel to the HK charger and to my surprise the HK is not accurate at all :/ go figure. The standby voltage is correct but the the loaded voltage is way off.
I retested the cells and now I'm sitting good, to my amazement.

4s GM
# Uv / Lv / A / IR
-1 3.47 / 3.02 / 5 / 90
-2 3.45 / 3.05 / 5 / 80
-3 3.47 / 3.06 / 5 / 82
-4 3.46 / 3.07 / 5 / 78

4s HW - Brand new so I'm using these as a base line.
-1 3.42 / 3.02 / 5 / 80
-2 3.41 / 3.01 / 5 / 80
-3 3.43 / 3.03 / 5 / 80
-4 3.41 / 3.06 / 5 / 70

This doesn't explain why cell # 3 always finishes charging first though. So I'm still confused. I'll take a look again tomorrow and double check the cell number is correct, because it looks like cell #1 would be the problem one.
I'm now charging my other 2 6s packs and will be able to test them in about 3 hours xD

 

[Hillhater]

to my surprise the HK is not accurate at all :/ go figure. The standby voltage is correct but the the loaded voltage is way off.

Never rely on a charger as a voltmeter..especially not a HK one. :lol:
And dont measure cell voltages whilst still connected to the charger.
You need to measure mV and mA accurately for IR testing.

 

[Ussyamoto]

And dont measure cell voltages whilst still connected to the charger.
You need to measure mV and mA accurately for IR testing.

I'm using the discharge function of the HK as a 5A load. I'm not trying to find the most accurate or actual IR. I just need a baseline to know the overall health of my cells compared to each other. I suppose I could connect my cycle analyst and see exactly how many amps and volts I'm testing, that's probably the most accurate I'll be able to get with what I have.

 

[Ykick]

Cells with low IR can still have less capacity (Ah) than cells with higher IR in the string which could explain why #3 hits HVC before the others. Capacity testing can reveal that information.

I wouldn't discount your measurements soley because of the HK charger but I would definitely try other meters and resistive load for the sake of comparison. Old space heaters, toaster ovens, clothes irons, stove/range elements, lights, etc., can make cheap test loads. Obviously, CA can provide power in/out information and a DVM can measure cell voltages. It never hurts to compare. That's why I buy and keep several cheap DVM's around to help verify measurements instead of relying on one meter.

However, if these cells are in 60-90mOhm range they're gonna be very, very saggy delivering power for an eBike.

I'm spoiled with low resistance RC Lipo which I cull out when they test about 20mOhm per cell or higher.

 

[Ussyamoto]

However, if these cells are in 60-90mOhm range they're gonna be very, very saggy delivering power for an eBike.

I'm spoiled with low resistance RC Lipo which I cull out when they test about 20mOhm per cell or higher.

Ya right now my sag under a full load of 25a for my whole pack is about 10v, 71v full charge to 61v then back up to 66v.

I'd like to go with LiPo but I don't trust myself and don't have an option for a proper safe charging setup.

I've got the 6s packs charging right now and will be able to test them in a few hours. I ran out of time last night and had to ride to work.

 

[Ussyamoto]

Just finished testing them all 1 after another and I can see now why the 6s cells are so messed up.

Battery Number Ir
4sGM / 1 / 100
4sGM / 2 / 92
4sGM / 3 / 94
4sGM / 4 / 98

4sHW / 1 / 90
4sHW / 2 / 94
4sHW / 3 / 96
4sHW / 4 / 86

6sG1 / 1 / 94
6sG1 / 2 / 94
6sG1 / 3 / 96
6sG1 / 4 / 98
6sG1 / 5 / 144
6sG1 / 6 / 100

6sG2 / 1 / 146
6sG2 / 2 / 168
6sG2 / 3 / 128
6sG2 / 4 / 136
6sG2 / 5 / 168
6sG2 / 6 / 128

The 6sG2 pack was the one that was dead shorted for 2-3min. I can definitely see the damage that caused now. Looks like cell 5 is bad as well in the 6sG1 pack. That explains why this cell always finishes first even when put in parallel with the 6sG2 pack.
I still can't see why cell #3 finishes first in the 4s packs in parallel, I'll have to look more into it.

These are not accurate IR numbers, they're just references for me to work off of. I'm using the 4sHW pack as a control because it's brand new. I'm going to keep checking 6sG2 over the next few weeks just to see if their getting worse using them of my commute.

I'm going to do a capacity check on cell #3 in 4sGM and see if that's why it finishes charging so quickly, because the IR doesn't look like the cause.

 

[Ykick]

As long as you can repeat the test and see similar results it's likely valid information. #3 is probably just lower capacity than the rest and thus will be the limiting capacity of that pack.

The way you're approaching this would indicate you'd probably be okay working with RC Lipo? Those 20C Turnigy 4S hardcase packs are a great format and value for many of us who go to the dark side. It's good to have respect for the fire potential but it's not quite all doom & gloom either.

Properly qualified RC Lipo cells outfitted with a good BMS can be a very stable, lightweight and relatively safe ebike battery pack. Just a thought....

 

[dnmun]

when a cell had high internal resistance then it is harder to push charge into the cell. you are measuring voltage and if it has a higher resistance than the other cells then for the same charging current, which it has to be, then the cell with higher resistance shows the higher voltage.

 

[Ussyamoto]

As long as you can repeat the test and see similar results it's likely valid information. #3 is probably just lower capacity than the rest and thus will be the limiting capacity of that pack.

The way you're approaching this would indicate you'd probably be okay working with RC Lipo? Those 20C Turnigy 4S hardcase packs are a great format and value for many of us who go to the dark side. It's good to have respect for the fire potential but it's not quite all doom & gloom either.

Properly qualified RC Lipo cells outfitted with a good BMS can be a very stable, lightweight and relatively safe ebike battery pack. Just a thought....

It's weird I just did another capacity test on the 4sGM cell 3 and it's at 9.9 Ah. I'm going to do a capacity test on 4sHW cell 3 next... I might have been shipped a dud?

Thanks for the vote of confidence, but I'd rather stay away from LiPo. I know they're cheap and powerful, but they have to be monitored while being charged and that they're safest being charged out doors both of which I can't really do.