2 pack faling behind with charging and get unbalanced (update 9/07/2020)


SLS BATTERY AUCTIONS are now live at www.batteryhookup.com EV, Lithium, LifePo4, Inverters & More

barry

Member
Joined
Oct 19, 2019
Messages
66
i hope you guys can help me.. i have 2 pack in my 14s powerwall faling behind with charging i had them with 0.007 balanced but after a few cycles the differece has gone up to 0.030 what could it be? maybe you guys had an experience with it?.
 

Korishan

Moderator
Joined
Jan 7, 2017
Messages
7,008
Are you using a bms?

How many cycles does it take to drift that much/little?
Generally speaking, drift can happen for lots of reasons. A cell that's very slow discharging, a connection is loose, missing/broken fuse wire to a cell(s), buss wire isn't large enough to handle the current, balance leads from bms not hooked up securely, etc
 

barry

Member
Joined
Oct 19, 2019
Messages
66
Korishan said:
Are you using a bms?

How many cycles does it take to drift that much/little?
Generally speaking, drift can happen for lots of reasons. A cell that's very slow discharging, a connection is loose, missing/broken fuse wire to a cell(s), buss wire isn't large enough to handle the current, balance leads from bms not hooked up securely, etc


Yes i am using a bms (ant 8s to20s 300 amp) all the voltages displays the same as what the mulimeter says.
The drift happens after 5 cycles i will look for broken glass fuses
Also all the cells are measured for IR and the were all very low (25 to 28 mohm) when i buils the packs.

you said "mutch/little" is the drif oke? will it be a problem?
 

Korishan

Moderator
Joined
Jan 7, 2017
Messages
7,008
barry said:
you said "mutch/little" is the drif oke? will it be a problem?

Some view <0.005V drift acceptable, others <0.01V acceptable.

It also depends on how many cells are in parallel. The more that are connected, the more drift that could occur. If it keeps happening, then you have an issue somewhere in your packs.
Another thing that could cause it to "seem" like there is drift, is if you did a test voltage while the charger/discharge was happening, or checking too soon after charge/discharge was disconnected. Don't do a measurement within seconds of disconnect. Wait about a minute to let the cells settle. Under charge/discharge, you could actually have slight voltage different from one end of the pack to the other (this is measured with one probe static position).

It's good that you are watching the voltages that closely, though. 3 digit readings can be pretty helpful in diagnosing issues.
 

barry

Member
Joined
Oct 19, 2019
Messages
66
Korishan said:
barry said:
you said "mutch/little" is the drif oke? will it be a problem?

Some view <0.005V drift acceptable, others <0.01V acceptable.

It also depends on how many cells are in parallel. The more that are connected, the more drift that could occur. If it keeps happening, then you have an issue somewhere in your packs.
Another thing that could cause it to "seem" like there is drift, is if you did a test voltage while the charger/discharge was happening, or checking too soon after charge/discharge was disconnected. Don't do a measurement within seconds of disconnect. Wait about a minute to let the cells settle. Under charge/discharge, you could actually have slight voltage different from one end of the pack to the other (this is measured with one probe static position).

It's good that you are watching the voltages that closely, though. 3 digit readings can be pretty helpful in diagnosing issues.


Thanks for your advise :) i will look in to it i will report back when i find something


Well i think i found something holy crap... :sick: :sick:


image_pachst.jpg
 

barry

Member
Joined
Oct 19, 2019
Messages
66
it stil hold voltage at 4.1 volts and 20 mili ohm with a hole in it :huh:
 

gauss163

Member
Joined
Jun 28, 2020
Messages
278
We need more info. How long (time not cycles) did it take for that balance drift? What type of balancer do you use? Did you do any type of matching when constructing the packs, e.g. by capacity, IR, or self-discharge?
 

barry

Member
Joined
Oct 19, 2019
Messages
66
gauss163 said:
We need more info. How long (time not cycles) did it take for that balance drift? What type of balancer do you use? Did you do any type of matching when constructing the packs, e.g. by capacity, IR, or self-discharge?


all the cells are matched in capacity and IR so that cant be the problem SD cells goes right to the recyle bin...
The BMS i use ia a ANT BMS.
All the other12packs stay perfecly in balance but i found the problem in one pack.... look at photo above :D
 

gauss163

Member
Joined
Jun 28, 2020
Messages
278
^^ Ah, that photo clearly shows the culprit. How in the world did that hole occur? Was there some prior damage to that spot, or was something hot resting on it during cycling? Did you solder directly to the cell?
 

not2bme

Member
Joined
Oct 16, 2017
Messages
497
Woah you caught it in time! Glad you asked the question!

Drift is never ok. There should never be a time when any of your packs should self discharge ahead of your other packs. If you do, then pull that pack out.

Deviation between the packs may change during the day, and that's OK. You will never create an exactly balanced pack. So at certain battery state or SoC during the day, especially when reaching full charge or low charge, one battery may reach that goal faster so if you look at the charge curve the voltage will shoot up creating this deviation from the rest of the pack. For example my battery for the last 30 days may deviate by 20mV depending on the charge level. You see a repeating pattern. Also my balancer does not activate until it reaches 30mV. This is how I know my packs would have an issue if suddenly my balancer starts to activate daily and I see the graph hit 30mV for no reason.


image_wzxhqv.jpg


Now your numbers and my numbers could be different, so maybe if your packs are more closely matched you will have a tighter deviation. Or if your packs are more loose then it might deviate more. For example it's ok to have a 100Ah and a 150Ah and 300Ah battery in series, but only as long as you take out of each pack the lowest common denominator, which is the 100Ah. So the pack even though has higher rated amps is only as strong as the weakest one. Take out any more than you risk ruining the smallest pack. Actually you would only want to take out 80% of the 100Ah pack, which means 80Ah is what you have out of each pack. But with that said, the deviation between these packs would be quite large depending on the SoC of each cell. When the 100Ah pack is depleted, the 300Ah pack would be at 60% charge and with a higher voltage. So not ideal and harder to manage but still a doable scenario.

But when you see a pack start to deviate from the normal, then you know you have an issue. So glad you noticed it.

On your cell that burst, I can only guess that since it was at the corner that you may have bumped it against something hard, creating a dent in the cell. This dent pushes the layers inside the cell into a pinch point that probably over time weakens and creates a short between the layers. That's why we commonly tell everyone to discard any dented cells. Now this is one possibility.
 

gauss163

Member
Joined
Jun 28, 2020
Messages
278
Please post a photo of the positive terminal of the cell with the ruptured sidewall. Are itsvents clogged by debris? Did any debris appear to erupt from the vents? Did you mess with the CID on any cell, e.g. attempt to reset it?
 

barry

Member
Joined
Oct 19, 2019
Messages
66
@not2bme

Thanks for your explanation and yes i caught it in time pfff... i do not know how it happend, the cell was not dented or damaged in some way what i could see when i installed it.
the strange thing is the voltage was 4.15 volt but its slowly goes down in voltage now
 

gauss163

Member
Joined
Jun 28, 2020
Messages
278
not2bme said:
[...] Deviation between the packs may change during the day, and that's OK. You will never create an exactly balanced pack. So at certain battery state or SoC during the day, especially when reaching full charge or low charge, one battery may reach that goal faster so if you look at the charge curve the voltage will shoot up creating this deviation from the rest of the pack [...]

The hourly variations of 0.01V or so shown in your graph likely do not correspond to SOC differences. Rather, they are more likely due to thermal drift in the voltage meters, and possibly also variations in IR, which will serve to perturbthe readingswhen measured under (dis)charge.
 

barry

Member
Joined
Oct 19, 2019
Messages
66
gauss163 said:
Please post a photo of the positive terminal of the cell with the ruptured sidewall. Are itsvents clogged by debris? Did any debris appear to erupt from the vents? Did you mess with the CID on any cell, e.g. attempt to resent it?

CID and fuse are intact :huh:i will take a photo.
there were no debris from the hole in the sidewall..

AND NO I DID NOT RESET THE CID THAT IS THE ******** YOU COULD DO ,i am not that stupid :D
 

gauss163

Member
Joined
Jun 28, 2020
Messages
278
barry said:
the strange thing is the voltage was 4.15 volt but its slowly goes down in voltage now

Yes, it will continue to self discharge due to (micro) internal shorts. Since it is still at high voltage it still has the capability to (violently) vent with flame, so treat it with caution. Neighboring cells may have been highly heated so may also have suffered internal damage.

It will be interesting to see a photo of the vent. Check to see if there is any debris under the vent that clogged it. Don't disturb the cell in any way. Try to preserve it in the name of science in the hope that someone will offer to do a careful analysis.

I've seen reports of such sidewall ruptures when internal shorts occur near the can, where gases can get trapped and be unable to reach the vent, or when the vents are blocked or damaged. I may have time to dig up some links later.
 

not2bme

Member
Joined
Oct 16, 2017
Messages
497
barry said:
@not2bme

Thanks for your explanation and yes i caught it in time pfff... i do not know how it happend, the cell was not dented or damaged in some way what i could see when i installed it.
the strange thing is the voltage was 4.15 volt but its slowly goes down in voltage now

Yes it's dropping and self discharging because of a short and giving out heat around that area. Energy can't be lost and needs to be transferred. So in this case it's transferring the loss through heat. That's the one thing that boggled my mind when someone made me realize that computers don't use any energy, and all of that is turned into heat. Yes it computed something for you, but it's a by-product and it's actually generating heat.


gauss163 said:
not2bme said:
[...] Deviation between the packs may change during the day, and that's OK. You will never create an exactly balanced pack. So at certain battery state or SoC during the day, especially when reaching full charge or low charge, one battery may reach that goal faster so if you look at the charge curve the voltage will shoot up creating this deviation from the rest of the pack [...]

The hourly variations of 0.01V or so shown in your graph likely do not correspond to SOC differences. Rather, they are more likely due to thermal drift in the voltage meters, and possibly also variations in IR, which will serve to perturbthe readingswhen measured under (dis)charge.

The chart is daily so each peak is when the packs are heading towards full charge during the day. I'm sure there's some variations in IR that would affect the voltage, but it's still due to the SoC.
 

gauss163

Member
Joined
Jun 28, 2020
Messages
278
not2bme said:
The chart is daily so each peak is when the packs are heading towards full charge during the day. I'm sure there's some variations in IR that would affect the voltage, but it's still due to the SoC.

SOC imbalances generally don't evolve under short (hour/day) timescales (unless you have a high-rateself-discharging cell, or huge temp imbalances). Li-ion cells havenearly 100% coulombic efficiency, i.e.charge_in = charge_out. So if they start out well-balanced they will remain that way over short periods since there are no charge losses.

But over much longer periods they can become (significantly) imbalanced due to differing self-discharge rates, and also due to various degradation processes.

The minor hourly variations you are seeing are almost surelydue to the reasons I mentioned in my prior post(a poorbalancing circuit may also play a role).
 

not2bme

Member
Joined
Oct 16, 2017
Messages
497
gauss163 said:
SOC imbalances generally don't evolve under short (hour/day) timescales (unless you have a high-rateself-discharging cell, or huge temp imbalances). Li-ion cells havenearly 100% coulombic efficiency, i.e.charge_in = charge_out. So if they start out well-balanced they will remain that way over short periods since there are no charge losses.

But over much longer periods they can become (significantly) imbalanced due to differing self-discharge rates, and also due to various degradation processes.

The minor hourly variations you are seeing are almost surelydue to the reasons I mentioned in my prior post(a poorbalancing circuit may also play a role).

Of course it does. If everything was equal then yes they would stay the same. But if one of my packs are not properly balanced at the exact same level then one pack will reach 100% faster than the other. Like I said before, if you have two packs in series that is 100Ah and one pack that is 200Ah. If at 50% SoC on both of them it reads at exactly 3.8565554V on both packs (i.e balanced at 50% SoC). When the 100Ah pack reaches 100% SoC at 4.20000002V, what would the 200Ah pack read? 4.2000002V as well? No. It will read somewhere in between. There's no logic that states that the SoC for all my packs will remain equal in all ranges.
 

gauss163

Member
Joined
Jun 28, 2020
Messages
278
^^^ So the differences are due to different assumptions. As I wrote: assuming they "start out well-balanced".... Are the packs in your graph very far off in capacity, etc?

Btw, one other point worth stressing. Chemistry also plays a role in balancing. If you create a pack using cells of different chemistries then you end of with a bit of a Frankenstein voltage vs SOC profile, no two of which will likely be the same. So even if their SOCs remain balanced, their voltages may not due to the differing voltage profiles. In that case, a balancer can do more harm than good in some cases - by trying to correct a SOC imbalance that doesn't actually exist.
 
Top