BMS destroyed my battery?!

Joined
Jan 10, 2020
Messages
47
Hi,

I recently built a 20s17p battery pack for my ebike using second hand samsung ICR18650-22P Cells (2200mah capacity when new).

The issue:
Since building the battery i have cycled (discharge-recharge) it approx 5 times, up till a week ago all went well, voltages were as I would expect, when I last checked battery voltage (after having the bike stored for approx a week without using it) the voltage was approx 52Volt (when it should be approx 70v since I had charged it before storing the bike only a week prior).


I checked the group voltages, 7 out of 20 groups are 0v (or close to it... 0v-0.32v).
The groups which are 0v are B2-B4 and B12-B15 (I started counting from the negative side of the battery).

The battery:
2021-03-20.jpg

NOTE: Prior to opening up the battery the bms's sense wires were insulated with an extra layer of insulation (duct tape) from the exposed cell ends/nickel strip for safety)

I openend up te battery pack and checked all cells/connections: All looks good (no corrosion or other damage to any cell or wire/connection).

Usage:

The batteries were only being discharged at approx 0.5C (they are rated for 10A max continuous current and I was using them at max 4A per cell during these 5 cycles). I don't live in a warm climate either so the cells definitely didn't get hotdue to my usage.

The cells did get quite cold approx 2 weeks ago (negative 10 celcius) but this shouldn't be an issue I think since from what I found only they only freeze at approx -40 degrees C (also if the low temperature was the issue I would expect all cells to have died, not just 7 groups).

My guess (what caused it):
I conclude that the issue is with the BMS since I find it highly unlikely that in so few cycles, so many cells from different parallel groups would 'die' due to the cells being in bad condition or something similar. What I think happened is that the BMS severely overdischarged the cells in these 7 groups. All wiring from the cell groups to bms is correct/intact (I have checked).

Do you agree with my conclusion? If not, what do you think is the issue that caused the 7 groups to die?

Unfortunately My BMS does not have bluetooth so I wasn't able to check its performance/the cell voltages during usage, when rebuilding/replacing the battery I will make sure to upgrade the BMS to a bluetooth capable one so I can check what it's doing..

Thank you very much!
 
How rigorously did you cap test the batteries before assembling the pack?

If it was a mixed bag not closely matched might not be the BMS fault.

I treat 80% SoH as the go/no-go boundary

with a really top notch model famous for good longevity maybe 70%

but really will have a very limited useful lifespan
 
@john61ct I capacity tested each cell in the pack. Minimum capacity 1850 max capacity 1950.. most of them approx 1900. I made sure to capacity match the cells so that the capacity of each parallel group was approximately equal . I also checked initial voltages of each cell before checking their capacity and made sure they were all above 2.5v initial voltage. I did not measure internal resistance (perhaps I should have )(I did not because I currently don't have a tester which is capable of that)
 
If the BMS did the damage, then the balancers for those channels are probably stuck on and drained those groups. This can be tested for by checking for voltage across the shunt resistors themselves (you'll have to hook up some test cells across those channels in place of the dead groups first; this can be done one at a time, but the BMS will still need to be powered by the pack normally so the balancer circuits have power to operate.

Balancers that are not on should have no voltage across the resistor (because there's no current flow). (but if the cells are dead, 0v, there won't be any current or voltage either).

If the balancers are the typical low-mA type, they might handle around 50mA average (at full voltage, it will decrease as voltage drops). At that rate, that would be 50mAh per hour, or 1.2Ah per day, drained from each bank. If it sat for a week, that would be about 8Ah. If your groups are 1800mAh cells, 17p, then that's about 30Ah. So to drain that in a week, it would take at least 180mA balancing current drainage on each to completely drain them. Probably at least 300mA, since it would have to be a lot higher at first to make up for the lowering current as voltage drops..
 
Cold comfort for OP

but this highlights the advantage of building packs so that the BMS is easily removed / replaced

also ideally giving concurrent access to the balance leads while the BMS is in place.

When the pack is not being cycled, disconnect the cells from any gadgetry, so they are isolated while in storage, even if only a few days.
 
john61ct said:
Cold comfort for OP

but this highlights the advantage of building packs so that the BMS is easily removed / replaced

also ideally giving concurrent access to the balance leads while the BMS is in place.

When the pack is not being cycled, disconnect the cells from any gadgetry, so they are isolated while in storage, even if only a few days.

Even better would be a setup where the BMS can be completely deactivated for storage, then reactivated for control of charging and SOC. It just seems wrong, in the 21st Century, to have to open a pack up and pull apart a big connector just to protect one part of your battery from another part while being stored. Do those fancy new Bluetooth BMS units have that capability?
 
@Goatman isn't 4.2 a normal maximum voltage for an 18650? I thought 4.2 or 4.25 was acceptable as maximum voltage. It could perhaps still be that the bms overcharged the cells a little but that shouldn't cause the cells to die in only 5 cycles I think.
 
you bottom balanced the cells at 2.5v then when you charge old cells with slightly different capacity and resistance to 4.2v or 4.25

they cant hold capacity/voltage up high and they can go out of balance right away and quickly

try charging each of your p-groups to 4.2v and let them sit overnight, they will drop to different cell voltages after 24 hours
i pick the lowest voltage of the p-groups like 4.15v?,( then ill bottom balance them again) multiply by the series count and thats my charging voltage, no bms and pack stays balanced
im thinking of using active cell balancers with bluetooth for my packs, no bms

just pull the nickel strip off your bad groups wait 24hrs and check each of cells voltages, see how many bounce back on their own and how many stay at zero

they also have a 2.75v lvc not 2.5v and if theyre like other samsung cells they dont like cold
http://www.datasheet39.com/PDF/845518/ICR18650-22P-datasheet.html
 
goatman said:
charging to 4.2v probably screwed them

Perhaps that would be true if cells were LifePO4 or nickle chemistry, however this is clearly not the case for OP.
4.2V is fine.
I do not trust BMS, so last 3 packs I have built I did not install a BMS.
My battery voltage/balance monitoring solution:
When I build my packs, I install taps with multiple RC-style 6S (two for my 12S packs) connectors.
I use the connectors to check cell balance/voltages from time to time.
If things get out of balance (never have yet...I use new identical cells) I plug up a 6S RC-balancer post-charge to bring things in line.
I "bulk-charge" my packs to 4.1V per cell, and discharge no deeper than 3.3V/cell using MPPP "boosting" Lipo solar controller/charger programmed to cutoff voltage of 4.1 * "S" of pack.
 
i see it in my testing where the capacity loss happens at the 4.2v end of the battery capacity

pushing amps into a cell at 4.2v where the capacity no longer is just seems like it would kill the battery faster

if i desulfate some batteries the capacity will return up at 4.2v end but more like 4.15v

so i agree with you about charging to 4.1v

in my case if i charge my batteries to 4.15v without a bms they settle overnight to 4.10v, my 25r's stay balanced and my 30Qs so far are staying balanced, no bms.all salvaged cells, if i charge to 4.2v they go out of balance

AFAIC charging old batteries to 4.2v kills them

you can see what im talking about in this thread

https://endless-sphere.com/forums/viewtopic.php?f=14&t=109601&start=25#p1622842

IMG_20210307_163327.jpg

IMG_20210307_163248.jpg
 
LeftieBiker said:
john61ct said:
Cold comfort for OP

but this highlights the advantage of building packs so that the BMS is easily removed / replaced

also ideally giving concurrent access to the balance leads while the BMS is in place.

When the pack is not being cycled, disconnect the cells from any gadgetry, so they are isolated while in storage, even if only a few days.

Even better would be a setup where the BMS can be completely deactivated for storage, then reactivated for control of charging and SOC. It just seems wrong, in the 21st Century, to have to open a pack up and pull apart a big connector just to protect one part of your battery from another part while being stored. Do those fancy new Bluetooth BMS units have that capability?

Unfortunately "completely deactivated" does not equal "absolutely no parasitic power drawn".

Think about it, how could it receive the signal to reactivate?

Now, there may well be manually switched units that **do** power 100% down, but that would very very rare, require thorough testing.

My advocating for complete isolation via unplugging, is not just for this reason.

Battery Murdering happens from "random" circuitry failures as well, and not just BMS. People tend to use cheap-chinese gadgets, and IMO full isolation while not in use, is the only way to protect the cells in storage.

Easy removal / replacement of the BMS also facilitates periodic stress / cap testing, balancing etc maintenance procedures.

Also, the BMS does not **need** to be "on the inside" requiring a difficult process to disconnect it.

It just needs to be "close electrically", in fact could be integrated into the controller - housing if not the circuitry itself one day.

With a balancing charger, or dedicated non-BMS balancing gear, and other precautions, the very need to use a BMS while discharging can be eliminated, once the owner has built up knowledge and confidence through deep familiarity.

After all "a BMS" is just a given set of functionalities, none of which need to be implemented in an all-in-one (usually cheap and failure-prone) circuit board.
 
maarten_almighty said:
@Goatman isn't 4.2 a normal maximum voltage for an 18650? I thought 4.2 or 4.25 was acceptable as maximum voltage. It could perhaps still be that the bms overcharged the cells a little but that shouldn't cause the cells to die in only 5 cycles I think.
I agree it is not necessary to go past 4.05 - 4.15V when charging - as far as the **cells** are concerned.

Depends on the charging C-rate, how long CV stage is held etc.

Unfortunately the way the industry has adopted **top** balancing as the mainstream default

and the crappy cheap circuitry used to implement balancing in 99.99% of balance chargers and BMSs

people have become conditioned to accept the tail wagging the dog

allowing their charging profile to be way too long & high V, thus stressful and reducing cell longevity

because "that's what the BMS requires" for ensuring the balancing process gets finished.

 
maarten_almighty said:
@Goatman isn't 4.2 a normal maximum voltage for an 18650? I thought 4.2 or 4.25 was acceptable as maximum voltage. It could perhaps still be that the bms overcharged the cells a little but that shouldn't cause the cells to die in only 5 cycles I think.
Yes this sub-topic is a derail as far as your case is concerned.

I believe your BMS did indeed murder your pack, but not due to the "4.2V is too high", that being stressful manifests by an overall reduction in longevity **long-term** impossible to quantify (or even notice) without rigorous testing using accurate equipment in lab conditions.

Note also, the BMS 99.99% of the time has no control over charging

nor should it.

The charger should handle the automatic termination, is not even aware the BMS is there.

THE BMS HVC functionality setpoint is much higher than normal charging, only there as a backup failsafe for when the charger's regulation circuitry **fails**

in order to prevent fires, last-ditch protection of the pack.

With a good charger, you can adjust the profile setpoints, ideally not just voltage but how long CV / Absorb cycle is held before termination.
 
It also is still possible there were crappy cells in the mix that a simple cap test did not reveal.

Of course, using a known-good quality model cell, genuine and A-grade, purchased brand new from a trusted seller

is 1000x better for longevity and reliability

than messing around with scrapped second-hand ones likely to be past EoL.
 
@john61ct please correct me if I'm wrong but I think that the high voltage cutoff in the bms should still come into play in some cases even when charge voltage is set correctly: example: let's image a really badly balanced pack, let's say 19 of the twenty parallel groups are at 3.5 volts and 1 out of twenty is at 3 volt. When charging the pack the cells that were at 3.5 volts will eventually reach approx 4.2v but the total pack voltage will still be relatively low (since there is one parallel group which is lower voltage) so the charger will keep on charging assuming the bms doesn't switch off the charging input due to at least one of the cells being above a certain voltage..
It makes sense that the over voltage protection would only kick in when a cell voltage is higher than it usually gets during charging (to make sure the protection doesn't trigger when it's not necessary) but not much higher than 4.25 I would think (in order to protect the cells from damage).
Perhaps I am using the wrong terminology here.. I assume most bms's also have an overcoltage protection which acts when the total pack voltage is above a certain value.. this protection would indeed not trigger during normal charging..
 
goatman said:
just pull the nickel strip off your bad groups wait 24hrs and check each of cells voltages, see how many bounce back on their own and how many stay at zero

Interesting suggestion, I am willing to try but I don't quite understand how that would work. I was assuming all cells in all of the seven parallel packs that are at or close to zero volts were trash/non recoverable. Could you please explain how cells which are at 0v or close to 0v could go up in voltage when disconnecting them from the rest of the cells in the parallel group(without charging them)(or did I misunderstand perhaps)?

Thank you
 
if 1 cell in the p-group is self discharging it will pull the others down to 0volts

you can think of the 0 volt cell as leaving the lights on in your car, it will drain the battery down
by separating the cells you remove the load (self discharging cells)

ive been trying for a week to zero volt some laptop cells so i can dispose of them and they keep bouncing back up in voltage, wont stay at zero volts, they kept bouncing back to about 0.96v

they were no good for a bike pack, i shouldve kept them to run a flashlight

i just grabbed them

IMG_20210321_195012.jpg

click on a 12v bulb, it wont light up but it will take it down to zero

IMG_20210321_195143.jpg

remove load and voltage starts climbing back up

IMG_20210321_195209.jpg

your cells that went to zero volts are probably no good for a bike pack anymore, they will fail

i would tear that whole pack apart, slug tape all the cells together to equalize them with each other for a week, take the tape off and toss any cells that arent the same voltage as the rest and you might be safe but you should do a resistance test to each cell to see if you have enough matching cells to build a pack with
 
goatman said:
if 1 cell in the p-group is self discharging it will pull the others down to 0volts

you can think of the 0 volt cell as leaving the lights on in your car, it will drain the battery down
by separating the cells you remove the load (self discharging cells)

ive been trying for a week to zero volt some laptop cells so i can dispose of them and they keep bouncing back up in voltage, wont stay at zero volts, they kept bouncing back to about 0.96v

they were no good for a bike pack, i shouldve kept them to run a flashlight

i just grabbed them

click on a 12v bulb, it wont light up but it will take it down to zero

remove load and voltage starts climbing back up


your cells that went to zero volts are probably no good for a bike pack anymore, they will fail

i would tear that whole pack apart, slug tape all the cells together to equalize them with each other for a week, take the tape off and toss any cells that arent the same voltage as the rest and you might be safe but you should do a resistance test to each cell to see if you have enough matching cells to build a pack with
So if I have 1 bad cell in a 9 parallel, the bad cell works as a parasitic drain on the other 8? I ask this because my old pack has 9 & 9 cells out in 2 p. The other 14 series are fine. In other words, I get 18 zero volt. And 126 cells ~ 3.5 v. Then the bms is not the cause.
 
RTLSHIP said:
So if I have 1 bad cell in a 9 parallel, the bad cell works as a parasitic drain on the other 8?
Yes

RTLSHIP said:
Then the bms is not the cause.
Not necessarily -- the BMS could have overdischarged those two p-groups.
 
RTLSHIP said:
goatman said:
if 1 cell in the p-group is self discharging it will pull the others down to 0volts

you can think of the 0 volt cell as leaving the lights on in your car, it will drain the battery down
by separating the cells you remove the load (self discharging cells)

ive been trying for a week to zero volt some laptop cells so i can dispose of them and they keep bouncing back up in voltage, wont stay at zero volts, they kept bouncing back to about 0.96v

they were no good for a bike pack, i shouldve kept them to run a flashlight

i just grabbed them

click on a 12v bulb, it wont light up but it will take it down to zero

remove load and voltage starts climbing back up


your cells that went to zero volts are probably no good for a bike pack anymore, they will fail

i would tear that whole pack apart, slug tape all the cells together to equalize them with each other for a week, take the tape off and toss any cells that arent the same voltage as the rest and you might be safe but you should do a resistance test to each cell to see if you have enough matching cells to build a pack with
So if I have 1 bad cell in a 9 parallel, the bad cell works as a parasitic drain on the other 8? I ask this because my old pack has 9 & 9 cells out in 2 p. The other 14 series are fine. In other words, I get 18 zero volt. And 126 cells ~ 3.5 v. Then the bms is not the cause.

yes

and if you have old cells lower your charge voltage to 4.1v instead of 4.2v
 
goatman said:
ive been trying for a week to zero volt some laptop cells so i can dispose of them and they keep bouncing back up in voltage, wont stay at zero volts, they kept bouncing back to about 0.96v

Get a bucket of water, dissolve some salt into it, then drop your batteries in. That will safely discharge them.
 
goatman said:
your cells that went to zero volts are probably no good for a bike pack anymore, they will fail
Agree

goatman said:
i would tear that whole pack apart, slug tape all the cells together to equalize them with each other for a week, take the tape off and toss any cells that arent the same voltage as the rest and you might be safe but you should do a resistance test to each cell to see if you have enough matching cells to build a pack with
No! Dead cells likely already killed p-groups, and now you want to parallel 1/3 dead cells to 2/3 live cells? He'd kill the rest of the pack.

As per the first quote, cells in any 0V p-group need to be recycled. Test the remaining 221 cells and rebuild the pack with remaining good cells.
 
goatman said:
and if you have old cells lower your charge voltage to 4.1v instead of 4.2v
Only as a function of capacity imbalance resulting in overcharge when bulk charging.
If top-balancing, this is not necessary.
 
fatty said:
goatman said:
and if you have old cells lower your charge voltage to 4.1v instead of 4.2v
Only as a function of capacity imbalance resulting in overcharge when bulk charging.
If top-balancing, this is not necessary.

i disagree, if youre pushing amps into a cell at 4.2v where capacity no longer exists, i believe it degrades cells faster and throws the pack out of balance even quicker. probably wrecks cheap bms's quicker also depending on how many amps the charger is pushing
 
Back
Top