How long is it safe to use an ebike battery?

Hi,

I own a 350W Bosch E-bike with a 36V 10S4P 400Wh battery made out of Samsung 29E cells. This battery is now 7 years old and has powered the bike for about 30'000km. It still works perfectly although its capacity is lower than new and it sags a bit. The battery has been used lightly because it is charged with a 4A charger (1A per cell which means something like 0.4C) and the discharge current was also limited to 16A (max 4A per cell whereas they're supposed to handle twice that amount in peak). I must be well over 500 cycles at the moment but since the battery still works, well I think about keeping using it.

Now I've read a few things about Li-ion battery fires and especially about dendrite formation on the anode which can potentially grow and short-out a cell internally. I guess this is a very rare occurrence but it seems the likelihood increases with age of the battery. Therefore, I'd like to know whether there is some maximum safe time over which a li-ion battery shouldn't be used anymore even if it still works perfectly because the risk of dendrites piercing the middle membrane in the Li-ion cell becomes too high??? Alternatively, is it possible to assess the level of dentrite formation in an old battery by looking at the battery under an electronic microscope? In anyway, I plan to invest in a metal box and charge the battery in it since I live in an appartment and cannot charge outside.

Thanks for your thoughts

Well, its tempting to say your battery is so safe you need not worry.

But after my house burned, I have a different standard for what % risk I'll take in the future. Pretty much zero, but I have a large suburban property, and keep my batteries in the old stove from when the house burned.

I would be particularly careful about the charging. I think your battery is a lot less likely to burst into flame once it is full, but when it gets to top of charge, that is when a bms problem can overcharge a cell and cause a fire, as happened to me with a cheap shoddy ali express battery.

Just sitting there, more likely to just slowly discharge itself when small dendrites form. Ver slow self discharge won't set it on fire. Test this by charging, then sitting overnight, then look at the voltage. Usually a pack will drop some overnight, but then it should stop, or at least get very very slow. Look at voltage again a day later. If its noticeably less, you do have self discharge happening. That would be time to retire the pack for sure.

Thanks for the information, really appreciated! If I understand it correctly, a slow self discharge over time could be indicative of a dendrite formation problem and then the pack should be renewed? Makes sense and that would be fairly easy to check. Currently the pack is not in use and is sitting with a partial charge of 36.3V. Instead of charging it up to 42V and then monitoring its discharge, does it make sense to monitor the discharge from the current state only (36.V)? I usually prefer to fully charge only just before using the pack but if I monitor it over several days, it could be best to start from a lower charge...or does this interfere with something (e.g. dendrite only forming near full charge)?

As charging is the most critical time when problems may occur, would that make sense to have some alarm on the charger if it outputs a voltage higher than 42V? I mean the BMS is supposed to stop charging if that ever occurs but the source of the problem remains the charger. So are there any simple external unit (voltmeter with alarm) that could be attached to the charger and open the circuit if the voltage exceeds 42V (or 42.2V)? That would act as an additional barrier in the event of a BMS/charger failure no? Basically it would be a voltage sensitive switch or something?

Anyway, it's probably time for me to invest in a metal box and a bag of sand. Currently I only have one of those Lipo guard charging bag but I guess this is not enough to contain a fire. I may buy a new battery anyway because the range of the old one becomes limited but it'll be good to keep it as a second battery (or use it in parallel with the new one using ideal diodes)

I'm more asking than stating , but wouldn't testing internal resistance be the clear way of knowing if a battery is still safe? If it's still within a certain percentage range of original DCIR , it hasn't "decayed" to an unsafe point , correct?

Resistance is a challenge to properly measure accurately.

Capacity much more straightforward.

If you really want to be safe, recycle at 80% SoH.

Push to 70% if you are careful and no other problem signals manifest.

After that IMO really taking your chances.

If your ESIR instrumentation is known accurate, and

you set initial benchmarks at commissioning time

and you are strict about exactly duplicating the protocol used, especially temperatures, keep the same gear ideally well calibrated, etc etc

then yes that is also a useful supplementary data point.

And of course "weakest link" applies, if one cell in a pack is substandard

without monitoring temperature

or using per-cell voltages while charging

a pack at 95% SoH overall could still go up in flames and burn down your house

due to that one cell going into thermal runaway

Ok thanks for all your replies. For now I'll have a look at the self-discharge rate to see if there's anything special. I also plan to open the battery case and measure the cell voltages to see if there's any imbalance that accumulated over these 7 years. This may also contribute to some range limitations. I'm not sure to know how to measure capacity without actually completely depleting the battery (which isn't good I guess) but I still need to do my homework on that one. Indeed I may renew the battery if the capacity has dropped below 80%. Probably that for the time being, the easiest way to improve safety is to charge in a metal box and keep a bag of sand nearby just in case.

Yes a cap test is a CC controlled discharge from 100% Full to some safe defition of 0% empty.

Since you might only do it every 50-100 cycles, not a major stressor, compared to improper care from ignorance anyway.

Again, charge it full, and let it discharge overnight. By morning it will settle on its " current real world max charge voltage" This will reflect loss of surface charge, which old cells can't hold anymore, plus any discharge by the bms of cells that get too full in the charging process. You will have some cells that got full, and some that did not. The end result is less, and will represent a good portion of your current loss of capacity from original.

In any case, by morning it should settle on a voltage that pretty much holds for a long time. If its losing another volt a day, you have dendrites causing self discharge. Letting it sit long enough in that condition, will drop cells below safe discharge voltages, and ruin the battery, and making the recharge dangerous.

I do agree, if the capacity is less than 70% on a range test, then that is also a good indicator that cells are very worn and need replacement.

Another good indicator is excessive sag under load. This becomes very apparent when the battery bms shuts the pack down at a higher voltage than when it was new. The old cells can't take the current, because of higher internal resistance. So if you used to get shut off at 28v and now it shuts off at 34, or whatever, that's a sign that the battery is worn out, even if still "safe" as in not exploding at the moment.

Lastly, 7 years lifespan is outstanding, and that alone may mean you should get new. Even if the thing is safe, inside an apartment, you are gambling your life, your neighbors babies life, etc. For god sake, stay awake while it charges.

That applies to the new one, just as much!!!!

Before you start checking the battery, be aware that some bosch battery bms boards stop working if the cells or balance wires are disconnected. They don't restart working if you reconnect the cells. This is to prevent people from installing new cells in their batteries.

dogman dan said:

Again, charge it full, and let it discharge overnight. By morning it will settle on its " current real world max charge voltage" This will reflect loss of surface charge, which old cells can't hold anymore, plus any discharge by the bms of cells that get too full in the charging process. You will have some cells that got full, and some that did not. The end result is less, and will represent a good portion of your current loss of capacity from original.

Click to expand...

Makes sense. Will do that.

dogman dan said:

Lastly, 7 years lifespan is outstanding, and that alone may mean you should get new. Even if the thing is safe, inside an apartment, you are gambling your life, your neighbors babies life, etc. For god sake, stay awake while it charges.

Click to expand...

I use a timer to charge the battery in the morning. The charging starts maybe 30 minutes before I wake up but last for about 1-1h30 so I'm awake when it reaches full charge. Not optimal I guess. I may try to set up the timer to start twice: once in the evening to have a partial fill-up, stop for the night so the battery stays at partial charge state overnight, and then restart the charging in the morning when I'm awake.

obcd said:

Before you start checking the battery, be aware that some bosch battery bms boards stop working if the cells or balance wires are disconnected. They don't restart working if you reconnect the cells. This is to prevent people from installing new cells in their batteries.

Click to expand...

Mine is okay in that respect since it's a Gen 1 Bosch battery. On that one, it has been shown the BMS can be disconnected safely (provided the correct sequence of disconnection is used) and reconnected later on. Many people have actually re-celled their Bosch Gen1 pack but I don't feel like I have the skills to do it and don't own a spotwelder. Problems only seem to occur with >Gen2 packs where Bosch implemented "suicide chips" to kill the BMS when it gets disconnected.

dogman dan said:

Again, charge it full, and let it discharge overnight. By morning it will settle on its " current real world max charge voltage" This will reflect loss of surface charge, which old cells can't hold anymore, plus any discharge by the bms of cells that get too full in the charging process. You will have some cells that got full, and some that did not. The end result is less, and will represent a good portion of your current loss of capacity from original.

Click to expand...

Yesterday evening, I fully charged the battery. Immediately after unplugging the charger (9.35PM), the voltage readout was 41.3V. This morning I checked the voltage again at 3.35AM and 6.45AM and the readouts were both 40.7V. So the voltage dropped by about 0.6V in about 6 hours and remained the same thereafter. This amounts to 0.6/41.3=1.45% of self discharge which includes some draw from the BMS, which sounds pretty reasonable for a 7 years old battery. Do you think this suggests that it is still "safe" (at least as far as dendrites growth is concerned)?

The Bosch batteries have been pretty trouble-free you would hear about it if there was a lot of fires like pedigo was having.

Just use it. It's a Bosch and thus well-made.

No amount of testing you can do can show that the battery is safe - just that it is unsafe. I have never heard of a battery fire from a Bosch battery, so I would not worry the slightest, but I guess we are all different.

A lot of people here use LIPO packs with no BMS. That is proper dangerous and really stupid, but people don't talk about fire hazard in those threads

A voltage drop between right after charging, especially that high a termination voltage

does not indicate "self discharge", which means lowering of SoC.

The only accurate test for the health of the pack is a CC load capacity test.

Ideally comparing to the exact same test taken when new as a benchmark.

Doing the same for ESIR will also help guide your EoL decision.

The pack has one or two thermal switches that when they detect heat it shuts everything down but doesn't mean it shuts it down before it's too late and is vented gas and let cell to cell to cell to cell caught on fire. But at the over designed of most bosch products and because they use it at a very low amp and you what have a 2 or a 3 amp charger
Have you tried to price a new battery ?

999zip999 said:

The pack has one or two thermal switches that when they detect heat it shuts everything down but doesn't mean it shuts it down before it's too late and is vented gas and let cell to cell to cell to cell caught on fire. But at the over designed of most bosch products and because they use it at a very low amp and you what have a 2 or a 3 amp charger
Have you tried to price a new battery ?

Click to expand...

I have a 4A charger so the charging current per cell is 1A since the pack is 10S4P. That's about 0.4C charging current which is pretty low. The price of a new Bosch battery in my area is about 400 euros which is 435 USD. This is pretty expensive for a 400Wh battery...but it is true that the Bosch BMS is truely reliable and has served me very well over 7 years! I haven't read about Bosch batteries going up in flames so if that is the price of safety, I'm sold. Initially I wanted to go the DIY route and build a 10S6P pack with guenine Samsung 29E cells, copper foil as bus bars retained with magnets (solderless) and use a bluetooth BMS. This would turn out 2/3 of the price of the new Bosch for 33% more capacity but I think I'd always feel uncomfortable about the BMS or cells getting loose etc. So maybe going with a new Bosch battery and keep the current one either as a spare or run it in parallel with the new one (using ideal diodes) would be the "safest" option.

likely out of balance more than suspected

get balance right

load + capacity test

29e relatively calm cell

keep an eye on

use if useful

kcuf said:

likely out of balance more than suspected
get balance right

Click to expand...

I still need to open up the pack an manually check balance. However, if it looks like out of balance, are there any ways to bring it back into balance using some simple and cheap things i.e. without having to buy a new balance charger etc? Basically are there any simple, cheap and reliable (passive) balance boards capable of a balancing current of e.g 1A instead of the few mA the BMS does? A good thing would be to have the ability to balance the pack using a balance-only board (not a fully fledged BMS, just balance capabilities) every now and then and without waiting for a full charge to start balancing. Do such boards exist?

Fully charge for 12 hours or more . This will give the battery a chance for bms it is very very slow to balance it may take twice as long. Check voltage let set overnight then check. did you in the past take it off the charger as soon as it turn green or to leave it on the charger for hours and hours giving a chance for the battery pack to balance through the BMS board.

yes open up

note cell group voltages

apply load doing same


i never regret buying

small rc charger

numerous uses

goto device for this situation

qualify and manually balance parallel groups


bleed balance pack bms very flaky

less than ideal balancing under perfect conditions

Sounds to me like it charged, balanced pretty good for any battery over 2 years old, and its still "safe" to keep using.

Again, after a year waiting for my burnt house to be rebuilt, nothing bigger than a laptop enters my home now.

Sag under load is another thing,, your battery is very old, and you will get much better performance from a new one. Surely you got your moneys worth out of it by now. But as it gets old, it will sag more under load, and shut down early, even if no bad cells are in it.

I still say soon as you can afford it, get a new one. So it goes the distance for you when you want it.