Bike cutting out when battery on 3 bars

[matthewslack]

I can half the maximum current by changing the C5 setting if that is a small enough load?
Obviously I can also manually
1 - put it into first gear
2 - not turn the throttle the full way
3 - put the bike on its side so it has no resistance

I got carried away watching womens curling semi final, sorry for delay!

That would be a fairly gentle load, so worth seeing what happens. Might need to run for a long time though.

What we might learn is whether the cell group lacks capacity compared to the rest, or whether it is just unable to supply high current.

All ways up, we already know the battery cannot do what you need of it in it's current state.

 

[Benjahmin]

From everything you've gone through - this battery is bad.
Replacing cells in a battery is very difficult and dangerous, even if you know what you're doing. From the questions you ask, you don't. I appreciate you are trying to learn but you're current state of electrical knowledge could lead you to getting burnt or worse. I'm not just scaremongering, have a look at this https://www.pedelecs.co.uk/forum/threads/this-should-be-illegal….42843/ -and these guys know what they are doing. Watch the sparks fly.
You ask Mathewslack where you can buy a socket as in the photo, it should be obvious he wired it himself. The fact that you ask rings alarm bells.
Do you realise that, if you have bare wires (red and black) coming out of a battery and they inadvertantly get touched together there will, at best, be a huge spark. At worst they weld together and you get a battery going into thermal runaway.

Bikemerchant, please, for your own safety, stop dicking about with this battery.
It is damaged and will only be further damaged by using it, by your own admission, as an electric motorbike.
It is not able to sustain this heavy usage.
Get in touch with this company:

eBike Batteries

ebikebatteries.co.uk

Ask to speak to Jimmy. Tell him what motor and controller you have. Tell him how you want to use the bike. He will arrange to have your battery picked up. He will re-cell it with quality cells capable of doing the job you want. Then have it sent back to you by carrier. All at a pre-arranged price. It really is painless, I know I've done it - 3 times.
It will not be cheap because what you are asking of the battery is heavy duty.

 

[wheeliepete]

Nope I use it as an electric motorbike (it's 40kg).
Is it safe to presume that the bad cell is in series 11?
If so would an option be to purchase 6 new high quality brand cells (one series) and try to find someone to replace them?

Why don't you take up the offer of the matching cells from the battery seller, then speak to Jimmy, who may be willing to replace them for you. I think he has done this for a forum member recently.

 

[bikemerchant]

Just thought of something extremely simple - there is a backlight attached to the battery that I can turn on with the flick of a switch.
I'll turn it on now and it should slowly drain it.
How long will I leave it until I check series?

 

[matthewslack]

Just thought of something extremely simple - there is a backlight attached to the battery that I can turn on with the flick of a switch.
I'll turn it on now and it should slowly drain it.
How long will I leave it until I check series?

Backlight will be tiny number of watts, so maybe 24 hours between checks.

 

[bikemerchant]

How much would I ideally drain it down to?

This morning - total 54.3V.
All series were 4.18V or 4.17V apart from 11 (4.10V).

 

[matthewslack]

How much would I ideally drain it down to?

This morning - total 54.3V.
All series were 4.18V or 4.17V apart from 11 (4.10V).

The ideal is a full discharge of the weak cell group. Minimum is enough discharge to see if 11 goes down faster under low load.

If this is the backlight load, it might be 1 or 2 watts, which isn't really enough. A tenth of the battery capacity takes about 2 to 4 days at that rate.

I use 100W or so because that's no more than 10 hours, and data every hour is enough for an adequate plot of what is happening.

 

[bikemerchant]

I can half the maximum current by changing the C5 setting if that is a small enough load?
Obviously I can also manually
1 - put it into first gear
2 - not turn the throttle the full way
3 - put the bike on its side so it has no resistance

I did this for almost 2 hours outputting 75W.
Current battery state is:
Total 51.8V
All series are 3.99 or 3.98V apart from 11 (3.92V).

So we have gone from a series variance of 4.22V Vs 4.12V at fully charged, to 4.20V Vs 4.12V, to 4.18V vs 4.10V, to currently 3.99 Vs 3.92v.

So difference between them has gone from 100mV, to 80mV, to currently 70mV.

 

[matthewslack]

I did this for almost 2 hours outputting 75W.
Current battery state is:
Total 51.8V
All series are 3.99 or 3.98V apart from 11 (3.92V).

So we have gone from a series variance of 4.22V Vs 4.12V at fully charged, to 4.20V Vs 4.12V, to 4.18V vs 4.10V, to currently 3.99 Vs 3.92v.

So difference between them has gone from 100mV, to 80mV, to currently 70mV.

75W is a decent load.

I'd keep going, if you have the patience. Another couple of sets of measurements would be good.

 

[bikemerchant]

Now, the total is 49.4V.
All series are 3.81V apart from 11 (3.72V) and 12 (3.80V).
So it seems they are being drained equally at this low load.
Note when it was failing last week the total V was higher than it is now.

Should I do another couple hours of this?
I could also ride it slowly, max limit is set at 700W but I could try to manually keep it lower.

 

[Nealh]

So the low group is pretty steady and decreasing in voltage at a similar linear rate, try a manual charge of that group to match the others then do a full balance charge. Once that balance charge is pretty even (check all cell groups again), disconnect the balance wire jst and discharge the battery, no better way then by riding. Then if all still looks good and voltage decrease is linear, would think about either changing the BMS or solder on another set of sense wires to monitor the pack better from the exterior.

 

[matthewslack]

Now, the total is 49.4V.
All series are 3.81V apart from 11 (3.72V) and 12 (3.80V).
So it seems they are being drained equally at this low load.
Note when it was failing last week the total V was higher than it is now.

Should I do another couple hours of this?
I could also ride it slowly, max limit is set at 700W but I could try to manually keep it lower.

What you are seeing is the cell groups following the discharge curve that was posted earlier in the thread by nealh.

They have come down the initial steep voltage drop, and are now moving across the gently sloping middle part.

The moment of truth comes when they get to the right hand end, where the voltage drop gets much steeper again. How the weak group behaves at that point tells us something about the capacity of its cells. If if goes down sooner, then that cell group has a capacity deficit, I.e. one or more of the cells has less capacity than the rest. If it goes down at the same time, then the problem is not capacity, but is brought on by high current demand.

It might take a while to get to the point when this becomes apparent. If you can continue, we'll have the best info to understand the problem.

 

[bikemerchant]

I had rode the bike battery as low as I ever had the day before it first failed this time around (1 of 4 bars showing on the display). I fully charged it overnight, rode it a couple km, parked it, then when I got back on it, it failed. Maybe this suggests a capacity issue that I ran it down too far?

I probably also should have mentioned this earlier, but the guy who balanced my battery last month alluded to a capacity issue. Photo attached that he sent me.
With English not seeming to be his native language and some of my questions ignored by him, I'll post the conversation here:


Him : We have tested another one of your cells over night and it’s down at 2932mah. We are going to have to test more cells to see what the story is because all cells must balance. So your battery won’t be ready today .

Me: That's the same series that was the problem (the 11th one)?

Him: No this is another cells from the pack .To do this properly we need to test more cells to see what the problem is .
If that’s we just replaced the group that was down with Samsung 35Es you would be back to us in no time with another issue .

Me: What tests are you doing?

Him: If you want it done correctly you will have to leave it with us or else you can take it back .We are not going to do a fast job because you will definitely be back to us again in the near future with a similar issue

Me: Can you explain what tests you need to do on the battery?

Him: We tested the capacity of a single cell we removed from the battery and it’s down at 2932mah.It should be between 3400-3500mah. We are going to remove more now and put them on test again

Me: What group is giving this reading? Because as you can see from my multimeter readings the only group that was close to this low a voltage was series 11

Him: The cell with 2932 is from group 11. Typos We have removed other cells from a different group in the pack and we are testing it now because these cells don’t seem like genuine Samsung 35E cells .
If the cells aren’t genuine we can’t put in genuine 35E cells because they won’t match . I’ll let you tomorrow what the story is ok .

Him: Hi, It was a balancing issue . They cells are not genuine Samsung cells , they are fake cells wrapped with a Samsung wrapper which is hugely popular. We didn’t have to change any cells . Your battery will work perfectly now , but the same issue could happen down the road . The only way to solve this problem permanently is the repair all the cells . But for now your battery will work perfectly.

Me: Thanks. How much would it cost to repair all cells?

Him: Don’t worry about that , it will work fine for a good while .
So use it until you have another issue and then you can contemplate charging all the cells .
It’s repaired and shrink wrapped now .

So the low group is pretty steady and decreasing in voltage at a similar linear rate, try a manual charge of that group to match the others then do a full balance charge. Once that balance charge is pretty even (check all cell groups again), disconnect the balance wire jst and discharge the battery, no better way then by riding. Then if all still looks good and voltage decrease is linear, would think about either changing the BMS or solder on another set of sense wires to monitor the pack better from the exterior.

So 5V500mA charger with magnet method to get #11 up to 3.81V, then connect the standard charger until it's fully charged?

What you are seeing is the cell groups following the discharge curve that was posted earlier in the thread by nealh.

Can you quote the post with this discharge curve? I can't find it.
Thanks.

 

[matthewslack]

I had rode the bike battery as low as I ever had the day before it first failed this time around (1 of 4 bars showing on the display). I fully charged it overnight, rode it a couple km, parked it, then when I got back on it, it failed. Maybe this suggests a capacity issue that I ran it down too far?

I probably also should have mentioned this earlier, but the guy who balanced my battery last month alluded to a capacity issue. Photo attached that he sent me.
With English not seeming to be his native language and some of my questions ignored by him, I'll post the conversation here:


Him : We have tested another one of your cells over night and it’s down at 2932mah. We are going to have to test more cells to see what the story is because all cells must balance. So your battery won’t be ready today .

Me: That's the same series that was the problem (the 11th one)?

Him: No this is another cells from the pack .To do this properly we need to test more cells to see what the problem is .
If that’s we just replaced the group that was down with Samsung 35Es you would be back to us in no time with another issue .

Me: What tests are you doing?

Him: If you want it done correctly you will have to leave it with us or else you can take it back .We are not going to do a fast job because you will definitely be back to us again in the near future with a similar issue

Me: Can you explain what tests you need to do on the battery?

Him: We tested the capacity of a single cell we removed from the battery and it’s down at 2932mah.It should be between 3400-3500mah. We are going to remove more now and put them on test again

Me: What group is giving this reading? Because as you can see from my multimeter readings the only group that was close to this low a voltage was series 11

Him: The cell with 2932 is from group 11. Typos We have removed other cells from a different group in the pack and we are testing it now because these cells don’t seem like genuine Samsung 35E cells .
If the cells aren’t genuine we can’t put in genuine 35E cells because they won’t match . I’ll let you tomorrow what the story is ok .

Him: Hi, It was a balancing issue . They cells are not genuine Samsung cells , they are fake cells wrapped with a Samsung wrapper which is hugely popular. We didn’t have to change any cells . Your battery will work perfectly now , but the same issue could happen down the road . The only way to solve this problem permanently is the repair all the cells . But for now your battery will work perfectly.

Me: Thanks. How much would it cost to repair all cells?

Him: Don’t worry about that , it will work fine for a good while .
So use it until you have another issue and then you can contemplate charging all the cells .
It’s repaired and shrink wrapped now .



So 5V500mA charger with magnet method to get #11 up to 3.81V, then connect the standard charger until it's fully charged?



Can you quote the post with this discharge curve? I can't find it.
Thanks.

It might have been a different thread. This is now almost qualifying as a saga!

Have a look on lygte-info.dk, or search 'lygte Samsung 35e'. They have tested these cells more than once.

Scroll down for the graph.

The alternative to monitoring and measuring voltages every hour or several, if you trust the bms to switch off as soon as one cell group goes to minimum allowed voltage, is to just leave the low load on until the bms switches the battery off. Then measure all cell groups, and the voltage of the ok groups points to where in the discharge curve they were when the weak one was empty.

Your email trail might trigger a thought or two later, but it needs reading more than once and then thinking about! My initial thought is they sound competent and knowledgeable.

 

[Nealh]

If the cells are re wraps/fakes as might be suggested by the battery repair guy you used then there is no knowing if any of the cells are new or what they are.
If re wraps then they could be 29E's rewrapped as the mah rating is about right.

 

[Nealh]

So 5V500mA charger with magnet method to get #11 up to 3.81V, then connect the standard charger until it's fully charged?

Easier to charge via the jst connector, solder a pair of arduino pins on the charger wires.
Buy a small strip like this.
12 Pin Header Pins Strip PCB 0.1" 2.54mm for Arduino, Breadboard | eBay

Break off a pair of pins as one with the plastic as the insulator, solder a wire to each pin and shrink the soldered joint. One then simply pushes the pins as pair into the correct corresponding jst port, the V+ pin in #11 cell port and the V- pin in #10 cell port.

 

[Nealh]

If you are going to keep this battery and use it, it looks like you will need to keep monitoring it and balancing it manually every so often when #11 wanders.
So you don't keep having to open the battery each time buy one of these 12s jst leads and solder it direct to the cells as indicated on page 1 and exit the jst connector to the exterior of the battery, omit/cut either the first or the last sense wire off so there are only eleven wires. The first wire preferably Black solder to 0v and then the others in order to #1 - #10, one needs eleven wires so that #1 has a V- position in case it were to need balancing.
JST XH2.50mm Connector Plug Wire Cable Cord 300mm + pcb mount socket | eBay

 

[bikemerchant]

Thanks guys.

Unsure if I should keep it and keep using it so, if the batteries are 29Es will my screen give an accurate reading of battery life while riding?
Also would depend on how often I would need to manually balance, I guess I would just check overall voltage with the multimeter after it's been fully charged every few days and if the total overall charged voltage drops then it's rebalance time?

solder it direct to the cells as indicated on page 1

I can't find where this is indicated but sounds like I would have to buy a soldering kit.
I also feel I would be more equipped to manually charge #11 with the magnet method as my JST feels like it's glued to the BMS.
I have a 5V 600mA charger which should do the trick? Then I just use the magnets to hold the wires in the places I usually put the multimeter probes when measuring #11?

 

[bikemerchant]

So the low group is pretty steady and decreasing in voltage at a similar linear rate, try a manual charge of that group to match the others then do a full balance charge. Once that balance charge is pretty even (check all cell groups again), disconnect the balance wire jst and discharge the battery, no better way then by riding. Then if all still looks good and voltage decrease is linear, would think about either changing the BMS or solder on another set of sense wires to monitor the pack better from the exterior.

I got a hold of some magnets today.
Readings this morning - Total 49.7V, all series 3.82 or 3.83V apart from #11 (3.73V).
I used the magnet method to bring #11 up to 3.83V, all other series remained the same.

Then I fully charged the battery as per normal, total 55.1V.
All series are 4.24 or 4.23V apart from 10 (4.22V), 11 (4.21V) and 12 (4.22V).

Is this even enough or should I manually bring all series up to 4.23/4.24V?
Or should I just ride now and re-measure? I can either ride limited to 700W or at full capacity.

 

[matthewslack]

I got a hold of some magnets today.
Readings this morning - Total 49.7V, all series 3.82 or 3.83V apart from #11 (3.73V).
I used the magnet method to bring #11 up to 3.83V, all other series remained the same.

Then I fully charged the battery as per normal, total 55.1V.
All series are 4.24 or 4.23V apart from 10 (4.22V), 11 (4.21V) and 12 (4.22V).

Is this even enough or should I manually bring all series up to 4.23/4.24V?
Or should I just ride now and re-measure? I can either ride limited to 700W or at full capacity.

I'd be happy at that - within a couple of hundredths of a volt.

Keep an eye on what happens in use, and hopefully you will find a way of living with the battery's limitations. Maybe also start secretly saving for a really good spec one further down the line.

Interested to hear how it goes.