Replacing a bad cell in a battery pack?

[Nick McCutcheon]

Hi all,

Unfortunately I just suffered a bad crash on my ACM2, which required me to replace the shells (I’m fine aside from road rash, but the wheel wasn’t...) The repair went well, and I even replaced my motor during the repair as it had been dented. 

During the crash, I noticed one of the battery packs had taken what appeared to be only a slight bit of external damage to the blue wrapping layer, as the side panel had come open in a few places. Turns out, when I go to charge the wheel now, it only charges to 90% and then the charger (ewheels fast charger) turns on and then off in a cycle that lasts around 2 seconds. My theory is that this is because the BMS keeps continuously shutting off and then turning back on as it tries to protect overcharging a bad cell. I haven’t opened battery pack fully yet to inspect, as I need to use this wheel nearly every day, but I suspect I’m going to need to either replace a bad cell (low cost, lots of time to fix) or get a replacement pack (high cost, very little time to fix), and I want to know if anyone on here has had luck with replacing bad cells. I’m fine with paying for a replacement pack if I need to, but I’d rather try to save some money before doing that.

Also, @Marty Backe, replacing the motor didn’t fix my axle play completely. So either I must not be tightening the nuts hard enough (I can’t imagine what sort of force it necessary in that case!!) or something else is going on here. 

Update:

I took the casing off of the battery. Two of the cells at the edge of the pack are clearly dented, and some of the links that connect the batteries to the BMS have torn and are barely still attached. I’ve deemed this pack unsafe, just to be sure. So for now I’m going to have a 650Wh ACM... It looks pretty simple, just desolder and remove the bad cells, and replace/resolder the new ones, right?

Update Update:

I bought 3 replacement cells (I only need two, but it's good to have an extra) and a ~$100 spot welder, so I'm committed at this point. I'm sure it can't be too hard. Also, it turns out one of the dented cells had completely broken off from the nickel strips. Yikes!

[Marty Backe]

24 minutes ago, Nick McCutcheon said:

Hi all,

Unfortunately I just suffered a bad crash on my ACM2, which required me to replace the shells (I’m fine aside from road rash, but the wheel wasn’t...) The repair went well, and I even replaced my motor during the repair as it had been dented. 

During the crash, I noticed one of the battery packs had taken what appeared to be only a slight bit of external damage to the blue wrapping layer, as the side panel had come open in a few places. Turns out, when I go to charge the wheel now, it only charges to 90% and then the charger (ewheels fast charger) turns on and then off in a cycle that lasts around 2 seconds. My theory is that this is because the BMS keeps continuously shutting off and then turning back on as it tries to protect overcharging a bad cell. I haven’t opened battery pack fully yet to inspect, as I need to use this wheel nearly every day, but I suspect I’m going to need to either replace a bad cell (low cost, lots of time to fix) or get a replacement pack (high cost, very little time to fix), and I want to know if anyone on here has had luck with replacing bad cells. I’m fine with paying for a replacement pack if I need to, but I’d rather try to save some money before doing that.

Also, @Marty Backe, replacing the motor didn’t fix my axle play completely. So either I must not be tightening the nuts hard enough (I can’t imagine what sort of force it necessary in that case!!) or something else is going on here. 

Have you tried charging the wheel with the stock charger, just to eliminate the charger as being the issue? Otherwise, that's unfortunate and you're in territory that's unexplored on my end.

There must be a mechanism in the axle bearings that can contribute to pedal sloppiness. I'm convinced now that that's the case with my Monster. Frustrating.

I'm glad that you're trying to keep the ACM2 - it's a classic now that you can't buy new ones 

[Nick McCutcheon]

1 minute ago, Marty Backe said:

Have you tried charging the wheel with the stock charger, just to eliminate the charger as being the issue? Otherwise, that's unfortunate and you're in territory that's unexplored on my end.

There must be a mechanism in the axle bearings that can contribute to pedal sloppiness. I'm convinced now that that's the case with my Monster. Frustrating.

I'm glad that you're trying to keep the ACM2 - it's a classic now that you can't buy new ones 

I did try the stock charger, it can’t charge the pack any more than the fast charger. I hope I’ll be able to replace the damaged cells. It’s still weird about my motor though, since the bearings are brand new as well. Your monster may have a bearing issue though, I’ve seen one other tesla owner with that problem. 

And of course! There’s no way I can give up on this wheel, it really is the best all-arounder like you’ve said. Pretty soon I’ll have enough spare/broken parts to build a whole new acm... 

[Marty Backe]

3 minutes ago, Nick McCutcheon said:

I did try the stock charger, it can’t charge the pack any more than the fast charger. I hope I’ll be able to replace the damaged cells. It’s still weird about my motor though, since the bearings are brand new as well. Your monster may have a bearing issue though, I’ve seen one other tesla owner with that problem. 

And of course! There’s no way I can give up on this wheel, it really is the best all-arounder like you’ve said. Pretty soon I’ll have enough spare/broken parts to build a whole new acm... 

Oh, you have the sloppiness with the new motor. Maybe you didn't tighten the nuts enough? Sure wish we had torque values for these nuts, otherwise you're afraid that you may break the axle if you crank on it too much.

[mrelwood]

2 hours ago, Nick McCutcheon said:

anyone on here has had luck with replacing bad cells.

A few of us have. I replaced a pair of dead cells and extended the 16S4P to 16S5P on my KS 16S, earlier a 16S2P to 16S3P on a Lhotz, and even rebuilt a whole block-shaped 16S4P battery pack from two flat 16S2P battery packs from two Lhotzes.

Replacing cells may indeed require a whole lot of work depending on where the dead cells are located, even for an experienced electronics repairer. It is also very stressful, as the cells and packs must be within 0.1V when connecting, and all the contact points visible it is way too easy to short a few. I was indeed able to be succesfull, but I have repaired dozens of guitar amps for paying customers, so I consider my experience level in soldering and electronic repairs pretty good.

Besides knowing how to make a trustworthy solder connection, the main issue in my mind is understanding the danger of messing up. Think about the amount of energy that is required to accelerate a 200lbs person to 50km/h incredibly fast several times, and still take him on a 50+ km trip. All that energy is right there in front of you to be messed up with.

[Nick McCutcheon]

1 minute ago, mrelwood said:

A few of us have. I replaced a pair of dead cells and extended the 16S4P to 16S5P on my 16S, and even rebuilt a whole block-shaped 16S4P battery pack from two flat 16S2P battery packs from two Lhotzes.

Replacing cells may indeed require a whole lot of work depending on where the dead cells are located, even for an experienced electronics repairer. It is also very stressful, as the cells and packs must be within 0.1V when connecting, and all the contact points visible it is way too easy to short a few. I was indeed able to be succesfull, but I have repaired dozens of guitar amps for paying customers, so I consider my experience level in soldering and electronic repairs pretty good.

Besides knowing how to make a trustworthy solder connection, the main issue in my mind is understanding the danger of messing up. Think about the amount of energy that is required to accelerate a 200lbs person to 50km/h incredibly fast several times, and still take him on a 50+ km trip. All that energy is right there in front of you to be messed up with.

It's good to hear at least this is doable. Thanks for the advice, especially about the 0.1V difference. Do you know how to discharge cells in case they come fully charged? Oh wait, I just remembered I have a headlamp that takes 18650's, I can just leave that on for a couple of hours  I bought a spot welder to avoid the pain that comes with trying to solder strips to the cells, so hopefully that will make things a bit easier. 

And yes, I am quite scared, but I will take all steps possible to make sure I don't accidentally short anything. 

[stephenbadger]

 mmm, new cells leave the factory 40% charged-up (storage charge) to minimise degradation ...my 2 cells worth, hehe!

[mrelwood]

A flashlight is good for individual cells, and I suspect you have an individual 18650 charger as well in case you empty them too far.

I have never used a spot welder, since I purchased my cells with spot welded soldering tabs.

A slight fear throughout the process is what will make you double-check and double-think everything you do. To me it sounds that you are well prepared and capable for the job.

One thing though is choosing the correct replacements. The worn old ones will make the new cells work a bit harder. I chose Samsung INR 30Q, which is what a more knowledgeable electronicist used for the 16S3P extension.

[Nick McCutcheon]

11 minutes ago, mrelwood said:

A flashlight is good for individual cells, and I suspect you have an individual 18650 charger as well in case you empty them too far.

I have never used a spot welder, since I purchased my cells with spot welded soldering tabs.

A slight fear throughout the process is what will make you double-check and double-think everything you do. To me it sounds that you are well prepared and capable for the job.

One thing though is choosing the correct replacements. The worn old ones will make the new cells work a bit harder. I chose Samsung INR 30Q, which is what a more knowledgeable electronicist used for the 16S3P extension.

Yep, I can just use the light charger (it takes two cells, so it’s perfect for this). Thanks for the reassurance. I already bought some of the same NCR18650PF cells that the pack is made of, would it be unsafe to use them? I may still be able to cancel my order.

[mrelwood]

20 minutes ago, Nick McCutcheon said:

I already bought some of the same NCR18650PF cells that the pack is made of, would it be unsafe to use them? I may still be able to cancel my order.

I honestly don’t know, but I would expect the originals to be a good choice even if some other cells might fit the situation even better.

[Chriull]

6 hours ago, Nick McCutcheon said:

Turns out, when I go to charge the wheel now, it only charges to 90% and then the charger (ewheels fast charger) turns on and then off in a cycle that lasts around 2 seconds. My theory is that this is because the BMS keeps continuously shutting off and then turning back on as it tries to protect overcharging a bad cell.

No. 2 seconds is too fast for this "cycle"

The hysteresis for the cell overvoltage protection should result in charger switching on/off somewhere in the range of weeks/month and not seconds...

6 hours ago, Nick McCutcheon said:

Update:

I took the casing off of the battery. Two of the cells at the edge of the pack are clearly dented, and some of the links that connect the batteries to the BMS have torn and are barely still attached.

Dented cells have to be replaced - good decision!

Unfortionately your BMS pcb got physicly violated  So some loose connection could be an explanation for the 2 second on/off cycle of the charger. Induced by some "thermal movement"?

In any case you should check your BMS PCB very very thorougly a couple of times with a magnifying glass for broken components/copper tracks!

Or making a good photo and magnify it on the monitor works great, too.

6 hours ago, Nick McCutcheon said:

I’ve deemed this pack unsafe, just to be sure. So for now I’m going to have a 650Wh ACM... It looks pretty simple, just desolder and remove the bad cells, and replace/resolder the new ones, right?

Update Update:

I bought 3 replacement cells (I only need two, but it's good to have an extra) and a ~$100 spot welder, so I'm committed at this point.

Li ion cells are not to be soldered, but spot welded. Try first on some old cells to get good and solid connections with your welder!

There was a link here once regarding battery pack building (the one with the fuse wires), were the author stated that soldering on one end of the cell is normally ok (because the tap is a bit seperated from the internals), but one side is in direct "contact" with the electrolyt and the isolating layers. So heating this end up with a soldering iron has to destruct the cell (at least a little bit).

In this link was also strongly recommended to put some rubber spacers between the shoulder of the cell and the nickel strips. At one side (?anode?) there is just this thin plastic cover/envelope preventing the shortening of the cell.

4 hours ago, Nick McCutcheon said:

 I already bought some of the same NCR18650PF cells that the pack is made of, would it be unsafe to use them? I may still be able to cancel my order.

Afaik the only recommendations i found for building battery packs is to use as good matched cells as possible - so the same model should be better than some other cell?

You could change your order to cells which already have the nickel strip welded! They should get better quality conections at the manufacturers plant than one could manage with a hobby spot welder? (Afaik one can choose between T,Z and U style strips - strips to the left on top, right at the bottom (Z), the same side (U) and to both sides (T)).

PS.: Since you are already messing around with your battery pack, you could think of putting wires from all battery connections to some outside connector! This will be very handy some time in the future to identify possibly degraded/aged cells!

[esaj]

On 4/22/2019 at 7:06 AM, Nick McCutcheon said:

It's good to hear at least this is doable. Thanks for the advice, especially about the 0.1V difference. Do you know how to discharge cells in case they come fully charged? Oh wait, I just remembered I have a headlamp that takes 18650's, I can just leave that on for a couple of hours  

The cheap ways to handle this I can think of:

-For discharging a single cell, get a 5-10W power resistor, like 3-4ohm (the maximum voltage of the cell is about 4.2V if fully charged) and just place it in series with the cell. It'll draw around 1-1.5A (depending on the cell voltage), just watch the voltage with a multimeter and remember that it "bounces back" a bit once the load is removed (the cell has internal resistance which will drop the voltage you measure during discharging, plus the voltage "recuperates" a bit over some time after removing the load). The resistor may need a separate (passive) heatsink; the typical "metal-cased" power resistors can handle maybe that 5-10W without a separate heatsink, but many sellers in places like eBay and Aliexpress may sell the resistors with totally nonsensical values like "50W" or "100W". If the resistor is less than the length of your finger and claims something like 50 or 100W of power, it may be able to take that, if screwed to a separate heatsink with thermal paste, but not on its own at least for a longer while  

-Place bleed resistors between the cells to be paralleled until they reach the same voltage:

With small differences (a few tenths of volt), you can "bleed balance" the cells to be paralleled through resistors, ie. put a "suitable" resistor between the cells in parallel:

The smaller the resistors, the more current they allow (so the faster the cells even out), but do check that the resistors can handle the power dissipation. In the above case, R1 has only about 2.1mW, which is nothing, R2 has about 8.5mW of power dissipation, so could use even smaller valued resistors (2.2 ohm or 1ohm). The currents can get quite complex to calculate, so use a simulator (like LTSpice in the above picture) if need be, but if it's just something like a few tenths of a volt, probably nothing bad can happen (from a current / resistor power dissipation point of view, of course be careful not to short the cells or such, the entire outer casing is the negative pole, also around the bulging top of the positive end )

A typical "normal" through-hole resistor is rated for 250mW (0.25W), so you don't need power resistors necessarily (won't hurt though, then they could handle even larger differences). Once the cells have evened out, you should be able connect them in parallel without the resistors and without any issues.

 

A better way would be to use a programmable electronic load that can be programmed to discharge the cell (with diminishing current) until a certain voltage threshold, but for a single pack/wheel battery repair this is overkill, as a good load will cost several hundreds. A cheapo (~$20-30?) chinese load could also work for small scale operation.

 

[GQS]

I've seen cell pack repairs done with solder (to existing welded tabs) instead of spot welding.  I believe spot welding the tab to the cell is to resist movement tears, but you can use slightly longer copper braid to solder two welded tabs together as long as you secure them. 

[Marty Backe]

2 hours ago, GQS said:

I've seen cell pack repairs done with solder (to existing welded tabs) instead of spot welding.  I believe spot welding the tab to the cell is to resist movement tears, but you can use slightly longer copper braid to solder two welded tabs together as long as you secure them. 

I soldered wires to the cell tabs so that I could connect two battery packs together. Felt like a very secure connection.

[esaj]

Soldering connections to cells has two problems in my opinion:

-You need quite a lot of heat to make a good connection

Lithium cells don't like heat, high temperature will hasten the permanent degradation of the cell, but likely not that much in a soldering operation, as it doesn't last long. The bigger risk in soldering is causing a thermal runaway-reaction in the cell, if the temperature raises enough (something like 125-150C / 260-300F), the cell may vent with flames, or in the worst case explode. With any luck, the protections in the cell will just cut the connection/release the pressure, but the cell will still be ruined.

-The connection has to withstand a lot of stress over time

Strong flux and enough heating on both the nickel strip/wire and the cell terminal will create a connection that's not just "solder sitting on top of metal" but actually causes a mixture (alloy) of the base metal(s) and the solder to be formed:

If insufficient heat/not "aggressive enough" flux is used, the connection will not be as strong. For lead-based solders, you need a minimum of around 200C (eutectic lead/tin-solder has a melting point somewhere around 173C if I recall correctly) on the metals to be connected, which is bad news for the lithium cell. For RoHS-compliant non-leaded solders, you need something like 250C. Higher heat and large tip on the soldering iron itself helps to heat up only the immediate area of the solder joint, if you use too low heat or small tip with not much surface-area touching the area to be soldered, you'll have to hold it much longer, and the heat will spread further than just the area to be soldered.

Even with well-done solder joint and no thermal problems (yeah, I call it a "problem" if the cell sprouts out a half-a-meter flame? ) with the cell, it's hard to tell how well it will handle all the vibration and shocks over time, since you can't see inside the joint. Electronic components are soldered in car and heavy-machinery ECUs and such, but they weight next to nothing (<1g),  an 18650-cell weighs something like 45-50g, if the cell has any "play" inside the pack, it could wiggle around just a little bit (like 0.1mm up and down or similar while riding, something you can't really tell by eye), and over time the solder joint might break (possibly while riding), especially if the soldered connection hasn't been that good to start with. Naturally hard crashes are another case where the connection gets stressed a lot.

Still, it could very well work just fine. But personally I trust welded strips to hold much better, they actually melt the two metals together on a very small area, and the heat stress on the cell is minimal because the welding spike is so short (milliseconds). A good middle-ground might be the cells that have pre-welded short nickel strips on them, you can solder a wire to the end of the strip with far less heat conducting to the cell itself, so no need to be that careful with the temperatures there.

[Chriull]

21 hours ago, GQS said:

I've seen cell pack repairs done with solder (to existing welded tabs) instead of spot welding.  I believe spot welding the tab to the cell is to resist movement tears, but you can use slightly longer copper braid to solder two welded tabs together as long as you secure them. 

 

16 minutes ago, esaj said:

...solder joint...

Still, it could very well work just fine. 

That was my opinion, too until i read the link https://www.electricbike.com/introduction-battery-design-2/  @EUCMania shared. There is written:

"Which brings us to the negative end of the 18650 cell. There is not much inside the negative end to protect the jelly roll from being hurt by heat (drain an old cell down to zero and cut it open for yourself, don’t trust what anyone says, even me). I simply cannot recommend that anyone solder anything onto the negative end. If you know someone who has done this and their house has not burned down, good for you. I STILL don’t recommend it."

So i changed my mind and imo soldering a 18650 does not work out without degradation and as you wrote risk of thermal runaway...

[GQS]

3 hours ago, Chriull said:

So i changed my mind and imo soldering a 18650 does not work out without degradation and as you wrote risk of thermal runaway...

Discharge the cell to less than 30% power, risk of thermal runaway resulting in outgassing and fire is prevented with low cell charge. 

[mrelwood]

Soldering directly to the cell would surely be hazardous. But a thin nickel strip will not heat the cell very much if the soldering takes place 1-2cm away from the battery. Solder sticks to the strips quite nicely, so the process is quite fast with a proper soldering iron. The only challenge I think was to get the two pre-soldered strips soldered together well, but I found a technique to heat both solders to melting point while adding flux. I'm pretty confident that my solderings will hold.

 

[Nick McCutcheon]

I just wanted to update everyone, I seem to have repaired the battery! I made sure each pack was at the same voltage before plugging them into the wheel, and the wheel charges up fully. I haven’t done extensive riding tests yet, but if something goes wrong (which hopefully should not happen), I’ll be posting it here. Thanks for all the help everyone, I really appreciate it