home made 18650 pack problem

[ic3wall]

Hi everybody,

I need opinions and advice from people with experience with pack building.

2 months ago, I built a 48V 7.8Ah pack from 18650 INR 13Q (milwaukee power tools). Last night, I was going for a short ebike ride and look what I found:




The pack is built as follow, 6 cells in parallel (4V) and 3 of these in parallel to get 12 volts, then 4 of the 12V in series to get 48V.

The failure happened at the junction of 2 6p strings. I dissembled the 12V string where it happened and checked voltage and internal resistance of each cell and didn't notice anything wrong.

There is a 45A fuse between the battery and the controller and the fuse was fine, so the pack "shorted itself"... The nickel tab acted as a fuse, which is a very good thing..

No idea what happened... ideas ?

[FastDemise]

The pack is built as follow, 6 cells in parallel (4V) and 3 of these in SERIES to get 12 volts, then 4 of the 12V in series to get 48V.

You should change it to this. Or to make it simple just 12S6P battery. My guess is that the metal bar touched the ground part of the battery and shorted. On the 18650 batteries the only thing positive is the small disc on top. If you were to remove even a little of the shrink wrap around the positive disc that's all the battery ground. Where the orange paper is on top. It's to keep you from accidentally contacting the positive with the ground creating a short. Seems you just created a short and your tab lost the fight. I would just replace the burnt battery with one that doesn't have th shrink wrap damaged or just add some tape and be on with life. Glad to see someone else with a 18650 battery pack.

[ic3wall]

The pack is built as follow, 6 cells in parallel (4V) and 3 of these in SERIES to get 12 volts, then 4 of the 12V in series to get 48V.

You should change it to this. Or to make it simple just 12S6P battery. My guess is that the metal bar touched the ground part of the battery and shorted. On the 18650 batteries the only thing positive is the small disc on top. If you were to remove even a little of the shrink wrap around the positive disc that's all the battery ground. Where the orange paper is on top. It's to keep you from accidentally contacting the positive with the ground creating a short. Seems you just created a short and your tab lost the fight. I would just replace the burnt battery with one that doesn't have th shrink wrap damaged or just add some tape and be on with life. Glad to see someone else with a 18650 battery pack.

yeah you're right, I meant "series".

ok, your explanation sounds good. Do you have a trick to prevent this? This is quite dangerous...

[FastDemise]

Just be good to inspect the battery before you assemble that there isn't anything damaged with the wrap. It appears you are able to weld the tabs on the battery so you shouldn't be damaging the battery with excessive heat to damage the wrap. When assembling keep the batteries pressed close together so you don't get all that excessive tab bending after welding which could press and rub. Just my opinion. I just collect the batteries preassembled in pairs and then just solder the pairs together. I sure wish I could weld them.

[amberwolf]

Another thing you can do is stuff thin plastic or similar non-conductive material between the edges of the cells and the tabs, on the positive ends of all the cells.

One problem I found when repairing my Vpower/CammyCC LiFePO4 pack is that when soldering tabs back onto the cells, even done quickly, it's possible to heat the shrink wrap enough on the edges to cause it to "retreat" or crack, or even flat-out melt, if the tab is close enough to it and the heat is applied in the "right" place. If it doesn't directly cause a short (probalby won't), it leaves that cell vulnerable to a short later, as the pack cells vibrate and shift around over time, especially if the tab is pressed against the end of the cell by whatever encloses the pack.

I suspect this is what kills some of the groups in various cylindrical packs.


In laptop packs I've opened up, there are often little O-rings of paper or plastic around that endcap, which would prevent this from happening.

[Kin]

I wonder if there would be any point is spraying on a few layers of thick gloss or something? If gloss is not enough, they have spray/paint on latex. It would be very annoying if you need to repair cells. But for the most part, if you want to measure voltages the paint could be poked through [I'd just keep balance leads]. Just wondering if anyone had a ye or maybe or NO to that idea.

[ElectroSurf3r]

I've seen this type of short on a number of 18650's, both my own and others' on the forum.
The problem is that the entire length of these cells are at the negative tab voltage, but there is usually only a thin layer of plastic shrink to protect them.
So, if an adjacent cell's tab makes it through the plastic, then it will short. Even the positive tab of one cell itself can cause a short with some downward force, wiggling, impacts, etc.

I prevent these shorts by simply wrapping some duct tape around the positive end of each cell. It's important to stuff one edge of the tape under the positive tab with a fingernail. This prevents the cell from shorting to itself, and the tape creates a nice buffer between cells so they don't short to each other.

[ElectroSurf3r]

haha AW beat me to the punch, and suggested this exact thing a couple posts above.
Hopefully my pic and info can provide some extra clarity.

For your pack, one wrap of tape around each 6p group will suffice. Just make sure to get it under the edge of all the positive tabs.

[999zip999]

It's easy if you heat of the tap and push down when soldering. Grounding the tap. It a weak area. You could use slime plastic between the pos. end. But that's a 180 cells. It shoots out a short flame as you hold it in your hand ( don't do this at home ), thru the vents. If you use a diamond cutting blade ( dremel ) the metal shaving can get in between. Like a battery and steel wool. Proof

[liveforphysics]

Those nickel tabs really don't conduct electricity worth a darn. If you're pulling more than 10amps, it's possible the strip got hot enough there (which would be a point of highest current density) to melt into the shrink wrap and hit the edge of the can.

[chroot]

That's why you did not use "O" foam paper alike or "O" thin plastic on top of positive prevent the copper tab reach the negative body. What's what my friend did made his 18650 and never had problem. I learned from him.

[ElectroSurf3r]

I tried making rings out of industrial wax paper and foam, and both were a PITA to cut, even for a small pack. Ended up settling on the duct tape idea, which has worked so far.
If someone makes pre-cut rings, then maybe, but duct tape still offers the advantage of a buffer between cells.

Luke- was that 10A quote on the nickel tabs intended for 1P strings? If so, that would explain my pack getting warm at WOT on constant uphills. I should throw a few more in parallel on my bike... Or just replace it with the self-balancing LiPo pack

[ic3wall]

Those nickel tabs really don't conduct electricity worth a darn. If you're pulling more than 10amps, it's possible the strip got hot enough there (which would be a point of highest current density) to melt into the shrink wrap and hit the edge of the can.

I don't believe that, 10 amps is not much.

People in this thread ( viewtopic.php?f=2&t=35897&p=523486&hilit=nickel+tab#p523486 ) are talking about 50-60 amps.

Also, the duct tape idea is not bad, but it requires to disassemble packs down to individuals cells... which takes forever, and then you have to re-solder everything.

[liveforphysics]

Those nickel tabs really don't conduct electricity worth a darn. If you're pulling more than 10amps, it's possible the strip got hot enough there (which would be a point of highest current density) to melt into the shrink wrap and hit the edge of the can.

I don't believe that, 10 amps is not much.

People in this thread ( http://www.endless-sphere.com/forums/vi ... ab#p523486 ) are talking about 50-60 amps.

Also, the duct tape idea is not bad, but it requires to disassemble packs down to individuals cells... which takes forever, and then you have to re-solder everything.

If you give me the thickness and width of your strips, I can calculate the resistance and heating per cm at 10amps if you like.
It might be much thicker than it looks and not be a big issue.

Copper's resistance is 1.68×10−8
Nickel's resistance is 6.99×10−8 (which is why it resistively welds so nicely.)

It has about 4x the resistive heating loss of copper with the same cross section. Or, you could look at it like, if your strip is say 0.2mmx10mm (2mm XSA), if it were in copper it would be equal to a 14awg copper wire. But in Nickel, it's right about exactly equal to a 20awg copper wire. (0.0333ohm/m)

[ElectroSurf3r]

I don't know, I bet the 10A quote is pretty accurate for a single tab.
The thread you linked is showing theoretical numbers, on tabs that are probably not the same size as yours. It's also calculating the maximum amps the tabs can handle, not the amps it takes to only heat and deform the plastic over time.
A simple fix for your pack would be to solder some copper strands across only the series connections. The parallel connections are not a worry, since they carry very low current.

[liveforphysics]

I don't know, I bet the 10A quote is pretty accurate for a single tab.
The thread you linked is showing theoretical numbers, on tabs that are probably not the same size as yours. It's also calculating the maximum amps the tabs can handle, not the amps it takes to only heat and deform the plastic over time.
A simple fix for your pack would be to solder some copper strands across only the series connections. The parallel connections are not a worry, since they carry very low current.

That would be a perfect fix.

Remember, hit it with a hotter than hell powerful and tinned iron, and scratch the nickel free of it's oxide layer (which is mild) before applying the iron. You need to be fast, the whole operation only lasting a few seconds, and this way you can keep the cell cool and undamaged.

[ic3wall]

A simple fix for your pack would be to solder some copper strands across only the series connections. The parallel connections are not a worry, since they carry very low current.

That's a very good idea, I could also spot weld 2 nickel tabs on top of each other on series connections. That would probably be less efficient than copper, but still it would help..

[ElectroSurf3r]

Oh and for the duct tape, you don't really need to break it down to individual cells. I think if you could break it down into your parallel strings of 6 cells, and just get under the 2 sides of the positive tab with duct tape, that should create enough space to eliminate most of the risk. Not to mention it will protect the groups of 6 from each other.

[pwbset]

On the 18650 batteries the only thing positive is the small disc on top. If you were to remove even a little of the shrink wrap around the positive disc that's all the battery ground.

This used to happen to me all the time when I worked with used konions. Even the tiniest little rip or defect in the green covering and ZAP! As others mentioned just adding a little duct tape or whatever before you tab weld can save headaches later.

How are those cells by the way? I just got an e-newsletter from Milwaukee this morning touting their M18 FUEL tools and Redlithium batteries. I used to use the 26650 emolis in the old V28s and they were pretty decent for ebike use.

[ic3wall]

On the 18650 batteries the only thing positive is the small disc on top. If you were to remove even a little of the shrink wrap around the positive disc that's all the battery ground.

This used to happen to me all the time when I worked with used konions. Even the tiniest little rip or defect in the green covering and ZAP! As others mentioned just adding a little duct tape or whatever before you tab weld can save headaches later.

How are those cells by the way? I just got an e-newsletter from Milwaukee this morning touting their M18 FUEL tools and Redlithium batteries. I used to use the 26650 emolis in the old V28s and they were pretty decent for ebike use.

they are pretty good, 1.3Ah, max discharge rate is 15C

only bad thing is that they charge at 0.5C...

[FastDemise]

So roughly a 2 hour charge time is bad? I got a bulk charger is only able to charge at .25C

Not sure I saw it but do you have any numbers on what your peak/average current draw is on your bike?

[ic3wall]

I don't have any numbers yet because I didn't have the chance to do any long rides... I finished my setup in early January and it's winter here. Now, I'm limiting current to 20A since Justin (ebike) told me that my motor (9C) would probably overheat over 1000W.