yes, you are not meant to plug in the balance wires via the connector until the rest of the connections have been soldered onto the BMS.
1 Like
Thank you DavidBanner.
yes, you are not meant to plug in the balance wires via the connector until the rest of the connections have been soldered onto the BMS.
Oh dear. I didn’t get that instruction.
Never have power on electronics when soldering if you can avoid it.
Definitely damaged, maybe beyond repair. It’ll have to be an extensive repair to make it trustworthy but it can be done.
I would suggest get them pulled and re heatshrunk
Thank you PXSS. How can you be certain? Is there a way to measure your conclusion?
Don’t know how bestech bms’es are designed, but on a supower bms balance wires should never be connected before battery + and - are connected or bms will fry.
1 Like
It’s my full time job. I design and build batteries for military UAVs. I’ve shorted plenty of batteries both on purpose and by accident and have honestly never seen batteries get hot enough to damage heatshrink in the manner yours is damaged without catching fire. Its either, they shorted out for a few seconds and are okay or they shorted out and burned.
Yes, there is a way but it’s not cheap. You need an RC charger that can measure IR. I use an iCharger 4010 Duo but that is definitely overkill for what you need. You can measure IR, capacity and balance the pack with it. Then you can compare that data to the manufacturer specs and see if you’re within the limits.
There is a way to measure IR manually but you need a good multimeter. Again, I use a Fluke 87V but that is overkill for you.
E: All the heatshrink damage seems to be on the negative terminals. Not one cell from what i can see on those pictures is damaged on the positive end. Did you damage these before hand or cut the heatshrink for any specific reason? It almost looks deliberate, it’s a little confusing.
1 Like
I think the most glaring problem here is that the positive wire is NEVER (in my experience) connected directly to the BMS, which what you’ve done here…
4 Likes
I would use a standard 80 amp fuse on the xt90 out and Tesla fuses between the balance leads 
for maximum protection 
Next time, solder on the balance leads unplugged and double check each grove with a volt meter.
Yep! I missed that edit. That would be a direct short across the battery. But the FET diodes should have protected the battery. Something is still wrong with the BMS, I’ll look at the spec sheet.
All the heatshrink damage seems to be on the negative terminals. Not one cell from what i can see on those pictures is damaged on the positive end.
Yes, I now suspect the damage came from the C- port. I assume the current flowing via the main neg lead damaged the subsequent neg terminals.
Negative.
I thought I did everything correct, up until:
- Plugging the balance wires in
- Connecting the main leads in the wrong order
- Connecting the main pos lead (disastrous)
This is the battery pack pre-soldering the BMS:
- Cardboard rings on pos terminals
- Each cell shrinkwrapped (clear)
- Hot glued together
- Nickel strips soldered to every terminal
Thank you Sebike.
Yes, it seems they’re the same in that regard. I don’t know how I missed that instruction. My BMS definitely fried, and would have burnt down the table, if not the house.
Thank you darkkevind.
It seems I connected the pos lead to the BMS by mistake. I’m devastated at the outcome. 
Still uncertain how it all happened exactly though, because:
- Not everything was plugged in (e-switch disconnected)
- P- disconnected
- Most importantly, while the main neg lead was connected to B- and the main pos lead was then mistakenly connected to C- it did not short immediately. Had it done so, it would have been obvious there was a direct short taking place. That is, I soldered it a couple of times, maybe even a few times, adjusting the angle of the wire slightly and adding more solder for a shinier solder joint.
[EDIT] I quote @Namasaki back referring to the same BMS model.
My case is similar to his, but without load connected, and without the charger connected. Even though I directly connected the main pos lead to the C- port of the BMS, the e-switch was off, and even if it switched on (somehow), the BMS really should have done the same thing it did for Namasaki (shut down immediately and protect batteries).
@PXSS I think you’ve identified a core problem, within the problem. Clearly, I’ve made errors, but it doesn’t make complete sense. In comparison to Namsaki’s case, my outcome seems far too drastic and inexplicable.
Yes, even if the main pos lead was connected to the C-, the BMS should have protected the battery. Right?
Are the FET diodes the black mosfet panels along the top of the BMS? If so, I recall them slowly erupting in smoke and igniting in flames _sequentially, from left to right.
I recall putting out the first lot of flames before dragging the board outside onto the driveway (that’s how the battery drooped out of shape, while hot glue had melted) and minutes later the mosfet panels / FET diodes smoked again and ignited into flames again.
See pic 3, on the balance lead side. The mosfet panels / FET diodes are more burnt than the other side.
Also, notice the B- port. The B- port / the main neg lead melted & popped off the BMS tearing some PCB. Something definitely went wrong on there, while the C- port / red main pos lead remain unscathed (apart from melted hot glue).
Perhaps this follows the logic that current flowed through the main pos lead > the C- port > the mosfet panels / FET diodes (and burning them) > throughout the battery (and burning some neg terminals) > ultimately burning the B- port to the point of disconnection. 
Thank you ZackoryCramer.
Yes, some have focused on the balance lead issue (being plugged in whilst soldering - clearly an error it seems), but I don’t understand how that caused malfunction. There was no heat, no spark, etc, coming from them whilst connected, and post-short they were unscathed. No melted glue. Nothing.
I would use a standard 80 amp fuse on the xt90 out
Interesting suggestion. Please elaborate.
That sounds fine, but I wonder is that not excessive to the BMS’s job, and the XT90’s job? And the 80a fuse would be to protect the anti-spark switch / VESC more than the battery, right? Which would also be a bit redundant on the anti-spark switch, right?
I quote Bestech’s BMS datasheet. See 80a max discharge current:
and Tesla fuses between the balance leads
What rating are the fuses? So,your suggestion would mean running fuses inline per balance wire, right? i.e. 8 fuses, per 8 balance wires. Again, that would seem excessive and redundant since the BMS itself should be regulating. I’d have thought that’s what the BMS is for.
Again, I quote Bestech’s BMS datasheet. See overcharge protection, over discharge protection, and over current protection:
@PSXX How devastating to read… 
But, yes, while I’m grateful for your input and I don’t doubt your qualification, but what you’ve written is not a reason for your conclusion. Can you perhaps elaborate and explain how heatshrink damage equals battery damage, and give examples of cases you’ve experienced?
Because as you’ve correctly noticed, the battery did not catch fire, and it’s still registering 25.7v (3.21v per cell), which looks fine to me? Very optimistic.
At 3.21v per cell (and no less), the BMS seems to have actually done it’s job. As mentioned earlier, it went from 26.2v (3.28v per cell) to 25.7v (3.21v per cell) during the short.
I quote the Bestech BMS Datasheet. See over discharge protection / over discharge detection voltage at “3.00V±0.05V”.
I hope you have some more optimistic news. Because in the condition that it is right now, I’d honestly like to repair it, plug it in and run it again as per normal. But I wonder what would happen? What are the pros and cons of doing so?
Giving trust to the battery pack(cells with strips and wires holding it together), the only way for a battery to malfunction is physical damage or unwanted load through the cells. The best way to do this is to use protected cells but they are a quater or more expensive and only available prebuilt in Panasonic NCR18650B. Or we can fuse all outputs which would be the BMS balance leads and outputs. Using a fuse like this one:
in series before either + or - terminals, we can prevent current over draw physically(the current bridge is burnt). We can further improve unwanted output in the BMS balance leads through 26 awg wires which starts to melt down after 3~4 amps(compared to the max balancing 200ma current draw). Yes they aren’t really Tesla fuses, I mean to say Tesla Style. I learned it here:
[UPDATE]
Balance wires via a voltmeter read:
- B1: ? – How do I measure this cell?
- B2: 3.28v
- B3: 6.38v
- B4: 9.61v
- B5: 12.8v
- B6: 16.0v
- B7: 19.3v
- B8: 22.5v – Why is not 25.7v (as read between main pos and neg leads)?
Amended the battery pack & BMS wiring:
- Repaired desoldered nickel strips
- Clamped the whole pack back into shape while re-heating hot glue obstructions (because bulges of glue were everywhere)
- Hot glued over the exposed unheashrunk cells (no cells remain exposed for risk of short)
- Unplugged the BMS balance leads
- Desoldered the problematic main pos lead at the C- and spliced out new wires to the charge port. Then heatshrunk the fat solder joint before finishing the XT90 plug.
- Cleaned the C- port and re-tinned it ready for connection to the charge port negative wire (not pos which contributed to the short).
- Added new nickel strip layers to the neg terminal at B1, and relocated the main neg lead to the outside edge of the pack (not the inside).
The original nickel strips are too hard to remove (super securely soldered, and reinforced) so instead I made this guy with layers of nickel strip in the opposite direction for the correct reinforcement.
@PXSS had me concerned about the main neg lead being too close to the pos terminals at B3, so I redirected it own the pack with ample hot glue insulation.
Ideally, I’d have rebuilt the entire pack as @PXSS suggested but it’s way too much work when everything is mostly OK. I’m also low on heatshrink for the 198650 cells. And, desoldering and resoldering the nickel strips will apply what I think is unnecessary heat to pack. In my mind 25.7v looks fine.
Oh… I’m so regretful, everything was tidy and quite pristine before the short. None of it was meant to be this sloppy. Had to learn the hard way, as usual. 
Thank you ZackoryCramer.
80a fuse: Brilliant. Yes, quite right, I’ll hook up fuses asap. I prefer the blade fuse design though, for easy replaceability while out and about.
Tesla-style wire fuse: 26awg melts at3-4 amps does it? Fascinating. I wonder if a similar inline resistor will work the same. But, replaceability while out and about will be difficult and inconvenient.
1 Like
go for a maxi fuse. Blue Sea Systems 80A MAXI Fuse https://www.amazon.com/dp/B000K2ILMC/ref=cm_sw_r_cp_api_zwdIAbQ66BNZK
Another tip. Do not use hot glue. Everything inside your enclosure is at the very least going to get to 50-60c, hot glue melting point is at like 80-100c? Not exactly sure. Either way, at 60, hot glue is ductile which isn’t good for keeping things in place.
I use amazing goop if I want something that isn’t permanent as it doesn’t go ductile with high temps and can still be removed. If I want something permanent then I use Thick CA.
E:
You’re supposed to measure between main neg and the balance pins. I assume you were measuring relative to pin 1 in the connector. Also, you should measure in the following manner:
Main neg - pin 1 Pin1-pin2 Pin2-pin3 Pin3-pin4 Pin4-pin5 Pin5-pin6 Pin6-pin7 Pin7-pin8
All of them should measure 3.XX volts and should have a max spread of 0.1v to be balanced properly.
From what you posted above: Cell 1:? Cell 2: 3.28 Cell 3: 3.10 Cell 4: 3.23 Cell 5: 3.19 Cell 6: 3.20 Cell 7: 3.30 Cell 8: 3.20 So you have an imbalance of 0.2v which is up to 15% imbalance.
1 Like
Sorry, I missed that you mentioned me quite a while back! How did everything go?
I have tried to read through the thread and it seems you’ve got good answers. It would be beneficial to you gaining some basic electronic knowledge. Then you would be certain how to measure voltages, connect battery packs etc.
The BMS has an over charge detection voltage of “4.28V±0.025V” (4.255-4.305v). Thus, since my 8s4p pack was registering 26.2v earlier in the day, it would have sent ~3.275v (26.2/8=3.275) into the BMS / battery, which is much less than 4.255v. More precisely, 0.98v less.
You wrote this. How I understand this over charge detection voltage is that it must reach those 4.28V (±0.025V) before it stops charging. Which it didn’t. The over current threshold is 200A which I do not think your battery is able to provide, and then that threshold was never reached either.
1 Like