[WARNING] Gotway Nikola 100v 1800WH Battery Fire

[alcatraz]

Opened them both up and probed them. 

I had plans before the arrival of the 2nd wheel to rebuild one of the packs as a learning experiment. Go from 40 to 60 cells.

[Planemo]

15 hours ago, travsformation said:

After that, I usually charged to 100% until I got the fast charger, and charged to 90% to extend battery lifespan, charging to 100% maybe every 3rd or 4th charge. But from what I've read, the benefits pf charging to 90% are negligible at best and not charging to 100% comes with risks (packs not being balanced) that I get the impression, outweigh the benefits.

I'm not answering for Chriull but IMO, yes the above is correct. Nothing wrong in going to 80 or 90% every now and then, but far from ideal long term.

15 hours ago, travsformation said:

What difference is there (if any) in terms of safety / battery longevity between charging with a selectable-current fast charger set to 3A, vs. using two OEM 1.5A chargers? The OEM charger don't switch to CV do they?

No difference between using 2x 1.5A or 1x 3A. Just because a charger may use higher amps in the CC phase doesn't mean the algorithms during the CC/CV transition, nor the CV phase itself are any different. They are not. OEM do also use CV AFAIK.

For info, the ONLY time I go to less than 100% (full balance charge) is 1. if I am charging on the fly during a ride, where the CV phase is disproportionately long compared to the CC phase (and the CV provides only the last 10% or less anyway) or 2. If my wheel has been left at say 40% and I decide I want to go for a ride but dont need 100% (this is another benefit of having a larger battery) so I might throw in an hour or so using a 6A charger which will give me around 70%. A 6A charger also means I have never needed to charge overnight. I don't use my wheel for commuting and I am never in a situation where I ride as soon as I get up. Given the choice, I would rather not charge the wheel whilst unattended or whilst the family is asleep. Maybe I am being overly cautious here but I am never forgetting just how much fire potential is in our huge batteries so I have made the personal choice not to do it as a rule, although I have done once or twice when camping IIRC, given the stock charger (the only one I had at the time) on a Z10 just takes so goddam long.

 

[Planemo]

15 hours ago, Lefteris said:

Well charging over a smart plug is a smart idea to say, you need 3 hours and 45 minutes to go from 30% to 100%, then tell the plug to switch off the power in 3 hours and 30 minutes, what would 15 minutes of power loss is to a wheel.

That is what i do with my Nikola Plus 100V with 21700 LG cells, also measure Watts and time taken etc.

The idea of a smart plug is valid for sure, but until you have accurately plotted the amps/voltage of the wheel/charger combo you are using, saying '15 minutes' is a bit of a shot in the dark. To use such small time scales you need to know to within 10 mins or less exactly what SOC the battery is in. For example, the last 15 mins in my combo is of such little difference to the previous 15 mins that its practically irrelevant. There IS a difference but it's so minute that I don't think it would make a great deal of difference in terms of battery longevity if thats your aim. Without being able ATM to refer to a previous plot I did, I think you would need to cut charging at least 45mins before full saturation if your aim is to avoid it.

And using any app to monitor voltage throws the figures out even more. From memory I think the wheel shows 100% at anything around 99v, which could be before the CV phase has even started. Probing the output of the charger for voltage and amps with a decent multimeter every 10~20 minutes from say 40% to fully saturated is the only way to build a decent enough plot to see whats actually happening and give enough data from which to make your decision on when to switch off the smart plug.

[travsformation]

9 hours ago, Planemo said:

I'm not answering for Chriull but IMO, yes the above is correct. Nothing wrong in going to 80 or 90% every now and then, but far from ideal long term.

Cool, thanks! 

9 hours ago, Planemo said:

No difference between using 2x 1.5A or 1x 3A. Just because a charger may use higher amps in the CC phase doesn't mean the algorithms during the CC/CV transition, nor the CV phase itself are any different. They are not. OEM do also use CV AFAIK.

I figured it was probably the case, but reckoned I was better off asking. Cheers!

10 hours ago, Planemo said:

For info, the ONLY time I go to less than 100% (full balance charge) is 1. if I am charging on the fly during a ride, where the CV phase is disproportionately long compared to the CC phase (and the CV provides only the last 10% or less anyway) or 2. If my wheel has been left at say 40% and I decide I want to go for a ride but dont need 100% (this is another benefit of having a larger battery) so I might throw in an hour or so using a 6A charger which will give me around 70%. A 6A charger also means I have never needed to charge overnight. I don't use my wheel for commuting and I am never in a situation where I ride as soon as I get up. Given the choice, I would rather not charge the wheel whilst unattended or whilst the family is asleep. Maybe I am being overly cautious here but I am never forgetting just how much fire potential is in our huge batteries so I have made the personal choice not to do it as a rule, although I have done once or twice when camping IIRC, given the stock charger (the only one I had at the time) on a Z10 just takes so goddam long.

Yeah, those scenarios are pretty close to mine too...the great thing about having wheels with this kind of range 

I think I've only charged it while sleeping once too, because it was completely empty and I needed it for first thing the next morning. Something I stay clear of if I can though...As you said, the consequences of an EUC fire are nasty, and even a small risk if better to avert! 

[AtlasP]

Sorry I only skimmed this thread/haven't read everything with a fine-tooth comb, so forgive me if this first question has already been asked/answered:

1. Are the only cells in question all 21700s? I have one of the later batches of 18650-based 100v Nik+'s (~Nov 2019 via ewheels)--is there any chance this is affected?
2. My 100v Nik would only get up to about ~99.1/98.9v (the reading on the side of the wheel) after a charge when I first got it, and now after ~750-1k miles the highest it will read after a charge is 98.6v. (This is all using the stock charger/no % cutoff.) Is this normal/common--possibly due to inaccuracy of built-in voltage measuring used for the display, or a natural progression of aging, etc--or is it evidence that I have a bad cell or cells that could be a problem?

Edited June 12, 2020 by AtlasP

[Michael Tucker]

33 minutes ago, AtlasP said:

Sorry I only skimmed this thread/haven't read everything with a fine-tooth comb, so forgive me if this first question has already been asked/answered:

There is a fire and board burn/occolation issue being discovered right now involving Nikolas, MSPs and Monster V3s. One possible connection between the problems is "used" or "recovered" 21700 cells from Panasonic/Tesla. 

[Chriull]

14 hours ago, AtlasP said:

My 100v Nik would only get up to about ~99.1/98.9v (the reading on the side of the wheel) after a charge when I first got it, and now after ~750-1k miles the highest it will read after a charge is 98.6v. (This is all using the stock charger/no % cutoff.) Is this normal/common--possibly due to inaccuracy of built-in voltage measuring used for the display, or a natural progression of aging, etc--or is it evidence that I have a bad cell or cells that could be a problem?

It could be

- the charger delivering too low voltage.

- the display showing too low voltage. I'd assume the app reports the same voltage?

- charger voltage could get more misadjusted over time.

- it could/be bad/degraded/misbalanced cells.

To find the exact reason (if something is really "off" or just the display beeing off) one would need to check a point after another...

If you charge your nik fully (2-3 hours after the light went green) does the voltage drop within the next hours/until the next day and then stays more or less constant?

 

[RagingGrandpa]

On 6/12/2020 at 4:56 PM, Michael Tucker said:

There is a fire and board burn/occolation issue being discovered right now involving Nikolas, MSPs and Monster V3s. One possible connection between the problems is "used" or "recovered" 21700 cells from Panasonic/Tesla. 

Battery packs have almost nothing to do with board issues. Gotway packs do not communicate with the ESC at all!
It's a testament to the cell quality that among multiple board failures involving extreme current discharges, the cells did not overheat.

On 6/3/2020 at 7:54 AM, WI_Hedgehog said:

It is unethical to keep spreading the Tesla car battery rumor without evidence.

 

Edited June 15, 2020 by RagingGrandpa

[davinche]

I posted some thoughts/questions I still have on one of @Mike Sacristan videos, but I thought I would post it here as well.

Quote

The question I still have is... if the Panasonic 21700s are still available, why did Gotway then switch to LGM50Ts for the most recent Nikola/MSP? Maybe my Nikola is an outlier (lucky me), but I've never crashed my Nikola and it is so far (the only reported) Nikola that have caught on fire. The whole thing is just puzzling to me.

So let's just think a little bit. These wheel resellers (loomo, wheel rider etc). Do you think they're more likely to put in better batteries for their mods or worst batteries? Then based on that position, we know they chose Panasonic 21700 cells for their mods. What does it all mean then?

So let's say hypothetically margins for these wheels are low: then these 3rd party sellers should be incentivized to use cheaper cells right? Which would mean that Panasonic cells are cheaper than LG cells. This then would mean that Gotway must have found some problems with these cells (but have decided not disclose what the problem is) and silently switched to LG cells.

Another scenario: let's say that because modded wheels are high risk, which means these 3rd party sellers are incentivized to minimize risk and use better quality cells. They choose Panasonic cells because it's "better" than LG cells. If that is the case, why did Gotway switch to LG cells? Is it because it's cheaper / lesser quality?

There are a million scenarios, but it still begs the question: Why did Gotway switch to LGM50Ts when Panasonic cells are still available? And why are these resellers using Panasonic cells instead of LGM50T (like what Gotway is using in their wheels now)?

 

Quote

And of course additionally: Let's say both Panasonic 21700s and LGM50Ts are comparable/interchangable. Then is it coincidence that all of the new wheels use LGM50Ts (V11, S18, 2020 Nikola/MSP)? And what does that say about the Panasonic 21700s if resellers/modders are using them instead of the LGM50Ts?

 

Edited June 15, 2020 by davinche

[alcatraz]

I'm guessing malfunctioning bms.

Or a bms with weak balancing currents running a capacity of cells way over what its rated for. Running LG cells could help only by being better matched from the factory. After they age they will be in the same boat as the Tesla/Panasonic cells, and balancing currents will be inadequate aswell.

Edited June 16, 2020 by alcatraz

[Cody]

The Cells inside a battery are not to blame. most are Exceeding standards (some allow to be charged up to 12v without bursting in flames)

- The battery PACKS manufacture is the problem-

[Aibolit66]

Yes, the problem is not in the 21700 cells, but in the location of the controller relative to the battery.  I have been repairing EUC since 2016 and can analyze it.

[Chriull]

55 minutes ago, Aibolit66 said:

Yes, the problem is not in the 21700 cells, but in the location of the controller relative to the battery.  I have been repairing EUC since 2016 and can analyze it.

You were able to identify the hot part? Wire, mosfets or even capacitor?

[alcatraz]

This is interesting.

Other gotway wheels with the same battery + control board configuration and spacing don't exzibit overheating issues. 

Why these wheels? What's different between the affected wheel models and unaffected models?

[Planemo]

20 hours ago, Aibolit66 said:

Yes, the problem is not in the 21700 cells, but in the location of the controller relative to the battery.  I have been repairing EUC since 2016 and can analyze it.

I'm not following your logic - you are saying that the batteries are giving off enough heat to assist in damaging the capacitors? I can't see that happening so I must be missing what you are trying to say.

[Chriull]

37 minutes ago, Planemo said:

I'm not following your logic - you are saying that the batteries are giving off enough heat to assist in damaging the capacitors? I can't see that happening so I must be missing what you are trying to say.

I've understood it the other way around - the motor wires/mosfets (or whatever in this hot area) "assist" the 21700 catching fire...

[Planemo]

Possibly, but I can't see that happening either. Theres a fairly stout piece of plastic to support the battery and if the cap was hot enough to burn through it, the cap and most of the board would already be toast.

I'm not saying the cap didn't do this of course, but I don't see how proximity is the culprit.

[Aibolit66]

On 6/24/2020 at 5:28 PM, Planemo said:

Possibly, but I can't see that happening either. Theres a fairly stout piece of plastic to support the battery and if the cap was hot enough to burn through it, the cap and most of the board would already be toast.

I'm not saying the cap didn't do this of course, but I don't see how proximity is the culprit.

Short circuit current creates an electric arc.  It looks like a welding machine.  Can plastic stand it?

[Aibolit66]

Here is an example of a Tesla, but battery cells did not catch fire.

[Aibolit66]

The choice is yours.  1. Continue to blame Panasonic elements.  2. Put a powerful fuse.  3. Coat the controller with a thermally conductive silicone compound.

[Chriull]

On 6/24/2020 at 4:28 PM, Planemo said:

I'm not saying the cap didn't do this of course, but I don't see how proximity is the culprit.

As @Aibolit66 tells the motor wires seem to be a "nice start" as culprit.

The melt the insulation - if they "touch" the capacitor they are heating him up directly and furthermore this creates a short between the potential of the wire and capacitors outside shell.

Like spotwelding the capacitor - not really recommendable.

Maybe this is possible too with the battery pack - melting through the piece plastic. Or the battery is ignited by the burning capacitor?

3 minutes ago, Aibolit66 said:

Put a powerful fuse

Plus short circuit protection in the BMS and some usefull current limit in firmware!

[AtlasP]

6 hours ago, Aibolit66 said:

The choice is yours.  1. Continue to blame Panasonic elements.  2. Put a powerful fuse.  3. Coat the controller with a thermally conductive silicone compound.

 It's like the Gotway glue-gun guy's wet dream...  

[Planemo]

8 hours ago, Aibolit66 said:

Here is an example of a Tesla, but battery cells did not catch fire.

Those pictures directly relates to an issue I have come across with an overheated capacitor, hence the reason I started a recent thread about it. I am still awaiting some info, but this further evidence (your pic) of overheating in the same area of the capacitor seems more than just coincidence. In my example, nothing shows any overheat damage except the cap, hence me pointing fingers at it. Its not my wheel, so I am waiting on further info.

[RagingGrandpa]

On 6/27/2020 at 9:49 AM, Aibolit66 said:

The choice is yours.  
1. Continue to blame Panasonic <cells>.  
2. Put a powerful fuse.  
3. Coat the controller with a thermally conductive silicone compound.

I missed your point...

Let's presume the failure mode that you asserted: 
"Short-circuit fault of a component of the control board (I don't mind if we assume it was a FET, a cap, or other components...), propagating to pack over-temperature and cell runaway."

How do you think potting the board prevents that?

Edited June 29, 2020 by RagingGrandpa

[Chriull]

1 minute ago, RagingGrandpa said:

How do you think potting the board prevents that?

The coating of the board i can't imagine as a solution either.

But the point with the motor wires - a long time desaster with gw wheels disturbed me long time. Especially with them beeing near to li ion cells/etc.

I don't really thrust in the big capacitors - i have the gut feeling they are all used out of spec (not only gw). But i'll need a bit more research do be really concerned about this...

About misusing mosfets as fuse i don't really want to.loose any words