LG Chem -LG-E63 cell discussion

A cell voltage of 4.185v is 97.3% of the nominal rating of 4.30v, which kinda matches the 94% overall capacity difference from the pushevs site. I assume there's additional buffer at the bottom end, so it looks like maybe 3% top and 3% bottom? Might not be symmetric but that seems reasonable, and it fits with the ~3% margin from other sources.

Based on this post by Anaglypta, the BMS update increases the sensitivity for voltage deviation from 1.0V (!) to 0.1V:
https://www.speakev.com/threads/kona-ev-recall-in-korea.153828/#post-2900846

Which is alarming. If one cell is higher than the others, that large a detection window could easily allow one or more cells to exceed 4.45V "danger zone" limit... up to possibly 5.30V! (That can't be right, can it?)

With the new update and a window of 0.1V, then even at a nominal 4.30V the max excursion is 4.40V, just a hair shy of the max limit, and hypothetically safe unless the cell is already damaged.

I'm no expert but it looks like the BMS update may actually be a fix, in that if everything operates properly no cell should experience gross overvoltage. I suspect the resistance testing mentioned in Anaglypta's post is intended to try and detect damaged cells for exactly this reason, and the greatly reduced charging power some have experienced is an attempt to balance cells well under 100% SoC in an effort to save the pack and avoid needing to replace it... if they can pull all the cells back into that 0.1V tolerance and keep them there, everything should be good to go. Slow/stalled charging might be more tolerable than a loss of range, after all.

As for what changed, perhaps the BMS programming is what changed? Maybe someone realized the problem over a year ago (but not necessarily the full consequences) and fixed the software, which made it to production vehicles starting March 2020. Pure conjecture on my part, so grain of salt etc.

 

Yes - the 1V made me wince when a saw it, but other numbers in the email (translated from Korean) look correct. If it really was 1 Volt then we've been driving around in something just waiting to happen!

John.

 

Not sure if you guys know, there have been fires in Chevy Bolt's recently too... they also use LG cells..

https://www.detroitnews.com/story/business/autos/general-motors/2020/10/13/us-investigates-fire-reports-chevy-bolt-electric-vehicles/114270540/

 

Yeah, I saw. Very hard to say if it is related in any way.

 

I wonder if the Bolts use the same cells as the Konas???

2018+ Bolt EV's have LG cells made in the US (I believe Michigan) while the Kona uses Korean made cells.. but.. I don't know if it's the same cell type..

 

Isn't NMC 622 the same as used in the Kona?

 

Yeah, 1 volt difference in that pack would be pretty horrendous. Surprising that the original
firmware wouldn't alarm at way less.

Does the cell model imply 63 amp-hours? That would make sense; giving a 5-ish %
overhead on "full" at 64 kWh and "really full".

_H*

 

I noticed that the service manual section on inspecting the battery quotes a maximum deviation of 40mV, though nothing about what to do if it's above that

 

Something that big could plausibly a value that was used during development that they forgot to set to a more sensible value for production

 

I can't see 1V - no way! There must be a mistake somewhere.

 

got to be 0.1 volt.. That would already seriously concern me..

 

I've spent an hour or two today trying to locate and browse through patents from Hyundai regarding this subject. In one dated Dec 2014 they propose a method for cell balancing a battery of 12 modules where the deviation range for initiating balancing is between 0.1 and 1.0 V. Now of course the patent only needs to describe an example of using the invention and the numbers may be entirely meaningless. Over 1.0 and the status of the car (driving or not?) is considered.

To get the idea of this particular patent you only need read the few paragraphs starting half way down the description. This scheme seems to attempt balancing on a continuous basis.

There is an absolute heap of information about balancing in general on EV forums as well. The main issue seems to be that if a passive (resistive) scheme is used the heat sinking capacity limits the effectiveness if used under high charge rates because it can't keep up. So, we don't know how the Kona's original scheme works, but if it's passive (and I'll bet it is), under DC charging it will have to take a break ( as we understand it now will) for this reason.

 

Not seen with our Kona built 24 June 2020.

Typical charge pattern (at 32 amps) is to 90%, then run it (locally) for however long it takes to reach 40% (usually 10 days).

Have only topped up to 100% on two occasions, no pauses in the charge event noted.

Edit: typo

 

The label reads 6980 for the BMS. Battery build date is 18 June 2020.

P/N is 37501 - K4054

I don't have Torque pro, so I can't pass on that data.

 

That's interesting. There are two separate part numbers for the battery pack:

37501-K4000 - what I believe the cars shipped with.
37501-K4054 - what you are getting now.

Both parts are 30k$, so I believe that they essentially represent something similar. But the newer part has no information about which cars it fits or anything else.

Now I am curious what they will give me for my 2019.

 

I wonder if the-Kxxxx is just the factory firmware version, so an old pack with the BMS update would be the same thnig