[NewsCoulomb]

In Korea (where LG and Hyundai both have strong backers in gov't), LG will pay ">50%" of the cost to replace batteries in 25k Konas in Korea (on another site, I read "2/3 of cost", and LG is probably going to split that with the manufacturer of the separator film).

In the US, GM should have a far stronger position in court vs LG Chemical of America. With precedent set in Korea, it should be a "slam dunk" for GM to hit LG with 2/3 of the cost of battery replacements. If I were GM, I would run, not walk, to court and file papers against LG.

Could GM really be pushing a software solution to avoid paying a small portion of the cost of battery replacements? At this point, anything GM does which may keep affected batteries on the road (including a GM software change) only transfers liability from LG to GM.

The question of whether the software change is good enough is moot. LG can be made to pay for new batteries. The more GM talks-up a firmware solution being good enough, the more it undercuts its future court case against LG.

I think the math is very simple on this. If less than 5% of Bolt EVs have faulty separators, replacing a few thousand modules/packs is several digits cheaper than replacing every Bolt EV pack. It's also less wasteful.

To me, the only justification GM would have for replacing all of the packs is if either all the packs are faulty or they feel the value of positive PR and advertising outweigh costs of replacing packs unnecessarily. However, $500 million buys a lot of positive press.

 

[Toyotiac]

GM has big EV plans with LG. Would GM be able to find a new battery partner if they stomp all over LG? Would they have the expertise to go it alone? Even Tesla partners with battery makers.

It is a balancing act. Just because they could win the battle, would they prevail long term by taking an adversarial position against a key supplier?

If SW can effectively detect the conditions, and if the actual failure rate is relatively low, why wouldn't GM proceed with SW solution for all, and replacements only when defective cells are discovered?

It looks like GM has a little over a month to validate that software and they will need absolute confidence in it. 1 fire from an undetected fault would be an epic fail.

 

[RPC]

Ha ha! No. But the failure mode of Lithium batteries (i.e. their propensity to burn), is well-known and understood from what's been learned through their use in smaller devices.

So here you are scaling everything up into these "monster" battery packs, and putting them in automobiles for people to drive around in. The stakes are very high - from both a safety and business-image standpoint. You want this to succeed because you say it's the future of your company. So then you need to do everything possible to ensure there isn't a single fire with these batteries! And that means you need extensive fire-detection logic in your BMS, and you also need to continue lab testing - even after the vehicles are out there - to try to uncover any weaknesses that may not show up immediately.

And I can already hear people reading this and saying: "it was only 4 battery packs out of 70,000 ... how are you going to find that?"

Yeah .. I'd buy that if these 4 vehicles all failed under different, and very peculiar, circumstances. But they didn't. They all failed pretty much the same way: sitting still, having been charged recently.

So ... cars can sit still and be repeatedly charged in a GM lab environment, right? They had 4 years to do this and neither they (nor their BMS software) were able to find this problem before one of their customers did?

I'm sorry, but they dropped the ball here. And now they're scrambling to do what they should've done at the beginning, or certainly could've done over the last 4 years, and it's taking them an eternity to come up with their (better-late-then-never) "solution".

If we assume each fire was due to one bad cell, that means we've had four failures out of twenty million cells in a population with an average age of two years. So a failure rate of 11.3E-12 per hour. If GM has 100 battery packs in extended testing and they're the same average age as the general population (assume GM sample each batch that arrives from LG) then the chance that one of them develops this issue is about 0.6%. So I'm not surprised GM didn't catch this in advance. The question remains: can they reliably detect a cell with dendrite growth with the diagnostics built into the battery pack?

 

[ARob]

It looks like GM has a little over a month to validate that software and they will need absolute confidence in it. 1 fire from an undetected fault would be an epic fail.

GM has a unique opportunity to watch what LG and Hyundai do. So far, the most favorable analysis of Hyundai's handling could probably be give a D or F grade. You are correct re: 1 fire from an undetected condition, that seems to be exactly the hole Hyundai stepped in to. Now, Hyundai may have forced regulators to demand pack replacements as a result of their stumble. GM held off on a solution, and if they can convince regulators they have a good handle on matters, they may not be forced to replace packs.

If the 90% temp solution and\or HTR\TCL is an effective interim solution, a few extra months of waiting for both the forensic analysis and precedents set by their partner LG on the Hyundai front gives GM more options, leverage, or confidence in their "final solution".

 

[Usain]

It looks like GM has a little over a month to validate that software and they will need absolute confidence in it. 1 fire from an undetected fault would be an epic fail.

Yes, and GM isn't the only one that needs to have confidence in their new "state of the art" software. (The "state of the art" thing cracked me up).

Doesn't NHTSA have to sign off as well?

 

[desiv]

Doesn't NHTSA have to sign off as well?

I could easily be WAY OFF on this, but I thought this recall was a voluntary recall by GM.
So they might not need NHTSA to sign off????

 

[electrons]

Guessing the new software fault detection algorithms track charge/discharge characteristics & cells give enough warning before a runaway fire event happens. Let's hope false alarms don't start bothering Bolt drivers constantly.

 

[TheLondonBroiler]

I worry about the process of detecting a cell that is experiencing separator puncture/failure/short circuiting. The Hyundai fix apparently stopped charging at certain high SOCs one or more times, looked for problems, then resumed charging if no problem was found, which apparently didn't work out too well.

I'm by no means a battery expert, but if a battery experiences even a minimal internal short, "the fuse has been lit," so to speak. We're not talking explosion (in most cases), but thermal runaway is a strong possibility when the cathode and anode are directly connected/short circuited.

When the "official fix" is announced in April(?), GM/LG should go into detail to prove to/inform customers and inspire confidence, as to why a software fix can resolve the issues in the Bolt, while a software fix failed in the Kona, because their battery cells originated from the same manufacturing facility and are very similar in specification (both supposedly NCM 622). GM allegedly contacted Electrek very recently with the information that the Bolt and Kona batteries have different cell separators. I've read at least one article that suggested that poor manufacturing "happenings" in that LG plant were the actual cause of damage and ultimately failure.

I have to wonder if @NewsCoulomb isn't onto something when pointing to cell specs being exceeded as a potential culprit, because Hyundai has apparently used the same cells in the most recent Ioniq and an electric bus, in addition to Renault using them in the Zoe 40kwh (according to PushEVs), without similar incident. The differentiating factor appears to likely be higher charge & discharge in the Kona.

We shall see what happens....

 

[boltage]

If faster than 1C DCFC accelerates the problem, then that may be why (a) the Kona has more issues, and (b) the Bolt with its current battery design is not going to get significantly faster DCFC.

The Kona's fast charging up to 75 kW is about 1.17C for its 64 kWh battery. The older Bolt's fast charging up to 55 kW is about 0.92C for its 60 kWh battery.

 

[Bolt2019]

Y'know...when I was shopping around early-ish 2019, the Bolt was available and the Kona was not far away. At the time the Kona was a few thousand more. Then the Canadian federal government announced a rebate on EV's and Hyundai promptly dropped their price to get under a threshold price. I kinda thought was sneaky. But they had that 30 or so extra km range. I didn't know anything about faster charging at that time. This is the first I've heard of it. I did know that the Kona was a modified ICEV. So there was that. I also have a long memory of the Hyundai Pony. So there was that.
Now we have their more fires although I'm not sure of the "per capita" rate compared to the Bolt. Now we learn they may have cheated on that range using really pretty much the same battery. Now we learn they seem to have cheated on charging speed. All of these little cheats just to get a small competitive edge.

I think I'm much more confident of my choice of an affordable-ish American car. Don't tell me government and regulatory oversight should be weakened simply to let the market play its game. I will continue to wait and see how they'll conclude this little recall adventure but I sure would have very little confidence in the Kona. Apple Pie and ChEVrolet all the way thank you very much. And US (and very tightly intertwined Canadian) rules of trade.

 

[electrons]

I sure would have very little confidence in the Kona.

Battery problems aside, the Kona is very good. They are using a better rear supension set-up vs. Bolt, a control blade multilink type in the Kona, and they redesigned the floorpan to fit the battery nicely. Same with the Soul EV at Kia. If done right, you can re-use an ICE vehicle chassis quite well.

The Bolt started from a Chevy Sonic-Trax ICE vehicle platform with a lot of new sheet metal stampings & suspension changes.

Kia will get past this and people get new batteries, assuming the U.S. gets the same Korean treatment.
Besides, 2021 Konas have a 10 year battery warranty and GM only backs the Bolt to 8 years. Older Kona EVs have a lifetime battery warranty! Non-transferable on sale though.

I worry about the process of detecting a cell that is experiencing separator puncture/failure/short circuiting.

I wouldn't know what signs of illness they can monitor for. Seems like they have to assume the separator failure doesn't tear all of a sudden(?). I'm a non-battery tech person; .. mech, aero, software, algorithm subjects only for me. Battery science is tricky.

Anybody know if they can or do have a hardware fuse & circuit structure similar to what NREL discusses? Note how they say it's "extremely challenging". Which might mean it's hit or miss, with lots of false-positives to drive Bolt drivers crazy.

 

[Eriamjh1138]

The more I think about it the more ticked off I am ..April are you freaking kidding me..5 months from when this recall started.

Impatient.

Agree! I'm not going to be satisfied with any SW fix. This is too much of a risk. This is ridiculous.

Then you should just sell your car or sue. Which one are you?

These reactions are unjustified.

 

[GJETSON]

Anybody know if they can or do have a hardware fuse & circuit structure similar to what NREL discusses?

They do not. Tesla has a fusible link for each of their thousands of cells. If it blows, that cell is out of the circuit. So it would no longer be charging or discharging. That may be enough to prevent a thermal runaway of the rest of the pack, but not necessarily. If it gets hot enough to damage surrounding cells you could still be in trouble.

 

[TotaledJetta]

A software fix for the battery fires reminds me of the fix Boeing used on the 737 Max. How did that work out for them? Of course Bolts don't fly.

 

[desiv]

I'm really not bothered by the proposed software fix. And if it doesn't work, we'll see what GM does.
Not worried at all really...

I wonder if that is partially because my first car was a Ford Pinto? Hmmm... ;-)

 

[GregBrew]

A software fix for the battery fires reminds me of the fix Boeing used on the 737 Max. How did that work out for them? Of course Bolts don't fly.

I don't understand your comparison. The Boeing fix has worked out famously.

The fix Boeing is adopting has been tested up the ying-yang to be safe. What started as one of the most reliable families of planes in history has become even more reliable.

If I was given the choice of planes to fly on a particular route, I'd choose a 737-Max-8...every time.

 

[shrox]

A software fix for the battery fires reminds me of the fix Boeing used on the 737 Max. How did that work out for them? Of course Bolts don't fly.

It has worked...

I flew on American's Boeing 737 Max and found the airline doing the bare minimum to inform passengers they were booked on the notorious plane

American Airlines has the largest fleet of Boeing 737 Max aircraft flying passengers in the US. But most flyers likely don't know they're on one.

www.businessinsider.com

 

[electrons]

Without knowing the details of the algorithms GM will use to detect bad battery cells before a fire occurs, it's hard to say if a pure software fix will really work long-term or not. It's certainly the cheap way to go.

Off topic a bit below:

TotaledJetta said:
A software fix for the battery fires reminds me of the fix Boeing used on the 737 Max. How did that work out for them? Of course Bolts don't fly.

I don't understand your comparison. The Boeing fix has worked out famously.

I think what TotaledJetta meant was the "MCAS" pitch-down trim (activated near stall) on the 737 Max8 was commanded via software to change the stick force gradient to make pilots not tend to float the nose upward too much near stall (high alpha), that's all. Actually pitch down trimming does do what its supposed to do, no problem with that part.

The real mistake is they allowed only 1 bad alpha (AOA, angle-of-attack) sensor to activate MCAS. That's the gist of the problem they got themselves into.

They then claimed it didn't impact runaway trim actions by pilots they were trained for anyway, all to avoid telling airlines they have to re-train pilots on simulators, to boost 737 Max8 sales. In reality, allowing only 1 bad alpha sensor to activate MCAS pitch down trim created a difficult human factors condition for the pilots, requiring them to flip a switch if they remembered to do it & they knew to recognize the failure condition, and manually trimming a difficult spinning pitch trim wheel, resulting in crashes.

The software fix recently implemented forces the avionics to use dual alpha (AOA) sensors to detect a bad sensor, and they limit how long pitch down trim can happen. That works. Should have done that in first place. Very stupid predicament.

 

[boltage]

The software fix recently implemented forces the avionics to use dual alpha (AOA) sensors to detect a bad sensor, and they limit how long pitch down trim can happen. That works. Should have done that in first place. Very stupid predicament.

They also made the angle-of-attack (AoA) display and AoA-disagree indicator standard equipment instead of an extra cost option that many airlines did not buy previously.

 

[boltage]

Without knowing the details of the algorithms GM will use to detect bad battery cells before a fire occurs, it's hard to say if a pure software fix will really work long-term or not. It's certainly the cheap way to go.

It is not a pure software fix, since, in the event of a detected fault, it is intended to alert the owner and GM that the battery or part of it needs repair or replacement. But obviously, if it works, it is less expensive than mass replacement or an inspection process that requires disassembly at the dealer shop.