General Motors says that I’m special.
Nearly 43,000 Chevy Bolt electric cars have been sold in the United States. According to General Motors, only one of those Bolts needed its battery pack replaced not once, but two times. And who is the winner of that unfortunate lottery? Me.
“You’re the one in a million second-event person,” Tim Grewe, chief engineer of electric propulsion systems at General Motors, told me in a phone call last week. “This is a terrible thing that happened to you. And even though it’s so rare, we never want it to happen again.”
The saga is long and detailed. So I’ve encapsulated it in a timeline. Below that, you’ll find an excerpted interview with Grewe.
The Story Begins
- Mar. 6, 2017: My 2017 Chevy Bolt Premier rolls off the assembly line.
- June 16, 2017: I receive the car in a three-year lease.
- Aug. 8, 2017: The Bolt abruptly stops, leaving me dangerously stranded on the side of the road.
- Aug. 22, 2017: GM diagnoses a low-voltage problem in Cell 25. The entire battery pack is replaced. My Bolt shows 1,746 miles on the odometer.
- Aug. 24, 2017: I post an account of what happened on PluginCars.com. I report that GM notified about 100 Bolt owners that they could face a similar issue.
- Sept. 11, 2017: GM Communications informs me in an email that the company made “improvements throughout the supply chain” to correct the cell low-voltage problem.
- Sept. 18, 2017: I learn of additional Bolt drivers (including a pregnant woman on a busy Los Angeles highway) who were stranded on the roadside but had not been notified by the company. GM says that OnStar data needs to be collected for some time to identify at-risk vehicles. Therefore, new Bolt owners (or cars on dealerships lots) could encounter the same problem.
- Sept. 2017 – Mar. 2018: Bolts with new battery packs appear to be reliable. GM works with LG Chem, the battery supplier, to make multiple changes to the cell manufacturing process. The company believes the problem is fixed.
Software Fixes for Earlier Warning
- April 2, 2018: Chevrolet issues a recall for drivers of 2017 Bolts to get a software update to provide more warning about any potential “cell low-voltage condition” and loss of propulsion. The software change allows cars to continue driving with a diminished range rather than forcing drivers to make an immediate stop.
- May 11, 2018: GM releases a new software update for all Bolt owners (not just for 2017 models) to “provide additional warnings.” GM later tells me the software “increases the accuracy of the range estimation in addition to providing more warning at low states of charge.”
- May 14, 2018: My Bolt receives the April and May software updates. GM tells me that the company is continuing to monitor all Bolts so it can “proactively contact owners to have their batteries serviced as soon as our diagnostics confirm batteries are suspect.”
Deja Vu All Over Again
- Oct. 2018: I begin to notice diminished range. I use a dongle-based tool provided by FleetCarma to pull more precise data from my Bolt’s diagnostics port.
- Nov. 20, 2018: In a post on InsideEVs.com, I report detailed range numbers about a 184-mile late trip that reveals my Bolt’s battery capacity is reduced from 60 kilowatt-hours to about 35 kWh.
- Nov. 21, 2018: I alert GM about the problem. Chevy Communications responds: “The team is looking into what could be the cause using remote diagnostics.” I am later told that the problems with my Bolt are communicated “all the way to the top” of the company.
- Dec. 3, 2018: I receive a standardized OnStar diagnostics alert stating, “An element in the battery has slipped below the threshold required for good performance. If unrepaired, it could cause a stalling situation while driving.”
- Dec. 2018 – Jan. 2019: I unsuccessfully try to wrangle Chevrolet and a local dealership to pick up the car for service. Chevrolet Communications tells me that my car is safe to drive – although with diminished range.
My Bolt was back in the shop in January -- to get its third battery pack.
- Dec. 2018 – Jan. 2019: I continue to pull data from my Bolt’s computer system, revealing that the threshold of low to high voltage is beyond the accepted threshold. (See screenshots below.)
- Jan. 16, 2019: I drive my car 25 miles to the closest Chevy dealership. Diagnostics show the offending low-voltage cell is No. 68. My Bolt receives its third battery pack. The car shows 11,940 on the odometer. I continue to pull data from the computer, which indicates a healthy battery pack and my range is restored to the expected level.
Trying to Understand
- Jan. 2019: After diagnosing my car’s second bad cell, General Motors – working with LG Chem – makes additional changes to the cell manufacturing process.
- Jan. 16, 2019: Six hours after I dropped off my car, Automotive News, reporting from its World Congress event, tweets this quote from GM CEO Mary Barra: “We've sold over 200,000 electric vehicles and we have yet to replace a battery pack.” Multiple owners on social media say that they’ve had packs replaced, some as recently as January 2019.
- Jan. 18, 2019: Chevrolet Communications tells me that, “Mary’s original statement was in reference to the wear out of a battery due to regular use. Upon further review, an estimated less than 0.01% of customer battery packs or sections have been replaced due to suspected wear out.”
- Feb. 1, 2019: I speak with Tim Grewe, chief engineer of electric propulsion systems at General Motors.
On Feb. 1, I spoke with GM’s Tim Grewe for nearly 40 minutes. Here’s how the conversation went – edited for clarity and brevity.
InsideEVs: Was the issue of the delta between low and high voltage of cells the same problem during both the 2017 and the 2018 incidents with my car?
Tim Grewe: It resulted in the same customer experience of reduced range and early capacity fade. When you do a 100-percent charge, and you have one cell not at the 4.1 level, like a cell at a lower 3.8 voltage, then that cell is not matched to the other cells in the pack. It was not the same quality issue as before although they were related.
On my second pack, Cell 68 had low voltage, thus creating an imbalance and low range.
Help me understand the difference between the two.
We have 280 cells in a Bolt. It has to do with the geometry of the cell. Even at a smaller level than the cell, there’s a lot going on. You have to make sure everything ages appropriately. Batteries expand, contract, and vary with temperature. It’s electrochemistry. And so there were two problems that took a while to manifest themselves. There were some very subtle things that happened in the supply chain and in the cell assembly that caused the inside of those cells to fade.
Do you test every single cell before it goes into a pack?
We do extensive tests. When you make a cell, it’s not energized. You have to put it through a formation stage. The anode and cathode formation phases take a long time. Then we have a quality system where we monitor them. They’re stored to make sure there are no subtle changes happening. We also have quality-control systems that monitor the pre-compression and post-compression phenomenon.
What do you mean by compression?
We literally press them together. We put highly engineered foam between them. Then we put them all together in the cell-module assembly. Each cell has a validated pressure throughout its lifecycle for hot and cold. Even with all the monitoring we did along the way, these got through all that because the phenomenon took longer to present itself.
When you say phenomenon, you mean something in the chemistry created a diminished capacity to hold the voltage?
Yes. So instead of having a 238-mile pack, you had a 155-mile pack.
That’s what happened the second time. The first time when the system detected the problem, the software said pull over and just stopped me. But the second time it just changed the range number on the indicator?
Exactly. With the prognostics using OnStar, there’s a lot of range of variability in people’s use of the car. So literally your article on InsideEVs was written right at the time when they would have been calling you to say this looks suspicious and bring it into the dealer.
Data after a charging event provides a clearer picture of low voltage.
Are you confident that someone else a month from now is not going to encounter a problem and say, hey, why didn't you notify me?
Highly confident. I'm not just saying that. You are the only one I know of in the second round.
There were a series of recalls and service bulletins. Did I miss any of them?
No. We want to be confident, but we also want to be humble because this is pioneering stuff. We’re watching with OnStar to protect your experience. We were the first ones out there with these high-range electric cars. We’re doing extra prognostic work to try to make sure things happen the way we expect.
We were so upset with your first experience that we got a fix out there right away, as fast as we could, to make sure the fewest number of customers would have your first experience because that was terrible. Between August 2017 and May 2018, we were trying to hit the problem as hard as we could to prevent the shifting to neutral and coasting down a hill.
Are you saying that it was partly my case from 2017 that led to new software development and the recall for all Bolt drivers?
Can you go into more detail about the differences between what happened in the first case and second case?
When you make a battery cell, there are a hundred steps. You take the aluminum. You take the copper. You laminate on a cathode material the nickel, the manganese, and the cobalt – and then you put the graphite on the anode. Then you have something called a separator that keeps those two from touching each other. And then you have to put a lot of those in parallel in each side of each and every cell.
There were different subtle manufacturing variance issues between the August 2017 and the January 2019 pack that caused it to lose capacity before the life of the car. It would be misrepresenting it to say it was one thing or another. It was subtle. Both resulted in capacity losses but for very different reasons. It's a long tech paper to go beyond that.
Can you locate where the problem was?
We know exactly what happened. Tool usage, maintenance schedules, and that kind of subtle stuff. It was LG Chem’s best practice. It was our best experience with previous batteries. We needed to do some things differently. Not necessarily better, just differently.
The way it’s packaged?
The way it’s manufactured. A battery cell assembly lab is sort of like a newspaper printing press. You start with these huge rolls of paper. And in this case, it's copper, and then it runs through this coating machine, and then you press it down. Then you heat it and cut it up. And then you fold it into a newspaper, and there are very subtle things in some of those process steps that degraded the capacity in different ways between the two different cells failures.
Have you made any changes in that process?
100 percent. Not only did we fix the issue, but we put extra monitoring to make sure it never happens again. In each case, we work directly with our supply chain. You fix it there and add a monitor in there. Where could a similar thing happen somewhere else in those hundred steps? And how do you make that additional monitoring and those steps as well?
It had to do with maintenance schedules and monitoring. We didn’t just attack just one operation that was the root cause of your cell problem. We read it across to all other similar operations, and we improved it and increased the monitoring of those as well.
Did these corrections happen only on the 2017 issue or the more recent case as well?
We made multiples improvements in 2017 and 2018. But it wasn't enough. Now the data shows the cells in your 2019 pack are super tight. That’ll be the proof that we finally nailed this thing. But we're going to continue to monitor with OnStar to make sure that’s true.
The problem in 2017 was experienced by about 100 people. Could one or more of those 100 people also suffer a second fault like what I experienced?
We haven't found one yet. You're special.