Double Stranding Leads to Replacement Battery, Software Change — 2017 Chevrolet Bolt EV Long-Term Road Test

2017 Chevrolet Bolt: Double Stranding Leads to Replacement Battery, Software Change

by Dan Edmunds, Director of Vehicle Testing

Note: As of April 5, 2018, GM will begin contacting owners of 2017 Chevrolet Bolt EVs and asking them to bring in their cars for a software update. We don’t know if this grew out of the account detailed below, but you’ll want to read the whole thing if you’re a Bolt EV owner.

It happened the first time on January 11 while editor Brent Romans was driving our 2017 Chevrolet Bolt EV to his home base some 235 miles away on the far side of Fresno, California. At the approximate halfway point in Bakersfield, he’d stopped briefly for a quick shot of extra juice at a DC fast charger. He’d unplugged with an indicated 174 miles of range in hand with just 113 miles left to cover. Piece of cake.

Traffic was lighter than it had been in L.A., and the pace of traffic was brisk. But the miles were coming off the range meter faster than Brent was driving them, so he reeled in his pace a bit. With 10 miles left to cover and 30 miles of indicated range remaining, the Bolt suddenly went into reduced power mode.

He tried to limp it home, but 45 mph was all the Bolt would give. A quarter-mile later, it wouldn’t even do that, and Brent coasted the Bolt to a halt at the edge of the dark freeway. A passing highway patrol officer came upon the scene and used his cruiser to shield the Bolt from oncoming traffic as they waited for the tow truck.

What had happened? One minute he had 30 miles of range, the next he had none. That’s what it seemed like, at least. He gave me a call as he sat waiting for the tow, knowing that I had once driven our Bolt 334 miles on a single charge, eventually plugging in to our dedicated 240-volt charge station with 14 miles remaining.

Did he see either of the two warnings that should have come up before it went into limp mode? They’re hard to miss. On my run, the first of them came when the battery gauge dropped to four green bars, which signifies 20 percent remaining. There was a chime and a visual "charge soon" warning that I canceled with a steering wheel button. The next came with two bars showing, now orange instead of green, which corresponds to 10 percent remaining. There was a more insistent warning and chime, but I was still able to dismiss the "charge soon" message and complete the last few miles to the office after noting to myself that it still read "soon" rather than "now."

Brent didn’t remember seeing either of these, which seemed odd. We talked about how the range is probably calculated, how the pace before his partial recharge in Bakersfield had been slowed by L.A. traffic, how his subsequent post-charge pace had been considerably more free-flowing. Had the computer’s prediction been playing catch-up to this change in driving conditions? It almost certainly was, but did that explain it? And why hadn’t he seen any warnings?

We decided on a wait-and-see approach. In hindsight, this was a mistake.

Brent had the Bolt towed to another nearby DC fast charger, where he filled it with more than 100 miles worth of battery charge before finishing the short remaining distance to his home. Over the next two weeks he continued to drive and charge the Bolt as needed, capping things off with another long drive back to the office from Fresno with an intermediate DC fast charge along the way.

Normal operation continued in the hands of others for the next two weeks, and then I received an OnStar diagnostics email dated February 7 that proclaimed everything to be A-OK. All systems got green marks, including the electric drive unit and the lithium-ion battery. Our battery-range calculation theory was looking good.

That thought lasted 24 hours. The very next day, February 8, the Bolt suddenly went into limp mode with Kurt Niebuhr driving in the leftmost carpool lane of a busy local freeway. With forward drive ebbing away, he somehow managed to thread his way across four lanes of traffic and guide the Bolt to a stop in the breakdown lane.

Kurt’s experience was pretty much a replay of the first incident except that he’d seen 60 miles of range evaporate before his eyes instead of 30 miles. He’d also had no more than a quarter-mile of reduced power at 45 mph before the Bolt quit altogether. At least he’d heard the office talk about Brent’s earlier incident, so he was aware of the urgency and dead certain about the lack of advance onboard warnings.

This time we had it towed directly to a dealership for diagnosis and repair.

We soon realized our troubles were the manifestation of a known issue that we had read about last August but didn’t think applied to our car. The accounts indicated that GM had informed a few hundred owners, but we’d had our car seven months at that point and had not been contacted. Surely the automaker would have told us if we were vulnerable, right?

The next morning, February 9, I checked the GM Recall Center section of the OnStar app, which plainly stated, "Currently, there are no recalls or programs associated with your 2017 Chevrolet Bolt EV," with our car’s VIN written beneath. Things didn’t add up.

We got in touch with a GM spokesman to get some answers, and he in turn put us in contact with someone with a deeper Bolt EV tech background. The following is my understanding of what happened based on these conversations, other accounts I have studied, and our own observations.

What is the problem?
It seems that early packs had a manufacturing process issue that was corrected with the supplier. It has to do with the voltage of the 288 individual cells in the pack. Normally, the voltage in each cell will remain even, but an unexpected voltage drop in even just one cell would cause an imbalance and prompt the battery management system to suddenly drop the remaining range calculation and cut power to reduce the strain. Our post-repair dealer paperwork supports this: "Battery pack cell #79 only 2.82 volts." The normal measurement is about 3.65 volts per cell. I’m no battery engineer, but apparently this is significant.

How many cars are affected?
I did not get a specific answer, except that this is a known issue that has long since been corrected in production. That implies a certain known production date and VIN cutoff, but it’s not as easy as that because it comes down to the date the battery pack was built, which may differ from the car’s assembly date. GM has said that only very early-production Bolt EVs were affected, amounting to less than 1 percent of total sales. What does that mean in specific terms if you’re a Bolt owner? Apparently, our car’s early build date of December 2016, not to mention its behavior, puts it in that group.

Why weren’t we notified?
I’m not entirely clear. The voltage drop manifests itself differently from car to car. Driving style and environment play a role, and the problem seems more likely to occur when the battery charge gets low. In other words, your early battery pack may not have a bad cell at all, or it may be susceptible but never develop the problem if you never run your battery down very far. Once started, however, the problem can progress rapidly, hence our 60-mile failure hot on the heels of the 30-mile one.

Also, certain early cars, including ours, are on a list for intensive back-end monitoring via OnStar. Some of those owners have been contacted. Others, like us, have not. The part that kills me is this: I was told that the automaker was only one day away from contacting us about it when Kurt had his problem. Really? What about Brent’s issue four weeks before that? Wasn’t that a big clue? It seems this monitoring business is not an exact science.

Now what?
Drive it. We have a new battery pack under warranty, and it was assembled after the fix was implemented. Good as new, so long as we’re talking about new after whatever date GM rolled out the battery fix. As for range and consumption, those are as good as we’ve seen before the troubles.

Shouldn’t this be a recall?
Funny you should ask. While I was writing this on Thursday, April 5, GM made a couple of significant announcements, which is great for them because you’d never believe what I had written instead.

While the automaker stops just short of calling this a recall, as of April 5, 2018, it is notifying all owners of 2017 Bolt EVs and asking them to come in for a software update (program N172127150). Do it. Letters will be mailed out soon, but this notification shows up today in the GM Recall Center of the OnStar user website and phone app of 2017 Bolt EV owners. Interestingly, the same Recall Center also contains a new entry for the program that outlines the battery pack problem itself (program N172104090). This last one dates back to September 5, 2017, and is something I would have liked to see a long time ago, but never did — even when I deliberately checked the very same section of the OnStar app a couple months ago.

The main feature of this new software update is a change to the point at which the car goes into limp mode. Instead of a quarter-mile of reduced pace at 45 mph before it shuts off, the new software can recognize a meaningful voltage drop early enough to give a driver limp mode for between 1 and 2 miles.

This may not sound like much, but consider Kurt’s crowded freeway scenario, in which he had to coax a dying car from the HOV/diamond lane to the shoulder across four lanes of traffic. It was pretty harrowing with a quarter-mile of drive. One or two miles would have made it much easier and safer. That said, cutting across multiple lanes of traffic at 45 mph is never going to win you any friends or be entirely safe. I wonder if they can set it up such that the four-way flashers come on automatically?

Wait. All 2017 Bolt EVs have the battery problem?
No, not all 2017 Chevrolet Bolt EVs have the battery pack issue. If you see that in your app, you are most likely still in the clear, but you’ll want to have it checked. The boring description sounds innocuous, but it is outlining what led to Brent and Kurt’s strandings. Besides, you’re going to visit the dealer anyway because you’ll want the newly announced software upgrade even if your 2017 Bolt EV is not an early-build car and has a healthy battery. Voltage drops outside the known issue have not been reported, but Brent and Kurt will tell you that you’ll want all the advance warning you can get in the unlikely event that things do go pear-shaped.

I’m frankly furious that this happened two times to one car, especially when this was a known issue and our Bolt was on a list somewhere for intensive monitoring because someone knew or suspected its battery was one of the early "before fix" units that was potentially suspect.

But I’m less irritated now that the issue is more out in the open. The new software fix will give drivers more maneuvering time, and even the battery issue is now visible in the recall center. If you get the notification, have your car inspected. The regular vehicle health report won’t spot it, and in our case, neither did the extra behind-the-scenes monitoring.

Dan Edmunds, director of vehicle testing, at 13,002 and 14,139 miles