Software updates and warranties

[father_of_6]

OTA software updates are a double-edged sword... on one hand we may get new features and improvements on our vehicles, but on the other hand sometimes updates introduce bugs or cause other problems.

On my 7 hour drive home from picking up my Model 3, a PT sensor failed. This left us with no cabin heat on a *very* cold day. Fortunately the seat heaters worked. Apparently this was a widespread failure (bad batch of PT sensors), and so Tesla's short term solution was to fix it with an update. I installed that update that night when I got home and in the morning when I drove the car the heat was working again. The update bypassed the failed sensor. Magical!

I got a call a few weeks later from my nearest service center and scheduled an appointment to replace the failed sensor. That whole experience was impressive.

Fast forward to today. I'd been holding off on updating because I wanted to keep my radar... but recently gave in and installed the latest update. Immediately after, the GPS location of the car was intermittently offset during navigated directions - this causes the car's map marker to appear off the road which then causes the directions to get confused at each intersection (turn right onto main street, turn right onto main street, turn right onto main street...).

I scheduled service for this problem and they pushed another software update, but said if this doesn't fix it I'll have to come in to have my GPS antenna replaced.

So here is the point of my long-winded post... I'm reasonably confident that this issue was introduced by a software update, yet they're suggesting replacing hardware to fix it. Should we be expecting to pay for repairs like this once we're outside of warranty? Is there a precedent for this?

My guess is that the software that dealt with this GPS map marker may have been rather hairy, perhaps needing to solve many edge cases and even interoperability with many different GPS hardware vendors.

Whoever rewrote the software probably missed some of those edge cases. Perhaps code was removed that was there to bypass hardware failures or deficiencies (like my PT sensors). I wouldn't even know there was a problem because they pushed a software workaround. But perhaps then a few years later and out-of-warranty they remove that software fix (even unintentionally).

How can Tesla expect owners to pay for labor and parts to fix a problem that was caused by software (or unveiled by removing a software fix)? What recourse do we have as owners? Do we really have to go to arbitration for these things?

It's not as though I can prove that they caused it with an update (even though I'd wager my left pinky toe on it).

 

[stopcrazypp]

The cabin camera causes interference with the GPS antenna because the cabin camera was not properly shielded. This showed up only in later updates because Tesla only recently (around mid last year) enabled the cabin cameras for more usage:
GPS issue since the past few weeks
An easy way to tell back then was this only happened when AP is on (because cabin camera was only used then). Lately Tesla have also enabled other cabin camera features, so you may need to also disable those to isolate the issue. May also try taping up the cabin camera, but not sure if it changes anything (might still be running and cause the same interference).

For some, later software updates (some were pushed by SC) did "fix" the issue (people suspect it just disabled the cabin camera again or reduced the frequency of it turning on), but for some only a replacement of the camera unit with a newer (presumably shielded version) fixed it.

If your SC is suggesting to replace GPS antenna (and you are paying out of pocket), make sure it is not this same issue, otherwise you may be wasting your money. Many in that thread reported replacing the GPS antenna did nothing (as it wasn't the issue in the first place). At least get SC to commit to a cabin camera replacement (to shielded version not the same old version) if GPS antenna does not fix it, so you are not charged twice.

I am also still on 2022.20.7 to keep the radar (trying to experience it for as long as possible until some feature makes it more worthwhile to switch to Vision and then I can do a comparison).

 

[father_of_6]

The cabin camera causes interference with the GPS antenna because the cabin camera was not properly shielded.

Thanks for this, I hadn't been following the issue much.

I suspect that if I were out of warranty that they would charge me for fixing this, even though it might have been caused by a combination of faulty hardware, faulty design, and a software update that changed how that hardware operates.

This just seems like new territory... Tesla changes the vehicle remotely and causes problems that owners then have to pay for. That's my concern.

 

[PianoAl]

I'm reasonably confident that this issue was introduced by a software update,

That's the problem, as you said--you can't be sure. With eight or so updates per year and several glitches per year, those two things must occur at the same time occasionally.

If they replace the antenna, and the problem is still there, would an out-of-warranty owner need to pay??

My GPS has been off occasionally. I'm assuming you've tried a reboot?

 

[father_of_6]

I did try rebooting, and the service center pushed another software update (rebooted after that, too). It didn't happen for a few days, but still happens intermittently.

That's the problem, as you said--you can't be sure. With eight or so updates per year and several glitches per year, those two things must occur at the same time occasionally.

I may have gone 4 or more months without updating, and in 2 years of ownership never once had this problem. Immediately after updating, the issue appeared. I actually feel like even the blue nav lines were better aligned to the maps before the update. Before updating, they were perfect... now it all seems slightly visually out-of-sync.

It's true the only concrete evidence that this was caused by software is that the problem appeared after updating.

My reason for creating this thread wasn't really about this GPS bug, there are already threads for that.

What I'd like to understand better is... do owners think that this is okay? Is there a precedent for this?

No other auto manufacturer that I'm aware of pushes software updates like Tesla does. What is the expectation when those software updates introduce problems that could cross over into hardware land and the vehicle is no longer under warranty?

In the case of this GPS bug, if it's caused by EMI from the interior camera, that's a design flaw or an issue with the original hardware, but if it's not caught until a software update after the warranty runs out, is Tesla going to fix it goodwill?

Tesla is in a position to push an update to our cars that cause us to have to pay them for repairs. Is the only protection we have against it arbitration after the fact?

This may stifle "innovation", but it would be nice if there were some legal precedent for this that puts the onus on Tesla for this type of repair. If I bring an ICE to a dealership for an oil change and they use the wrong oil which then causes engine damage, I would expect them to repair the engine at no cost to me (although I would totally not expect them to admit fault).

 

[jjrandorin]

There is an existing thread in the software subforum (where this discussion likely should be, imo) on this topic:

Should Tesla break out-of-warranty cars with software updates?

Intentionally bricking all out-of-warranty cars at the same time would obviously lead to a PR catastrophe and expensive lawsuits. This will never happen. But if Tesla breaks just some cars in somewhat minor ways, would not this lead to added service revenue (resolving issues) and car sales (for...

teslamotorsclub.com

 

[yeahthatguy]

It's true the only concrete evidence that this was caused by software is that the problem appeared after updating.
My reason for creating this thread wasn't really about this GPS bug, there are already threads for that.

What I'd like to understand better is... do owners think that this is okay? Is there a precedent for this?

No other auto manufacturer that I'm aware of pushes software updates like Tesla does. What is the expectation when those software updates introduce problems that could cross over into hardware land and the vehicle is no longer under warranty?

In the case of this GPS bug, if it's caused by EMI from the interior camera, that's a design flaw or an issue with the original hardware, but if it's not caught until a software update after the warranty runs out, is Tesla going to fix it goodwill?

Tesla is in a position to push an update to our cars that cause us to have to pay them for repairs. Is the only protection we have against it arbitration after the fact?

This may stifle "innovation", but it would be nice if there were some legal precedent for this that puts the onus on Tesla for this type of repair. If I bring an ICE to a dealership for an oil change and they use the wrong oil which then causes engine damage, I would expect them to repair the engine at no cost to me (although I would totally not expect them to admit fault).

It very well could be new code that's causing the error to appear, but I'm not sure what you are looking for in terms of case law or legal precedent. I think this is another instance of needing to take a different approach then one would with a traditional car manufacturer, think of the car as more akin to your phone/PC than your ICE car in the sense that when there are issues, ALWAYS start with the software.

If you have an iphone and call up support/go in for support one of the first things they'll do is reboot the device. Then they will factory reset the device, then depending on the circumstance either push a previous software build, or push a new software build to test. Once they have been able to reproduce the issue after a factory reset (no user data/configuration changes to blame for the issue) or on different software builds we can be somewhat confident its a hardware fault.

This should be the same approach taken by any tesla owner then encountering an issue that's possibly software related. Open a ticket with support, have them remotely read the logs, it can be very obvious if the hardware is causing a kernel panic is similar. However if the logs do not contain the smoking gun, a previous build or a preview build should be pushed to the car and we attempt to reproduce the issue. Once its been determined that the issue persists among different builds its a reasonable jump to start looking at hardware.

If you have a headache and go to the Dr's and their response is its brain cancer and we have to operate, I think most folks would go get a 2nd opinion from a specialist. If your furnace goes out and the repair guy says it'll be 20K for a new one, I think most folks would get a second quote. Same should be said for our cars. We all know Tesla is haphazardly managed, each SC has a different approach and most of the time they are running around with their heads cut off with more to do than time and bodies to do it. I would not doubt in the least they'd take one look, say its hardware, give you an estimate and move on to the next. If service is pushing you for some physical part replacement that's not covered under warranty, push back. You have every right to have the same diagnostic data the techs are looking at, there is nothing so secret about the car that they can not or should not share with you the results that confirmed its a hardware issue.

As much most things these days, you have to be your own advocate because no one is going to fight for you.

 

[stopcrazypp]

It very well could be new code that's causing the error to appear, but I'm not sure what you are looking for in terms of case law or legal precedent. I think this is another instance of needing to take a different approach then one would with a traditional car manufacturer, think of the car as more akin to your phone/PC than your ICE car in the sense that when there are issues, ALWAYS start with the software.

If you have an iphone and call up support/go in for support one of the first things they'll do is reboot the device. Then they will factory reset the device, then depending on the circumstance either push a previous software build, or push a new software build to test. Once they have been able to reproduce the issue after a factory reset (no user data/configuration changes to blame for the issue) or on different software builds we can be somewhat confident its a hardware fault.

This should be the same approach taken by any tesla owner then encountering an issue that's possibly software related. Open a ticket with support, have them remotely read the logs, it can be very obvious if the hardware is causing a kernel panic is similar. However if the logs do not contain the smoking gun, a previous build or a preview build should be pushed to the car and we attempt to reproduce the issue. Once its been determined that the issue persists among different builds its a reasonable jump to start looking at hardware.

The problem is Tesla doesn't push previous builds, I have never heard of SCs doing so and not even sure if they have that ability. For some issues, they can push custom changes (with cooperation with engineering at HQ) but I have never seen mentioned someone ever getting them to push a previous build.

 

[yeahthatguy]

The problem is Tesla doesn't push previous builds, I have never heard of SCs doing so and not even sure if they have that ability. For some issues, they can push custom changes (with cooperation with engineering at HQ) but I have never seen mentioned someone ever getting them to push a previous build.

Historically they have had that ability, I know because it was done to my vehicle when our state AG threatened to sue tesla on my behalf, granted this was early 2021 at the height of the pandemic on much earlier software.

Now at various points in the software lifecycle they will have schema updates or similar that you can not revert as they fundamentally change some aspect, but there is nothing magical about tesla's firmware, going from XXXX.XX.1 to XXXX.XX.2 and then going back to XXXX.XX.1 should not be impossible. Going from software ver. 11 back to 10? unlikely as that is not a minor version change. Now whether or not one would need a jtag to reprogram subcomponents that have been updated is another story.

I'd like to see Mr. GreenTheOnly's take on the matter.

Now, please draw a distinction between "possible" and whether tesla will. Tesla didnt discount cars until they did, they didnt offer free supercharging again until they did, etc. Able to and will are vastly different.

 

[PianoAl]

What I'd like to understand better is... do owners think that this is okay? Is there a precedent for this?

Right, it's a interesting issue. There is a precedent, but I'm sure it's extremely rare. That is, a manufacturer makes a recall repair (physical) that unmasks another physical issue with your car. Yeah, extremely rare.

Here's a possible way of looking at it, using your case as an example. Let's say the GPS antenna was faulty from the start, but the problem was masked by the way the software worked. New software, and the problem is unmasked. Bad luck if it happens after the warranty has expired.

Compare that with this case: A lousy driver has a car in which the directional signals were faulty from the start, but he never uses them. After the warranty period, his new girlfriend (software upgrade ) drives the car and discovers the problem. Bad luck.

Not saying we should look at it that way. And note that if buggy software masked the problem, that's Tesla's fault.

 

[father_of_6]

And note that if buggy software masked the problem, that's Tesla's fault.

Or... if Tesla's software *fixed* a hardware issue, and then subsequently unfixes the hardware issue after your warranty expires, my opinion is... Tesla broke something that worked and they should pay for repairing it.

 

[father_of_6]

There is an existing thread in the software subforum (where this discussion likely should be, imo) on this topic:

Should Tesla break out-of-warranty cars with software updates?

Intentionally bricking all out-of-warranty cars at the same time would obviously lead to a PR catastrophe and expensive lawsuits. This will never happen. But if Tesla breaks just some cars in somewhat minor ways, would not this lead to added service revenue (resolving issues) and car sales (for...

teslamotorsclub.com

Thank you for pointing that thread out, it very much hits on the point of this thread, although with a different (perhaps adversarial) approach.