-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
-
Tesla's Supercharger Team was recently laid off. We discuss what this means for the company on today's TMC Podcast streaming live at 1PM PDT. You can watch on X or on YouTube where you can participate in the live chat.
Wiki Sudden Loss Of Range With 2019.16.x Software
- Thread starter Dutchmeeuw
- Start date
No fiddling with CAN bus until my CPO warranty runs out.
The 2020.36.11 S85 is my car, I know that the battery was well conditioned. Approx 100km drive on German Autobahn up to 150 km/h and Supercharger as navigation destination.
If SMT is attached to TeslaLogger, SMT data will also be shared and plotted in those diagrams, see here Charging
The 2020.36.11 S85 is my car, I know that the battery was well conditioned. Approx 100km drive on German Autobahn up to 150 km/h and Supercharger as navigation destination.
If SMT is attached to TeslaLogger, SMT data will also be shared and plotted in those diagrams, see here Charging
Tesla can't really just change the firmware of the BMS without updating almost everything else on the vehicle to accept the BMS's changes. Even ignoring the functional change with CAN messages that have been updated and such, the bootloader on the BMS reports a hash of the BMS firmware to the rest of the vehicle, and it will reject it with a "Software update required" error if it doesn't match the hash expected by the rest of the vehicle. In the case of the BMS, this would be super obvious because the HV battery wouldn't be allowed to engage, the 12V would quickly die, and the car would be dead.
There also is no in-car mechanism for updating module bootloaders on the S, so faking the hash isn't really practical. (This does exist on the Model 3/Y, only for the BMS bootloader, as far as I'm aware.)
As for forced updates in general... I know they've done this in the past on a handful of occasions. I've been a target of it myself, prompting a Twitter response from Musk back in the day. (Some time later a somewhat disgruntled ex-Tesla employee on a hacker forum claimed responsibility for it, but can't confirm.)
There are also instances where some cars were never properly marked as "delivered" (or otherwise got unmarked), and thus were considered to be owned by Tesla by the firmware update system. Vehicles owned by Tesla that receive updates get the updates initiated automatically overnight in most cases. (There's fun videos of people being in service center parking lots in the middle of the night watching headlights flash and such during the updates. haha)
I've only heard of a handful of (unconfirmed) accounts of Tesla forcing updates on customer vehicles. In the few cases of this I've personally examined, the logs clearly showed someone physically at the touchscreen accepting the scheduled install (and when asked, the owner confirmed they used the vehicle at the time indicated)... so I think this is more likely more akin to the "I swear I was hitting the brake" defense against sudden acceleration. Just, in this case, the user is hitting the button to schedule the install without realizing instead of canceling the incessant popup at some point.
I've not examined any vehicles related the the issues in this thread that show Tesla forced an update of any kind.
TLDR: No, Tesla is not secretly updating the BMS firmware of cars on older firmware versions.
There also is no in-car mechanism for updating module bootloaders on the S, so faking the hash isn't really practical. (This does exist on the Model 3/Y, only for the BMS bootloader, as far as I'm aware.)
As for forced updates in general... I know they've done this in the past on a handful of occasions. I've been a target of it myself, prompting a Twitter response from Musk back in the day. (Some time later a somewhat disgruntled ex-Tesla employee on a hacker forum claimed responsibility for it, but can't confirm.)
There are also instances where some cars were never properly marked as "delivered" (or otherwise got unmarked), and thus were considered to be owned by Tesla by the firmware update system. Vehicles owned by Tesla that receive updates get the updates initiated automatically overnight in most cases. (There's fun videos of people being in service center parking lots in the middle of the night watching headlights flash and such during the updates. haha)
I've only heard of a handful of (unconfirmed) accounts of Tesla forcing updates on customer vehicles. In the few cases of this I've personally examined, the logs clearly showed someone physically at the touchscreen accepting the scheduled install (and when asked, the owner confirmed they used the vehicle at the time indicated)... so I think this is more likely more akin to the "I swear I was hitting the brake" defense against sudden acceleration. Just, in this case, the user is hitting the button to schedule the install without realizing instead of canceling the incessant popup at some point.
I've not examined any vehicles related the the issues in this thread that show Tesla forced an update of any kind.
TLDR: No, Tesla is not secretly updating the BMS firmware of cars on older firmware versions.
Jeez. I never claimed that there was a difference or what method is involved.
I have just been trying to help people to distinguish when something is a hardware problem or when it's not.
People are on this forum all the time freaking out about, "I've lost 9 miles off the display over the last two weeks. I think my battery is dying!!!!1!!1!!" And we have to reassure them that this is probably just drift of the estimation, and they don't need to worry about it. But I am simply pointing out that the behavior is a little different if the display drops 20+ instantly. That's not normal estimation drift, and it is a fairly well known hardware problem that is worth a service call.
That's it.
But all people wanted to do was disagree and argue with me. So nevermind trying to help anymore.
And I do hope my comment didn't sound aggressive or unappreciative. I was actually trying to agree that really the exact cause might not matter to many and it isn't in doubt that it happens.
I think that any factual and careful observations have a lot to offer.
As you observe, the noise can - if not challenged or clarified - obscure the real issues.
Last edited:
Here it comes .. And now the main stream media is running negative articles and looking into battery fires . .
. . for GM, Ford, Hyundai and BMW. No mention of Tesla.
Auto Makers Grapple With Battery-Fire Risks in Electric Vehicles
Well they Force Downgraded me from V8 to V9 without me accepting anything. If fact, I had WiFi disconnected on the car and the drivers seat power disconnected, as at one time we believed that could stop updates. I had been refusing the update for a very long time, and certainly never accepted it. One day maybe a few months after others reported being force downgraded on to V9, I go out to the car and when it wakes up it does some blinking and I'm on V9 and my WiFi is back on. I assumed I would have to enter the code again for WiFi to come on again, obviously they stored the code, wish I had thought of changing it. I have yet to hit the gas instead of the brake.
BigNick
Infamous Fat Sweaty Guy
I was running 2020.36.11 when charging at the rates referenced in my previous post (and still am, I haven't received a 2020.40.x update.)
I don't know if 2020.40.x would be faster or if what I was experiencing is now considered normal.
At one point I would peak at 128 kW (over a year ago, but after 2019.16.x), and on a trip to Luray Caverns, I was still pulling 115 kW after about 5 minutes plugged in at 14% SoC. I only remember this because I have a screenshot on my phone. This was 2019-05-31 at the Strasburg, VA Supercharger. The charge had finished before we finished eating our lunch at Denny's.
Would love to examine the logs from your car at some point, because this whole interpretation of this scenario sounds pretty unlikely.
First, turning off WiFi only persists "until the next drive."
Second, WiFi settings used to never be forgotten pre-V9. You could "remove" them, but they were just marked as removed. This behavior changed in V9, and removing them actually removed the settings. I personally had several networks I'd previously removed re-appear when updating to V9+.
Additionally, WiFi obviously isn't required for updates. Once you see the update popup (as you say you "had been refusing the update for a very long time"), your car has already downloaded the update and staged it... so connectivity is moot at this point and is in no way required to "install" the update.
Finally, disconnecting the driver seat power won't do anything in most cases, as there have been no updates to the driver memory seat module for quite some time on most cars. So an update wouldn't fail because of this since maybe about 2015. The update process was finally streamlined to only bother checking modules that have new firmware to install vs the currently installed version in order to speed up installs. The installer itself always comes from the update that is being installed, and thus even if you were on V8, your car would use the updater associated with the version being installed... one that would have ignored the missing module that didn't need an update anyway.
So... yeah, sorry: You, or someone else in your vehicle, hit the "schedule install" button instead of dismissing the popup.
Can easily Occam's Razor this one away, just like the "my car suddenly accelerated" nonsense:
Did Tesla go out of their way to force an update on a handful of vehicles where people had been rejecting updates regularly (but not all such vehicles), requiring either manual intervention for each one at a higher internal technician level (to the best of my knowledge not many people at Tesla even have access to pull off a forced remote update), or made complicated changes to their entire firmware update backend to do this...........
Or did someone just eventually hit the wrong button on the touchscreen?
Would bet on the latter, and I'm sure logs would back that up.
Last edited:
So am I to believe that Tesla can't make the log look like I pushed the install button? Did all these people make the same mistake after doing all they could to not end up on V9?
Member Connected to Wifi? Modem Update already downloaded?
HankLloydRight No LTE No
swegman No LTE No
Electroman ? LTE Unsure
Cars_cars_cars Yes 3G Yes
mvotb No ? Yes
Vern does not have Yes
Quote by wk057
"Tesla can and does force updates, and they do it in a way that looks the same in the logs as if you scheduled it yourself. I've seen it done to at least a half-dozen cars. Not really sure the reasoning, but they definitely do it. The can also force it from garage easily, which is how they do it on Tesla-owned cars to auto update them, but I've not seen that method used on a customer car."
Have you changed your mind?
I wish they would publish the battery manufacturer as that might be more useful knowledge to have. From my recollection, Samsung makes the BMW battery and LG makes the GM battery.
Thats... not a lot of people. Plus, as noted, WiFi doesn't matter. If you're connected to 3G or LTE you can and will receive updates.
Yeah I have, and actually I clarified this in my post above. They've manually done this to a few people, such as myself, for various reasons. Every other case I've examined related to forced updates were either due to user error (hitting the install button) or an error on Tesla's side related to the car being marked as undelivered. The latter was fixed years ago. And as of something like 2016 (as far as I'm aware) it takes a technician with a reasonably high level of access on Tesla's side, of which there are not too many, to be able to pull off a forced update... and really, those folks have better things to do.
Huh. This will be really interesting seeing @wk057 argue against himself--in the earlier thread saying Tesla does force updates and make the logs look like the user did it, but now saying that never happens.
Unfortunately my posts are still on "moderation", so this will (eventually) show out of order with replies and such along with my slightly earlier, but delayed, post addressing this above ^, but didn't touch on the specifics this particular point, which is of course valid criticism. My apologies for the popping-into-existance-out-of-order-posts.
Back when I was looking into this issue before, I did originally find evidence that Tesla had forced an update on a vehicle doing so in a way that made it appear like the user scheduled the update in the logs. This, however, turned out to be an artifact of the method used to remotely do this and not intentionally trying to fabricate logs. I have a much better understanding of the proprietary logging format today than I did previously, and the method they used is actually easily differentiated from an actual user pressing to schedule.
Hope that clears this up.
No, it certainly doesn't clear it up. What you just described is how a person could detect the difference.
But in post #13631, what you did was really arrogantly talk down to @mvotb , dismissing his idea by telling him it certainly couldn't and didn't happen, and implying he was an idiot or liar:
"So... yeah, sorry: You, or someone else in your vehicle, hit the "schedule install" button instead of dismissing the popup.
Can easily Occam's Razor this one away, just like the "my car suddenly accelerated" nonsense:"
And then when you were caught by him quoting you, admitting before that Tesla CAN and DOES push updates without user authorization, you seem to backpedal, saying that yes, they can do that, and by inspecting the logs, it's possible to tell if that's what happened. But still no apology for your attitude and assumptions.
So no, this is still a case of really bad hypocrisy and being insulting and then trying to excuse it away.
I generally appreciate how informative and knowledgeable your posts are, but you f&@$ed up big time here and got caught. You knew this was something that happens, but were really mean in telling someone it's not possible and didn't get away with it. Geez. Just own up and apologize.
Similar threads
