Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Software Update 2018.6

This site may earn commission on affiliate links.
Funny you should say that. I've been experiencing the exact same thing.

Car is very new, no alignment problems, and the left-side bias is huge. And, it always happens (100% repeatable) in the same spots on certain roads. In other places it is fine.

The dash display shows the car hitting the left lane line, the wheel will vibrate with the lane departure warning, and the side sensors light up red as it almost grinds me into the divider! So the car "knows" what it's doing is wrong in three different ways, yet autopilot is super determined to do it, and keep doing it, anyway.

Something in the motion control side?

This problem is new with 2018.4+ for me.
I am experiencing the exact same thing and it all began with the 2018.4 series. And I thought the left lane hugging was a problem long since resolved. Guess not. From a safety of driving perspective this series of releases has been very, very disappointing.
 
  • Informative
  • Helpful
Reactions: R.S and crashvt
So I got the update and I am still experiencing the parking sensor failure. I no longer have confidence that Tesla even knows what the cause is or has any idea how to fix it.

I've been told that the "next firmware" will fix this for the past 2 months. I'm sure they will eventually fix it but it could also be hardware that they are blaming on firmware. Its just one of those intermittent issues that is hard to pinpoint.
 
As a software developer, I would bet dollars to donuts that there isn't a hard six (or eight) week "schedule" for fleet-wide FW update rollouts.

Correlation is not causation.

I think when they have an update that's has some significant new features to be released, or critical bug fixes, and has been test-rolled out to a small number of cars, they do a fleet-wide release "when it's ready."

Just because it looks like a scheduled release program doesn't mean it is. For instance, the HO HO HO easter egg was hastily rolled out one day before Christmas. Luckily it didn't introduce any regression bugs (well, that we know of so far -- ha ha )

As a software engineer, what I would bet is that something like 6 to 8 weeks is the practical lower limit for time between releases: not due to development effort, but due to testing/QA being the big time/effort sink. There has to be a complete test suite on the software: how long does it have to run? (given the complexity & regulatory environment, I'd guess *days*). How about a longevity test (looking for bugs that only show up after running for some time): I'm thinking a long weekend, as a minimum, and a week would probably be closer to the mark.

As to a "limited release" before going fleet-wide: great idea, but you want to choose your beta testers so that they *test* the release (and provide timely feedback).
 
  • Informative
Reactions: pilotSteve
As a software engineer, what I would bet is that something like 6 to 8 weeks is the practical lower limit for time between releases: not due to development effort, but due to testing/QA being the big time/effort sink. There has to be a complete test suite on the software: how long does it have to run? (given the complexity & regulatory environment, I'd guess *days*). How about a longevity test (looking for bugs that only show up after running for some time): I'm thinking a long weekend, as a minimum, and a week would probably be closer to the mark.

As to a "limited release" before going fleet-wide: great idea, but you want to choose your beta testers so that they *test* the release (and provide timely feedback).
So do we assume the three versions of 2018.4 were minor corrections pushed out to update things to a small group for feedback? I think that was the most updates to a weekly version ever, and we assume (!) that .6 is the final version of this. Maybe.
 
I've been told that the "next firmware" will fix this for the past 2 months. I'm sure they will eventually fix it but it could also be hardware that they are blaming on firmware. Its just one of those intermittent issues that is hard to pinpoint.

That’s the problem with OTA updates in my experience. They are great if they add functionality(sometimes they also break stuff, but that’s not the intention), but they are a lazy excuse for not fixing things, right away.
 
I figure this is worth mentioning: My S100D w/ AP 2.5 was in the SC yesterday and due to some weird glitches I was having with 2017.50, they forced an update. It was only updated to 2018.4.7 3ee0f17.

Sign that .6 had some bigger issues and stopped rollout?