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

17.24.28 released - improved 'silky' update from 17.22.46?

This site may earn commission on affiliate links.
I saw conflicting rumors about it being in the US, and it has a handover cutoff of 37mph or something around there.

Audi has been talking the "Piloted Assist" talk for years. We even heard this with the A4 and Q7 having substantial lane keeping abilities and that turned out to be an utter disappointing combination of a low-speed-two-follow-car ping pong and late accel/late braking mess. And 3 second hold when coming to a stop.

I'll believe these other systems when I see them. Would absolutely LOVE to be proven wrong, and would love for Tesla to have some on-the-road competition to motivate them to drive AP2's timeline more aggressively.

It is speed limited but it will be level 3 (it will be rolled out only in Germany in late 2017 and then to the US is my understanding). That being said, I think everything sounds great in theory but often fails to match the hype when revealed. Audi's system likely will be the same (and I'm also thinking of Caddy's eye sensing LiDAR mapped system).
 
  • Love
Reactions: chillaban
It is speed limited but it will be level 3 (it will be rolled out only in Germany in late 2017 and then to the US is my understanding). That being said, I think everything sounds great in theory but often fails to match the hype when revealed. Audi's system likely will be the same (and I'm also thinking of Caddy's eye sensing LiDAR mapped system).

I'm still kind of miffed that Audi's ultrasonic automatic parking has been available in Germany for literally 5+ years but still has not arrived in the US except by accident in a Golf config that was kept the same in EU vs US. The rest of them have repositioned, cost-reduced parking sensors that were no good for automatic parking.

But yeah, I'd love to see the system put to the test when on the market. Speculation about things that are supposed to be out "soon" has been a very disappointing hobby, as someone who has been spending the last 5 years and attempting to buy cars with advanced automation features.


To give Tesla credit, even AP2 v17.17.4 is easily better than any other TACC and lane holding system I've tested. The 2017 E class Drive Pilot I would give third place behind Tesla's AP1/AP2 implementations. They are smoother in certain situations but also come dangerously close to clipping offset vehicles and not reacting quickly enough to traffic changes.

And I think a lot of that expertise comes from Tesla. If EyeQ3 were as simple as hooking in 2 cables and mounting a camera / radar, I would expect the market to be flooded with excellent AP1-like systems.
 
  • Informative
Reactions: croman
Dude I question exactly that: do we have different performances because of camera positioning and dynamic cropping?

As I typed above this my s75 just completely swing into another lane on 24.28 on the highway on a straight lane. Oh boy

I think there's a set of potential aspects that make this difficult to unravel:

1. Potentially bad QC on the camera position/orientation.
2. Different areas in the country having different styles of roads.
3. Different expectations by different people.

I don't think any single one of these covers all cases(the example you gave is not covered by #3, for example, and we've seen both problem and lack of the same problem reported by people living in the same area). Tesla really needs to up their game on the QC side of things, but #2 and 3 are going to give them trouble for a long time. There's no easy solution to unravelling them.
 
  • Like
Reactions: buttershrimp
Are you ready to lose your mind?

My car was 17.17.4, loved it. It was on shop so I had a loaner 17.24.28 which I expressed my frustration here already.

Now, my car has 17.24.28 just got from SC and it's driving like a lady. Smooth. I just drove same version on loaner 30 min ago and it's a horrible experience.

I got the AP2 loaner w 7 miles, mind has 13,000 miles. Could this be the difference?

I'm impossibly curious now.
 
Are you ready to lose your mind?

My car was 17.17.4, loved it. It was on shop so I had a loaner 17.24.28 which I expressed my frustration here already.

Now, my car has 17.24.28 just got from SC and it's driving like a lady. Smooth. I just drove same version on loaner 30 min ago and it's a horrible experience.

I got the AP2 loaner w 7 miles, mind has 13,000 miles. Could this be the difference?

I'm impossibly curious now.

After 5 minutes you are ready to declare victory? I can't believe anything you are saying. As @dpilone said -- AP doesn't even work without calibration which takes between 35 and 100 miles at certain speeds with well marked roads. Secondly, why the hell would Tesla loan out a new car when they have tons of CPOs with plenty of miles on them?
 
After 5 minutes you are ready to declare victory? I can't believe anything you are saying. As @dpilone said -- AP doesn't even work without calibration which takes between 35 and 100 miles at certain speeds with well marked roads. Secondly, why the hell would Tesla loan out a new car when they have tons of CPOs with plenty of miles on them?

Make that 50 miles on sections 24.28 couldn't handle on loaner.

I don't know how the car thinks maybe how I drove the loaner during calibration messed it up?
 
  • Funny
Reactions: croman
Make that 50 miles on sections 24.28 couldn't handle on loaner.

I don't know how the car thinks maybe how I drove the loaner during calibration messed it up?

I don't think driving style affects calibration. It just needs to see road markings for a while to ensure it is calibrated properly. I believe a SC can also calibrate or test the calibration. At any rate, I do not think you can make conclusions about how any firmware operates after one drive or even one day. So much varies in terms of time of day, road conditions, lane position, etc.

Hopefully you are still satisfied with AP and believe it is better than the loaner after a week. Incidentally, is your actual car AP2?
 
Are you ready to lose your mind?

My car was 17.17.4, loved it. It was on shop so I had a loaner 17.24.28 which I expressed my frustration here already.

Now, my car has 17.24.28 just got from SC and it's driving like a lady. Smooth. I just drove same version on loaner 30 min ago and it's a horrible experience.

I got the AP2 loaner w 7 miles, mind has 13,000 miles. Could this be the difference?

I'm impossibly curious now.

Now that is interesting. In my list, that leaves only #1 as a possibility. Tesla needs to solve these QC problems ASAP. Their software guys are hitting it out of the park only to be undercut by crappy QC.
 
I don't think driving style affects calibration. It just needs to see road markings for a while to ensure it is calibrated properly. I believe a SC can also calibrate or test the calibration. At any rate, I do not think you can make conclusions about how any firmware operates after one drive or even one day. So much varies in terms of time of day, road conditions, lane position, etc.

Hopefully you are still satisfied with AP and believe it is better than the loaner after a week. Incidentally, is your actual car AP2?

Yes sir it's a Model S75 AP2

I don't think I have a way to make myself clearer how 24.28 on loaner AP2 was irritatingly useless.
 
Now that is interesting. In my list, that leaves only #1 as a possibility. Tesla needs to solve these QC problems ASAP. Their software guys are hitting it out of the park only to be undercut by crappy QC.

Everything is a combination of software and hardware. Sometimes I will spend 30-50% of my year as a software engineer working on something that could be fixed in hardware, but at the end of the day, it was a better decision overall to add a software solution. It's not about blaming one department or another, the car as a whole is Tesla's product and it should be a mutually beneficial decision where issues are addressed. At the end of the day, if it's a cropping problem, everyone shares the blame in deciding to go down that road, and the quality of the software that implements it.
 
  • Like
Reactions: Rafaelm
Some things have changet at least with the AP backend: on my 2014 AP1 with 17.24.28, the speed limit DB / detection has changed from earlier firmwares: I keep getting lots of faulty detections. AP Speed restrictions are now applied to even more roads, and often wrong: 70 km/h in 80 km/h zones, detecting 50 and 60 km/h local roads as 80 km/h (which is the legal default) and resetting speed limit in between signs etc...

I got my new X 1 week ago. The car was delivered to my house. It only had 35 miles when I got it. AP was available already

Maybe something was changed
 
All i see in this thread is the follow:

Poster A: AP2 is so much better!
Poster B: AP2 has regressed/doesn't work!
Poster C: AP1 is so much better, here's a video
Poster D: AP1 has regressed so much!

Repeat pattern.

Part of this is the lack of release notes at least for AP1. I installed it yesterday, and the release notes don't say what's changed. It just said enhancements and bug fixes. That was it.

The only thing I noticed was all the easter eggs were in a handy pull down after I pushed the T on the center console (to view the version).