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

Software Update 2018.48.x

This site may earn commission on affiliate links.
I'm also having to reboot my MCU1 every day now. Pin to drive keypad seems to move around the screen and no longer be central. I've also had it freeze. Bluetooth is a mess. Delays answering after I click answer call with the wheel button and the muted issue also for me. Web browser will actually load a web page for me now. But after viewing a few pages seems to crash and go blank.
Tesla really, really, need to get some better software validation procedures in place before they unleash these dodgy updates on us all. Amazingly they've re-introduced the not being able to log in to Spotify bug again, that was only fixed in the last update and it's back, what the!!
I don't have Twitter but if anyone does can they tell Musk to do a Steve Jobs and walk in to the MCU programming team and tear shreds off them, they seem to have lost their way.
 
They need a QA department, testing all hardware and software revisions. And I do mean actually testing it. not just sitting with a virtual machine or whatever in the office with top wifi connection.
Tesla MCU QA department deciding on release stability.

1515790985810.jpg
 
Having done my emulator time this smelled like fish from the beginning. This deserves an incredible amount of attention. I wonder if they are subtly trying to gauge the backlash from MCU1 owners when real features start to be differentiated. For the record the Tegra2 platform is not as shitty and behind the times as many might think. In my opinion the UX could be radically improved from its current state.
This is pretty possible if not likely the case!
 
  • Like
Reactions: bollar and am_dmd
It works for me. 2018 MS.

2) same here
3) They were fairly responsive. If you press and hold the driver's seat, all of them will go off. There's a communication via LTE, cloud services, and then the car, which has to send a response back that it has actually been done. There should be something that blocks the user from taking further action on one seat until the change is confirmed, if the button has been idle (not 2 consecutive presses).
4) Yeah it does nothing which is kinda sad.
6) Must be buggy software which doesn't account for lag in services, and the state is out of sync.
7) yup sync error.
Yea, I tried it again, it seems the app is fighting with the car what the setting should be, resulting in a completely undefined setting where the app and GUI and the actual controls are completely out of sync (e.g. app and big screen in the car shows steering heat wheel is on, but the steering wheel heat doesn't know that, so it really isn't). A true example how Tesla releases their software, slap something together which looks pretty, let the customers figure out what doesn't work. Honestly, really sad. I work with automotive companies and yes, they take their sweet time and it's frustrating sometimes, but at least they think about testing the stuff they release so it works for a large majority of users.
 
Last edited:
  • Like
Reactions: rnortman
It works for me. 2018 MS.
They need a QA department, testing all hardware and software revisions. And I do mean actually testing it. not just sitting with a virtual machine or whatever in the office with top wifi connection.
I don't think the guys who code and verify the UI ever run it in a real car. Let me give you an example. In 2013, I submitted a bug about the fact that whenever I turn on my headlights automatically, when I try to get out of the car I get warning beeps and a message which says my range will be affected, which is not true because the headlight turn off as soon as I lift off the driver's seat. So, a few months later they fixed, it, the warning text was gone but the warning tones were still there! My best guess is, whoever got assigned the bug just suppressed the visual warning but they didn't even have a speaker connected to realize they only suppressed the display of the warning, not the full warning including the dings. About 3 years later the visual warning returned (someone noticed the hack in the code?), now is shows both the useless warning and the dings. The car trained me to ignore any warnings on exit - one time somehow the car didn't shift into park (maybe I didn't press hard enough on the stalk, I don't know) and as I was getting out it dinged at me to tell me that the car is still in drive, I plain ignored it because warning dings on exist are just normal with Tesla (on entry too, I get some warning which was not able to catch on the screen because it beeps and then goes away before I ca read it). Lucky the car automatically shifted to Park as I was getting out, but still, why have useless warnings?
 
Anyone seeing charging problems on S or X? My X has been stable for over a year charging to 80%, goes down to 77 and then back up to 80 when in the garage on the HPWC.

After 48.12.1 first it went to sleep ‘too deeply’ and charged to 97% (great) and now is down to 75 without starting a charge.

They can’t seem to leave things well enough alone. I rebooted as I always do after loading the software and again after the 97% fiasco. Still waiting to see where it starts charging automatically but was wondering if anyone was seeing changes like this.

I changed to ‘energy saving OFF’ to see if that helps as I’ve also been getting the issue 3’s have been seeing where ‘saving’ lets my MCU 1 sleep like an MCU 2 to the point the app can’t wake it up. (Yes, they don’t have the setting but they go too deeply asleep. Our 3 does that.)

As was said, QA really is getting weak.


Edit: charging started at 74%. Let’s see where it stops.... and it did stop at 80... wonder WTF it did when it loaded the software and charged.
Because it went to sleep it didn't follow its own rules? I had this sort of thing months and months ago, where it seemed to make no sense that it didn't charge right when it couldn't reach the mothership, something you'd think would be 100% a local function. Arg....
 
Last edited:
Honestly, really sad. I work with automotive companies and yes, they take their sweet time and it's frustrating sometimes, but at least they think about testing the stuff they release so it works for a large majority of users.
As do I and this is what kills me watching what Tesla do, I know the software validation procedures some of the other OEM's go through which does mean they are slower but the big difference at the moment is if they don't get it right first time then it could cost them money or risk losing a customer for life. This is because they require all owners to bring the car in to the dealer to get updates, just imagine as Tesla owners we didn't get over the air updates. So you'd waste your time going to the dealer to get a bug fixed only to have to return a month later to get another bug fixed that got introduced with the last update! And the cycle repeats. The problem is because Tesla don't need to do that there is less importance put on getting it right the first time because they can just deploy another over the air fix.
To me it seems like they don't have a good system in place to ensure that fixed bugs don't reappear. I know not everyone has Spotify enabled vehicles but we've gone from a previous update where users could not log in to getting that fixed with an update to now with the last update we can't log in again, that stinks of the software team not doing thorough version control and someone has put some old code back in.
I don't care how many Easter eggs they want to add in, I think it is cool they like to set themselves apart from the cold stone walls of traditional auto makers in that sense, but not when there is functional issues with how the MCU operates and as I said in a previous post, those Easter eggs don't write themselves, and when the code seems to live on a knifes edge with stability and everything remaining functional adding in 'clutter' does not help the situation or appease the minds of those that need to reboot the MCU regularly to have it operate as intended.
As it is on my car with that last update the I've lost Spotify access, the MCU is unusable for the first 5 mins of a drive because it is totally unresponsive (which can include reverse camera), someone has decided to change USB music to listing songs in alphabetical order (well, I can't use Spotify so back to USB). Sorry but it is sloppy software validation that allows this to happen, the software department needs some better guidance, this isn't a phone app, it's a $100K+ car for some of us who bought the big banger.
 
Last edited:
To me it seems like they don't have a good system in place to ensure that fixed bugs don't reappear. I know not everyone has Spotify enabled vehicles but we've gone from a previous update where users could not log in to getting that fixed with an update to now with the last update we can't log in again, that stinks of the software team not doing thorough version control and someone has put some old code back in.
Usually in a structured software team, for each bugfix, an effort is made to reproduce the error programmatically in an unit test. Then that unit test is one of many tests that is automatically executed for each version you do. That way old bugs don't find their way back unnoticed.

Doubt someone has put old code back, but there are clearly lack of quality measures on the software.
 
  • Informative
  • Like
Reactions: croman and P100D_Me
ChrML, imagine the MCU operation could be open sourced, I run an open source Android build and it is so much better than any OEM effort a company has attempted to improve. I understand why the 'brain' of the car could never be open source but I think it would yield amazing results in stability and functionality.
 
  • Like
Reactions: bollar and KArnold