Chủ Nhật, 30 tháng 10, 2016

Firmware 7.1 part 30

  • 1/1/2015
    guest
    With over 9000 miles on autopilot including a recent cross country trip, I have experienced the autopilot triggering the lane departure warning on multiple occasions, sometimes more than once on the same day. Not to say that it is frequent but neither is it rare. Beyond the obvious that it seems to depend on the nature of the curves than on the quality of the lane marking, I don't have enough data to detect a pattern.

    However, I don't think there is anything seriously wrong with the programming. Actually, the behavior makes sense when you take into account modular program design. While they are processing the same video data, the lane departure warning and the autosteer decisions are made by separate modules with different design goals. Lane departure just decides where the car is relative to the lane at this instant. Autosteer is adjusting the steering to keep the car centered in the lane using a feedback loop while avoiding the overcorrection that leads to oscillation (ping-ponging) around the center of the lane, even while the lane is curving. These very different algorithms should be in agreement about how to position the car most of the time, as seems to be the case, but nothing requires them to agree all the time.

    I don't expect autosteer use the input from the lane departure warning. I would expect that side collision avoidance, as a separate module, would override the base autosteer, when needed. Getting close the edge of the lane is not an emergency as long as there is no vehicle or barrier nearby.
    �
  • 1/1/2015
    guest
    Thanks for the explanations. I am the opposite. I know that bugs gets introduced. The media player running overnight issue was fixed and broken again and fixed again, etc. several times from what I read. Seems like you stand a good chance of the update breaking it again if it is working. I also know they have pushed updates that remove features like auto-lowering of the suspension (boy, were the first updaters mad when they discovered that!). So for me, if my car is working the way I like I am happy to keep it that way, at least until I have had a chance to come here and find out from you guys what effects, good and bad, intentional and unintentional, the latest update has.

    Thanks, first updaters, for being the guinea pigs and sharing your findings with the more cautious among us!
    �
  • 1/1/2015
    guest
    My pleasure. You're welcome!
    �
  • 1/1/2015
    guest
    Well, Canada... ;)
    �
  • 1/1/2015
    guest
    I would buy that statement if the sensor suite was capable of detecting all potential hazards outside the lane, including drop offs, soft shoulders, and low curbs, but it is not. I think many people find that auto-steer is unduly cavalier about the sides of the road.
    �
  • 1/1/2015
    guest
    Yup, you're right!
    �
  • 1/1/2015
    guest
    I was really posting mainly because I'm still somewhat annoyed with how sensitive the lane departure warning is on the undivided roads that are so common where I live, and how Tesla refuses to offer a sensitivity setting for this option, in spite of many people offering similar complaints and recommendations. I even went so far as to make the following video, and send it to Tesla, over a year ago, to demonstrate the problem.



    "Thirteen Lane Departure Warning alerts. Listen carefully. They can be heard at approximately :18, :25, :31, :46, :55, 1:03, 1:20, 1:29 1:43, 1:52, 1:57, 2:16, and 2:21. Obviously I was driving in a way intended to generate these alerts, to demonstrate the fact that the system is overly sensitive. P85D, Firmware version 6.2.2.4.236."
    �
  • 1/1/2015
    guest
    I agree with you to a point. Use of autosteer is nominally "limited" to divided highways, which rarely have those hazards, unless you are in a construction zone on the highway (and those are common). That said, Tesla also enables the autosteer on "other" roads, albeit sometimes with speed restrictions based on the software's interpretation of location and speed limits.

    I like to push the technology as much as the next person, but I am on alert and not trusting the software when I do so. I appreciate that Tesla treats us like adults, permitting us to push the envelope. That said, I feel that Tesla is a bit cavalier in publishing an optimistic marketing spin on the current capability while keeping the lawyers happy by making us agree that the driver remains fully responsible. More transparency from Tesla engineering about situations that challenge the technology would help more people understand when it is most appropriate to apply the technology and when to be extra cautious, rather than getting people frightened by bad experiences.
    �
  • 1/1/2015
    guest
    Wow, that is crazy.
    �
  • 1/1/2015
    guest
    Hi Andy:
    I have noticed that in your video, your county or town is using "micro paving" to caulk up cracks in the highway. So does my town and county.

    That said, I have noticed that at times, the linear "micro paving" lines will trigger the lane departure warning subsystem. I speculate that the cameras view the darker micro paving lines as lane lines somehow. This was evident in some of the areas of your video as well.
    �
  • 1/1/2015
    guest
    The "micro paving" or tar strips can trigger a lane departure warning in certain lighting conditions.
    �
  • 1/1/2015
    guest
    (This is similar to a comment posted on Youtube with the same video.) What I cannot tell, as a driver, is where the car is relative to the lane line when the lane departure warning (LDW) is sounded. The driver's visibility is limited. I can see the creases on the car's hood, but they are quite far from the edge of the car, and are curved as well. So how does one know the actual location? In other words, I am at a loss to say if I think the LDW is overly sensitive, or not. (Leaving aside erroneous detection of markings as mentioned in th preceding post.)
    �
  • 1/1/2015
    guest
    That is definitely not what is going on here. I've driven these roads in all lighting conditions, and the issue is the system is simply too sensitive.

    By the way, to address Art's comment, only the roads in the second part of the video have that micropaving. I don't think you'll see that on the road in the first part of the video.


    One thing I did as a test was I found some lane markings that were "scored", to create a real "rumble" when tires drove over them. I could easily get the Tesla lane departure warning to go off without having my tires on the lines. The "scoring" of the lines actually extended a bit to the inside, so the Tesla system was alerting well before any wheel touched a line. Perhaps some people want this level of sensitivity. But many of us don't, and would prefer to be alerted much less, and only in much more serious situations, like when our wheels are actually crossing the lane markers.
    �
  • 1/1/2015
    guest
    Andy:
    I watched your video again and in many of the "warnings" micro-paving is present. Not all. but many.

    Were you on full autopilot with lane keeping? IMO the car seems to wander from left to right and back again. If you are using lane keeping, then that is the type of road I personally would NOT use it on. Too many side streets with perpendicular entrances for my taste. But I am really cautious I suppose.

    I would agree with you that I want to be alarmed much less in this case of yours. Like ...not once even.
    �
  • 1/1/2015
    guest
    I disabled lane departure warning for my driver profile. It was annoying in city driving where the road is shared using rules that do not line up with lane painting. When you flow with the traffic there is no danger in crossing a lane marker so I shut it off.
    �
  • 1/1/2015
    guest
    I agree. This is now my 3rd car with a LDW system and I turn every one of them off. Except the advantage with Tesla is you still get side collision warning/mitigation despite turning LDW off (if you veer too close to a car beside you and the ultrasonic sensors pick it up, you'll get an alert)

    I found the systems to be generally unhelpful because they tend to kick in when I'm intentionally departing my lane like making room for a motorcycle or following the correct fork on the freeway.
    �
  • 1/1/2015
    guest
    And you are the only one, at least on Tesla Firmware Upgrade Tracker Web App is just one report (you?), but 10 more after that on 2.24.102 :-o
    Also the last version on teslalog was 2.24.102 (beside some 103.1.x versions, some V8 tester I suppose).

    2.26.103 is probably a custum build just four you. ;-) :roll:
    �
  • 1/1/2015
    guest
    Probably a rude message just for WK057 embedded in the code:)
    �
  • 1/1/2015
    guest
    I turned mine off after a few days as well.
    �
  • 1/1/2015
    guest
    Curious, have you seen any efficiency improvement a la +10% Energy Efficiency after 2.24.102 Update
    �
  • 1/1/2015
    guest
    Same here. I wanted to like it. I wanted to use it. But it annoyed the hell out of me on narrow two lane roads with "false" alarms. BTW, the lane departure warning on my previous 7 series BMW was just as annoying and got turned off too.

    Mike
    �
  • 1/1/2015
    guest
    Diagnosis: Emotional attachment to a (delightful) aluminum can with wheels.
    Financial Analysis: Appears to be able to afford several substantial upgrades (if offered).

    I dearly love my car and all, but one of the things I am ready to do at any time is say goodbye.*

    It's just a car.

    Buy a new one... if (when) you can afford to, I say trade it in, once there are enough new features out.

    This is a technology platform, after all.

    * for a newer Tesla only.
    �
  • 1/1/2015
    guest
    It's a special one just for him which has been sanitized lol
    �
  • 1/1/2015
    guest
    No there's more than just him, @wk057 posted that he got this version too a few days ago. Their may be more that haven't updated Tesla Tracker yet.
    �
  • 1/1/2015
    guest
    I had .102 from July 3. Picking up my car from service tonight and just saw the notification on my phone that I have a SW update. Take it that .103 is waiting (unless I'm the first 8.0!)
    �
  • 1/1/2015
    guest
    Please inform us on your progress!
    If your version is higher than 7.1, I'll be parking at the Amsterdam HQ (SeC and SuC) first thing this morning. :cool:
    I'm currently on .102, so will drop by at that SeC within the next week anyhow. But .103 was of course already spotted, so chances for 7.2 of 8.0 are slim.
    �
  • 1/1/2015
    guest
    Ah, update fever! I understand.

    I wonder what it is like to have a car that didn't improve* regularly? I find my memory fuzzy for that period of life.

    *usually
    �
  • 1/1/2015
    guest
    Haha, that's probably sound advice. My first car was a Honda Civic and I (tastefully) modded the hell out of it. It was a sort of therapy and way to learn about it. Then one day, a distracted, speeding, self consumed, lunatic of a pregnant woman drove up a bike lane, through a red light and totalled my car, a taxi, and her decrepit van. I vowed never to mod cars again, and you just drew attention to my impulse.
    �
  • 1/1/2015
    guest
    Went to the service center but car would not connect to wifi there :-(
    IMG_1237.JPG
    �
  • 1/1/2015
    guest
    I'm starting to think the same thing. I spoke to someone locally who was just updated at the same service center... and received 2.24.102.

    I have not. P85D is just as big of an energy hog as always. Just did a 55 miles round trip and got 360 Wh/mi driving normally. 366 Wh/mile lifetime efficiency over 28.3k miles. 367 Wh/mi over last 7.5k miles. So, yeah... so much for the P85D being more efficient than the P85. :rolleyes: (For comparison, wife's P85 life average is 326 Wh/mi @ 30.5k miles)
    �
  • 1/1/2015
    guest
    Totally bizarre if they created a package just for you.
    �
  • 1/1/2015
    guest
    Picked it up - still at .102. For some reason they decided to reapply it. Sorry to get everyone's (including my) hopes up. :(
    �
  • 1/1/2015
    guest
    Consider using "Ignore" function liberally

    Maybe different driving style also? more ludicrous oriented driving style? !
    �
  • 1/1/2015
    guest
    ...but then again, my understanding from his mods and former posts is his MS *is* unique in the world of Tesla. Besides, I don't believe anyone here has ever been able to predict with 100% certainty exactly what Tesla is doing related to update frequency, numbering schemes, availability, or exactly what every update is for, anyway. All we can do is second-guess after-the-fact. :p
    �
  • 1/1/2015
    guest
    Tesla has hidden torque sleep in the range mode setting. Turn on "Range Mode" and you will experience a roughly 7% efficiency advantage over the RWD models. For whatever reason, I think Range Mode can only be turned on if the car is in Sport mode. I really don't understand why torque sleep can't always be running. It shouldn't even be an option.
    �
  • 1/1/2015
    guest
    Blasphemer! You're willing to sit around waiting, perhaps reading War and Peace, while your car spins up the second motor to give you
    maximum power when you mash the pedal at 70mph [120kph, for our Canadian friends]? Your time is clearly not very valuable! ;)
    �
  • 1/1/2015
    guest
    Haha, it's valuable, but I guess I suck at reading cuz I can't seem to read a long novel in a dozen milliseconds :p
    �
  • 1/1/2015
    guest
    My P85D's lifetime average is 365 Wh/mi and I've had Range Mode set to ON ever since it was available concurrently with INSANE (soon to be LUDICROUS) mode.
    �
  • 1/1/2015
    guest
    I have a 90D so no "sport mode" but when you say there's 7% efficiency advantage what do you mean, Rated Range?
    �
  • 1/1/2015
    guest
    That's really odd and doesn't seem normal, especially if you have a P85 reference that goes on the same roads and trips and it's beating you in efficiency, and by what seems like a large margin. Is your alignment off?

    I mean achieved efficiency (not projected) -- as in actual watt-hours per mile/km.

    Here's my reference for the details of how this test was conducted:


    Their methodology seems reasonable to me.
    �
  • 1/1/2015
    guest
    Okay. The reason I ask is w/r/t range there's only a .4 mile gain on a full charge. Less than a half a mile didn't seem worth it to me.
    �
  • 1/1/2015
    guest
    Just another reason to use percent over range ;-)
    The referenced empirical data suggests that range mode on dual drive cars adds around 20 miles / 30 km's of extra range for a full charge. Your results may vary.
    �
  • 1/1/2015
    guest
    Can you point me to your reference? I'm not seeing that kind of improvement on my car, not even close.
    �
  • 1/1/2015
    guest
    Keep in mind the comment came from Calgary. It probably does NOT make that much difference in SoCal.
    �
  • 1/1/2015
    guest
    I don't have a P85 reference, this P85D is my first Tesla (thought I've had it for over a year and a half).

    I'm pretty sure I don't have alignment issues as I've had it aligned 3 times since I've owned it: once when my rear suspension was replaced, once during the yearly service, and once after the radar was realigned (causing driver assistance errors).
    �
  • 1/1/2015
    guest
    My reference is based on the 7% quoted from the video in my previous post here:
    Firmware 7.1

    Not sure how it would make a difference. In fact, the way torque sleep works is by using the more efficient front motor. In adverse conditions, like snowy winter roads, the rear motor would be kicking in for stability and acceleration.

    I thought you said your wife has a P85. Maybe I misread. If that's your only Tesla and you've got that high consumption, it may be just the way your driving or the locations (up and down hills). Have you actually tried going on a stretch of road (say 5 or so miles) with another Model S driver to compare?

    The alignment was the only thing I could think of off the top of my head that hopefully is an old issue. I guess the ultimate way to know is check your tire wear. The rears will always wear the inside slightly more than outside, but the front do the opposite (they wear the outside more than inside when turning). So after the tires have visited the front and back, they should have even wear across the tire.
    �
  • 1/1/2015
    guest
    I normally get caught up on a thread before posting, but I'm not doing that this time, so apologies in advance if anyone has already pointed this out.

    That video was made over a year ago, long before Auto Steer was available. I pointed out in the notes of the video that I was driving in a way to induce the lane departure warnings, as I was trying to make a video that would demonstrate the problem. So I was purposely driving close to, but not over the lane markers to show that the departure warning sounds unnecessarily at times. If it so happened that in some of the instances there was some micro-paving, I'm sure that's more a coincidence than anything else.

    Now that auto steer is available I agree that these are not the kind of roads it works well on, and I don't use it on these roads. In fact, I said that in my post that started this part of the discussion, here:

    I'm glad you agree that the car is alerting too much in the video. Those roads are fairly narrow, and keeping the car dead center in the lane, which is what the current lane departure warning system seems to require, becomes annoying over time. A sensitivity setting would solve the problem.
    �
  • 1/1/2015
    guest
    Range mode can be turned on in Insane Mode and Ludicrous Mode.

    Also the efficiency benefits of torque sleep are not limited to range mode being on, as per Jerome Guillen. Those benefits are increased with range mode on, but some of the efficiency benefits of torque sleep exist even with range mode off. So while I only watched a bit of the video you referenced, the validity of any conclusion reached in that video is questionable if they were operating on the premise that range mode is required for there to be any efficiency benefits from torque sleep.
    �
  • 1/1/2015
    guest
    My wife and I drive in much colder climates than either of you guys, and have a much better Wh/mile number, so I guess we're just driving like wimps! Through 29,483 miles our lifetime Wh/mi is 325. (Photo is from today's trip. I still record certain specific trips we take.)

    I'll add that the main trip that we each take is in my case the near 60 mile trip shown, and in hers a few miles less, with approximately thirty to thirty five minutes of each of those trips on a 65 MPH interstate where we each set the TACC to 70. So while we may be driving a bit on the wimpy side, we're not driving 55 on the highway or anything. :)

    Trip.jpg
    �
  • 1/1/2015
    guest
    Yeah, I've not seen sub-300 Wh/mi in ages, except in places where it's significantly downhill for an extended period (like from Makoto's in Boone to my house... loss of like 2300 ft elevation). Even then I don't get anywhere near what others seem to get.

    I do drive with the AC on ~69F, and with the flow of traffic in the fastest lane, usually. My speed of choice locally is usually 78 MPH on the highways (speed limit 70). But even the last time I *really* tried to get good efficiency I basically tailgated a tractor trailer at ~62 MPH or less for ~50 miles and still only got 254 Wh/mi. Maybe I should do some more controlled testing and possibly get Tesla to investigate a bit depending on what I come up with.

    Yesterday I drove over 100 miles, pretty conservatively. No hard accelerations or passing or anything. Here's what the car says:

    EV_tripCharge_Miles,111.038
    EV_tripCharge_kWh,36.440
    EV_tripCharge_whpm,328.180

    I tried one way with range mode on and one way with range mode off... made no real difference (95% interstate).
    �
  • 1/1/2015
    guest
    Just received 2.26.150
    �
  • 1/1/2015
    guest
    Got it last night. But just brought the car in this am for 25K mi service, so not enough time to notice anything new.
    �
  • 1/1/2015
    guest
    Updated to 7.1(2.24.102) yesterday. Lost Slacker & had to reboot.
    Any word on what 2.26.150 changes?
    �
  • 1/1/2015
    guest
    Oh wow! Just got another update, on 2.28.19!! Thats two back to back

    Capture.PNG
    �
  • 1/1/2015
    guest
    @maximus16 - I see you have an 85. Can you please note the power output around 50% SOC next time you go to a SpC? Trying to get more data on a possible taper algo change.
    �
  • 1/1/2015
    guest
    Ha! I haven't wanted to spend the $ on the Power Tools app :(
    I am a cheap ass when it comes to that. lol!
    �
  • 1/1/2015
    guest
    SC to 100%, did 4 0-60 launches with max power at 342KW
    When I go to Dallas again, I usually SC to 50-60% and can get some data then
    �
  • 1/1/2015
    guest
    Wow, people are sure getting a lot of those 2.26 / 2.28 builds!
    �
  • 1/1/2015
    guest
    note that you do not need any app for this. You can read the charge voltage and amperage directly from the charge session screen.
    �
  • 1/1/2015
    guest
    Just finished installing 2.28.19. Surprise! Nothing new in the release notes. Didn't notice any changes, but haven't driven it yet.
    �
  • 1/1/2015
    guest
    See if AP is still there j/k. :)
    �
  • 1/1/2015
    guest
    It is. But now you have to pass an IQ test on screen before you can activate it. :)
    �
  • 1/1/2015
    guest
    When you shift into drive, there's a big Consumer Reports pop-up ad on the main display. All you have to do is check the "I understand" checkbox with your finger, then hold down the "I Agree" button for 20 seconds while the timer counts down. Easy peasy.
    �
  • 1/1/2015
    guest
    Lots of folks here have been keen causing firmware updates to occur faster than normal via SC, so I thought someone may find it of interest how at least one MS in the "normal fleet" has been updated to the latest and greatest -- skipping so many of the dot-releases being reported. My last update to 2.20.30 was via home WiFi on 5/26, and I just now completed update to 2.28.19 also via home WiFi which I received notification of earlier this evening -- some 6-weeks later. (Yes, it's been recorded on the tracker.) Only change in release notes for me is what has been reported previously and I had previously received -- when activating Summon from the stalk, a selection of direction is required on the CID. Ho hum for anything new or exiting. I'm going to bed. ;)
    �
  • 1/1/2015
    guest
    I was on the same firmware as you and also got FW update notification today. Update took 28 minutes and on 2.28.19 afterwards. I was at the Rocklin SC yesterday and it never connected to Tesla Service. I don't know if the consensus was it connects automatically or you have to manually tell it to connect but I selected the Tesla Service profile and hit connect and it never did. While I was waiting for my car to charge I went in and talked to someone to replace missing valve stem cap probably due to rotation I did recently. I asked if there was any way for users to connect to the Tesla Service AP and he told me it was password protected. I'd assume that information is already in the profile in the car already but it sounds from the conversation that it will ask for password.

    Anyway I never got a firmware update until the next day but it sounds more like a soft rollout since you got it the same day as me and coming from the same FW.
    �
  • 1/1/2015
    guest
    Of the 18 or so releases i have installed over the past year and a half, I have not yet received one that wasn't also sent to hundreds of other cars on ev-fw.com, so I'm willing to bet that 2.28.19 is a wide release.
    �
  • 1/1/2015
    guest
    Yes, 2.28.19 seems to be a fast-and-furious release, going out to lots of cars yesterday and today.

    Thanks to everyone who are helping keep EV-FW.com up to date with their firmware versions! Keep 'em coming.
    �
  • 1/1/2015
    guest
    I'm feeling left out, I'm still on 220.xx
    �
  • 1/1/2015
    guest
    I'm scared to update now. Don't want to hobble my autopilot.
    �
  • 1/1/2015
    guest
    No worries. Still works.
    �
  • 1/1/2015
    guest
    I noticed from EV-FW.com that only cars with autopilot are reporting the update to 2.28.19:

    Build Model Trim #Updates AP NAP
    2.28.19 S 70 3 3 -
    2.28.19 S 70D 10 10 -
    2.28.19 S 85 3 3 -
    2.28.19 S 85D 8 8 -
    2.28.19 S 90D 9 9 -
    2.28.19 S P85 3 3 -
    2.28.19 S P85D 8 8 -
    2.28.19 S P90D 4 4 -
    2.28.19 X 90D 1 1 -
    �
  • 1/1/2015
    guest
    I've been on 2.20.30 since May and I've just received notice to update, presumably 2.28.29. I've immediately installed every build since I bought the car 16 months ago, including seven updates after the initial release of v7.1.

    I'm inclined not to install this one right away, or at least until there's some user feedback in this forum. There are three reasons:

    1) The seven updates since 7.1 was released have not brought improvements or added value. Maybe there have been summon improvements but I can't use summon given the setup of my garage. I believe regeneration has definitely deteriorated, based on when I have to lift the pedal to coast to a particular intersection. The point is that except for major function releases (next 8.0) my current attitude toward updates has become worry about loss of function, not excitement over potential new function. This is not the way it's supposed to be.

    2) There is an element of protest over Tesla's refusal to document the content of updates, other than to re-release earlier release notes. I would be satisfied if they just said "bug fixes" if that's what it truly was.

    3) This is the first update notification I have received since the AP crash became a big deal. My fear is that in this and coming releases, AP functionality / usability is more likely to be reduced than increased. This is my protest against that too.

    Does this seem like a reasonable stance? Obviously I would want v8.0 functionality improvements if they are significant. Also, I would not avoid taking the car to the service center just to avoid updates.
    �
  • 1/1/2015
    guest
    The Tesla Trojan Horse update - right when you think there were no changes Consumer Reports flips a switch remotely (after everyone has installed it) and Autopilot is renamed to Advanced Adaptive Cruise Control with Steering Assist (AACCSA) and it only operates on a closed course on private property.

    </sarcasm off>
    �
  • 1/1/2015
    guest
    I would have to assume that, since AP was/is an advertised, explicitly paid for feature, Tesla Motors will not disable the feature you paid for or 1) they'll need to write a lot of people a check for the difference and/or 2) buy back the cars of people who ask as they did not deliver as promised (note what VW is going through right now with their diesel disaster).
    �
  • 1/1/2015
    guest
    Sillydriver, I have exactly the same concerns and feelings. You are not alone. Even though I fear loss of autopilot functionality, I am nevertheless downloading the update. Fingers crossed that I don't regret it.
    �
  • 1/1/2015
    guest
    I don't think you serve yourself or Tesla well by being paranoid. They know way more than you -- you can only guess. And there's no evidence that the combined guesses of the owner community are worth anything.

    Seriously, if you don't trust the company then you shouldn't own a car like this -- one that puts enormous control over every aspect of its use in the company's hands. It makes no sense at all.

    I got my update in the wee hours of this morning and installed it immediately. I accept that it's Tesla's best shot at getting it right.
    �
  • 1/1/2015
    guest
    I just got it thus morning too
    �
  • 1/1/2015
    guest
    BTW, I just got back from a 40 mile interstate round trip on 2.28.19 - there is no change in AutoPilot functionality that I can tell.
    �
  • 1/1/2015
    guest
    Just guessing, here.
    What if the enhanced AutoPilot algorithms for better radar detection of perpendiculair moving objects etc are already in this release, but just running in the background? I.e. not really 'piloting' the car, but already testing those algorithms in real life situations. There was the concern of too many false positives, remember?
    �
  • 1/1/2015
    guest
    Thanks NOLA_Mike. I really wish Tesla would advertise bug fixes in their release notes so we know what was completed. As a software engineer myself, I know the value of a well-written release note document.

    I understand that enumerating the bug fixes that were done could be alarming to most people -- but still.
    �
  • 1/1/2015
    guest
    Not directed any anyone individual. Just a general thought / observation / personal hope:

    As some owners now begin to hold-off applying firmware updates for fear of (in their opinion) negative changes to AP, I appreciate their POV, but hope the subject does not become another distracting and repetitive side-bar to the generally more useful insights found here and in related threads -- like was the case with V7 UI changes some others took issue with, stopped accepting firmware updates, and told us about over and over and over across TMC. It's when I learned what an ignore list was for. ;) In my perfect world, unless there is significant new observations or input, folks would say it once, do what they need to do, and let the rest of us move on -- or given TMC elects to have such broad subject mainstream threads like this one, perhaps create a separate thread for like minds to discuss a more specific thought as much as they desire.

    To each their own with their own MS decisions. Peace. Me? I plan to continue applying each and every update I'm sent for the good and possibly bad that may involve, as I decided I'd do before ordering my MS last Fall.?
    �
  • 1/1/2015
    guest
    It will also make bug fixes appear more slowly. Not a trade-off that serves Tesla or its customers well. Personally, I'm a software geek who would love to pore through all the change log details, but I recognize that satisfying my curiosity just isn't in anybody's best interest.

    I am heartened that Tesla doesn't make poor business decisions in response to pressure from people like me.
    �
  • 1/1/2015
    guest
    Along those lines, it would be a colossal PR disaster for Tesla to unilaterally disable AP without any press release or announcement or any sort of damage control well in advance of turning it off.

    I don't have AP, so I don't have a dog in this race, but I really do not think Tesla would ever disable AP in any car, ever. They might enable more nag screens or more aggressive features to make sure people aren't treating it like autonomous driving, but they won't just turn it off one day.
    �
  • 1/1/2015
    guest
    Blind faith aside, here's the other side of the coin:

    The failure to provide release notes for each release does not inspire confidence.

    Finding out by accident what may or may not have changed is not optimal.

    If there's nothing to review before applying the patch and nothing to review after applying the patch, then why do I need to be involved in applying the patch at all? Just push an alert to the car afterward to tell me that there may or may not have been changes of an undisclosed nature.

    The simple line, "Minor bug fixes and enhancements", often seen with many iOS and Android app updates, has value; it clarifies that you are unlikely to experience any significant functionality changes.

    We shouldn't have to guess, in other words. More importantly, SvCs shouldn't have to guess. Blind faith is fabulous, but it doesn't pay my liability insurance premium. We as early adopters tolerate the risk just fine. For the Model 3 effort, I would expect that the simple line referenced above will quietly find its way into an alert window both prior to and after the OTA install process. No opt-out, as obviously if Tesla makes an update available for your car, the expectation is that you will install it.

    In the end, best practices exist for a reason. Adding a simple line directly before and after the install for review so you know what you're not getting, at the least, and know what to not expect, at the least, is a cheap and easy win for Tesla.
    �
  • 1/1/2015
    guest
    I am doing the current update now. I will most likely drive 175 miles today so I will see if there are any changes.
    �
  • 1/1/2015
    guest
    I'd need a little more detail than "minor bug fixes". Categories like "improvements to range algorithm", "navigation accuracy" etc would be useful.
    �
  • 1/1/2015
    guest
    Agree with most of what you say, including there being value in even a simple "squashed minor bugs in something-or-other" statement in release notes if more info were not provided (which I of course would prefer.)

    ...not to sidetrack your broader thought, but to this specific part of your comment, I believe Tesla needs to ask the owner to install the update. The issue is the updates can take a significant amount of time (1hr+) to install and your vehicle is unusable and the process cannot be interrupted for the duration. In my case, I now have only my MS, and I schedule updates when I am 99.9% sure I won't need to go anywhere for a couple hours, and hopefully an emergency will not arise.
    �
  • 1/1/2015
    guest
    They may introduce themselves to liability if they publish too much detail about bug fixes for autopilot since fixing something is an acknowledgement that there was a problem. I.e. If a bug fix addresses an autopilot issue that resulted in or contributed to someone's car accident, lawyers will have a field day.
    �
  • 1/1/2015
    guest
    Agreed. Once the work has been tested, it takes no time at all to provide those useful pieces of information - assertions/FUD to the contrary above aside.

    The minimalist case above (one line that provides value as much or more for what it does *not* say as for what it does) is indeed the epitome of a baby step. A start. Anything would be better than copying over the release notes (see single Summon reference in 2.26.102, for example) from some previous release.

    The information we get from Ingineer's posts is helpful because at least it may indicate, dependencies aside, what subsystems *haven't* been modified.

    The level of detail you reference would be well-received by the Model 3 crowd, and again, once the QA is done, takes almost no time at all to include.
    �
  • 1/1/2015
    guest
    Claiming this is what you would like without any indication of what you would pay for it is fairly meaningless. Yeah, I'd like unicorns and rainbows with every update too, if they're free.

    So how much is it worth to you to have release notes? Two days of delay per typical release? A week? 5%? To me any delay would be a loss, and anybody who thinks release notes come for free is kidding herself. Just consider the words "vetted by legal".
    �
  • 1/1/2015
    guest
    Help me out here. I just received a notification of a f/w update available, but have been traveling through Yosemite Park and Valley the last couple of days and even at my hotel there is no data available or wifi to connect to. So how exactly did the car download the latest version with no connection at all? I thought from what I've read on here that the notification appears once the latest version has been downloaded. Has it taken a whole week in jigs and jags throughout spotty service to download packets and finally it's done?

    On another note, AP was stellar all through the park. There are newly painted lines on the roads, and with the speed limits and AP limits, I was able to serenely enjoy being driven Disney-like through the majestic scenery and was able to enjoy every bit of it without having to worry too much about the road. And when speed limits changed, as they often did coming into slower areas, AP kept me right on track. Yes, I did keep my hands on the wheel mostly, and yes, I did long for that BIG X picture window. Other than that, thank you Elon for your vision.
    �
  • 1/1/2015
    guest
    One way to dispel FUD is with transparency, and with experience. Having seen release notes prepared and released, including with the fearsome loop(s) through legal, the level of detail referenced above is relatively straightforward. Half a day, tops, as the default is *one line*. Keep in mind that many owners don't even get every release; ergo, the opportunity cost is low by default.

    ---

    Dear Tesla,

    How long does it take to vet existing release notes?

    How long would it take to vet "Bug fixes and minor enhancements."?

    Cheers,

    Yer customers

    ---

    This doesn't have to be hard.
    �
  • 1/1/2015
    guest
    It didn't. There are places in the Valley that get data signal, notably at Tunnel View. Spend any time there?
    �
  • 1/1/2015
    guest
    Aha! Got me. Yes, and even tried to hike up to Inspiration Point for the best view so I was there a while. It was hot! Thanks for the insight.
    �
  • 1/1/2015
    guest
    Assuming you received the update-available notice and then started the process, what you're doing is installing the update which was already downloaded.

    This is not mere harping on terminology: it's a point of information I hope some may find useful, as I did when I learned it.
    �
  • 1/1/2015
    guest
    I experienced the same things during a recent trip, and then learned a bit of a lesson.

    Was parked at Otter Cliff in Acadia NP (Bar Harbor/Mount Desert Island, Maine). No connectivity to the car there and maybe a bar of Edge (2G) to the phone. Up pops the yellow alarm clock.

    So I figure, hey, in here for a bit, and applied the update. The car was not a happy camper, I've noticed, when away from connectivity for awhile - 4 levels deep in a parking garage, out in the boonies, same thing - odd error messages. Sometimes scary error messages - such as "Update failed - Call Tesla" or similar, with a dead car that won't start. That ain't fun in the middle of a NP, although in fairness Acadia is at least within walking distance of a town. Fortunately, a couple of restarts later and all was well.

    So l've learned, and had I been thinking/paying attention this would have made sense anyway, to not apply updates unless there's some connectivity - fewer error messages that way.

    As for AP, I too was pleasantly surprised especially during multiple loops through Yellowstone during the same trip. Likewise found that both TACC and Autosteer enhanced the in-park experience. TACC was helpful during bisonjams, and Autosteer did an admirable, if far from perfect job on sketchily-marked 2-lane park roads. Definitely monitored speed, which is somewhat of an unsung feature of Autosteer.

    Regarding connectivity, it doesn't take much to get an update. Meaning that if you were at a high point in the park, sometimes there's enough of a signal to be sufficient. Maybe review a cellular coverage map of Yosemite to see if there were some opportunities. Yellowstone is woefully deficient in terms of coverage and it's very much on purpose, as it turns out. But still, there are a few (5) designated spots plus some high points. Not at the main hotels, which is amusing.

    Haven't been to Yosemite yet, but am glad to know that it is an AP-friendly NP :).
    �
  • 1/1/2015
    guest
    I think we have had this debate before right? In any case there are ways to word things to avoid that but yes, if I had to wait an extra day to find out what was done in this *general* way I would. Apple, Microsoft and many, many other serious companies have figured it out.
    �
  • 1/1/2015
    guest
    My X failed the update twice and now has loaded 2.26.150. I wonder if I will gat another update to 2.28.19 later. My S has installed 2.28.19.
    �
  • 1/1/2015
    guest
    Unsure is there is anything going on unique to your MX. You may want to check the MX forums just in case, and you can always send an email to Tesla or give a ring to your SC. Within a day or so, they can push an update to you or assist if there is some unique problem your MX has encountered. Good luck.
    �
  • 1/1/2015
    guest
    I would do a system reboot to ensure everything is clean
    �
  • 1/1/2015
    guest
    I got 2.28.19 pushed out to my car in Abu Dhabi. Normally I only get the updates that are pushed out to the whole fleet. My last update was 2.20.45. My nearest SuperCharger is now in Dugopolje, Croatia - only 2,514 miles away.
    �
  • 1/1/2015
    guest
    Well, I guess you won't be using that one very often.
    �
  • 1/1/2015
    guest
    I was told that every car gets the same full download, then the car extracts the particular version for it.
    �
  • 1/1/2015
    guest
    Firmware 7.1
    �
  • 1/1/2015
    guest
    Haha. Tesla keeping promises. That's a quaint idea. Be assured, they will do what's they think is best for future sales. A rollback is by no means out of the question.
    �
  • 1/1/2015
    guest
    Lookup what happened with Smart Air Suspension in 2013 after a media freakout about a car hitting a tow hook. Tesla admitted that even though chances of it happening again were astronomical, they still disabled it to show that they are proactively doing something. They never put that in any release notes, nor even admitted they are doing it on purpose when the first few owners contacted them about why Air Suspension doesn't work anymore while driving (it worked while parked and raised back up when driving) - this is so people don't skip the update. I asked for a refund of my Air Suspension and all I got back is "sorry, we don't do that". I came close to declaring my car a lemon as it was brand new and service center couldn't " fix" the air suspension which under my state law would qualify as a lemon car. SAS got re-enabled months later after the titanium plate retrofit. Bottom line, it is not at all unlikely that given enough media pressure, Tesla will disable auto-pilot, pretend they don't know what happened, then tell you they are working on a solution with some deadline they are going to blow by months.
    �
  • 1/1/2015
    guest
    Thanks.

    It's kinda funny because this is probably how it was explained a few months ago by one of the two or three people that are intimately aware of how the updates are done. After a few months those of us that simply read about it forget the fine detail. Like I forgot that's it's the patch that's different and not the build. Not a good thing to forget in a forum full of software people. :)

    To me that wasn't the important part of what I was trying to to convey. What I was trying to convey is the server is triggered to send an update (by a request from the car), and that the update contents would be different from one car to the next. My understanding is the update patch is built on demand.

    I was trying to support the argument that it was theoretically possible for the update priority to be higher at the service center. Where a firmware request that originated at a service center wifi could have a higher priority than a request that originated elsewhere.
    �
  • 1/1/2015
    guest
    In terms of believing something 100% that's related to firmware of the car there are only two people to believe.

    There is WK057, and Ingineer

    Is there is anyone else that should be added to the whitelist for firmware content/update-method?

    I'm on the greylist so only believe half the stuff I say.
    �
  • 1/1/2015
    guest
    It's software so, theoretically, anything is possible. There could be code that recognizes particular VINs and does something entirely different for each one. Choices could be made at random. Choices could depend on the phase of the moon. In theory, they could do anything at all.
    �
  • 1/1/2015
    guest
    Darned locals.
    �
  • 1/1/2015
    guest
    That wasn't the case for Hong Kong owners. They had it disabled for a while due to government then had it reenabled later on.
    �
  • 1/1/2015
    guest
    It was really about addressing one phenomenon that people were experiencing. Something that seemed like it wasn't a random event. Where I was trying to explain that it was likely the intended behavior and this is how they likely did it.

    As to software I would recommend never saying "anything is possible" to a software developer. :)

    i'm a hardware engineer so I just chuckle at it.
    �
  • 1/1/2015
    guest
    I think you'll find the three superchargers in Jordan are much closer.
    �
  • 1/1/2015
    guest
    Try summoning with your key fob.

    Summon-by-fob went away after this am's update. UPDATE: The setting requiring continuous keypress was toggled "ON" by the update. Setting it to "OFF" solved the issue.
    �
  • 1/1/2015
    guest
    Possibly just a coincidence, but I got a very early forward collision warning today after having installed the 2.28.19 update last night. Driving along at 35 mph and a car braking about 50 yards ahead of me set off the warning. I have it set to early, but I've never had it warn me that early!
    �
  • 1/1/2015
    guest
    Mine works. Do you have "continuous press" off?
    �
  • 1/1/2015
    guest
    It was off before the update. Update changed the setting.

    Stuff like this would be very helpful to see in release notes whenever they change a setting back to default or otherwise change its meaning.
    �
  • 1/1/2015
    guest
    Probably not intentional. In fact, I think a previous release did the same thing.

    As for the early collision warning, I'm set to "early" to and that has happened to me periodically before this release--probably not something new.
    �
  • 1/1/2015
    guest
    And easily gleanable from regression testing notes. Presuming that their QA teams are not the redheaded stepchildren of the organization as is often the case, and that sufficient regression testing is in fact being done. Even smoke testing would catch some of this stuff, but by then it's a little late.

    For all of the silly FUD put forth as an excuse for why release/change notes are not given, there's a huge (dare I say potentially yuuuuuuge) amount of liability in *not* keeping your customers in the loop when the UI is changed out from under them.
    �
  • 1/1/2015
    guest
    Before you all go blaming Tesla you should know my setting of "off" didn't change with this update. So maybe not everyone had this problem?
    �
    • Người đăng: vào lúc

    Không có nhận xét nào:

    Đăng nhận xét

    Đăng ký: Đăng Nhận xét (Atom)