Tesla's Screen Saga Shows Why Automotive Grade Matters
thedrive.com
thedrive.com
In the end, I never really got to finish the project because of the rabbit hole it became.
But remember, I'm a hobbyist. I have an electronics background, but that's about it. To see a huge gigantic company of this size and reputation pursue this kind of mistake that has real impact on paying customers is astounding. I mean, I like Musk and the work he's doing, but this doesn't seem right (just a personal opinion). I hope the customers don't have to bear this for Tesla's mistake.
It's a conscious decision, probably due to cost/production volume concerns. Musk had pulled this kind of tricks before with lithium-ion batteries before (remember the early Tesla fires?), they were initially using consumer grade laptop batteries (with a highly energy density/weight ratio at the time but more prone to fire and explosion comparing to other manufacturer's automotive grade lithium-ion battery chemistry, granted Tesla might have improved.)
And honestly, the cost of automotive thermal grade 2 certified central display for 17-inch TFT screen was probably (still is I suspect) cost-prohibitive for a low margin auto industry.
What's really astounding to me is the "cabin overheat protection" feature. As the article pointed out, what this really is Tesla trying to prevent the non-automotive grade screen from failing, even though it's touting for protecting dogs and children at extremely hot summer days. It regulates the cabin temperature at +40C, but that's the temperature children's organ will start failing.
While the article seems well researched and very informative otherwise, 40C is the internal body temperature at which a child's organs will start failing.
The internal air temperature of the car (regulated to 40C in this case) will of course have an effect on the passenger but does not necessarily lead to them reaching a body temperature equivalent to the ambient air because the body possesses cooling mechanisms.
40C still seems problematic for children though. Quoting https://www.uptodate.com/contents/heat-stroke-in-children:
> Evaporation is the principal mechanism of heat loss in a hot environment, but this becomes ineffective above a relative humidity of 75 percent. The other major methods of heat dissipation; radiation (emission of infrared electromagnetic energy), conduction (direct transfer of heat to an adjacent, cooler object), and convection (direct transfer of heat to convective air currents); cannot efficiently transfer heat when environmental temperature exceeds skin temperature (typically 35°C or 95°F).
Actually it is a very well recorded phenomenon. Just because you and the people you work with may have adapted well does not mean that it is universal by any means. People die in heat waves regularly; normally diabetics with peripheral neuropathy who can't sweat enough to drop their internal temperature or people with heart problems who can't take the additional stress of lowering their temperature. [1] [2] [4]
"Heat related mortality is similar in hot and cold parts of western Europe and in hot and cold parts of the United States. This implies that the populations of hot regions have adjusted by physiological or other means to their hotter summers."[3]
Also, the metric for organs failing at 104 is entirely correct, just not with regards to external body temperature. Your organs will undoubtedly fail if your internal temperature is 104F/40C. I've had sepsis with a temperature of 40.9C and I would have died if it wasn't for quick treatment from the doctors.
Quick anecdote; if you're ever in the situation where your body temperature starts to get that high, scream and shout until you get an actual doctor on hand - I had nurses who had no idea of the severity of my condition and would have left me to die through shear negligence. I was lucky enough that the surgeon for my leg injury decided to check how it was healing, and rushed to call in a team of 4 or 5 doctors to stabilise me as I went into shock from hyperthermia. Spent the next 24 hours covered with cold towels and ice swapped out regularly by some very apologetic nurses.
[1] https://emj.bmj.com/content/22/3/185
[2] https://news.nationalgeographic.com/news/2013/07/130716-heat...
Then it's not correct in this context, where everyone else was talking about external temperature.
The title reads: Prevent Child Deaths in Hot Cars. It's advisory for child death in hot cars. Your anecdotal doesn't invalidate any of the health advisories in the link below.
[1]: https://www.healthychildren.org/English/safety-prevention/on...
The planet is warming up because we cant provide enough clean energy to feed demand.
I can't help but feel that having everyone's car air-condition themselves when not in use, so that you have a screen large enough to play space invaders, electric or not, is really bringing us closer to solving that.
40C is very survivable. Lots of places on earth regularly hit 40C
For overheating, I think they can. A much higher surface area to volume ratio means even a small amount of sweating can give a lot of cooling.
"Dog mode" keeps the car at the temperature set by user, typically 20-25C
Opel Ampera (~Chevy Volt) has a plastic ball race in the electric motor, that fails at ~ every 100k km. The US market had a metal ball bearing initially, the EU market always had a plastic one. GM/Opel saved like 1$ on this ball bearing, but it costs the owner hundreds to replace it every couple of years, as it takes hours to replace it. This is a major drivetrain issue IMO, and it was not handled by a recall (at least not in Europe)
The same car has a steering lock with the cheapest carbon-brush motor that you can get, that wears out over a couple of years. When it fails, the car is not driveable, and you have to replace the whole steering lock. Again, no recall by GM/Opel.
These things are not built to last, the design goal is to be as cheap as possible.
To execute the same thing with hardware credibly, you have to accept the idea of a mass recall and replacement of faulty parts. That works for many companies like Anker, which goes out of their way and proactively replaces parts.
You can't apply the software deployment model and be nonchalant and brusque when customer issues arise. The combination just serves to turn off the masses.
This is the top of the top, Musk himself, saying "screw that quality requirement, this is good enough", and plowing forward regardless (and having it bite them later).
Who knows how many other components they were able to cut corners with for now. And how many will start to fail in the coming years.
Telling apart "rules" that exist because nothing else works from those that are just inertia is difficult.
>When Tesla performed the tests, the electronics ended up working just fine.
Tesla's environmental testing just failed.
Doesn't necessarily indicate good things about the longevity of a new Tesla.
This is in addition to things breaking which weren't designed to break in the first place.
Asking probably because my desktop pc went out with a bang last night, waking me up. It's probably the accumulated dust that killed it. You'd need all terminals and wires with potential isolated from air, no?
That desktop, had it been coated, would probably still be working, assuming the dust did not trigger thermal trouble.
That's a thin coating. Automotive tends to dip in a heavier resin-type coating. At the other extreme, there are ultra-thin coatings used on some mobile devices.
It's amazingly hard to design electronics to survive thousands of cycles of 100C thermal cycles.
Though the main failure mode was due to EMI as coil and spark plugs generate all kinds of crap.
I saw a video card short out that way.
I think Tesla made the right decision because while it's certainly costing them a lot of money to deal with, the screens in most cars (especially back then) were so bad that it probably would have been worse for their brand.
The most likely outcome is that they will figure it out, and much earlier than any of their competitors.
But now the "good will" repairs are over, the smug attitude of the article is more or less deserved.
In this case, Tesla simply wanted to be ahead of the curve. Worked for a while, then the known failure rate of the parts caught up.
Whether or not it was worth it depends on the strategic intent and whether or not that bet paid off. We'd really need more information to make that call. And Tesla is not going to give anyone on this thread access to the strategic and financial information necessary.
But yeah, with or without Tesla, obviously the screens would have existed.
Disagree. The most likely outcome is that many folks who paid a lot of money for a car end up with a car with a screen that's flawed in some way. (I own an S that has the yellow box.)
It is also possible Tesla "figures it out" earlier than their competitors; but that's less likely as the former is now a certainty.
[1]: https://www.nasa.gov/sites/default/files/atoms/files/public_...
One of the purposes of using the "right rated" parts is that you get to off-load some of the testing and screening that you might otherwise wish to do to a component vendor (who is better equipped to absorb the price of testing, and because they have tighter integration with manufacturing, a better ability to achieve specifications).
Like I totally get the idea of using lower cost components. But sometimes they are actually the solution. And sometimes finding the problem is incredibly expensive, if not in $$$, then at least in PR.
Americans love to shit on gold plating, but they also love to shit on greedy corporations minimizing the bottom line.
The issue here (as a sibling post points out) is QA.
The problem SpaceX had was using industrial grade parts when they should have used aerospace grade parts.
I was looking at the details of the problem, and suggesting that those parts were neither aerospace grade nor industrial grade, but something lower still.
Non-rigorous testing is a reasonable thing for "industrial grade". I am not convinced that being far under half of the rated strength is a reasonable thing for "industrial grade".
I paid the $6 and sent mine USPS certified return receipt just to be sure. I recommend you do the same.
People who bought an iPhone shortly after launch paid $600 with no idea that the app store would appear and become what it did. For all anyone knew, the iPhone would go the way of the Newton or the Zune. Two years later, the same phone was retailing for $200 with a proven track record and an exploding ecosystem of apps.
What being an early adopter doesn't mean is that the item may not be fit for purpose. If early iPhones had never taken off, and the product line was cancelled, that's your bad luck; you took a gamble when you bought it. If the screens had just not been rated for mobile use and had leaked LCD fluid into people's pockets, that's not part of the gamble.
Tesla's appeal to early adopters, but that's not an excuse for product defects.
Further more, it seems to be relatively new screens that suffer from the yellowing problem. It doesn't seem to be a common issue in Teslas built before 2016 or so.
I remember paying $300 at launch I think...that was the AT&T locked one but back then it was easy to unlock.
The $600 ones weren't available until much much later when they started selling off contract unlocked phones directly.
I only started paying more for iphones later (off contract, unlocked, from China, etc...).
Source: https://www.nytimes.com/2007/09/07/technology/07apple.html
https://i.redd.it/kv4y1j3gjut21.jpg
This WILL kill someone. It's not a question of if, it's a question of when, and it's intentional.
Maybe the people responsible for that decision will see the inside of a jail cell, or maybe they won't- but the fact of the matter is that the presence of this "feature" is negligence at best, much like this case where bricking your car when something unrelated to its core functionality dies.
These are 'sports-car' problems. If I was shopping for a vehicle to fill the role as my primary transportation these flaws would exclude Tesla from my checklist, and I have a hard time believing that i'm the only one.
A 500 dollar Geo Metro from the mid 90s will get me to the hospital when I need it to. It'll get me away from an evacuation area. It'll drive me away from the forest fire. A Tesla is more glamorous, but when dealing with transportation being a means to someones survival i'm gonna stick to KISS -- and I wish that Tesla had the same idea in mind with regards to many things about their cars.
(keep the ambulance/911 comments to yourself. It was N/A in the situation I found myself in.)
Maybe I’m wrong, but the Telsa forums were absolutely livid when a poster used the word brick. His car was in state where it couldn’t start or upgrade again.
Are the upgrades user initiated? I've had (or seen) enough productions and live events degraded or interrupted by the windows autoupdater to be terrified of the idea of "isn't in use" updates.
Perhaps some updates need to be done overnight?
Sadly I highly doubt that and if someone goes to jail it will be some poor engineer that was forced the implement it that way.
Just look at what happened to Uber after they actually killed someone with their self-driving car because of their disregard for safety from the top in order to save time and money.
Next up, Boeing.
Boeing has a pretty terrible history of safety violations, especially when money is involved. They also have a long history of trying to cover their involvement until the last possible second.
I can think of a few incidents, but one of the really atrocious incidents was Boeings' response with regards to Lauda Air Flight 004.[0]
If Niki Lauda hadn't been so assertive about Boeing making public demonstrations of the maneuver in question they would have never admitted fault. It wasn't until sufficient celebrity pressure was applied to them that they admitted to flaws that had been well known and swept under the rug to save cash.
Now, cue the folks that compare the number of Boeing planes in the air with the number that have fallen out of the sky.
This number is irrelevant when we're talking about a company with a history of 1) : ignoring problems, 2) : refusing fault for their own responsibilities, 3) : covering up flaws that are costly to repair , and finally 4): modifying or cleaning debris fields of downed planes before national inspectors, possibly with intent to reduce corporate blame.[1]
[0] : https://en.wikipedia.org/wiki/Lauda_Air_Flight_004 wiki article mentioning Boeings' role in post-incident investigation, delay of public information, and fault refusal until Lauda asked for a demonstration
[1] : https://www.youtube.com/watch?v=E42NN1rU93o an interview with Niki Lauda where he mentions the behavior of the Boeing ground crew with regards to the debris field at the site of the Thailand LaudaAir Flight 004 reverse-thrust incident.
Solar panels have the same problem, and they have "expansion joints" to deal with it.[1] Hard to do that for one seamless screen.
[1] https://www.pveducation.org/pvcdrom/modules-and-arrays/therm...
Automakers have taken notice that people are using their phones for an increasing share of infotainment functions.
The tough, time-consuming parts are working through the supply chain to build out the connectivity and "happy coexistence" of phone and car. Also, I've yet to find a UX which blends physical, voice, and touchscreen controls while accounting for their strengths, weaknesses, and learnability. Tesla included.
As long as the mounting plate is sufficiently isolated from the rest of the car's movement, they shouldn't need to suffer any vibration in excess of what they'd endure under normal use and transport conditions.
Of course, it's still a better situation for consumers: if the iPad does die, you can replace it that same day with an off the shelf part (or you plug your phone in for that day instead), it's trivially expensive compared to what head units of similar type cost (it's actually even cheaper if you can use a bargain-basement 50-dollar Android tablet instead of an iPad), and it's something that you can upgrade as you get the money to do so.
Plus, it also heavily discourages manufacturers from pulling a Tesla and refusing to start your car before you update the infotainment software, which is also a shockingly anti-consumer trend.
(Take the issue with fires, for instance - Teslas catch fire far less frequently than conventional cars, but every time it happens it's global front-page news.)
I see a lot of "they should have done X and they didn't so they're wrong" but not a lot of "they should have done X but they did Y and Y worked worse than X by these measures".
If this article is accurate (let's assume it is, those quotes from Musk will get the lawyers coming so fast if they were false), then it just proved that this actually is the case! Their arrogance that "pah, who needs automotive grade?" has now come back to bite them in the ass...
Parked gas cars don't usually burst into flames.
Usually not, no. But every now and then, they do, just like electric cars. It's hard to say which type of vehicle burns down more often, the sample size is small and there are so many more petrol powered cars than electric powered cars.
Electric and petrol powertrains have different failure modes. An electric car can fail while being parked, which is less likely to happen on petrol cars (and if so, it's usually caused by an electrical failure in the car). However, petrol powered cars have a higher chance of bursting into flames while in operation due to the compination of high temperature components and pressurised fuel being close to each other.
A car (being petrol or electric) has a high density energy storage. Whenever that enegry can no longer be contained, bad things happen. Personally, I'd prefer my car burning down while being parked over bursting into flames when I'm driving it.
2.4M results for:
Parked car burst into flames "mercedes" -tesla -bmw
1M for
Parked car burst into flames "tesla" -mercedes -bmw
0.5M for
Parked car burst into flames "bmw" -mercedes -tesla
Unscientific I know but surely an indication that at least some parked cars do burst into flames. Not usually I'll grant you, but a bit more frequently than one might think.
I agree with the comments here that Tesla should make it right for the customers though.
OTOH physical knobs and switches are much easier to operate without looking at them. Which is particularly important in a car.
Do you know that almost all (if not all) other auto-manufacturers made the calculation that they better lie to their customers about the emissions coming out of their cars?
Who's really cutting more corners?
> In fairness, other automakers have had issues with touchscreens as well. Just last year, Ford settled a class action lawsuit over a host of problems with its MyFordTouch systems and had previously extended warranty coverage for its SYNC systems. Cadillac had to recently revamp its CUE touchscreen system after numerous issues including "the dreaded spiderweb cracks." But [like, it's totally different for Tesla for reasons not having anything to do with screens or reliability][1]"
Basically, screens suck. Everyone's screens break. Apple and Samsung have shipped junk screens too. But Tesla's bad screens are an existential crisis. The auto industry hates them so much...
[1] Paraphrased.
All screens suck but only one manufacturer put their entire control panel in the screen... so yeah...
That's not honest journalism.
Now that they no longer do, what do you expect to happen?
I've been in hotter places than that as a child and my organs are still functioning. Not that convincing a tale, got to be honest.
- electric vehicles
- long range
- self driving
- big screens
I really wish they or somebody else focus on just the first two parts - long range electric cars. No need for the AI or other fancy stuff.
I also have a small touch infotainment screen in my car. Works great for calls, music and navigation - that is all what I need it to do. I have not had a single problem with it for about 4 years.
I think it is incredibly unsafe that things like heat and radio controls have been shifted to non-tactile touch inputs, rather than analog buttons and knobs. But I am an old fogey that keeps at least one hand on the wheel and my eyes on the road.
At least Toyota Avalons do it in a smart way: the display for the back-up camera is embedded in the rear-view mirror. You can only see it when you are backing up.
Any control in my car that I can't operate while wearing those mitts is essentially useless to me.
The primary issue with touch is the difficulty many people have using it by feel, and or "muscle memory" alone. Typically, one needs to identify the UX element desired, then touch, and then often verify the thing happened or select after that.
Where it is modal, another whole layer of that interaction is needed, and or driver needs to retain more state info. "Which screen is it on, and is it currently available for me to use?"
Steering wheel controls can be better, but many of those do toggle through various modes and the driver needs to glance at a small screen in the dash cluster.
Doing that is generally superior to a touch only UX though.
On many cars, low importance functions are clumped into these kinds of UX. Higher order things get their own control.
I drive a lot of different cars. New, old, and Tesla.
I thought the Tesla was spiffy, but did not think much of the big touch panel for reasons I have given here.
My very favorite drives are on cars with great controls, and all of it working with dim interior lighting. Just enough to see what needs to be seen, the rest easily operated in a tactile way, night vision intact.
This is not to say touch is not useful in cars.
I think it can be and is.
This is to say touch has a very long way to go before it begins to seriously challenge well designed buttons, levers and switches.
Is it one screen in a million that fails or one in ten? There is no way of telling from the article.
@leggomylibro: To reply to your deleted comment, you asked if the vehicle needed 1800W to keep itself conditioned but not maintain state of charge. It's closer to 1440W, as a 15A 120V circuit is derated to 12A (80%) for continuous duty cycle. Coolant pumps can be power hungry!
@akozak: I believe it is Tesla specific. I am unaware of the feature being present on other EVs.
Not because they'd replenish a meaningful amount of range per day, but because they'd provide better-than-ICE unpluggability for stuff like long-term parking, cabin ventilation, battery maintenance, telematics modem sleep current, etc.
>Tesla appeared to mostly fix this problem with its "cabin overheat protection" feature (which it sold as being to protect dogs and children, despite the fact that it held the temperature at +40C which is about the temperature where a child's organs will start shutting down)
But as the linked explanation says, +40C is the internal temperature which becomes dangerous. This is not the same as the dangerous air temperature. Humans both produce heat by metabolism, and remove heat by sweating. Danger is a function of temperature, humidity, duration of exposure, hydration, clothing, general health, etc. and although I don't know exactly what is safe for children, I point out that a great many people live in climates where air temperature routinely exceeds 40C for several hours at a time.
If you want to express outrage about abdication of responsibility at least be talking about the relevant feature.
Cabin overheat protection is there to protect things left in the car like electronics, and theoretically requires no user interaction.
Dog Mode allows the user to dial up the desired temperature and then walk away. This is the actual dangerous feature, because the belief is that the air conditioner will keep the temperature at the level you want it. This despite a history of at least one police dog dying every month in the USA due to equipment failures, and theirs dog cars have specially modified environmental controls to maintain and monitor the temperature, including opening windows as a last ditch temperature control effort. Even those systems fail sometimes.
And here is Tesla setting people up for failure using a single air conditioner that is monitored remotely using an internet connection that is flakier than filo pastry.
It is a good question if, while in dog mode, if the car detects the temperature exceeds the set point, will it lower the windows or try to notify the owner?
Leaving your dog in the car for an extended amount of time is not a good idea, and in many places is illegal. And yet, sometimes the safest place for your dog is in the car.
Personally I haven’t used it yet, but I like the option of a big display showing the interior temperature of my dog is in the car without me for a few minutes.
FWIW is typical failure mode for an air conditioner is for the condenser to fail, so the conditioner ends up over-heating air that hasn’t been cooled by the condenser yet, resulting in the interior temperature hitting 60C in minutes, so eg a police dog (typically a large-ish dog like German shepherds) will be dead or dying when you come out to check on it fifteen minutes later.
The systems used in police dog vehicles include redundant components, temperature monitoring and emergency actions like opening windows to allow cool air to flow. Police around the USA still lose about one dog a month with those specialist systems in place.
It’s only a matter of time until someone posts on the Tesla reddit that the car killed their dog and the fans will dogpile on the poster telling them it’s their fault for leaving the dog in the car in the first place.
People will leave their dogs in cars on occasion. I believe there are plenty of times you have to do that actually as a matter of safety, where there isn’t another option or a second person to stay with the dog in the car.
But a car isn’t perfectly safe for dogs, and perhaps hundreds of dogs die each year in hot cars. (Apparently a million dogs are killed in collisions each year, but I guess that’s besides the point)
So as a feature to keep dogs comfortable and safe, it’s a feature that is likely to protect dogs’ lives on hot days.
Yet, as you say, and I agree, it’s entirely possible a hardware or software failure will result in a dog’s death.
So they are protecting dogs lives, but in doing with a general purpose automotive AC could also risk a dog’s life in the event the system fails. In addition the existence of the feature may also contribute to a complacency around leaving dogs in cars that otherwise could have safely remained with the owner.
Does that mean it’s wrong to ship the feature? FSD, in a similar sense, has the potential to save tens of thousands of lives a year. It will also undoubtably cost some number of lives a year as well.
If we are unable to tolerate catastrophic failure of a system, you necessarily lose all the benefits of the system. To me it seems morally wrong not to ship the system which will save orders of magnitudes more lives than it might cost. That’s true whether we’re discussing dogs or humans.
It’s funny it’s hard to draw a precise line, but I’m sure for example the FDA has dealt with this exact calculation. No medicine is without risk, and to win approval is not to show a medicine is risk free. Their is some threshold where the risk is worth the net benefit.
In the case of an AC system in an automobile, it is already a mission critical system which is highly reliable by the way. When I had a cooling fan die on my last car a couple years ago I was unable to use the AC in my car on a 100°F day. It was amazing how hot the car got, and how quickly. If I was on a remote highway it would have been life threatening, to go from 72°F climate control to 100°F+ with no water on hand.
Does that happen when you have no engine?
It's probably sold as a feature "you'll find your car comfy when you go back", or even "your children will be fine even if you're a horrible parent" and doesn't say anywhere "disabling this feature may cause the screen we provided to melt in the sun, and we don't provide a replacement, as maybe next we'll have a way to fix it". And I don't think the limitation is supply chain, it's just that they want to keep the costs low for the technical debt.
So a few touchscreens have yellow lines. The Tesla touchscreen is still clearly better than any other car's touchscreen and if they'd done it this guy's way they wouldn't have been able to ship it at all because the part simply wasn't available then.
"One of the best cars in the world" is obviously subjective, and there are some ways in which Tesla is absolutely leading. But all of the innovation on display in Teslas don't earn them a free pass on build quality issues.
So they tried adding software to keep the cabin below 50°C even when the car is parked, but either it doesn’t work all the time or people turn the feature off, of someone is parked in the sun long enough to run the battery below 20%.
The trade off was a higher rate of warranty repairs in order to build the car they wanted to build.
I think in the case of the Model 3, partially how they solved it is through special glass they used above the front seats. I’ve been wondering why the glass is so unique above the front seats but not the same tint for the rest of the roof glass. It’s very interesting — water droplets on the glass above the front seats look almost blood red. Now I’m guessing it’s filtering to block IR or other solar gain which might otherwise burn out the screen.
I'm not sure whether it's fair to say they "cut corners," but I'm not sure it's not fair. The other alternatives could have been finding a supplier willing to work with them on an oversized screen that met at least Automotive Grade 4 standards, or being willing to redesign the instrument panel to take a smaller screen. (Or two smaller screens.)
Warranty repairs? Tesla has described these as a small number of "goodwill repairs" and summarily begun denying them even to new car owners as "cosmetic, won't fix".
The replacement screen looked great at first - indistinguishable from the original one.
But over time it has developed exactly the same issue: a yellow band/border/ring around the outside of the screen. I wonder if this is a wider issue within the LCD industry, rather than just being specific to Tesla.
The yellowing will happen to all screens eventually, just heat speeds it up a lot.
Sorry to all the yellow screen owners who won't get them replaced, and RIP that dude that crashed into the concrete barrier.
https://medium.com/@MidwesternHedgi/teslas-driver-fatality-r...
Whatever the numbers it's managed to save, Tesla's buggy self-driving software has killed a few people for sure. How would you like that seal of approval on your product?
now i know why these screens aren't more prevalent but still...tesla's dashboard looks superslick
Some quick math, if an industrial part costs $10, and an automotive part costs $20, then for ten cars you save $100. Assuming that 20% of all your cars are faulty, that's a $80 savings translation. With the remaining money, you could spend $40 on R&D for a screen at the same price point with higher reliability and still come out on top. However, at some point, you will reach a time where the number of replacements that don't justify the design. For a bootstrapped company, I don't think it's a bad bet, but it is time critical to find an alternative before they start losing money.