I wonder what Tesla think is the expected service life of their vehicles? I would imagine the average person on the street would expect to see a service life of >20 years.
My worry with a large number of modern tech heavy cars will be that while the centralisation of a lot of these controls and features into one panel may be convenient to the user, when these inevitably fail (as in this case) - what can the owner do when the parts are no longer available, or if the manufacturer is out of business? Mechanically the car may be sound - but if the system responsible for the interaction between the vehicle and the owner is borked and impossible to replace does that mean the car is a complete write off?
I think back to some of the beaters my friends owned when they started driving, with broken HVAC switch gear, burned out audio head unit wiring harness, etc... while mechanically sound (enough to be road legal...), issues like this didn't hinder or prevent driving the car, and they continued in service for years in this state.
> My worry with a large number of modern tech heavy cars will be that while the centralisation of a lot of these controls and features into one panel may be convenient to the user
Centralization of controls into a single touchscreen is most certainly not convenient to the user, because touchscreens suck [1]. What is convenient is the thoughtful deployment of physical controls in ergonomic locations.
[1] Touchscreens are really only a good choice in situations where there's extreme space constraints, or as an enhancement to existing controls for very specific interactions where they make sense.
In my experience touchscreens are great for passengers and very dangerous for the driver, unless the car is stationary.
All commands must be around the wheel. An alternative that could work is voice commands, if the car can really understand the driver. Good luck to foreign drivers, tourists with rental cars or people living abroad, unless they can configure their own language.
Years ago, I was in Germany with a Serbian, Japanese, and someone from Turkey. We each took turns trying to get the German navigation system of our rental to understand us, and then we finally pulled over to operate the very painful non-touchscreen navigation system.
I don't think the parent was talking about complicated systems like GPS, that you should probably set up before driving, and yes, they are often a pain, touchscreen or not.
A touchscreen is a dangerous interface for the myriad of simple mundane systems like heat up/down, AC on/off, fan more/less, sound more/less/next.., heated seats, windows... things that you may need very much want to do while driving.
Just image if this happens on the new Model S Plaid and Model X. Both are suppose to have the gear selector on the touch screen with the car guessing weather you are in drive, park, or reverse based on your surroundings. If you screen dies and it guesses wrong they you are stuck.
I really don't think they get that much right. Everything above the drive train is subpar for vehicles much cheaper than theirs, and now it's become clear that they're less reliable too.
They also moved the indicators to the steering wheel in what looks like touch capacitive buttons. That also seems like a mistake to me over an actual stalk.
Is that what those circle things are on the wheel?
Musk's reply sounds like their AI will automatically turn the left/right on and you have to over ride with a touch screen - which to me sounded like a center console... Either way based on the many posts/videos on their current ML I don't think either is a good idea.
It should be illegal to make a gear selector with no tactile feedback. People have died from poorly designed gear selectors. Having the car guess the direction you drive in also sounds like the most braindead thing ever. It won't be long before someone dies from this.
People have died from poorly designed gear selector levels.
Anton Yelchin, the actor that played Chekov in the recent Start Trek films died because his Jeep Grand Cherokee has a "joystick" style control for the Park gear selector. It wasn't obvious that it was properly engaged. He got out, the car rolled down an incline, and pinned him against a brick wall until the life was squeezed from his body.
Elon Musk likes to joke that cars are two-ton death machines, but this stuff is no joke.
No, the 2015 Grand Cherokee shifter doesn't randomly change gears. There are known cases of it being accidentally hit from a passenger since there is no locking mechanism. That said, it is standard procedure to engage the parking brake and anybody that doesn't is a bad driver.
That car had two indicators for park, on the center console and the dashboard which Yelchin ignored. Plus, he never engaged the parking brake, which is standard procedure. His death was his own fault.
a) My much cheaper car engages the brake automatically when I stop it and turn the engine off.
b) To engage the brake, you have to push the lever, and then it moves back to its original location, like a joystick. This means that there's no positive tactile feedback that the brake is engaged
To quote:
"The problem is that the "Monostable" design doesn’t provide any meaningful feedback about what gear you’re in — it returns to the center position after each shift. To completely confirm if you’re in drive or park or reverse, you have to look at either the LEDs on the shifter (often covered by your palm) or the digital display in the instrument cluster. This has confused thousands of people, led to over a hundred injuries, and now potentially a death. And it’s all because of a design that prioritizes screens over switches."
The 2014-2015 Grand Cherokee was recalled because of 304 vehicle rollaway incidents, 117 crashes, and a death.
There is no excuse for this kind of "too clever" design.
People drive while tired. People drive in emergency situations. People drive while distracted by screaming infants.
You can't blame dangerously bad user interface design on the users. Essential safety settings should be designed so that the operator falls into the pit of success instead of stumbling into the pit of deadly failure.
The recall didn't involve any physical change, it was a firmware update that engaged the parking brake when the transmission is shifted into park. This shifter design came over from Mercedes, where it was used successfully.
A motor vehicle is a heavy machine which requires proper operation because it can cause serious injury and death. This is no different than operating a forklift. People need to start taking responsibility for their actions. If folks don't read the operating manual and can't successfully operate a motor vehicle, they have no business driving.
That's a set of goalposts that move automatically to render any argument against it invalid.
No matter how counter-intuitive, error-prone, or difficult the control interface of an automobile is made to be, you can always say: "Well, if you can't figure it out, don't drive!"
Would you fly on a plane designed with this attitude? "It'll drop out of the sky if you accidentally bump anything, without a noticeable warning, but pilots that can't handle flying shouldn't be in the cockpit anyway!"
But seriously: would you get on a 737 MAX without the MCAS fixed?
Would you get in a 737 MAX with an unfixed MCAS a decade after the MCAS incident, when people have forgotten? With a new pilot that had never heard of the two specific crashes?
Or would you insist in flying in a plane designed not to crash into the ground in ways that's counterintuitive for the pilots to deal with?
Similarly, would you let your relatives drive a 2014 model Jeep Grand Cherokee that hadn't been recalled and had the parking brake changed to be automatic? Would you trust your Grandmother to check the tiny little light every time that will stop her dying, or would you take it to the shop for her so that she doesn't have to?
I read (or skim) car operating manuals, but that's because I'm technically minded, and curious. I know a lot of people who don't bother, and haven't read a car manual in their lives: if they have "no business driving", there would be a lot less drivers on the road.
All aspects of driving should be made as intuitive as possible, from the operator controls to the fonts on road signs. Adding more mental work makes a dangerous task even more so.
There are US regulations on this. [1] "If a steering-column-mounted transmission shift lever is used, movement from neutral position to forward drive position shall be clockwise. If the transmission shift lever sequence includes a park position, it shall be located at the end, adjacent to the reverse drive position."[1] So the P-R-N-D-L order is required by US law. GM used P-N-D-L-R at one time, which led to some unexpected shifts into reverse.
Ford had to recall and fix the 2015 Lincoln MKC. That had a column of buttons in the order P-R-N-D-S-I, the last two being Sport and Ignition. Ignition was the start/stop button, and if pressed while the engine was running, immediately stopped the engine. This tended to happen when someone was trying to go to "sport" mode and accelerate. Worse, it looked just like the other buttons.
My guess is that the hardware engineer that originally designed the board picked a reasonable part for a traditional auto OEM. Then at some point software engineering decided to log all the things. Nobody actually looked back at what the storage system longevity would be with this new lifecycle.
It's a dumb mistake, but not exactly an uncommon one. I made the same mistake early on and I have contract engineers that propose inappropriate storage solutions to me still on projects. Some very insistently. I think a whole lot of people that work on the embedded side of engineering treat storage as a magic perfect box.
The hardware folks picked fancy stuff that was not intended for automotive use. Look up the screen yellowing for an example — no manufacturer would sell a screen in the size Tesla wanted that was rated for hot car interior temperatures, so Tesla went with one that wasn’t.
1.4 drive writes per day is a shit ton of a lot of writes to hit a drive with daily. this is a computational not an automotive concern, & a pretty frelling heinous oversight at that.
I wonder if a “cycle” is actually a full drive write. If it is, then they have ~11GB/day, which is nuts, at least in the pre-dashcam models. But I can easily imagine that a cheap eMMC part has poor wear leveling and eats a “cycle” after much less than 8GB.
It's not, the issue was there is no wear leveling, so it's 1.4 writes per day. This is approximately 2 trips a day, 5 days a week. Reading between the lines I believe the car "eats into" the 3000-5000 cycles everytime it is cycled on.
> I would imagine the average person on the street would expect to see a service life of >20 years.
For a car? Are many people driving around in a car from 2000 or earlier? When I search for cars that old or older it's 0.4% of the used market. Clearly the vast majority of cars do not last this long and I think you have an unreasonable expectation.
This sounded unlikely to me, so I checked a local site and got 4.9% >20 years old in the used market. A lot of people drive their old cars beyond the point that they are not worth re-selling, so I'd expect that to be an underestimate of how many cars are that old.
The article linked suggests the average age of a car on the road in the USA is 12 years - if you're then selling cars that are designed to obsolete in 6 years, that has potential consumer law implications.
Why is that an unreasonable expectation? There are far fewer mechanical parts to fail. Moreover, Elon Musk has announced their batteries can last a million miles. What's the point of the battery lasting a million miles if the car can't? Sure, I may have to replace the motors but I wouldn't have to expect to do that until at least 250,000 miles and it would be a lot cheaper than buying a whole new car and would be worth it if I knew I could expect to get 250,000 miles out of it.
We're flying airplanes that are 30-40 years old. 18 wheel trucks are routinely driven a million miles. I think we need to adjust our expectations that cars need to be replaced every 5-10 years because they're no longer reliable. That's not a wise management of natural resources.
Because they clearly don't often last that long in practice.
Are you saying you wished they lasted longer? Ok yeah that'd be great... but I mean they don't at the moment you can see that for yourself, so expecting it is an unreasonable expectation.
You can expect that cars should last a hundred years if you want... but you're going to be disappointed.
A car accident will dramatically shorten the life of a car. Depreciation will determine the average lifespan of most cars. That is approximately 12-16 years
> Depreciation will determine the average lifespan of most cars.
Depreciation has little relevance to the majority of used-car buyers who need a vehicle, not an investment. They drive it until it can no longer be driven.
What is more relevant to the lifespan of a car is durability and availability of parts. For example Honda shared parts between the Civic, CR-V and Integra so they are relatively easy to obtain even going back to late 90s models.
That doesn't impact how long the cards would have lasted. They're basically moot in this dataset because they were eliminated before wear and tear could kill the car.
"...about 80 percent of all Toyota vehicles sold over the past 20 years are still on the road today." [1]
Used cars are an entry point into an automaker's brand, and they do indeed want their cars to last as long as possible. Furthermore, reliability is a pretty large factor in resale value, and by extension, lease rates.
Well said! I tend to drive cars until they stop working. My first car was almost twenty years old and I loved it. Even if I'm not planning to drive a car that long, I still want the option when I buy it and would have a hard time buying something that felt temporary.
You probably aren't finding cars this old on the market because their owners drive them into the ground.
We were raised to follow this practice and it's hard to imagine doing anything else.
My current car is 20 years old and I have owned it for about a decade. All the other cars I previously owned were also about 15+ years old. I'm not interested in buying a car with a shorter lifespan.
I daily drive a 1997 Dodge Ram pickup, with roughly a quarter million miles on it, and it still works perfectly.
My previous vehicle was a 1984 Ford F350, which I purchased for $800 in 2017 and then subsequently drove from Seattle to North Carolina with only one breakdown (that I could fix roadside).
With proper maintenance and repair, most vehicles can last several hundred thousand miles and 30+ years. The lack of representation of these vehicles on used car search sites is that most of these inexpensive vehicles are sold private-party. Look at your local Craigslist site for examples.
20 years is getting out there but it's not uncommon. A car can certainly be maintained and kept in use much longer than that. The only thing really terminal in the northern climes is rust, once that gets a foothold it's hard to stop. Everything else can be maintained and/or repaired.
My newest car is a 2009, purchased in 2018 for $4,000. Everything works except the tire pressure sensors (the batteries in the wheel sensors have died; I'll replace them at next tire change).
My oldest car is a 1969 but that's a toy, so I don't really count it. My oldest car that I use for transportation is a 1983. The air conditioning has failed on that one, but I live without it.
You maybe can replace the coin cell lithium batteries. The sensors are not built to be servicable but I got it done by scratching off the rubber/glue/expoxy stuff and soldering in replacement cr2025's + sealing it afterwards again. Youtube has some videos
I think looking at the used market might not be a good way to measure, because 20+ are worth essentially nothing, so they don't get sold a lot :) but that doesn't mean they don't continue working. Think that many people cannot afford changing car, even more so when the current car is worth nothing on the market.
Personally one of my cars is a 93, and the other is a 00. Both are DDed half of the year depending on the weather. But, I'm an ethusiast so not exactly the norm.
In rural areas (read lower income) it's not uncommon to see late 80s, 90s vehicles out and about. Some are in good shape, others not so much...
The income factor is a big deal. Rust is also much less of a factor in some regions than others.
Outside of the US, some countries have heavy import taxes on new vehicles, meaning that the secondary market dominates and older vehicles are more likely to be overhauled than junked.
In Tesla's environmental impact report [1], they estimate the average car in the US is driven 12,000miles/year for 17 years before it's scrapped. They amortize over this period for their lifecycle C02 impact of a Model 3.
From my own research, this number is a good guess, though it's region dependant.
This doesn't match the OP which claims Tesla's are only engineered to last 6 years before expensive maintenance.
Up until I bought my 2017 Highlander I was driving a 2001 Golf which was purchased in 2008. It was still running well in the end. I just couldn’t see myself replacing the clutch when I was already sick how poorly it fit my lifestyle. If I didn’t have kids I’d have kept it on the road for sure.
It had 300k km on it and ran just fine. Newer cars run forever with basic maintenance. I sold it for $1k and the buyer planned to give it to their teen as a commuter, so I assume it’ll be on the road a while yet.
"Research by R.L. Polk says that the average age of a modern vehicle is 11.4 years, while the average length of time drivers keep a new vehicle is 71.4 months — around 6 years."
Not sure why you're going after me, but I believe it's relevant because if I'm judging the expected lifetime of a car, it makes sense to know the age of an average car. I looked it up after reading the parent comment so I shared the information for others to benefit from it as well. The post has a couple of upvotes so presumably others also found it relevant.
Yes. Until a week ago I used to drive a car built in 2000 and the average age of cars in my country is 14 years, so cars of 20 or more years are not exceptions.
The sad thing is I can barely find some parts for that 21 year old car that is mostly in perfect running order, but several small buttons don't work anymore, most of the display (showing basic info like outside temperature, fuel consumption etc), rear windshield blade, so at some point good running cars are abandoned because of lack of parts, not because they really need to. Replacing with newer cars is not just the financial cost, but the associated energy and pollution costs that manufacturing processes have.
For electric cars I have huge life expectations, 30 years or more with just batteries and tires as consumables, but the car body and electrical engine should last a very, very long time. There is no reason to accept less than that.
This attitude is what destroys the planet. We need to build things that last long and can be repaired well. It uses much fewer resources than throwing stuff away and making new things.
> Are many people driving around in a car from 2000 or earlier?
A fair number, though not a lot because an accident or neglected maintenance will often kill them first. Though the average age of a car on the road has been climbing quite a bit recently, so 20+ years being common probably isn't unreasonable to see during the lifespan of cars sold today. Well, maybe not for the Teslas, but...
> When I search for cars that old or older it's 0.4% of the used market.
That's because there is little demand outside of collectibles, so they mostly aren't worth selling; if you've got a car that old, you are probably going to drive it until it's ready for the scrapyard, not reselling it.
It depends entirely on your market. It's not that the cars don't last that long, it's that cars this old are commonly shipped out to places with a lower standard of living (and maintenance requirements/testing).
That doesn't seem to bad, all considered. 5% would be the natural number for any one year out of 20, to get to 0.4% from 5% in 20 yrs should be about 8 yrs of mean life per car, assuming exponential decay. I'm fairly sure this thought process is missing some fudge factor to make it exact, but seems about the right ballpark.
The point is you are paying for a car over 7 years, with the expectation that you will drive it longer than that. I'm not sure how old you are (as that may color your experience), but I'd expect a high end, expensive car to last 20 years, especially if it does not have all the parts of an ICE car.
Your service points on an EV are going to be tires and brakes. If the car battery is rated for 1 million miles, how many years is that?
The central control system is a Touchscreen which will be exposed to a variety of heat/cold/Sun exposure how long should it last?
Not everyone can afford to buy a car every 7 years or WANT to do so.
The fact that newer cars are more complex and seemingly expected to have a shorter life has lead to an odd situation where parts availability and price may actually be far better for older cars than newer ones, especially due to the aftermarket.
Is this true? From my memory, many car brands were only good for 100-150k miles when I was a kid. Now it seems like just about all cars made in the last decade are expected to be viable to 200-300k miles
It is funny, how here in Russia there are a lot of talks that modern engines need rebuild after 100K-150K of KILOMETER (not miles), we could not buy car with "million-worth" engine anymore!
It is very common view, that modern cars are much more safe, comfortable, ecologically sane, faster, and even reliable in first 1-2 years, than old ones, but they will break up and require very expensive repair (engine re-sleeving or re-boring with full rebuild) much, much sooner than older ones.
What is interesting to me, that USA users write about new cars that they expect 150K+ miles from them and it was not so in the past. Completely opposite to what users think here.
If Tesla doesn't come up with a diplomatic way of allowing third-party repair they will be the company that forces regulators to add regulation. It's one thing to be out of a personal computing device due to vendor policy but when enough people get burned by a bricked car then all hell will break loose.
> It's one thing to be out of a personal computing device due to vendor policy
IMHO, the only reason people tolerate that is because the culture was habituated to it in the 90s/00s, when personal computers were improving so quickly that you'd likely want to replace yours even if it wasn't broken.
Cars are so expensive and the culture has been habituated to them having increasing life and reliability that I don't think a car company could pull off shortening their usable lifespan without burning their reputation. Until full self-driving capability gets to a deployable state, I don't really see any technological advancement that could ease people into a faster refresh cycle.
I think we'll see leasing pushed harder and harder. The car will be a monthly payment (which is already the case for a lot of people, whether they lease or buy) and after 5 or 7 years it's simply replaced with the current version, and the old one reclaimed/recycled. That gets manufacturers off the hook for having to support old cars with repair parts and services (the latter being especially important for manufacturers like Tesla that don't have dealerships). Fisker is pushing this "ownership" model with their new EV.
That was partly because it was cheaper to upgrade some old parts than buy a whole new machine. You could squeeze an extra couple of years out of a machine with a CPU or GPU upgrade.
Replacing broken parts was less common. No one really expected to just maintain the machine to the same specs for several years.
Random aside, I wonder which is better in an apocalypse, gas or electric?
At first I thought "gas wins" because there will be no power grid. But then I thought "maybe electric wins" because getting more gas might be impossible where as getting electricity I can get from solar cells, water wheels, hand cranks, etc... Maybe I takes 2-3 days of charging for every N kilometers of travel but that's better than zero when you run out of gas.
Maybe Mad Max is showing the age of its source material. All their cars should be electric. They'd even go faster!
Why not both? Plug-in hybrids (PHEVs, range extended EVs, that general class of vehicle) give you the flexibility to run on whatever you can find, and also function nicely enough as a reasonably efficient generator for electrical power.
But if you have to pick one, no question, electric. The supply chains for refining and delivering gasoline far, far exceed the supply chain for charging an electric: Sun, solar panels, inverter, car. Probably some batteries on the house side in there, but if you're charging during the day when the sun is out, those are more or less skipped if you have enough panel.
Mad Max has plenty of problems. "Oh no, guzzoline is in short supply, let's use it as fast as we possibly can!" is not a particularly useful way of approaching resource limits. On the other hand, it makes for some absolutely spectacular cinema! I love that Fury Road was almost entirely practical effects - the giant vehicles driving across the desert look absolutely right, because... well, they were giant vehicles driving across the desert. I understand the Doof Wagon (gigantic wall of speakers on wheels) tended to get stuck an awful lot, though.
Hybrids are false flexibility because if the government posts the national guard to close gas stations, hybrids can only go a few miles or less in some cases on electric power. Electrics can charge at home or at friendly places along the way and with long range, but for hybrids, with 2-30 miles of range, it’s not going to help.
At worst a hybrid is basically a gas car with naive virtue signaling. At best it’s a slightly more fuel efficient gas car.
I'm referring to the 20-40 mile range class of PHEVs - Volt and crew. We're pretty rural, and the 35 mile range our Volt has will get us into town and back comfortably without lighting up the gas engine once.
I have a 2013 Nissan Leaf. If I could shove a Tesla Model 3 sized battery in it I would drive it until every last molecule dissociated. I dread the expensive, spyware riddled, buggy, and probably shorter lived future of the automobile.
Same goes for appliances, which are becoming encrusted with features and less reliable.
A snappily branded, advertised on podcasts, D2C dumb-fridge, dumb-oven, dumb-washer-and-dryer, and dumb-TV company aimed at YoPros would be well positioned to make a mint.
I strongly believe this will happen soon and as a consumer I can’t wait for it. Especially once it turns into a gold rush like the D2C mattress space and everyone is competing to see how much money they can lose to acquire a customer.
Some of these already exist in certain niches. For example, there are a fair number of "chef-style" ranges with no smart features. I have a gas unit with all analog controls and its only electric features are the oven light and the piezoelectric igniters for the gas burners. I expect it will last close to forever though, somewhat ironically, it did cost more than the smart-feature-laden ovens you'll find in most consumer stores. There are also commercial fridges without lots of extraneous features though they're often not ideal for consumer applications due to the way they look, the amount of noise they make when cooling, and their size.
I'd love to see more such "dumb" options in the TV and auto markets but I suspect we overestimate the demand for such things because we HN readers are largely tech-savvy. I think the ideal would be dumb devices with open APIs so that third-party smart features could be added, but that's of negative benefit to manufacturers (lack of vendor lock-in, potential for brand damage due to poorly-implemented third-party features, etc.), so I'm not optimistic about that happening.
An EV with great range (e.g. 300mi), good performance (doesn’t have to be a sports car), and classic controls.
You could go real minimal on computers. You’d just need a microcontroller to monitor speed, temp, charge, battery health, etc. Basic cruise control of the 1990s type would be nice but not essential. A radio with a USB plug for a phone to play music and good speakers is all you need for an entertainment system.
Basically I’m describing an electric 90s Civic with mid range options.
Oh and no cellular modem or spyware.
Put all the money into drive train, battery, and reliability. Design it to have easily swappable and upgradable batteries and to last for 25-35 years.
You’d have one customer at least.
You could probably make a pickup version with killer torque and be like Tesla for red states. Rednecks will be sold on the electric thing once they see what instant full torque means.
So what exactly are the frills that you're setting aside? XM radio? Regen, DC-DC converter, all of these things require more than just a single microcontroller.
Modern engine should meet California emissions. It might require a modern engine control computer and all that jazz; which may be less serviceable but whatever. I'd take a modern computer controlled fuel injected what not over my old carburated v8 that was easy to adjust, but needed adjustment all the time. Just starts every time for me thanks.
But, ignoring engine adjustments, what I'd love in a car is things like using the most common lights; and making the lights easy to access. Because changing the lights is a normal and necessary thing to do.
Don't make it take 30-45 minutes to change the (12V) battery, because changing the battery when it dies is a normal and necessary thing to do.
Basically, design around ease of maintenance, more than making it look cool and hip. I'm not necessarily going to do all the maintenance, but if important things are easier to get to, it saves me money, because it saves my service shop time.
I'd love it if it wasn't 2 cm off the ground, so you could get under it to change the oil; but I'm not going to win that, because I'm exaggerating a bit, but low clearance helps fuel efficiency, and fuel efficiency is required. I'd also love less of the under body panels to help with airflow that get banged up and then fall off or drag and cause terrible noise.
It is normal for cars to outlast many of their parts, which are replaced through periodic service (every x years or x,000 miles.)
Sophisticated, user-facing computing is one of the newer parts of cars and so we don’t really know what to expect yet as far as longevity relative to the car itself. However, we should not necessarily see the need to replace computing devices, or parts of them, as prescribing the life of the entire vehicle.
There's no justification for memory that will hit a wear limit in a few years to be integrated into a couple of layers of components (it's on the controller chip inside the screen control module).
If it was a relatively accessible memory card, it'd be fine for it to have a service lifetime.
It's not just the labor, the repair is to replace the module (rather than just the memory).
They didn't initially design it to fail so quickly, a software update accelerated the wear of the card.
So there's lots of things going on.
People would similarly get pissed off if a $10 mechanical part that was embedded in a $300 mechanical part consistently failed in 3 years. Especially if the failure was due to a change in a fluid spec or something.
Yes, sorry, I meant to ask if you knew the labor to replace the entire module so I could compare it to a mechanical part of equivalent cost.
I realize there's anger but I'm trying to understand how much of a leap this is beyond a mechanical failure that people don't like, and view as an engineering failure, but nevertheless expect and endure when owning cars of a certain age and mileage.
If you don't have the information, that's okay. I'm looking for it (unsuccessfully so far.)
The biggest problem with old[er] cars is that they don't compare to modern safety systems like large crumple zones and advanced airbags. There are many examples where even a 1-year gap in make year will greatly increase the safety of passengers:
I drove an 1983 Land Rover. The joke is that the other vehicle is your crumple zone, and I’m quite conscious that if I hit anything I will break my nose (again) on the steering wheel.
I think every major company that sells consumer based product wants to move to the iPhone model (atleast what I call it). They want you paying them constantly, always, monthly. Pay Tesla 450/month and when an update to their car arrives you can pay 475/month to now have the “latest”.
The pressure for a company to have everyone paying them monthly is just to high. You’ll get pulled into their “ecosystem”, where it’s harder to switch. Why sell a car once per 10 years to a dealer who marks it up and has the interaction with the customer. Electric cars already have less maintenance (theoretically). Tesla doesn’t want you going away from paying them, no company does, and they are all fighting to prevent you from not paying them.
This works great in a software dominated world, or in a world where Moore's law (or something like it) makes the device (though useful) obsolete. However, cars fit in a category that people consider to be a durable good. For decades, now, the expectation is that the car does not become landfill material until long after 10 years of use. If basic functionality is either a) unobtainable; or b) ridiculously pricey before those 10 years are up -- you've just lost a customer. For all their GPS and voice-commands, the Teslas, at root, remain a car. Buyers still expect that the point-A-to-point-B functionality can safely be maintained during the typical lifespan of a car -- and not that of an iPhone.
> However, cars fit in a category that people consider to be a durable good.
It’s not because people consider it a durable good. It’s because people can’t afford or aren’t willing to pay the amount for a new car if an alternative is available.
Many people aren’t going to shell out for a $35k Tesla that only lasts 10 years / 200,000 miles and is worth near zero at the end of 10 years if they can get a Toyota that will last 300,000+ miles, 20+ years, and is still worth a few thousand dollars.
Cars are the second most expensive purchase most people make in their life, they literally can’t afford the luxury of a new one willy nilly (for cars brands aiming to be mainstream brands).
Tesla has a history of being somewhat hostile to third party repairs, so I worry there will be a long between-period when this becomes obviously needed, and when legislation and/or market forces create a third party ecosystem despite Tesla’s attitude.
Even with ICE cars there's been a long history of manufacturers attempting to lock out third-party repairs and modifications, yet there are plenty of companies who specialise in hacking around those limitations.
In the case of an electric car, which is simpler in a lot of ways, it seems more likely that parts will just be replaced with third-party compatible equivalents; which might naturally lead to a Ship of Theseus situation...
The difference had been that most cars did not have a direct connection to the manufacturer. Tesla has the ability to disable third-party parts and accessories remotely unless you prevent the vehicle from connecting.
This is already a thing on some newer Toyotas. For example, if you wanted to replace a FCW camera, you'll need to subscript to Toyota's service to get something signed so the replacement camera will work with your car. "ECU Security Key".
As someone who bought a used 2010 low mileage element. The touchscreen radio/gps is slow and uniquely terrible in a way I had forgotten 10 year old tech is.
It’s not heat but only radio/gps thankfully. Looking for a drop in replacement has been challenging as it’s not clear how one will work with the back up cam and steering wheel controls.
> I wonder what Tesla think is the expected service life of their vehicles?
Their basic warranty is 4 years, and the drivetrain warranty is 8 years. If you buy a used vehicle direct from Tesla they will add on 1 year to the original warranty, but that's it.
That isn't much longer than the warranties you can get on gas-powered cars, and their used warranties seem a bit worse than what you can get from "certified used" programs from other car companies.
In theory, an electric car should last much longer than a gas car, but Tesla's warranty isn't any longer. I bet the price tag to replace Tesla battery packs out of warranty is shocking.
It's simply a good indicator for how long the manufacturer is willing to commit to fixing their own mistakes for free - the point is that Tesla matches or exceeds the duration of their hardware guarantee relative to other automakers. Actual repairability is, of course, a different metric.
EFI should probably be compared with its electronic counterpart in an electric car... not a carburator. The transmission should be compared with nothing because an electric car doesn't even need one even though Elon tried to put one in Teslas. Also, an electric motor is simpler then a combustion engine and usually is a lot more reliable.
EFI is much, much simpler and reliable than carburetor. It has much less moving parts and exact dimensions, it don't need to be re-jetted for different conditions, etc.
Yes, carburetor is more reparable, if you know what to do, but EFI is much simpler.
The thing that's confusing me, is that apparently people who think things without moving parts and adjustments are "simpler", are basically looking at things from an end user perspective. Your response seems to confirm that.
But it's not obvious to me what simpler for an end user has to do with more reliable. I mean, my smartphone usually dies in about a year or two. Actual phones in olden days didn't.
So I still don't get why electric cars would be considered simpler in a sense that's relevant to reliability in expectation.
Carburetors need adjustment and maintenance, but some cars with ECUs and EFI from the 80s are getting old enough that the electronics are actually failing. So how do you define "simple" and "reliable"?
Modern electronics is very reliable. Yes its failure mode typically is all-or-nothing, you could not fix failed EFI control block with epoxy putty and ducktape, only with another block, but it is dead simple — reattach several connectors, and it's all. I'm traveling on motorcycles long distance, and I'm happy, that I could have box size of cigarette pack which will fix my problems with EFI/ECU on the road. But, really, I'm paranoid, I don't need one. And I've owned carburetor motorcycle — problems with carburetor were NORMAL situation, not exception like "lightning hit my motorcycle and fried up all electronics".
Exact sizes and moving parts ARE problem.
Yes, EFIs from 80s (35+ years) are deteriorating due do deteriorating of conformal coating and soldered joints. But how much times jets and membranes of actively-used carburetor should be replaced in 35+ life span? Especially, if gasoline is only so-so? Why replacing parts of carburetor is Ok and replacing EFI control box is not?
You could not buy new EFI, but could buy repair kit for carburetor? It is sad, but it is not problem of technology per se and, again, is not about simplicity, but about vendor lock-in. It is another (but valid!) argument.
And I have opposite experience with modern electronics: I'm still using my first (!) smartphone. As main one, not as backup. I'm using my workstation around the clock for 6+ years (I'm not turn off it for nigh, it works, really, 95% if these 6 years non-stop), my home NAS is built on platform which is 7+ years old (yes, HDDs are much newer, but HDDs are MECHANICAL!). My LCD monitor is 11+ years old (!) and still in perfect condition, and my PREVIOUS monitor is still used by my wife without problems...
It is all anecdotal evidences, but I've replaced a lot of sprocket and chains on my bicycle over last 6 years. I need to replace sprockets and chain on my motorcycle each season.
Yes, if only one transistor on my MoBo dies, I'll need to replace MoBo (and CPU, as it is old and you can not buy MoBo for it now) as one piece, I could not replace transistor as chain or brake pads (grr, really, I can, but it is not very common), but still my computer is much more reliable than my (expensive? almost top-of-the-line) bicycle and my motorcycle, and my EFI-equipped motorcycle is more reliable than previous one, with complex carburetor with many fine passages, calibrated jets, special membrane & stuff.
> I wonder what Tesla think is the expected service life of their vehicles? I would imagine the average person on the street would expect to see a service life of >20 years.
I've wondered about that myself. Because comments I've made that an electric drive train should have an expected service life of 30 years seems to draw a lot of ire from somewhere. Probably the automotive equivalent of found the google engineer.
the average person does not think a car will last over 20 years... on average people keep theirs for closer to 10... what are you basing that idea off of?
If the average person expect their car to least 10 years then a lot of people are expecting their car to last more than 10 years.
Actual average age of cars on the road in the US is around 12 years. If you are not in the rust belts, it is common to still see cars from the nineties. They are usually driven by people who could not afford a newer car or could not afford major expensive or hard to get parts.
> If the average person expect their car to least 10 years then a lot of people are expecting their car to last more than 10 years.
Right, but how many of the people that expect their car to last more than 10 years are buying new cars? The signalling I get from Tesla is that they don't care a whole lot about the used market (as evidenced by this issue).
I think consumers don't realize how much they actually care about the used market. If you buy a $60,000 car that you sell for $20,000 when you're done, you paid $40,000 to drive the car. If you buy a $60,000 car that you can't resell because it's dead or too risky for buyers, you paid $60,000 to drive the car.
Unfortunately, that cost doesn't manifest until years later. Roughly now, based on how long Teslas have been around and their expected lifetime number. I would not be surprised to see Tesla start losing orders if they can't support the used market.
> I think consumers don't realize how much they actually care about the used market.
I think consumers care very much about the value of what is typically their 2nd most valuable asset. They see it when they discuss trading in a car for a new one. Or when shopping for a used car. The information is very readily available.
The people not caring about it consciously prioritizing luxury over resale value.
The revamped Model S and Model X still have big central touchscreens (now in a landscape orientation, rather than the previous portrait-oriented screen), but Tesla went even further by eliminating the turn-signal and gear-selector stalks. Tesla CEO Elon Musk said the car will simply "guess" which way the driver wants to go based on sensor data and navigation maps. Tesla is expected to include backup controls as well, leaving drivers to use the touchscreen for selecting drive, and steering-wheel controls for turn signals.
That is terrifying, and this sort of thing is why I don't seriously consider Teslas now I'm looking to buy an EV.
On top of that, at least where I live, you must use your turn signal to indicate changing lanes at least 100 feet before making the change. The car cant possibly know what I'm thinking.
It's not just that feature, though. I wouldn't (in fact, didn't) buy a Tesla because of the whole attitude to the control system: let's put everything on a single screen in the interest of aggressive design minimalism. Their approach to self-driving scares me witless. I hate the childish easter eggs and the gimmicky doors. I love the EV technology, but I really dislike the car they've wrapped around it. That new feature is just another example of the general approach.
I ended up with an EV that is just a normal car that happens to be electric, and it's great. I'm in the market for another EV now, and despite the fact that the Tesla X is technically the best solution for what I want to do (towing) it's not even in the running because of all the rest.
This is exactly the free marketing Tesla expected with releasing the renderings with the yoke.
There are also renderings with a traditional wheel buried in the HTML.
The outrage drives dozens of articles and a hundred million clicks, making the car the most talked about new vehicle for the next several weeks.
“You won’t believe this one ridiculous thing about the new Tesla!”
Mission accomplished. I would bet $100 that the Model S ships with a physical way to activate turn signals on the steering apparatus.
Aside from turn signals, I think they will try to do away with the physical stalk to “gear shift” and it will ultimately be an improvement.
As-is the right hand stalk is overloaded with functions for cruise and autopilot, on top of gear shifting, which makes the current incarnation sub-optimal in several specific usage scenarios. So it doesn’t surprise me they are looking to change how it functions.
In addition Tesla used a cheap eMMC with a limited number of write cycles.
Tesla did reduce the excessive logging and even implemented a notification when the car suspects it is failing. I think they even extended the warranty, but of course reducing the logging will push some failures outside of the warranty period.
This recall thankfully fixes this problem permanently for those who could be affected.
Or made the EMMC storage user-accessible and replaceable the same way that the LTE radio is (it’s hidden behind an easily removed panel behind/under the main display)
I have an original retina MacBook Pro, ordered the day they were announced, before the presentation was over, in fact. Works fine. I have ThinkPads that are even older. Tesla was cutting corners, no doubt about it. As the regulatory burden increases, they'll adjust. It is an interesting idea though, a first-order example of regulation moving a market.
I agree they were cutting corners, but comparing it to a laptop isn't that useful because as far as I understand it the main issue with screens in cars is the high temperatures they can be regularly exposed to.
It's a fair comparison when you take into account what they are designed for - a laptop isn't designed to be used in a place where it can get to the temperature that can kill.
A screen in a car needs to be designed to take into account the environment, and it looks like in this case it wasn't.
That isn't the issue, the issue is the limited life cycles of flash memory, 3000 to 5000 cycles.
The article says that in paragraph four:
The government agency identified the source of the problem as the touchscreen's NVIDIA Tegra 3 processor and its integrated 8GB eMMC flash memory chip. When the chip reaches "lifetime wear," it can no longer maintain file-system integrity, causing failure of touchscreen functions, according to the NHTSA
There are plenty of people who drive classic cars built from 1950 to 1975. Some people drive even older cars. I once saw a Ford T stopped at a 7-11 to buy gasoline.
Will that Tesla S be the same in 2045 to 2090, or even in 2120? It's looking like the answer is no. The only ones left will be mounted on platforms in car museums. There will not be auto shows featuring real Tesla cars, because the electronics will be dead and impossible to fix. No reasonable part replacement will be possible because of digital signatures.
We'll see some car bodies that get hollowed out and then draped over the innards of newer cars. Somebody already got an unsupported Tesla S running again by installing a gasoline engine, even cutting through the floor to install a transmission tunnel.
> No reasonable part replacement will be possible because of digital signatures.
Bah. Just strip out the electronics and replace them with something newer and more open. Keep the body and the motors, replace the batteries and the computers.
> but Tesla went even further by eliminating the turn-signal and gear-selector stalks. Tesla CEO Elon Musk said the car will simply "guess" which way the driver wants to go based on sensor data and navigation maps.
That's impossible. Maybe an AI could guess left/right turns, but signals are also used to signal merges; if I need or want to merge, how could the car possibly know, and know which direction? In a normal vehicle being operated responsibly, the only indication that a merge is going to occur is the driver engaging the signal.
(Though I admit, many a driver doesn't start to signal their merge until their car has half left the lane. A signal then is often useless: I already know they're merging because the car is drifting recklessly into another lane.)
I’m not even sure about the turns. If I suddenly decide I want to stop at a shop on my route home where I’ve never gone the car wouldn’t begin to signal until I’ve slowed, which is better than nothing.
Unless the car is intended to be completely autonomous, it presumably supports non-autonomous driving, and therefore should have a freaking physical turn signal control.
If it's intended to be totally autonomous, why does it even have a steering wheel?
I recently talked about this with a friend who is an engineer at a large traditional car manufacturer (working on sensor technology for self driving). It came up out of a completely different discussion (I had not heard about touch screen failures).
He said that Tesla is well known for using general consumer grade parts (he mentioned the touchscreen as an example) instead of automotive parts (which have typically much longer expected livetimes). That allows the to push the price down but also to iterate faster.
The worrying bit is not really that Tesla does it, but that the mandated strategy at that manufacturer is to only compare themselves with Tesla (all others don't count), so I think we will see a drive to lower quality from all manufacturers to emulate tesla. This is a well know premium manufacturer.
If you are interested in the potential tragedy of modern electronics longevity, look-up “tin whiskers”.
My prediction is that the RoHS regulations that brought us lead-free solder will go down in history as the cause of massive environmental contamination around the world. In other words, exactly opposite what they thought they would achieve.
How’s that missing turn signal stalk supposed to work when you’re supposed to signal turn well in advance of actually turning? If you’re using GPS, then it might be enough but people miss turns all the time, change their minds due to traffic conditions and the information in GPS systems is not always correct.
The revamped Model S and Model X still have big central touchscreens (now in a landscape orientation, rather than the previous portrait-oriented screen), but Tesla went even further by eliminating the turn-signal and gear-selector stalks. Tesla CEO Elon Musk said the car will simply "guess" which way the driver wants to go based on sensor data and navigation maps. Tesla is expected to include backup controls as well, leaving drivers to use the touchscreen for selecting drive, and steering-wheel controls for turn signals.
I suspect many of the people who buy Tesla's are the sort of people who trade in their cars for a new one every 3-5 years. So really it's the second hand buyers who get shafted with no warranty protection at all.
> Tesla will replace the VCM daughterboard with an "enhanced" eMMC controller. The recall is expected to begin March 31, and Tesla will notify customers of when to take their cars to a service center to have the work performed free of charge.
Its not the touchscreen its the emmc. Yeah that sucks but its not like they're bricking old cars with expensive or unserviceable parts.
>It could have been a cartridge or easily serviceable module.
Especially if they are arguing that it's normal for the memory to fail in 6 years. I can't imagine all of my center console buttons failing after six years and requiring a whole new center console to get up and running again.
I'm really curious to see how Tesla's reputation will change over time. As they scale and have to provide cheaper cars (assuming they want to go after a larger market anyway, which they've publicly stated for over a decade), and have to deal with more customers who are less forgiving than early adopters, and also have to deal with legacy decisions from their early years of cars failing and all of the weird service issues and NDAs and bricking cars that are resold and have to make good on their FSD promise and (I could go on - my point is not to hate Tesla but merely to point out they are still in the early stages of being an adored consumer company because most consumers currently cannot afford them) well, I think Tesla is going to earn a similar reputation as other car manufacturers as time goes on.
Yeah, I know Tesla sells things besides cars. But I think their reputation will be based on things they sell to the public, and I think most people will judge them on their cars.
> have to deal with legacy decisions from their early years of cars failing and all of the weird service issues and NDAs and bricking cars that are resold and have to make good on their FSD promise
I honestly think they were hoping to make that up in volume. As it gets cheaper to make cars at scale, some portion of the savings goes to fixing up the issues with the cars they already shipped. They haven't been able to scale their pipeline, though.
> I think Tesla is going to earn a similar reputation as other car manufacturers as time goes on.
Long term, I don't think it's going to go well. "Move fast and break things" is a bad principle to apply to a car. I can use DDG or Bing if Google doesn't work, but I can't summon a Kia when my Tesla won't start. I don't think they can use that velocity to create features consumers want enough to deal with the instability. They might strike gold elsewhere though where the instability is less of an issue. It might be okay for self-driving semis to break down, if they're cheap enough that you can afford extras to use when some break.
It will be interesting, they're having to learn hard lessons about support and customer service that other car companies learned half a century ago. The question is can they learn those lessons faster than the other car companies can learn to build comparable EVs.
I would say they're going to have a hard time of it.
But it's a good candidate for someone looks for a reliable car. Someone hit me when the car on auto pilot. My hands are on the wheel but it's really hard for Tesla to take back control. The car was shaking and I have to apply a force to the drive wheel to force it went straight.
With Ford adopting Android Auto & many others doing the same, I wonder how many cars are built to last more than 5 years before they become insecure digital trash buckets.
Not true. If you dig deeper, you’ll find Tesla was originally charging Model S and X owners $2500 to replace failed MCUs. Only under regulator scrutiny did they concede and issue a recall.
Owning both a Model S and X, I’ve seen Tesla first hand attempt to evade covering warranty work on these models.
Friendly reminder to opt out of arbitration when you buy your Tesla (when the VIN is issued).
I read a bit of that. It seems really weird that these letters are sent to Tesla. So if some lawsuit WAS going to happen, Tesla can say they never got the letter. There's something really fishy about Tesla.
There is a letter linked from Tesla outlining why it doesn't consider this a defect:
* The memory should have lasted 6 years not 3.
* The NHTSA is unreasonable when it requires the system to last the expected lifetime of the car (12 years).
It might just be ass covering, but currently the official stance from Tesla is that a car that falls apart after 6 years should be enough for anyone, because anything else is too hard.
My car is about 10 years old and it's fantastic. Granted, it's only done 99,000 km (~61500 miles). I imagine it'll continue to serve me for years to come (unless I get a bad case of the car fever) and then another owner for even more.
The irony is that hybrid/electric vehicles are being owned longer than fossil-fuelled vehicles before being sold.
This is definitely a Tesla mistake, probably simply owing to lack of corporate experience and long-term planning. Could even be poorer access to (specialized auto-grade) parts owing to their small scale in the past.
It's especially bad when you consider one of the main selling points of an EV is simplicity and reliability. Or at least that's one of the main arguments I've heard over the years (there are lots of other reasons obviously).
I don't know what you'd call it exactly, maybe extreme sloppiness, but part of the problem is that the unit is doing excessive logging to the disk without any mitigations. That doesn't seem hard to fix.
But worse than the intentions is the fact that Tesla doesn't treat these things are design flaws that should be addressed even out of warranty. Is it right hat they feel its ok to profit from these flaws?
Flash wear leveling and expected lifetime are pretty basic engineering requirements when you work with eMMC based embedded devices.
Every company I've ever worked at has done this analysis and also set up metrics to catch a component at the system level causing heavy flash wear by excessive logging, caching or other write heavy operations.
Just turning off the enormous amount of crap written to syslog would probably have added years to the eMMC life. Why did it take them so long to do that?
If you drive a car with the Tegra CID and current software, you get the distinct impression that they really wish all those cars were dead.
It's intentional. Tesla wasn't covering these repairs under warranty and are upselling people with failed MCUs to MCU2 upgrades which cost $1500+labor.
If you search around, you will also find countless posts of people being denied warranty work for their Tesla's or fixes only being billed as good will after the customer fights back.
Tesla didn’t care because there was no impact on them for their poor logging configuration until a regulator stepped in (as the costs were borne primarily by the customer).
They probably also do want cars with lifetime supercharging dead (which are primarily MCU1 vehicles).
My worry with a large number of modern tech heavy cars will be that while the centralisation of a lot of these controls and features into one panel may be convenient to the user, when these inevitably fail (as in this case) - what can the owner do when the parts are no longer available, or if the manufacturer is out of business? Mechanically the car may be sound - but if the system responsible for the interaction between the vehicle and the owner is borked and impossible to replace does that mean the car is a complete write off?
I think back to some of the beaters my friends owned when they started driving, with broken HVAC switch gear, burned out audio head unit wiring harness, etc... while mechanically sound (enough to be road legal...), issues like this didn't hinder or prevent driving the car, and they continued in service for years in this state.