The company says in the documents that the front windshield wiper motor controller can stop working because it’s getting too much electrical current. A wiper that fails can cut visibility, increasing the risk of a crash. The Austin, Texas, company says it knows of no crashes or injuries caused by the problem.
In the other recall, a trim piece along the truck bed can come loose and fly off, creating a hazard for other motorists.
Tesla says in documents that the trim piece is installed with adhesive, and that may not have been done properly at the factory.
There’s a lot of Tesla hate on build quaility on Lemmy/Reddit, but they actually have some of the most reliable power trains and exceptional software. There was some problems back in the very early days of the Model S with the 85kw/h packs, but they’ve moved beyond that now.
Given this is part of the power train, and software, it’s right in line with their expertise.
I wouldn’t trust any software from any manufacturer to steer for me, at least not in such a way that I can’t easily disable and use a mechanical manual backup.
Also, steering isn’t “power train.” If you’re gonna lump it in with something, it fits closer to the [alleged] self-driving system.
So I looked it up, and you’re right, steering isn’t part of the power train. Looks like its it’s own thing.
I don’t think it’s right to lump it in with the L2 system though as it’s taking your input and translating it into the appropriate movement. The L2 driving system decides to go left or right and will send the same signal you would by steering left or right. I guess it’s just it’s own thing just like power steering is it’s own thing.
Obviously this is first gen tech in cars, but it’s been around for quite awhile in aviation with no backup mechanical link, we haven’t all died yet.
Tesla’s system is triple redundant, but that doesn’t guarantee something won’t go wrong, only time will tell on that one. Maybe we learn triple isn’t enough and the NHSTA mandates quad?
I do have a feeling though, that within 20 years or so, it’s almost all going to be steer by wire. It’s safer in the event of a crash as you don’t have the steering shaft in front of you. It also saves space from not needing all the mechanical linkage. I imagine insurance would have higher rates on cars that don’t have steer by wire as well due to increased risk in accidents.
Edit: I would add though that motors are part of their specialty, and the steer by wire system is using 2 of them, so they do get some of their existing expertise on that.
Edit: I guess the triple redundancy is on the steering input. Obviously with only the 2 motors, thats only double redundant.
Exactly: the same signal. If the electronics controlling it receive one input from the steering wheel and a different input from the self-driving computer, are you sure it will prioritize the steering wheel input in every single possible circumstance? 'Cause I’m not!
The difference between this and regular power steering perfectly illustrates my concern: the way power steering works is that it assists the driver’s movements by amplifying the force that you’ve applied to the wheel. If it fails, you can still steer the car; it’s just harder. (I know this from personal experience BTW: the power steering in my old pickup truck is out right now. I haven’t fixed it yet mostly because I’m still deciding whether I want to keep it or downgrade/simplify to a non-power steering rack.)
In contrast, if something goes wrong with this system, it is very unclear to me that the driver could override what the car wants to do, no matter how much force you apply to the steering wheel. Or, for that matter, if turning the wheel would be effective at all: you might end up just sawing the wheel left and right with no effect whatsoever on the way the tires are pointing.
I don’t like those failure modes! At least in a mechanical steering system, for it to fail completely like that would require something like a tie rod breaking or the splines in the steering column shearing off – in other words, metal ripping apart that (a) shows warning signs you can easily inspect for (e.g. deep rust or cracks on the tie rods), (b) you probably notice happening because it makes noise, and (b) probably happens kinda gradually rather than instantaneously because steel is ductile.
I’m not fully opposed to self-driving, by the way: it’s just that (a) I want the system to be Free Software so I can inspect and trust the code, and (b) I want it to be coupled to the steering column with a belt or a clutch or something that can slip and allow me to mechanically override it if I yank hard enough on the steering wheel.
First of all, aviation has vastly more stringent oversight than cars do, in terms of manufacturing regulations, maintenance regulations, and pilot regulations.
Second, fly-by-wire passenger jets are also just categorically different not because it’s flying vs. ground transport, but also because it’s public transport vs. an owner-operated private vehicle. If I’m already entrusting my safety to a pilot or bus driver anyway and they decide fly-by-wire or drive-by-wire is acceptable, that’s one thing. But when I’m the one operating the thing myself, it’s entirely another.
That’s an interesting point there, but have you considered that even with a mechanical link and current safety features, it can still override you? I unfortunately almost drove into someone at very low speeds in a dark rainy parking lot, but the cars safety systems overrode me thankfully. I don’t think they would have been injured it was so slow, but just to show that nowadays with cars you don’t always have full control. In that case it was the brakes not steering, but modern cars can and will prevent you from changing lanes into someone in your blind spot for example.
All of it, mechanical link or not, ultimately comes down to software now, and I’m not sure there would be any real difference between your steering wheel input overriding a self driving system with or without a mechanical link. Putting torque on the wheel while in these semi self driving modes disables the self driving features, but that’s software that disables it when you take over. What if that software failed and you were now fighting the self driving car also trying to steer and as you tried to steer it put equal power against you thinking the steering was rough?
Also with steer by wire if there is a catastrophic failure, turning the wheel will simply do nothing. E.g if both motors simultaneously die, or somehow all 3 signal wires at the steering wheel are cut. Edit: I believe it’s dual battery redundant as well, so the HV and 48v would both need to die to kill steering.
I’ll be honest with you: all but one of the half-dozen (which is too many, BTW) cars I own have manual transmissions, and half of them don’t even have ABS, let alone any other fancy electronic nannies. I mention that to help explain the extent to which I am fundamentally Not On Board with anything that interferes with my manual control of the car. (I’m also a Linux user and a DIYer, which are some more clues to how much of a control freak I am: I expect my property to be exactly the way I want it to be and do exactly what I want it to do, and nothing else.)
Don’t get me wrong: I wouldn’t mind having radar cruise with lane-keeping for long trips on the freeway, but only if such a system were fail-safe enough that even if it were stuck on, yanking on the wheel hard enough would get the car to turn. I would absolutely insist on the maximum torque the self-driving system could apply being much less than the strength of the human driver. I don’t know if that’s the case in late-model vehicles or not, but if it isn’t, I would consider those vehicles to have an unsafe design.
Well that’s fair. You know what you want, nothing wrong with that.
I really don’t know on this either. Even if you can overpower it though in a worst case fighting you scenario, I imagine that would be a pretty fucked up driving moment.