-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
What is going on with those Model 3/Y Accidents in China / Paris ?
- Thread starter Nabush
- Start date
pikeman
Member
I was thinking of all the accidents the FAA investigates and the ultimate determination. It might have been caused by a number of root causes, like cardiac arrest, but it’s the human element that’s the real cause. Not the aircraft. Same with Tesla. I think there’s many people just waiting to pounce on Tesla as the problem when it’s “ the nut behind the wheel “ who is actually responsible.
Again, while it may be that the driver is ultimately at fault, we need to look at exactly WHY they made the mistake(s) they did. For example, Air France 447 crashed into the Atlantic because of an incorrect pilot response to a mechanical issue (pitot tube icing up) that caused the autopilot to change the way it operates (it went from what's called 'normal law' to 'alternate law'). The change in autopilot mode was not immediately noticeable (design flaw), but the aircraft crashed because the pilots thought the autopilot had more control than it actually did (pilot error) but they were led down that path due to poor design.
rpiotro
Active Member
I think they just mishandled the situation. 5° nose up and 80% power is the prescribed cure for the situation they found themselves in. Their artificial horizon was still working.
Airbus and Boeing have very different design philosophies. Airbus wants to take the pilot out of the loop as much as possible. Boeing just the opposite.
That's a good point. Getting back to why people mistake the pedals in a Tesla, I do think there are contributing factors on top of people being idiots.
1. Light accelerator pedal pressure causes the car to brake at certain regen levels
2. No acceleration noises (like many other ev brands) or engine noises
3. Drivers trusting auto emergency braking features without realizing the circumstances in which it will turn off
Now I'm not saying that Tesla needs to change anything, but I do think the design of the car contributed to how drivers reacted, where they totally freeze up and are unable to understand and correct their mistakes.
Yes. It's obvious they mishandled the situation. But the change to alternate law was not pronounced enough to really clue the pilots to the fact that the aircraft would no longer prevent a stall. Couple that with the fact that the pilot's and co-pilot's controls are both electronically controlled joysticks and not the traditional interconnected yoke where one pilot can get feedback on what the other pilot is doing, so they were working against each other (one pulling the nose up inadvertently while the other was trying to push the nose down) which prevented recovery from the stall.
Apparently there is a new petition filed with the NHTSA claiming the steering hardware causes a voltage spike causing the accel pedal sensor to miscalibrate and causing real SUA. Breaking: NHTSA Petition Claims Tesla Sudden Unintended Acceleration Is Real (but Fixable)
the petition: https://www.autoevolution.com/pdf/n...d-acceleration-is-real-and-curable-217525.pdf
the petition: https://www.autoevolution.com/pdf/n...d-acceleration-is-real-and-curable-217525.pdf
stopcrazypp
Well-Known Member
The petition is from the same guy (Ronald A. Belt) which has thrown around many different theories on SUA (some of which seem plausible at face value, for example his previous regen braking theory), which have been disproven and denied by the ODI, with ODI saying he was misinterpreting data. The article doesn't mention that at all, but you can see the name of the guy. As such I would take this with a huge grain of salt.
The procedures he is doing reminds me of the Toyota SUA case, where one side hired an expert that used resistors to get the sensor to respond. The internet went crazy about an electronic cause, but no government ever accepted that explanation (the only issues governments accepted were floor mats and mechanically sticky pedals, both of which were found in the real world). The mechanism he used this time is more plausible than the resistor stunt done in the Toyota case, but given it is from him and he seems to have an axe to grind (as such he may miss some obvious factors that make his scenario impossible or implausible), I tend to not trust what he says until it is further verified.
Edit: His whole premise is based on a 1.65V calibration signal Tesla uses in an ADC and that when that dips below 0.33V and if the car happened to be doing an ADC calibration (he is not clear however how often this happens), it can cause the signals to seem like acceleration. He simulates a dip by injecting a voltage at the chip, which is ok for testing purposes, but it makes a whole lot of assumptions about circuits upstream. First the 1.65V is supplied by a 5V regulated supply. He says the signal can dip below that when the 5V supply dips to 1V. He further suggests that the 12V supply can dip to low levels (down to 2V) which can cause the 5V supply to dip that low. However he never demonstrates this on a Tesla (even though he claims it can happen due to inrush current to the steering motors)!
On page 16, the evidence he uses for a 12V supply dipping is from a GM recall that showed a power steering motor can cause the voltage to dip during a slalom maneuver. He doesn't actually show testing on how voltages respond when a steering motor is maxed out on a Tesla. At minimum he first needs to demonstrate how the 12V supply responds, then how the 5V regulated output responds (given there may be capacitors, etc upstream that would buffer any dips), then down to the 1.65V signal.
The entire premise also is dubious given most SUA incidents happen with no steering input. He says it's from slow parking maneuvers that causes a current inrush to the steering motor, but his own example (from GM) is from rapid slalom maneuvers, not from slow parking maneuvers.
Last edited:
This is another paper written by somebody with nothing better to do. (Or maybe he is being paid by short sellers?)
The prior two papers were written in the same way, positive that they identify the issue, but both were proven false.
The article said he proved that it could happen, but just like his prior papers he described a way you might test his theory, but he does no actual testing. It is all just him thinking about things based on information from other people. Like including a graph of the 12v supply in an ICE as proof...)
@wk057 do you want to take another look at this guy's work? (Is the DC-DC converter limited to 200A and can be overwhelmed by the EPAS?)
Gauss Guzzler
Safety Score = 42
I don't think he actually did that, he just described that you could test it that way. I doubt he has actually ever touched a Tesla.
I would think a better test would be to hook up a Iarge 12v load that you could control to stimulate what he thinks the EPAS is causing.
Gauss Guzzler
Safety Score = 42
Funny that he vomited up 47 pages of drivel but couldn’t find 5 minutes to actually test his theory with a volt meter.
stopcrazypp
Well-Known Member
I saw the picture on page 20 with a wire to pin 44 on the chip and I assumed that was what he did himself. But you are right that the way he worded it suggests he might not have done it himself.
Yep, that would be the way I would test it too. That would eliminate a lot of assumptions about a whole range of circuits in between. Even with this test, the other thing to test would be if the steering motor (or other real car loads) actually would produce such a spike and for how long.
Yeah, I think that picture, and all of the other inverter pictures, are from Damien Maguire's videos of his hacking efforts to make it so Model 3 drive units could be used in EV conversion projects.
Just like with his prior papers, I don't hink he tested anything himself. As I recall one of them would have been as simple to test as disconnecting the brake switch, but yet he couldn't be bothered to do that to prove his theory. (Probably because he knew it would prove his paper false.)
Tha_Ape
Active Member
You'd think if this was problem you'd see it happen a lot more with millions of Tesla on the road
This is once again not how any of this actually works, nor is it how it's even implemented on the inverter. I don't have time to dive into all of the details this evening, but suffice it to say I skimmed the document linked and it looks to be basically the same bogus claims about voltage drops with much more fluff than the last paper.
I love how a scope graph of 12V drops on an ICE car with EPAS is used to try prove a point.
lol. Something tells me an ICE alternator isn't going to be able to respond to large loads as quickly as a digitally controlled DC-DC converter with ~20 kHz+ switching. Oh, right, there's never any such voltage drop on a Tesla for that reason. At worst the DC-DC will allow the 12V rail to hit about 11.5V when purposely cycling the lead acid battery... which isn't even a thing on newer cars anymore, so the rail is kept pretty constant on those.
Oh, and something like 10-15 other modules on the car send 12V readings to the gateway for logging, and 3 or 4 of those will latch any dips < 10.8V and ensure that makes it to the log regardless of polling rate. (The only time I've ever seen dips like that were with a failed DC-DC, or on a car not supported by HV... clearly not the case if you're speeding away.) I've never seen any 12V dips in any of the logs I've examined from "SUA" claim vehicles from any module. In fact, I explicitly looked at these datapoints when debunking the previous "paper" that claimed the same nonsense.
Happy 4th.
Edit: Addressing this specific:
The EPAS maxes out at about 40A draw, and only in short bursts. The DC-DC is rated for 200A continuous, but can do WAY more than that in short bursts (I've pulled 400A from one for several seconds).
Most vehicles don't see more than ~100-150A load even in the worst possible situations (EPAS going, fans maxed out, seat heaters on, radio blasting, etc).
I love how a scope graph of 12V drops on an ICE car with EPAS is used to try prove a point.
lol. Something tells me an ICE alternator isn't going to be able to respond to large loads as quickly as a digitally controlled DC-DC converter with ~20 kHz+ switching. Oh, right, there's never any such voltage drop on a Tesla for that reason. At worst the DC-DC will allow the 12V rail to hit about 11.5V when purposely cycling the lead acid battery... which isn't even a thing on newer cars anymore, so the rail is kept pretty constant on those.Oh, and something like 10-15 other modules on the car send 12V readings to the gateway for logging, and 3 or 4 of those will latch any dips < 10.8V and ensure that makes it to the log regardless of polling rate. (The only time I've ever seen dips like that were with a failed DC-DC, or on a car not supported by HV... clearly not the case if you're speeding away.) I've never seen any 12V dips in any of the logs I've examined from "SUA" claim vehicles from any module. In fact, I explicitly looked at these datapoints when debunking the previous "paper" that claimed the same nonsense.
Happy 4th.
Edit: Addressing this specific:
The EPAS maxes out at about 40A draw, and only in short bursts. The DC-DC is rated for 200A continuous, but can do WAY more than that in short bursts (I've pulled 400A from one for several seconds).
Most vehicles don't see more than ~100-150A load even in the worst possible situations (EPAS going, fans maxed out, seat heaters on, radio blasting, etc).
Last edited:
rpiotro
Active Member
I'm thinking follow VW's lead. ID.4 Do you think it would help?
PianoAl
Active Member
Yes. If you've ever been a passenger with a younger driver or someone who's going too fast and realized that you've been pressing your foot against the floor, you know how automatic this can be.
