Tesla "logs" vs reality... potential problem?

[mongo]

Doesn't matter.. why in either case would there ever be a reason to ram it at full throttle? I agree that the driver should be in control, when it is reasonable to assume the driver is, in fact, in control. Sudden 100% throttle is not a reason to assume that. I'd say the ONLY time anyone ever plants their foot (in a Tesla) full to the floor is when they are launching. Even passing there's no reason to put your pedal to the metal. In those cases, it is perfectly sane for the car to be given first reading and only allow it if the sensors show clear ahead. That's just plain common sense stuff. What "edge case" could you possibly come up with otherwise?

I'm originally from norther Michigan, sometimes the road you are turning on isn't plowed so the only option is to back up and ram the plow induced drift. (admittedly you might not need the full power of Tesla to pull off that maneuver)
Or there is pile up, a uhaul tow behind trailer is in front of you, a semi is bearing down from behind. What do you do?
Carjacking? Regardless, my inability to think of the scenario does not mean the scenario does not exist.


The accidental flooring may be part of reason for chill mode. The instant power is a little disconcerting to me.

 

[bcsteeve]

When you say TPS are you referring to the Throttle Potion Sensor on the throttle body?

Neither: Throttle Position Sensor (I'm sure you didn't really mean "potion" though lol)

Did those have stepper motor drive? (antoehr layer of redundancy)

None, no.

For accelerator pedal on drive by wire, there are two potentiometers, one increases and one decreases as the throttle is pressed (can also have different supply voltages). If they don't track, the ECU throws a DTC (P2135, P2138 are related), and may go into a limp mode (had to replace wife's when it went bad).

Not necessarily. For the MB product I worked on, in particular, the voltages were offset and not inverse and had the same supply voltage. The second potentiometer was there purely as a failover, rather than a failsafe. Limp mode would not be triggered by the TPS fault, nor could you even throw a TPS fault without disconnecting both. Good design? Well, I think it was appropriate for the car. As I've said, it has been my experience that I've never seen a system where the outputs are tracked, monitored, etc. I'm sure some do. I have significant experience in terms of design time, troubleshooting, oversight and production management however it is limited to 3 makes and 4 models only.

Electronic Throttle: Where's The Cable?

No offense to Mr. Weaver intended, but I question some of his research. Much of what he writes is partially true for some manufacturers, but it is by no means universal. As I said, I've done work [independently] on Mercedes, [employed] Renault and [oversight] Toyota. None of the systems I've worked on mirror what that article describes, at least not exactly. He's taken bits and pieces and jumbled them together in an attempt at a technical article. Perhaps that's appropriate - I don't know his intended audience. Take, for example, this bit:

Some manufacturers use a separate electronic throttle control (ETC) module that is wired to the throttle body and PCM. Other manufacturers have their throttle body wired straight to the PCM. In both cases, information is sent over the CAN-bus or serial data lines for faster communication between the various components.

That most closely describes the Mercedes product. However, throttle data was not "sent over the CAN-bus [...] for faster communication between the various components". First off, what "various components"? Its really just two... the TPS and the ECM. And they are directly wired analog out to analog in. It doesn't even use CAN for that communication. The ECM then processes that and sends a signal on the CAN for diagnostic purposes, but it is indirect. CAN is infinitely slower than a direct analog connection, so I'm not sure why he thinks it would speed things up.


Anyway, I didn't intend to derail this conversation into a technical one on speed control. The fact is, I have no idea how Tesla does any of what they do. It has just always bothered me that they scapegoat customers citing "logs" that may or may not exist, and it bothers me more how the media, presumably insurance companies, and apparently the general public just takes their word for it, particularly now that I feel - from the conversation I had with an employee who, admittedly, may have been talking outside of her paygrade - they've demonstrated to me personally that the logs can't be trusted. At least not 100%.

 

[bcsteeve]

I'm originally from norther Michigan, sometimes the road you are turning on isn't plowed so the only option is to back up and ram the plow induced drift. (admittedly you might not need the full power of Tesla to pull off that maneuver)
Or there is pile up, a uhaul tow behind trailer is in front of you, a semi is bearing down from behind. What do you do?
Carjacking? Regardless, my inability to think of the scenario does not mean the scenario does not exist.


The accidental flooring may be part of reason for chill mode. The instant power is a little disconcerting to me.

I have no doubt the chill mode is related, absolutely. They're spinning it as a "feature", which is fine... I suspect it is a liability move. Now, at least, you'll have to choose to be able to ram buildings, so if you don't choose "chill" you're on the hook for it. I'm going to guess that eventually it is going to be the default.

Ramming snow isn't going to be 100% throttle. And if a semi is bearing down on you, the extra 3 feet you're going to gain by making your problem the person in front of you's problem isn't going to make the difference. Carjacking? What? The scenario doesn't exist because it can't. And even if it did, you don't give airline passengers the ability to open the door for the 0.00001% chance that's the best thing to do. You don't not have safeties on guns because of the 0.00001% chance having it causes a problem. We have safeguards all over the place that in some conceivable situation cause problems, but on the balance they do more good than harm. Seatbelts, airbags, etc etc. The argument has always been - briefly - that we shouldn't have it because "what if...". The same goes for this. So far, some people and property has been damaged because of the way it is currently. Nobody - I suspect - has avoided damage because they were able to go full blast into an object in front of them.

Incidentally, Tesla in the past has come out on the right side of "can we, should we?". It is now standard for the car to not overheat and potentially kill children/pets/elderly left in a car in the summer. They had the ability to do the right thing, and they did it. Yes, you can shut it off, and perhaps that's the compromise with what we're discussing here. Perhaps the default should be "don't launch into walls" and you can shut it off if that bothers you... but then its fully on you.

 

[mongo]

Neither: Throttle Position Sensor (I'm sure you didn't really mean "potion" though lol)[None, no.
Not necessarily. For the MB product I worked on, in particular, the voltages were offset and not inverse and had the same supply voltage. The second potentiometer was there purely as a failover, rather than a failsafe. Limp mode would not be triggered by the TPS fault, nor could you even throw a TPS fault without disconnecting both. Good design? Well, I think it was appropriate for the car. As I've said, it has been my experience that I've never seen a system where the outputs are tracked, monitored, etc. I'm sure some do. I have significant experience in terms of design time, troubleshooting, oversight and production management however it is limited to 3 brands and 4 cars only.

No offense to Mr. Weaver intended, but I question some of his research. Much of what he writes is partially true for some manufacturers, but it is by no means universal. As I said, I've done work [independently] on Mercedes, [employed] Renault and [oversight] Toyota. None of the systems I've worked on mirror what that article describes, at least not exactly. He's taken bits and pieces and jumbled them together in an attempt at a technical article. Perhaps that's appropriate - I don't know his intended audience. Take, for example, this bit:

That most closely describes the Mercedes product. However, throttle data was not "sent over the CAN-bus [...] for faster communication between the various components". First off, what "various components"? Its really just two... the TPS and the ECM. And they are directly wired analog out to analog in. It doesn't even use CAN for that communication. The ECM then processes that and sends a signal on the CAN for diagnostic purposes, but it is indirect. CAN is infinitely slower than a direct analog connection, so I'm not sure why he thinks it would speed things up.


Anyway, I didn't intend to derail this conversation into a technical one on speed control. The fact is, I have no idea how Tesla does any of what they do. It has just always bothered me that they scapegoat customers citing "logs" that may or may not exist, and it bothers me more how the media, presumably insurance companies, and apparently the general public just takes their word for it, particularly now that I feel - from the conversation I had with an employee who, admittedly, may have been talking outside of her paygrade - they've demonstrated to me personally that the logs can't be trusted. At least not 100%.

Yeah, typo.
Didn't mean to use that article as anything over than verification of the dual slope design of the accelerator pedal sensor, could have been written better.
Were you working with the pedal sensor as well as the TPS?

Does the discussion have two aspects?
Do logs show 100% of what has happened? Sounds like no.
Did what is in the log happen? No definitive data so far to indicate that it didn't. (and no reports of the car going to 100% when not parking, but I'll concede that could be a SW edge case)

 

[bcsteeve]

No definitive data so far to indicate that it didn't.

If we're talking about the unintended acceleration still.... how can you argue that the person's input needs to always be thought of as authoritative, yet when it comes to "what happened" we don't see the driver's account as definitive? He was there. He says he didn't plant the pedal down. I don't see that as any less reliable than suspicious logs

Anyway, I have to go. I'm late. It was an interesting discussion, thank you.

 

[jeffro01]

Look, I don't care what your qualifications or credentials are. They have absolutely zero bearing on your original issue or your take away from that. You're openly questioning Tesla's integrity when it comes to logging because of an odd circumstance you experienced. It's not that I don't get why you're asking the question in a broad sense, it's that you refuse to accept the reality that you're looking at a software bug, nothing more.

I trust the logs over any drivers account of anything because drivers are known to lie, they lie consistently, the lie ALL THE TIME when it comes to what did or did not happen... Sometimes these lies are deliberate in an attempt to cover up behavior, other times they are not intentional and are simply misremembering events which is a well documented psychological phenomenon. Either way, I trust the logs without question over any real world account.

Jeff

 

[mongo]

I have no doubt the chill mode is related, absolutely. They're spinning it as a "feature", which is fine... I suspect it is a liability move. Now, at least, you'll have to choose to be able to ram buildings, so if you don't choose "chill" you're on the hook for it. I'm going to guess that eventually it is going to be the default.

Ramming snow isn't going to be 100% throttle. And if a semi is bearing down on you, the extra 3 feet you're going to gain by making your problem the person in front of you's problem isn't going to make the difference. Carjacking? What? The scenario doesn't exist because it can't. And even if it did, you don't give airline passengers the ability to open the door for the 0.00001% chance that's the best thing to do. You don't not have safeties on guns because of the 0.00001% chance having it causes a problem. We have safeguards all over the place that in some conceivable situation cause problems, but on the balance they do more good than harm. Seatbelts, airbags, etc etc. The argument has always been - briefly - that we shouldn't have it because "what if...". The same goes for this. So far, some people and property has been damaged because of the way it is currently. Nobody - I suspect - has avoided damage because they were able to go full blast into an object in front of them.

Incidentally, Tesla in the past has come out on the right side of "can we, should we?". It is now standard for the car to not overheat and potentially kill children/pets/elderly left in a car in the summer. They had the ability to do the right thing, and they did it. Yes, you can shut it off, and perhaps that's the compromise with what we're discussing here. Perhaps the default should be "don't launch into walls" and you can shut it off if that bothers you... but then its fully on you.

All valid points. I appreciate your gun analogy because I would prefer a safety on the 0-5 acceleration ramp.

Do cars need more than 150 HP? Is the acceleration ability of the Tesla cause to make it the only vehicle I would be aware of that prevents you from flooring it? Maybe the answer is yes.

If we're talking about the unintended acceleration still.... how can you argue that the person's input needs to always be thought of as authoritative, yet when it comes to "what happened" we don't see the driver's account as definitive? He was there. He says he didn't plant the pedal down. I don't see that as any less reliable than suspicious logs

Anyway, I have to go. I'm late. It was an interesting discussion, thank you.

That's why I would love to see that Tesla logs both accelerator voltages each control cycle, along with the code from ADC to log (and the schematic, and the wire harness).

Good talking with you!

 

[mongo]

Look, I don't care what your qualifications or credentials are. They have absolutely zero bearing on your original issue or your take away from that. You're openly questioning Tesla's integrity when it comes to logging because of an odd circumstance you experienced. It's not that I don't get why you're asking the question in a broad sense, it's that you refuse to accept the reality that you're looking at a software bug, nothing more.

I trust the logs over any drivers account of anything because drivers are known to lie, they lie consistently, the lie ALL THE TIME when it comes to what did or did not happen... Sometimes these lies are deliberate in an attempt to cover up behavior, other times they are not intentional and are simply misremembering events which is a well documented psychological phenomenon. Either way, I trust the logs without question over any real world account.

Jeff

In this case this issue is that the vehicle was reporting low tire pressure for a decent length of time (such that one would expect it to be in the logs), but the person at Tesla said the logs showed no tire pressure issues for the last 2 days.
Depending on phrasing, meaning, and interpretation, that can make a person doubt the veracity of the logs or the people reading the logs.

Contrast:
"Logs would not show that type of fault"
"The logs are not showing a fault"
"The logs show the fault does not exist"

If you are looking at a fault on your screen while talking to this person, your take away would be very different depending on which line was spoken.

 

[bcsteeve]

Look, I don't care what your qualifications or credentials are. They have absolutely zero bearing on your original issue or your take away from that.

Agreed, I only mentioned it when the discussion moved to speed controls, that's all. My "take away" from it is based solely in logic, admittedly with imperfect information.

You're openly questioning Tesla's integrity

Gasp! How dare I question the supreme whateverness of the infallible Tesla [yes, I'm rolling my eyes]

Either way, I trust the logs without question over any real world account.

That's really sad. Trusting anything "without question" is sad, but particularly something that is so limited in its scope and so easily manipulated - or at least interpreted - to suit one's interests.

Make no mistake. Tesla is first and foremost a public company who's primary interests are profit for their shareholders. They don't deserve "trust". Millions of people were justified fans of VW, and why not? They did what they did because of corporate interests, not because they are inherently evil or untrustworthy as people... they are (as all corporations are) untrustworthy by the nature of the system.

 

[bcsteeve]

That's why I would love to see that Tesla logs both accelerator voltages each control cycle, along with the code from ADC to log (and the schematic, and the wire harness).

Good talking with you!

That's how all manufactures should do it. I said earlier that I "doubt" Tesla does that. I shouldn't have said that. My presumption is that they probably didn't want to reinvent the wheel, but they've shown they do that all over the place. Plus, maybe all high end cars do it that way or similar. I have to say every single car I've done cruise work for was in the sub $30k category where perhaps minor cost savings make their way into engineering more than they should.

I'm curious what you do for a living. You're well versed in a broad range of topics it seems.

 

[jeffro01]

Agreed, I only mentioned it when the discussion moved to speed controls, that's all. My "take away" from it is based solely in logic, admittedly with imperfect information.

Gasp! How dare I question the supreme whateverness of the infallible Tesla [yes, I'm rolling my eyes]


That's really sad. Trusting anything "without question" is sad, but particularly something that is so limited in its scope and so easily manipulated - or at least interpreted - to suit one's interests.

Make no mistake. Tesla is first and foremost a public company who's primary interests are profit for their shareholders. They don't deserve "trust". Millions of people were justified fans of VW, and why not? They did what they did because of corporate interests, not because they are inherently evil or untrustworthy as people... they are (as all corporations are) untrustworthy by the nature of the system.

Yawn... Your entire argument is based off of a freaking software bug which you've taken to ludicrous (yes that was intentional) heights in order to draw a non-existent conclusion which you're using to confirm your own bias. The fact remains the logs do not lie, people do. As such, I'll always trust logs over what any person claims to have happened.

Funny how you call my position sad but you didn't even attempt to dispute the reality that people lie and computers don't. It's okay, I get it, you're seeking confirmation of your own bias and I'm simply not giving it to you.

Jeff

 

[woof]

My money's on whomever was reading the logs didn't look at the right ones, or didn't understand what the logs actually showed.

 

[verygreen]

And once again, that's entirely your assumptions/imaginations/whatever. You have no idea when/what it does or doesn't log. From what she told me, it *sounded* like it is continually sending all information, not just abnormal ones. Again, perhaps she misspoke. I can't know that.

The car absolutely does send a whole bunch of logs even when in park. The logging is somewhat reduced as various parts of the car "Sleep", but they are regularly woken up for various reasons (less and less regularly with prolonged parking. Ever noticed how you leave a car parked (unplugged) and it loses a lot of miles the first couple of days and then starts to really sip power?)

I didn't notice the warning until I left the garage. I'm not certain if it was on when I got in the car, or if it took moving before they faulted. But at that point, I was late and I saw it wasn't *that* low, so I did my driving for the day and topped up when I got home. I'm going to guess that we're talking 4 to 6 individual trips for a total of 20 to 30 kilometers over the course of about 4 hours.

The warning certainly does not reset right away after you inflate the tires back, you do need to drive some distance (for it to even register the increase in pressure - which you obviously did).

I trust the logs over any drivers account of anything because drivers are known to lie, they lie consistently, the lie ALL THE TIME when it comes to what did or did not happen... Sometimes these lies are deliberate in an attempt to cover up behavior, other times they are not intentional and are simply misremembering events which is a well documented psychological phenomenon. Either way, I trust the logs without question over any real world account.

I would tend to trust the logs too. There's just one small problem - have you ever seen any logs from Tesla, or just some statements from them supposedly based on logs they have (and contain things they claim)?

In my own experience related to failed brakes they always carefully avoided all conversations about logs and mostly immediately steered the conversation towards "the system does not indicate any faults after we test-drove your car".

 

[Tiger]

Tesla is only as good as its logs are.

 

[bhzmark]

nothing in the logs to indicate a problem

there is a disconnect between what is actually going on in our cars

Tesla logs also wouldn't indicate whether the driver was having a medical event or a paranoid delusion.
But that doesn't mean it didn't happen. It just means it wasnt logged and transmited to Tesla headquarters. I don't remotely see what the big issue is here. Especially since you claim the tire pressure was fine all along so there wouldn't have been any low-pressure event to actually accurately report back to Tesla headquarters.

At most it seems there was a local warning giving just an Orange alert when the tire pressure wasn't actually very low . Is that really all the cause of this entire thread?

What did service say after you took it to them?

 

[TomVdP]

There's NO CASE anyone can conceivably come up with, I'm sure, where sitting parked in front of a wall warrants the full use of an accelerator.

Insurance fraud!

 

[Ken7]

Unless I missed it, isn’t there an obvious point we’re missing here? Nowhere did I see that Steve actually checked the tire pressure with an external tire pressure gauge of known accuracy. This would have immediately shown whether the tire pressure sensors were reporting accurately or not. If they were, then the conversation could have shifted to why the car’s diagnostics screen was reporting an issue that didn’t exist.

At that point it might have been obvious in the phone call that Tesla’s data was correct and the issue was not a discrepancy in accurate data, but rather a potentially faulty onboard diagnostic system.

It would seem then that the mothership might not be plugged in to the onboard diagnostics screen, but rather the actual data stream (my terminology may be off here) that is consistent with ‘no problem’.

BTW, I got my first tire pressure warning a few days ago with a temperature drop here in NY. The first thing I did was check the left front with a tire pressure gauge I knew was accurate. It reported the same 37psi as the sensor. The other tires were 38 & 39, so perhaps 37 is the trigger? Despite the manual saying the warning won’t clear until you drive 20 minutes at a speed over 25mph (or something close to this), I filled the tires to near their correct pressure (42 vs the 45 called for) with a portable compressor, and as soon as I backed out of the garage, the warning light went out. That surprised me.

 

[mongo]

Unless I missed it, isn’t there an obvious point we’re missing here? Nowhere did I see that Steve actually checked the tire pressure with an external tire pressure gauge of known accuracy. This would have immediately shown whether the tire pressure sensors were reporting accurately or not. If they were, then the conversation could have shifted to why the car’s diagnostics screen was reporting an issue that didn’t exist.

At that point it might have been obvious in the phone call that Tesla’s data was correct and the issue was not a discrepancy in accurate data, but rather a potentially faulty onboard diagnostic system.

It would seem then that the mothership might not be plugged in to the onboard diagnostics screen, but rather the actual data stream (my terminology may be off here) that is consistent with ‘no problem’.

BTW, I got my first tire pressure warning a few days ago with a temperature drop here in NY. The first thing I did was check the left front with a tire pressure gauge I knew was accurate. It reported the same 37psi as the sensor. The other tires were 38 & 39, so perhaps 37 is the trigger? Despite the manual saying the warning won’t clear until you drive 20 minutes at a speed over 25mph (or something close to this), I filled the tires to near their correct pressure (42 vs the 45 called for) with a portable compressor, and as soon as I backed out of the garage, the warning light went out. That surprised me.

An external tire pressure sensor, would validate reported pressure, but has no bearing on what the vehicle is seeing/ reacting to. The issue was not the accuracy, but rather the use of the data.

Here are some relevant snippets placed in a more chronological oredr with my 3rd party interpretation at the end.

A couple of days ago the temperature dropped significantly and my pressures were 38 psi in the front and 39 in the back. The two front ones were complaining. So I figured 38 was the threshold.

I didn't notice the warning until I left the garage. I'm not certain if it was on when I got in the car, or if it took moving before they faulted. But at that point, I was late and I saw it wasn't *that* low, so I did my driving for the day and topped up when I got home. I'm going to guess that we're talking 4 to 6 individual trips for a total of 20 to 30 kilometers over the course of about 4 hours.

I was sitting in my car with the head unit complaining that two of my tires' pressure were low. What was odd is that all four tires were indicating 42psi on the screen, but the two front ones were in orange and the alert at the bottom was on constantly. I didn't really know what to do because how can I correct a problem when it is reporting that there is, but isn't, an issue? So I called Tesla. The representative assured me there was nothing wrong and nothing in the logs to indicate a problem and there was no indication from the car's tire pressure sensors that anything was wrong. Huh? Yes, I tell her, in fact there is an indication because there is - clear as day - an alert telling me there's a problem and two indicators telling me they are low, but at the correct PSI.

She had no answer for me other than to say I should get it checked at my next service.

. Their logs indicate "all good" and in fact she told me straight out that there is no current alert about low pressure, nor had there been for the past couple of days. But there was (first accurately and later inaccurately).

The concern is based in that the car was/ had been reporting an issue, and the Tesla rep said it had not over the time period in question. Could be communication breakdown (how long is a couple of days?), could be logging issue.
If the rep was more a car person, they might have said "The vehicle is current showing correct tire pressure, it can take a few miles of driving to clear the warning." (like @verygreen mentioned)

 

[Ken7]

An external tire pressure sensor, would validate reported pressure, but has no bearing on what the vehicle is seeing/ reacting to. The issue was not the accuracy, but rather the use of the data.

IMO, before any discussion is begun, you'd want to know what the accurate data actually is. Then you can enter into a discussion about discrepancies between what's reported by the diagnostics and what Tesla logs see. To me it's of far less concern if Tesla logs are agreeing with the accurate data, that there is no tire pressure problem. It's a much greater concern if there actually is a validated issue that Tesla logs are not indicating. Just MO.

From my perspective, once that discussion is underway, the more immediate concern is why the diagnostics are reporting a problem that doesn't exist.

 

[mongo]

IMO, before any discussion is begun, you'd want to know what the accurate data actually is. Then you can enter into a discussion about discrepancies between what's reported by the diagnostics and what Tesla logs see. To me it's of far less concern if Tesla logs are agreeing with the accurate data, that there is no tire pressure problem. It's a much greater concern if there actually is a validated issue that Tesla logs are not indicating. Just MO.

From my perspective, once that discussion is underway, the more immediate concern is why the diagnostics are reporting a problem that doesn't exist.

I think we are talking about two slightly different but interwoven things.
Case 1 (this thread so far):
Vehicle's handling/ logging/ error reporting of the TPMS data it is receiving and logging consistency. Given that TPMS is saying X, what should happen?
Case 2 (what I think you are referring to):
Is what TPMS saying accurate? Is it behaving correctly?


There had been a fault (low pressure), which was resolved (tire inflated), but had not cleared (car not driven after fill). OP contacted Tesla, logs were indicated to not show original fault or current display of fault that OP was looking at (potential lack of feedback from UI/ logging/ rep communication issue)

In the words of the great Samir Nagheenanajar "why does it say paper jam when there is no paper jam?. Or in this case. "Why are you saying there is no fault when it is saying there is a fault?"