REVERSE means FORWARD??

[Crissa]

It's possible it's a misapplication of the hold to stop but I'd expect that to be overwhelmed by any slope. I've seen motors with regen controllers accidentally go the wrong way when there's a capacitance issue in the charger. Lastly, it's possible to be between the magnetic registers for the rotation sensors and the alternating flux frequency creates backwards motion for a moment, or that the sensor thinks it's moving correctly but it's not - but while that's happened in cheaper motorcycles and and robotics, I wouldn't expect it in a Tesla because their motors are not built of spare parts.

All of these seem like they should be much too small or dealt with the myriad of sensors a Tesla has.

-Crissa

Sponsored

 

[CyberGus]

Two others beside the OP said it has happened to them tho.

Of course, but extraordinary claims require extraordinary evidence.

Every software developer’s first question is: “ how does it recreate? Figure that out or bugger off”

 

[averyk12]

Of course, but extraordinary claims require extraordinary evidence.

Every software developer’s first question is: “ how does it recreate? Figure that out or bugger off”

Hilarious to put this under “extraordinary claims”.

 

[scottf200]

Of course, but extraordinary claims require extraordinary evidence.

Every software developer’s first question is: “ how does it recreate? Figure that out or bugger off”

haha, I understand ALL too well. As a software developer I have worked on plenty of support issues in the field and if you can receate it then we can narrow down the circumstance and then recreate it in-house for WAY easier solution finding.

HOWEVER, if there is a significant problem of going forward when the user thinks it is suppoed to go in reverse, that $hit could run someone over or cause serious damage. That is not a bugger off thing ... and the UI is doing something funny ... that is the 'tank' run-ith something over.

OFFICER ... I swore I put it in reverse ... result

 

[Stuck4ger]

haha, I understand ALL too well. As a software developer I have worked on plenty of support issues in the field and if you can receate it then we can narrow down the circumstance and then recreate it in-house for WAY easier solution finding.

HOWEVER, if there is a significant problem of going forward when the user thinks it is suppoed to go in reverse, that $hit could run someone over or cause serious damage. That is not a bugger off thing ... and the UI is doing something funny ... that is the 'tank' run-ith something over.

OFFICER ... I swore I put it in reverse ... result

I agree, I’d place my money on miss-swiping (which I have done) than a true hardware or software issue.

 

[CyberGus]

That is not a bugger off thing ... and the UI is doing something funny ...

I agree, I’d place my money on miss-swiping (which I have done) than a true hardware or software issue.

I've mis-swiped myself, but the OP is alleging that the vehicle did not move in the direction indicated by the UI.

The Tesla software team, if they respond in the customary DevOps manner, will dismiss this as "operator error" since something like that is "not possible". ?‍

Unless and until someone can get video evidence, this issue will be dismissed as hysteria or FUD. And even then, it will be dismissed as a Deepfake until there are multiple reports from reliable sources.

 

[Stuck4ger]

Agreed. If it happens again pull out the phone and make a video. And make a service request, Tesla can pull the data and see if it’s real or not. If it is real, they’d be the first to want it fixed.

 

[CyberGus]

Let me give you a real-world example of software debugging in critical systems:

A friend recently had a high-tech pacemaker installed to address a congenital heart defect. The surgeon programmed it post-op using a wireless transceiver, but my friend insisted that the Bluetooth option then be disabled. He works in IT Security and is understandably paranoid about it.

Every day thereafter, he is jolted awake in the middle of the night with his heart racing. Each time, he checks the clock and it is exactly 2AM. Every time. He was convinced this was a malfunction, and a life-threatening one at that.

When he took his concerns to his doctor, the device was scanned and no errors found, nor any 2AM activity logged... he was told that he must be having nightmares, and should try to relax more and drink less coffee. He complained quite forcefully after more 2AM jolts, but the surgeon insisted that he'd pored over the literature and there was no documented function that would do this, and nothing scheduled to happen at 2AM, but agreed to a conference call with the manufacturer and their programming team.

After a lengthy back-and-forth, one of the developers admitted that there was an undocumented daily self-test scheduled at 2AM, but that it "cannot possibly be responsible for this." Oh really? Why not? "Because the only time we send full voltage to the wires is during provisioning to test the impedance, and we never do that after the provisioning is completed."

But what if Bluetooth is disabled?

"Well, that's shouldn't stop the provisioning from being marked as complete. I mean, I'm looking at the code right now, and..."

awkward silence

"Oh. Shít."

 

[CyberGus]

Here's my hot-take. The high-level control flow should go something like this:

User sets the direction
- command sent to drive unit
- drive unit queried for current state, and UI set to match

but instead, they are probably doing something like:

User sets the direction
- command sent to drive unit
- UI set to user choice

and there's a rare but non-zero chance that the "command sent to drive unit" is lost or otherwise fails. However, the UI is already set to unconditionally match the User Choice, and is now a disconnect from what is actually in force. The UI is physically and logically separate from the drive electronics, as you can "double-thumb" reboot the UI while in motion.

To recreate this, you'd need to switch between forward/reverse repeatedly until the rare "lost command" occurs, nudging back and forth each time to confirm. An arduous chore.

Anyone experiencing this should certainly report it to Tesla, but I'm skeptical that they'd dedicate the significant resources needed to recreate or debug unless there was ample evidence of this being a true defect and not operator error.

 

[Friday]

Since the OP was in a very specific place when this happened, couldn't they go back and recreate the scenario and experiment? Like Chuck's Left Turn?

 

[TyPope]

I've mis-swiped myself, but the OP is alleging that the vehicle did not move in the direction indicated by the UI.

The Tesla software team, if they respond in the customary DevOps manner, will dismiss this as "operator error" since something like that is "not possible". ?‍

Unless and until someone can get video evidence, this issue will be dismissed as hysteria or FUD. And even then, it will be dismissed as a Deepfake until there are multiple reports from reliable sources.

I suppose one could do the following:
1. STOP!
2. Enable FSD.
3. Disable FSD.
4. Hit the mic to submit an audio report and say "The UI is indicating forward but my Cybertruck is going backwards. Please advise."

It may not be in the exact right spot for a complaint like this but at least someone at Tesla would hear it and they have telemetry.

Sponsored