My car is not sleeping – now fixed

see below for an update to this post

Since the 2020.24.6.1 software that Tesla pushed to my Model 3 in late June, my car has not been sleeping correctly. I know this because of the before and after data that TeslaFi (a 3rd party app I use) showed me right away after the update that my car was not sleeping.

What is TeslaFi? It’s a cloud based server side data logger which amasses data from my car, you see, Tesla provides an API which which gathers and aggregates data you can use to understand the energy use and recharging times and costs, showing, over time, the dreaded battery degradation, knowing when the car is awake vs asleep, alerting if the car is left unplugged or a window is down for more than x minutes, tracking mileage if you need it for business or tax reimbursement, tracking the car if used by other people (i.e. kids)

Since the TeslaFi data was showing the car mostly in idle throughout the night instead of asleep, along with the corresponding higher battery consumption during the night, I put a ticket into Tesla, got a mobile ranger out to my house, and he plugged in and found API requests from a 3rd party every minute. Well that sounded like TeslaFi, but I hadn’t worked enough with the TeslaFi support staff to understand exactly how TeslaFi dealt with the sleeping mode of the car. I had explained to the ranger that I had disabled TeslaFi (and other apps) by changing my user/pass which invalidates API tokens, but he explained that sometimes the tokens don’t really become invalid. I understood that checking the “vehicle list” in the API doesn’t come to the car, but it looked like it was now. Only I should have asked, why just my car? Many TeslaFi users should have been complaining about this.* My ticket with Tesla was closed but I then worked with TeslaFi support and found out so much more about how they behave and interact with the car. Tesla shouldn’t have really dismissed this (and I suspect others) as a well known “app” which keeps pinging the vehicle, thus keeping it awake.

What I know now, having not only removed all 3rd party API use, removed user/pass from my own Tesla app, removed the app, even moved the car to “not my garage” so the WiFi and LTE environment is different, is that something is waking the car, usually just after it goes to sleep. This is what TeslaFi was reporting to me right after the 2020.24.6.1 software release. Frankly, while my configuration settings were not “perfect”, they shouldn’t have been contributing to this situation. Common sense would dictate that something changed with the software release, because the behavior of the car changed. And I should have been more insistent.

So, I have a new ticket open with Tesla, I’ll have to have a ranger hook up to the car again, only this time, I’m armed with a specific API request for this “vehicle list” which I can invoke from a web browser. E.g., we can do an experiment, where I can invoke the API, he can see it or not. My expectation is he won’t be able to see it, thus proving or disproving the point that Tesla has been making all along, that 3rd party apps, if you use them, are the problem.

Part of why I believe this is not the case now is that TeslaFi and I put the car into a very long interval polling mode where API requests were done at 20 minute or longer intervals, and the car should be asleep for those, especially in the middle of the night (you see, normally I have it configured to allow “deep sleep” mode where from 11pm to 6am, no vehicle interaction is done.) So at 6:01am I should normally find the car asleep, but as I say, this has not been the case. Even in my everything has been removed situation, the underlying question is why is the car losing 6-10% of charge overnight? Because it’s in IDLE, obviously, but why?

Update August 2020 – A Tesla Mobile Ranger replaced my autopilot computer because they saw Ethernet port errors in the logs. Unfortunately, this did not resolve the sleeping issue when he left and the car had 24 hours to “do it’s thing”.

Update later in August 2020 – a Service Center visit resulted in a sensor in the driver’s seat being replaced. This was the Driver Occupancy Sensor. Like the autopilot computer, this did not resolve the issue. They have ordered a new seat for me, believing the wiring harness is the issue. One would think this is easy to test, driver in the seat, driver not in the seat…

Later – Well, the seat replacement did not fix this and I took the car back. It’s September now, and I’m pondering what to do, because after the seat replacement failure, Tesla investigated further and simply said, it’s a 3rd party app, but we don’t know which one…

Thinking further on this, I think I had mistakenly left Teslafi on for my account, which also has a Model Y on the account, no problems sleeping there… Anyway, I think they see the vehicle list API request at the account level and simply say, it’s a 3rd party app, you figure it out. So, what I really needed to do was disable Teslafi at the account level, both cars, and then see what they say.

So, I’ve done this, for 5 days (disabled both Teslafi accounts, pointing at the same Tesla account) and provided detailed dates, times, specific repairs, etc. to a very helpful support person who forwarded the whole mess to Advanced Engineering. I heard back from them last week (Oct 15) and they want to have to car back so they can look again, something in the hardware is obscuring what is happening in the car. So, I bring it on Oct 23, with Teslafi completely disabled, both cars, and we’ll see what gives.

Late October – so, hopefully, I’ve got a conclusion now… After 2 weeks, the car is sleeping. Tesla Service Center replaced a voltage controller for the driver’s seat. It was periodically dropping voltage to the seat occupancy switch. This appears to solve the problem, car now going to sleep on its own. Previously they replaced the switch and the entire seat suspecting the wire harness, but neither was not the solution. Tesla advanced engineering was instrumental in resolving this. I had to provide them, through support, with a very exhaustive list of actions taken prior to this.

* A few TeslaFi users are complaining about strange sleep patterns and it seems very similar to my own situation, but going back to the end of 2019 and certainly on other software releases. There have been issues with “hot” rear cameras (energy usage) and HV Battery firmware questionable behavior, but I think we’ve not yet seen the end of this. I asked some of these users if they had two cars and I’ve gotten 3 “bites” so something is going on with this commonality. Notice below, the Model 3 is online and the Model Y is Asleep. The other users all say the first entry, the older car, AMP in my case, is the one not sleeping.

Vehicle List API response: Array ( [response] => Array (

[0] => Array ( [id] => 32847053162823755 [vehicle_id] => 438746238 [vin] => 5YJ33333333333333 [display_name] => Amp [option_codes] => AD15,MDL3,PBSB,RENA,BT37,ID3W,RF3G,S3PB,DRLH,DV2W,W39B,APF0,COUS,BC3B,CH07,PC30,FC3P,FG31,GLFR,HL31,HM31,IL31,LTPB,MR31,FM3B,RS3H,SA3P,STCP,SC04,SU3C,T3CA,TW00,TM00,UT3P,WR00,AU3P,APH3,AF00,ZCST,MI00,CDM0 [color] => [access_type] => OWNER [tokens] => Array ( [0] => 17a9f08445f2214c [1] => c91a61248ca6580d ) [state] => online [in_service] => [id_s] => 32847053162823755 [calendar_enabled] => 1 [api_version] => 10 [backseat_token] => [backseat_token_updated_at] => [vehicle_config] => )

[1] => Array ( [id] => 31695657956927095 [vehicle_id] => 2130063041 [vin] => 5YJYYYYYYYYYYYYYY [display_name] => Volt [option_codes] => AD15,MDL3,PBSB,RENA,BT37,ID3W,RF3G,S3PB,DRLH,DV2W,W39B,APF0,COUS,BC3B,CH07,PC30,FC3P,FG31,GLFR,HL31,HM31,IL31,LTPB,MR31,FM3B,RS3H,SA3P,STCP,SC04,SU3C,T3CA,TW00,TM00,UT3P,WR00,AU3P,APH3,AF00,ZCST,MI00,CDM0 [color] => [access_type] => OWNER [tokens] => Array ( [0] => 74952f4bcd414992 [1] => 7b24edad1afecc5f ) [state] => asleep [in_service] => [id_s] => 31695657956927095 [calendar_enabled] => 1 [api_version] => 10 [backseat_token] => [backseat_token_updated_at] => [vehicle_config] => )

) [count] => 2 )