Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Software Changes in 2019.5.15 f5def7e Preventing Sleeping?

This site may earn commission on affiliate links.
Honestly, 3rd party app is most likely the one to blame here

My feeling at this point is that it may be an interaction between this version of code, and the use of their published APIs.

Most of the trusted, well built 3rd party apps are designed in such a way to allow the car to sleep. They offer certain capabilities such as detecting when the car wants to sleep (the car stops collecting certain metrics, such as temperature) and backing off for a period of time until the car sleeps. Its plain to see in the logs that I've captured that the App backs off, but when it returns, the car is still not sleeping.

In the mean time, TeslaFi reported it tried more than 6 times today to allow the car to sleep without success.

Here is an example of how it looks now, after attempting and failing to sleep:

Screen Shot 2019-03-19 at 7.12.16 PM.png


Here is how it looked prior to the update on March 11th:

Screen Shot 2019-03-19 at 7.14.32 PM.png


Additionally, the Apps can be configured in a way that it does no polling at all during certain hours of the night. I've configured this as well. Yet still, the issue continues!
 
Last edited:
I've sent an email to Tesla about this issue as it's very clearly something to do with the code starting in/around 2019.5.15. I don't know if/when I'll get a response but I have my doubts that anything worthwhile will come from it. I don't have the bandwidth to waste my time making a Service Center appointment for both vehicles when this is a software issue.

My Model 3 has been updated to 2019.8.2 last night but it has been struggling to sleep (except for the hours set in TeslaFi's deep sleep settings). Nothing has changed anywhere for both my vehicles except for doing software updates. I've also tried doing the various resets of the cars, including powering off, and changing my Tesla.com password. The only connectivity to the API besides TeslaFi.com (which had been able to get the car to sleep prior to 2019.5.15) is my Tesla app on two devices and neither of us is actively using it anyway.
 
  • Like
Reactions: GolanB
New clue last night --- I had a suspicion that the algorithm that Tesla is using to keep the car charged may have changed in this release. Yesterday before retiring for the night, once the batteries got to 90% I lowered my target SOC from 90% to 80%. Sure enough, the car finally went offline at 4:55 AM this morning, and it has been offline for more than 5 hours. At the same time, TeslaFi did not register it falling into sleep.

Screen Shot 2019-03-20 at 10.26.05 AM.png


This is the first time since upgrading to 2019.50.15 that the car has actually gone offline while at home.

I'm wondering whether the charge cycles are somehow interacting with TeslaFi's polling cycles. I may need to manually calculate Phantom Drain this weekend when I disable TeslaFi as a test. In the mean time, it appears as if there is a new software release being deployed (2019.9 8282d10) to the fleet.
 
New clue last night --- I had a suspicion that the algorithm that Tesla is using to keep the car charged may have changed in this release. Yesterday before retiring for the night, once the batteries got to 90% I lowered my target SOC from 90% to 80%. Sure enough, the car finally went offline at 4:55 AM this morning, and it has been offline for more than 5 hours. At the same time, TeslaFi did not register it falling into sleep.

View attachment 388444

This is the first time since upgrading to 2019.50.15 that the car has actually gone offline while at home.

I'm wondering whether the charge cycles are somehow interacting with TeslaFi's polling cycles. I may need to manually calculate Phantom Drain this weekend when I disable TeslaFi as a test. In the mean time, it appears as if there is a new software release being deployed (2019.9 8282d10) to the fleet.
So you are thinking this has to do with the current SOC needing to be higher than the target SOC? This is very curious. I just moved both cars down to 50% and will see if this changes anything.

EDIT: It doesn't seem to be working for me. I made the change 40 minutes ago and it's still doing the attempt and failure to my every 5 minutes setting. Hopefully, there's a fix for this.
 
Last edited:
7C0B1D35-53DA-4C03-BA92-C20FC88E9EA9.jpeg
So you are thinking this has to do with the current SOC needing to be higher than the target SOC? This is very curious. I just moved both cars down to 50% and will see if this changes anything.

EDIT: It doesn't seem to be working for me. I made the change 40 minutes ago and it's still doing the attempt and failure to my every 5 minutes setting. Hopefully, there's a fix for this.

So far, in my case the model 3 has fallen asleep multiple times since moving target SOC below current state of charge.

In the last attempt, I completed a drive at 8:50 PM, and the car fell asleep approximately an hour later at 9:46 PM.

I also upped the time to try sleeping to 25 mins, but the car was asleep in less than 15. I noted a new feature coming in TeslaFi that I hadn’t seen before in the settings “Do not check the vehicle state during sleep attempt *COMING SOON*”.
 
So far, in my case the model 3 has fallen asleep multiple times since moving target SOC below current state of charge.

In the last attempt, I completed a drive at 8:50 PM, and the car fell asleep approximately an hour later at 9:46 PM.

I also upped the time to try sleeping to 25 mins, but the car was asleep in less than 15. I noted a new feature coming in TeslaFi that I hadn’t seen before in the settings “Do not check the vehicle state during sleep attempt *COMING SOON*”.

Changing my Time to Try Sleeping from 5 to 15 seemed to allow both cars to sleep. I am going to attempt dialing it down to see if I can minimize the possibility of lost drives. Even setting it at 5, I would lose a drive or two at least once a week or so, so hopefully I can find that sweet spot.
 
Changing my Time to Try Sleeping from 5 to 15 seemed to allow both cars to sleep. I am going to attempt dialing it down to see if I can minimize the possibility of lost drives. Even setting it at 5, I would lose a drive or two at least once a week or so, so hopefully I can find that sweet spot.

Polling the API's (versus push notification of state change) is probably the main reason we've got discrepancies in the data set. Architecturally, it would be better if Tesla would create a notification subscription system (publish/subscribe bus), but until that happens we'll have to make do with the current system.

I've seen in the raw data logs over the last 24 hours that the charging cycles seem to be more frequent now - meaning, that the car appears to wake up, checks it's SOC and begins to charge more often than it used to. This keeps the SOC closer to the target setting, rather than drifting by several percentage points, but I think this may be one of the main reasons we may be seeing interactions with TeslaFi.

Here is what it looks like on my end - rather than falling asleep, it bounces back into a Charging State - this every 1 hr and 15mins.

Screen Shot 2019-03-22 at 11.16.11 AM.png


In the prior release, the car may sleep for 8 hours or more, and only begin charging again once it wakes up, or if the car is taken for a drive. I suspect Tesla is changing the charging algorithms here in order to keep a constant SOC. As an example, this morning, the car woke up and charged four times.

Screen Shot 2019-03-22 at 11.22.44 AM.png


This more frequent charging is certainly new behavior in 2019.5.15.
 
Always reboot your car after an update (hold both thumb wheels until screen goes black).... this has been recommended to me by multiple Tesla reps, it's fixed pretty much any odd issue I have following an update.

Brandonee, I believe what we've discovered here is more of a behavioral change in the underlying code itself in this release. Rebooting, as well as powering down completely and bringing the system back up (which I've tried) has not modified the behavior, and a number of us have reported changes in sleep behaviors.
 
Changing my Time to Try Sleeping from 5 to 15 seemed to allow both cars to sleep. I am going to attempt dialing it down to see if I can minimize the possibility of lost drives. Even setting it at 5, I would lose a drive or two at least once a week or so, so hopefully I can find that sweet spot.

You can log in to teslafi and cancel the sleeping before you drive if you know it's trying to sleep. You have to remember to do it though...

Edit: I have it send me emails when it tries to sleep.
 
Is your Sentry Mode enabled? If it is, the car never sleeps. Just idles all day.

Not in my case; I rarely use Sentry Mode, as I park in an area with existing night-vision equipped cameras. As an aside, I suspect that as more Model 3's hit the road; people may be attracted to setting off Sentry Mode for fun. In addition to the brief show, it may become a prank against Telsa owners since they'll come running back to their car once they get a notification. In NYC, most people ignore car alarms - they are hardly a deterrent at all. Sentry mode will probably be more useful in cases where there is damage to the vehicle while parked.