Keep phantom drain under control

[AlanSubie4Life]

Update, the screen does not accumulate all negative losses, aka gains.

Remaining miles went up by 10, 151 to 161, “consumed while parked” went from 18.6 to 18.1.

Pretty buggy. So this means actually 8 miles over 48 hours, some of which was real useful use (about 3-4 I would say).

So 4 over two days. A bit high but in ballpark especially given BMS guesswork!

Nothing but sleeping today. Next capture in a few days will show a wake up of course when I opened the door.

View attachment 907742

View attachment 907743

Update:
Car was left unplugged. It slept peacefully & unmolested from Thursday until tonight. I opened the door around 9:30.

Lost 3 miles (161 -> 158).

So, I have gained 7 miles over the last 6 days while unplugged (see above that I was at 151 when parked Tuesday at 6:30PM). (With no driving in those 6 days.)

Since the beginning of the experiment I have lost 11 miles (the screen says 20.8 but that is wrong as proven in pictures above) while parked over about 7 days. 4 were probably useful miles. So about 1 rated miles per day of useless drain.

Car goes about 40 hours between wake cycles of about 1.5-2 hours.

This would suggest that very roughly, idle mode is 3mi/5hours *234Wh/mi = 140W.

Anyway that concludes this experiment. Massive BMS shifts do occur. It makes for a confusing picture!

1) Ignore what happens in the first 10-20 hours after parking! But note the end state after the delay.
2) Measure after that. (Delta measurement required.)

Maybe they’ll clean up the erroneous display of the data but not holding my breath.

 

[pdx_m3s]

My rated miles at 100% are up to 300. From a low of 285 or so. May be related. Anyway.

Any special sauce in recovering lost capacity? I’ve never done any intentional form of cell balancing, and my car has only been over 90% a handful of times in 4 years. 218/240 miles right now. Probably not very accurate. I’m curious where my capacity really is at.

 

[stopcrazypp]

Any special sauce in recovering lost capacity? I’ve never done any intentional form of cell balancing, and my car has only been over 90% a handful of times in 4 years. 218/240 miles right now. Probably not very accurate. I’m curious where my capacity really is at.

Some people say this is a waste of cycling (pack does not need to do this to balance), but I have tried charging to 100%, unplugging and letting it sit overnight and had it climb a few miles. Probably letting it sit at 100% lets BMS get a more accurate reading (same reason why Tesla now recommends to do this regularly for LFP batteries). But this was in warmer weather. In this season, probably the effects of the cold will trump anything you do.

 

[AlanSubie4Life]

Any special sauce in recovering lost capacity? I’ve never done any intentional form of cell balancing, and my car has only been over 90% a handful of times in 4 years. 218/240 miles right now. Probably not very accurate. I’m curious where my capacity really is at.

No. Just lowered my charge limit and charge every day (not for this experiment of course). Probably the most important factor is updating the software.

The projections from 55% (I went higher for this experiment) seem to match well with what I actually get when I charge to 100% (which is rare but does happen).

218/240 miles right now. Probably not very accurate.

It is the best estimate there is, in any case. Car does not pull any punches. Usually pretty close, within a few %.

 

[AlanSubie4Life]

Update:
Car was left unplugged. It slept peacefully & unmolested from Thursday until tonight. I opened the door around 9:30.

Lost 3 miles (161 -> 158).

So, I have gained 7 miles over the last 6 days while unplugged (see above that I was at 151 when parked Tuesday at 6:30PM). (With no driving in those 6 days.)

Since the beginning of the experiment I have lost 11 miles (the screen says 20.8 but that is wrong as proven in pictures above) while parked over about 7 days. 4 were probably useful miles. So about 1 rated miles per day of useless drain.

Car goes about 40 hours between wake cycles of about 1.5-2 hours.

This would suggest that very roughly, idle mode is 3mi/5hours *234Wh/mi = 140W.

Anyway that concludes this experiment. Massive BMS shifts do occur. It makes for a confusing picture!

1) Ignore what happens in the first 10-20 hours after parking! But note the end state after the delay.
2) Measure after that. (Delta measurement required.)

View attachment 909554View attachment 909559View attachment 909555View attachment 909556View attachment 909557


View attachment 909560View attachment 909558


Maybe they’ll clean up the erroneous display of the data but not holding my breath.

Definitely some bugs here in the software.

Lost 16 miles today from 198 to 182 in about two hours. Car was mostly asleep (I did get things out of it several times, hence the spikes).

This was after supercharging previous day, and did not lose anything overnight. This occurred after a 20-30 miles drive. Cabin was warmed. Parked in garage.

My rated range is 301 miles at 100%. Previously it was 285-290 or so.

Perhaps not a coincidence!

Anyway seems clear there are some major bugs here - the car is definitely not using any energy in two hours, that seems clear. Just not in any way physically possible as can be clearly seen here (car asleep nearly all the time). You can see the 16-mile step on the right hand side. 1PM to 3PM is where it occurred (I checked).

Both before and after the step, rated range projected to 301 miles.

 

[AAKEE]

Just for notification, I think you got grip of this it in the thread. If not I could be a part of the expkanation.

During a drive and just after a drive, the SOC is only estimated by “last known nominal remaining reduced by user energy”.

If the last known nominal remaining was wrong, due to a wrong SOC or wrong capacity estimation the estimated SOC will be wrong after a drive also.

If the BMS overestimate the capacity the SOC will be overestimated also when parking. The amount of the overestimation is of course depending on the amount of used energy. Longer drives -> more overestimation.

The BMS /the car will adjust this after parking, a big part of the adjustment will happen quite quickly but the fine tuning needs sleep for at least a few hours until the OVC is stabilised.

I only have grip of pictures of the opposite, underestimation and the change in SOC.

After a 240km drive, 31% when parking.

After 20 minutes, 33%.

Another drive but the sleep period:
When falling asleep after ~ 20 min: 53%

After about 9 hrs, 53.29%.

0.29% would be about 1.5km or 1mile, increase in my case but if the BMS was overestimating, there would most probably be a loss of the same range (if the overestimation was equal to the underestimation).

 

[AlanSubie4Life]

Just for notification, I think you got grip of this it in the thread. If not I could be a part of the expkanation.

During a drive and just after a drive, the SOC is only estimated by “last known nominal remaining reduced by user energy”.

If the last known nominal remaining was wrong, due to a wrong SOC or wrong capacity estimation the estimated SOC will be wrong after a drive also.

If the BMS overestimate the capacity the SOC will be overestimated also when parking. The amount of the overestimation is of course depending on the amount of used energy. Longer drives -> more overestimation.

The BMS /the car will adjust this after parking, a big part of the adjustment will happen quite quickly but the fine tuning needs sleep for at least a few hours until the OVC is stabilised.

I only have grip of pictures of the opposite, underestimation and the change in SOC.

After a 240km drive, 31% when parking.
View attachment 911366

After 20 minutes, 33%.
View attachment 911368

Another drive but the sleep period:
When falling asleep after ~ 20 min: 53%
View attachment 911372

After about 9 hrs, 53.29%.
View attachment 911370

0.29% would be about 1.5km or 1mile, increase in my case but if the BMS was overestimating, there would most probably be a loss of the same range (if the overestimation was equal to the underestimation).

Yes, understood this is re-estimation. However, clearly something is broken.

You can see above I drove a little under 32.3 miles for the interval in question ( 2-3 miles of which were from Supercharger to home the prior night - and there was no adjustment after those).

This was all done at a reasonable consumption of 346Wh/mi, so used a reasonable amount of rated miles, 346Wh/mi*32.3mi/234Wh/rmi = 48 mi.

This is reasonable - I charged to 247rmi and so that would have brought me to 199rmi (I saw 198rmi).

So, metered usage seemed reasonable. It was not undercounted, and rated mile use seemed exactly correct. So why remove 16rmi? After a SECOND sitting period (the first one removed nothing).

I definitely believe it is a BMS thing - there is literally no other possible explanation, but these adjustments are very large - and I would conclude that I don’t have anywhere near 300rmi of energy at 100% since these adjustments always seem to take place.

And unless they take place on the fly (and probably even if they did) it is a substantial enough shift that it would be a bit of a surprise on a road trip.

 

[AlanSubie4Life]

Another example clearly proving that negative values are not reported. I am clearly at -5 or -6 miles consumed (5-6 miles generated) since last charge - right there in the picture.

No significant temp changes or anything here. Pretty steady 55-60 degrees from garage to stop, etc.

 

[AlanSubie4Life]

Also does not correctly account for use in Park, and allocated BMS adjustments to Sentry Mode (this was about 2.5 hours of use, which it claims used 9.2 miles (for Sentry - I turned it off at work today!), lol ).

All a steaming pile of garbage. Guess it will be fixed at some point. Definitely very useful as is, but you have to know that it is not showing you what it says.

Can see all the segments. Green segment is to work. No step afterwards. Remaining stops totaled less than 2.5 hours.

 

[KArnold]

Had an interesting data point. My car is a '17 100 KW Model S.

My normal phantom drain is 2% to 3% in a night. I drove about 150 miles arriving home about 1:10 AM with 10% indicated. Ambient temperature was about 40 F and in the garage. I didn't plug it in as I planned to go to the SC first thing in the morning. When I woke up, my car indicated 1%!! I plugged it in and bumped it to 15% before going to the SC.

I use Tessie to monitor my car and it has a "do not wake" option that does work. Attached is the Tessie log for that time. It went from 10% at 1:10 AM, to sleeping from 01:19 AM through 02:19 AM. At that point it was down to 3% - all in 1 hour of sleeping. after that the phantom drain was closer to normal. All things considered it looks like BMS adjustment but that's quite dramatic. It has been a few months since I've taken a long-ish trip, if that matters.

Sentry, HVAC, and pre-conditioning was off. The 12-volt which was changed in November 2020 (2+ years). The car sleeps sporadically.

Again, to my knowledge it was a one-time occurrence but I'll be watching closely. I've never seen anything even close to this in the past.

I hope the 10% indicated when i arrived home was accurate and not a BMS anomaly. It would be scary if it was off that much while driving. I'd hate to lose 7% in 1 hour of sleeping (or even parked) if on a trip.

Is the PARK energy app option not available on legacy MS? I don't see the option. I'm on 2022.40.30.10 (latest FSD).

Any other suggestions welcome.

 

[AAKEE]

I hope the 10% indicated when i arrived home was accurate and not a BMS anomaly. It would be scary if it was off that much while driving. I'd hate to lose 7% in 1 hour of sleeping (or even parked) if on a trip.

I would guess that your BMS is overestimating the capacity.

I haven had a S but as it looks, if you constantly have a adjustment downwards in SOC overnight, this could indicate that the BMS thinks the battery capacity is bigger than the real value.

A longer drive with a downward adjustment of the SOC after a sleep is a clear indication that the BMS ocerestimate the capacity.

When driving the BMS can not really measure the real SOC. It is estimated by last known nominal remaining - the used kWh during the drive.

When parking, the car sleeps (or ”rests” if sentry is on), and the open circuit voltage can be read (if sleeping).
This OCV voltage = shows the real SOC, and an afjustment to the SOC is made if this differs.

 

[AlanSubie4Life]

It would be scary if it was off that much while driving.

Good news is you have a buffer of x% (not sure what) so it is not quite as bad as it seems. Pretty large adjustment though!

s the PARK energy app option not available on legacy MS?

Probably not.

 

[AlanSubie4Life]

Also does not correctly account for use in Park, and allocated BMS adjustments to Sentry Mode (this was about 2.5 hours of use, which it claims used 9.2 miles (for Sentry - I turned it off at work today!), lol ).

All a steaming pile of garbage. Guess it will be fixed at some point. Definitely very useful as is, but you have to know that it is not showing you what it says.

View attachment 912156View attachment 912168
Can see all the segments. Green segment is to work. No step afterwards. Remaining stops totaled less than 2.5 hours.

Final data point on this, did a metered charge (watched very carefully the last couple minutes to know exactly when it completed) to determine efficiency of the charging to see how accurate it was:

Conditions, 64 degrees steady temp, in garage at about 64F.
Charged for 1hr19min 40A@235V:
12.38kWh available. Just going to convert to rated miles for convenience.

For 2018 Performance it is 234Wh/mi (displayed; constant is 245Wh/mi of course):

12.38kWh/ 234Wh/rmi = 52.9 rmi avail.

49 rmi were added => 93% efficiency (too high!)

After car sat in garage plugged in for 1hr45min, rated miles added went to 41 (175rmi=>167rmi). This 8 mile loss showed on energy screen of course.

41 rmi => 78% efficiency (too low!)

After my first drive, after a wait in the garage at 64F, rated miles increased by perhaps 6 (hard to say; was not monitoring actual count and have to read off of Energy Screen graph, and it is not tabulated in the loss screen of course).

So net loss of 2, so real energy added was:

47rmi => 89% efficiency (about right based on EPA data). Maybe it is as high as 90-91%.

Of course there could be other adjustments, but I recharged.

This illustrates the basic way this works. As posted elsewhere, just a random BMS walk. A full accounting would require more careful observation over time. This would require an app which tracks this (I am not aware of one - none of them seem to track gains, though maybe they have been updated).

It also explains why people do not get to their set charge limit (they do, but then the BMS removes a couple %).

This doesn’t seem to be a strongly temperature-dependent phenomenon, though perhaps there is some - does not seem to matter from 50-65F though which is all I can test.

 

[Gbrlcbra]

This is how I eliminated phantom drain on my 22 MYLR. I took delivery in late 2021 and for the first year never really noticed any drain. The car was still new to me etc. My living situation changed and I couldn’t plug in like i usually do, this is when I started noticing the drainage. In between, there was days that it was under control and days where it was just unbelievably odd why I would lose 8% overnight. The car showed standby mode. First I opted out of sending Tesla any updates that my car could send them to “help improve service” that was my starting point. Since I have my own private parking space with only my wife parked next to me I disabled sentry while at home, work etc. That seemed to help but the car was still draining with standby mode. The next step was disabling WiFi, and that seem to eliminate standby for me. The next two - three days there was no phantom drain, nothing. I had the same charge for 17 hours or so with absolutely 0 drain. Then the drain came back, it was back to the head scratching. In between all this my wife took delivery of her M3 and we shared keys and profiles on the car. Anytime she’d open my car her profile would take over and stay until I’d use the car again. I noticed that her profile had every option that I disabled on my car and that’s the reason the drain was present. Now I make sure my profile is set at all times and make sure the WiFi is disabled at home and the phantom drain is near zero for me. I think the biggest culprit is WiFi that keeps the car on standby. Especially if signal is weak it can be constantly trying to connect and keep you car on without shutting off. I upgraded to a phone with 5g , all my updates are done with my hotspot through the phone with no issues. Hope this helps.