Forums

Vampire drain

Vampire drain

That thread about software uodate 4.6 got me thinking - the vampire drain must be a complicated problem, otherwise they would have fixed it a long time ago, right?

Apologies if I've missed earlier discussions about this aspect of it, but - given the amount of energy lost overnight, it can't simply be due to drain from electronics, or even battery heating (especially in the summer months), can it? Doesn't it have to be something in the chemistry of the battery itself? And doesn't that make it a very very tricky problem without an easy solution?

Also, I'm sure this calculation has been done elsewhere, but - during summer months, when you keep the car plugged in, approximately how much energy does it take from the charger to keep itself topped off? That's aside for any driving usage obviously.

Brian H | 12 juli 2013

Use volkerize.com to search for vampire. Or rent this movie.

July10Models | 12 juli 2013

An earlier firmware version had really good sleep mode option. At close to freezing temperatures my car only lost 2 miles of range in 12 hours. The sleep mode will be back hopefully sooner than later. The difficulty may be in idiot proofing the sleep mode. with all the battery management in power off sleep, the car would depend on the user to remember to wake it up and plug it in. With no alarms and nothing running to notify Tesla, some owners will undoubtedly abandon their $1000000.00 car for the better part of a year until it bricks subsequently blaming Tesla. My hobby Lithium Polymer batteries are charged to fifty percent and stored for six months. some of which I've had for twelve years.

ChristianG | 12 juli 2013

Well it's not rocketscience.. if it was, the spaceX guys could help.

J.T. | 12 juli 2013

From the "Punch List" thread. (Walter is WAlter Franck of Tesla)

In response to when #44 fix vampire load will be coming, Walter replied as follows:

Regarding the ancillary drain of Model S. Over the course of the next several months the firmware will continue to be updated to reduce the power consumption of the car’s computer systems. We strive to strike the right balance between minimum energy consumption and ability to have the car always ready and responsive. In our next major firmware release coming this summer, the car's computer systems will use half the power they currently do when the car is off. By the end of this year, they will use about 1% of the power that they do now when the car is off. The loss of range when the car is off has absolutely nothing to do with energy needed to heat, cool or otherwise do anything to the battery pack. The battery pack simply doesn't consume energy when the car is off nor do the systems that manage it. All of the "sleep" energy loss is going to onboard computer systems and providing the useful benefit of keeping them ready to start-up at a moment’s notice when the driver returns. As noted above this energy consumption will be almost completely eliminated over the next 2-6 months.

Docrob | 12 juli 2013

It baffles me why this has been so slow to fix. This car is supposed to be the embodiment of efficiency and the low power consumption is one of the major draws of the car, if they know how to fix the vampire drain why is it not priority number one? instead the fix is being trickled out painfully slowly.

tobi_ger | 12 juli 2013

This has probably a lot to do with the embedded systems and their l*nux derivate to be programmed/extended for such power-saving measures across multiple systems. Certainly man power and skills in that area needs to build up as well. And then: testing, testing, testing. :)

stevenmaifert | 12 juli 2013

As I've said in so many other threads about this, I have declined the updates so I could stay on 4.1 to keep the sleep mode. The car only loses 2 miles of rated range per day and runs without issue. One wonders what the problem was in the first place and why it's taking so long to bring the sleep mode back. Some say it took too long for the car to "wake up". The car is drivable immediately and it takes about 30 seconds or so for all the functions of the touch screen to be fully operational. Could you live with that, especially when you have the option to use or not use the sleep mode if you want the car fully ready from the start?

ir | 12 juli 2013

Having worked on embedded power management for cellphones, I can tell you that it is akin to cutting off a person's head, freezing it and surgically re-attaching it back later. You need to reconnect everything back exactly as it was before waking the patient!

You are literally killing the system by powering off its "body" (CPU & attached HW) but keeping the "brain" (RAM memory) alive in self-refresh. When the system wakes, it needs to re-initialize all of the HW peripherals exactly the way they were and synchronized with the software in RAM perfectly. Otherwise, you get a lot of weird system instability, odd behavior, random crashes, etc.... This is why they pulled the deep sleep feature when lots of owners complained about problems after leaving their cars unplugged overnight.

Another problem with cryo-freezing people and computers is it's hard to make sure nothing important is lost while suspended. If power to the right subsystems are not maintained (eg. properly setting-up RAM self-refresh) data will gradually erode over time (eg. when car is left unplugged overnight as many complained). Perhaps as a dirty hack Tesla could be backing-up RAM to internal storage, thus there is no space left for our music (see another thread where Tesla PR says they've pulled this feature because USB drives offer more storage).

And to address another misconception: not everything is powered off. Many hardware peripherals such as touch screen, WiFi, Bluetooth and especially 3G modems are little computers themselves. They can be programmed to handle many mundane tasks (eg. answering wireless scans and maintaining cell tower connections) using little power and only waking the main system under certain conditions (eg. when the Tesla app requests system status). It should quietly wake, quietly handle the request and go back to sleep.

RedShift | 12 juli 2013

Ir,

They can add a 'boot from non volatile memory' option. NVM can hold the code.
Not hard for a system that Tesla is using. An always on logic can initiate boot at a predetermined time, configurable by the user.

I speak as a HW verification guy who worked on Cellphone full chip verification.

Paul Koning | 12 juli 2013

"ir" is right that this can be a complex problem. It need not be, if you plan for it in advance. I work on data storage systems that need to handle power failure -- same sort of problem. It's not trivial but it isn't really all that hard if you do it right. For example, reinitializing the peripherals the same way as last time is easy if you design the initialization machinery to work in a single consistent way. Yes, if you start out with a modified PC, the job may be harder. But a purpose-designed embedded system handles these things.
From what little I have read, it sounded more like there were some bugs in this area, and the simplest fix was to remove the feature for a while. I understand that trade-off and it makes sense to me.

Bob W | 12 juli 2013

I think it's a tough problem because the car's convenience features all require some power to work right.

  • The GSM radio must draw some power to listen for remote access requests.
  • The charge port radio stays on so that it can listen for the UMC's "Open Charge Port" command command, even when the car is locked.
  • The door handle radio draws power to listen for the FOB.
  • The door handle sensors draw power to sense when they are touched.
  • The clock must be powered to support timed charging and proper time display,
  • The display computers keep running for some time even when they go dark (you can see this in a bright light). That's why they come up instantly when you tap them.
  • The cooling fans use power to automatically cool the display computers when required.
  • The streaming audio cache remembers where you were when you hit pause, so that the music restarts instantly. If stored in RAM, that RAM requires power.
  • The USB ports stay lit and powered so that you can find them and charge a cell phone inside the car when it's dark.
  • Soon the WiFi radio will stay on all the time to feed map updates, music libraries, etc. via your home LAN or mobile hotspot.

There are probably a number of other microprocessors monitoring everything from tire pressure to coolant temperatures that stay powered continuously whether the car is on or "off".

One would think that they could use 100-200W of "shore power" to power all of these devices when the car is plugged in and parked, but I don't think the car's design can support that. Unlike the heater and A/C (which can draw shore power through the main battery), everything else mentioned above is powered by the 12V auxiliary battery. That's the battery that is draining too fast, and it only gets recharged periodically from one source: the main battery. That's the source of most of the vampire load we see on our rated range meter.

Perhaps if you connected a trickle charger to the 12V battery terminals under the nose cone overnight, you would find vampire load (against the main battery) would just about disappear. It's worth an experiment. It's also not worth the hassle unless you need to leave the car parked and unplugged from the UMC for a very long time.

michael1800 | 12 juli 2013

I realize the Tesla's computer ecosystem is much small than the PC world, but I lean towards understanding. After years of S1 and S3 refinement, hibernate and sleep modes still don't work quite right on any of my computers (without digging into the command prompt to find culprits and modify the driver checkboxes). It's not the same thing, and not for the same reasons perhaps, but it obviously isn't a simple solutions without the forsight ir and others mentioned.

stimeygee | 12 juli 2013

So you're saying that the entire drain is computer systems? Does that add up - that up to 12-15 lost rated miles a night, is equal to how much energy those computer systems would use in that time? Seems like a lot to me (although admittedly I'm basing that on basically nothing.)

shamrockceo | 12 juli 2013

It does seem to be plausible that it could use that many miles. To think about it another way, if you were to drive across the country for 8 -10 hours or so, it seems reasonable to me that, over the course of all that time, you might lose 12-15 miles of range due to the basic hybernation features of all the onboard systems, even if you weren't actively using any of them.

Similarly, all of those things are running while your car sits there while you sleep at night, so the 12-15 miles seems unreasonable simply because in your mind, the car hasn't moved, so why the power loss.

I might be wrong, but that's how I rationalize it to myself, anyway haha.

jchangyy | 13 juli 2013

Just to add one more data point. I have 4.5.

I was away for 11 days. Started with 231 miles (85kwh). Came back to 147 remaining. Sat in my garage with average temp around 90 degrees. So, loss of about 8 miles per day.

Sleep mode would be nice to have if I were to be gone for a month or so.

Andre-nl | 13 juli 2013

8 miles per 24 h. @300 Wh/mile, that is 2.4 kWh/24 h = 100 W vampire load.

stevenmaifert | 13 juli 2013

@Andre-nl - That's about right. Now multiply that 2.4 kWh/24h by 365 = 876 kWh/yr X .15/kWh (Tier 1 rate in San Diego) = $131.40. Prior to getting my ModS and Leaf, my average power consumption was about 400 kWh/mo. My headline: TESLA MODEL S WASTES ENOUGH POWER IN A YEAR TO POWER MY HOUSE FOR TWO WHOLE MONTHS! I know, I know... they're working on it :)

jonesxander | 13 juli 2013

@andre and steven

ouch!

Andre-nl | 13 juli 2013

@stevenmaifert,

I wasn't trying to say 'it is ONLY 100 W'

I added the calc to get an idea of the power draw, since some didn't believe it were only the computers but also had to be self-discharge of the battery. 100 W for a few computers looks plausible to me.

876 kWh is enough to run my house for 6 months.

Brian H | 13 juli 2013

Since sleep in 4.1 cut the losses to 2 mi/day, that amounts to about 650 Wh, or 28 W or so. Trivial.

Hi_Tech | 29 juli 2013

I've just gone through an upgrade to v4.5 release 1.33.61. Still have the battery drain issue. I've been loosing about 8-10 miles/day. Though this accounts to about 3% per day of the whole battery pack, it still is a huge amount (3% of 85kWh = approx. 2.5kWh/day = 100Wh/hr).

To put it differently, my active laptop uses about 15-20Wh/hr. So if the arguement is that a "resting" Tesla requires approximately 5 laptops worth of power running 24 hrs/day to monitor the system, I don't really buy it. I'm sure they can easily make these fixes. I thing a loss of about 2-3 mi/day is much more reasonable.

That said, I really love my MS!!!! :) I've had it for about a month and have marketed it about couple dozen people already!

Reilly McHugh | 29 juli 2013

Mine has been sitting still for the past couple weeks and have only been losing about 4-5 miles a day. Not sure what software I am on...cars at the bodyshop, they leave unplugged during weekdays and plug in over the weekend.

jbunn | 29 juli 2013

Really hoping for a resolution to the vampire load issue. I only get to charge once per week. Starting at about 175, by the end of the week, I've lost 50 to 70 miles without driving at all. It's like having a car with a hole in the gas tank.

Olof | 13 augustus 2013

I have a Perf Sig VIN 1249 with the latest software, seems to loose 15 miles a day.

I had a loaner P85 VIN 8400 something it lost the exact same and then I had a loaner 60 VIN 15000 something also lost 15 miles a day. How come you guys only loose 8-10 miles?

Can it have something to do with the key? Our garage is valet parking and the attendants always leave the key inside the car. Does the car draw more power with key in car?

Bob W | 13 augustus 2013

@olof, does your range display say "Rated miles" or "Ideal Miles" below the number? Most quote only rated miles since ideal miles isn't a useful number in real world driving. You could be losing 15 "ideal" miles a day.

If that's not it, do you have anything plugged in to your USB ports? They stay powered up even when the car is powered off. A simple charge cable with nothing on the far end but another USB connector shouldn't draw any power, but a USB hub could drain the battery more.

Bob W | 13 augustus 2013

@olof, how cold does it get in your parking area at night?

riceuguy | 13 augustus 2013

There is a simple explanation...at full charge, the range drops between 12-15 miles per day. This starts to taper off as the charge gets lower (IIRC, you lose a fixed percentage of the remaining charge each day...I didn't understand the science behind it), but if you initially park your car without charging at say a half charge, it will lose less per day than if you park it with a full charge. The drop is more dramatic the lower you get as the car starts to take its own measures as well. Hot or cold temperatures impact these numbers as well, as much of the energy consumed by vampire drain is to heat or cool the battery to preserve and protect its long-term ability to hold a charge. In addition, all computers as well as the 3G remain on currently, with only the screens turned off, so yes, it's like running 5 laptops 24x7.

Supposedly this was to be addressed by end of summer "in the next major release" but fall is likely more realistic.

Olof | 14 augustus 2013

@Bob: Rated miles. It is about 75 degrees in the garage.

Bob W | 14 augustus 2013

@riceguy, no the battery level doesn't matter. The computers and other devices consume about 100W continuously when the car is off (2.4 kWh/day). The rated range drops by 1 mile for each 287 Wh consumed (source. 2400/287 = 8.4 miles per day.

I don't know why @olof is seeing 15 miles lost per day if nothing is plugged in to the USB ports and he is on 4.5, as that seems quite high.

Something harmless to try is to reboot both the center console and the 17" display to see if that makes any difference. I would also try taking the fob out of the car, especially if "auto-present" is on.

Software 5.0 (coming soon), is supposed to reduce vampire drain quite a bit (at the cost of slower "boot" time and much slower phone app. access time).

Renegade | 14 augustus 2013

Will be interesting to see how much the new update reduces vampire drain.

http://www.teslamotorsclub.com/showthread.php/20140-Firmware-5-0

Hope those who update will report back and let us know in the next couple days.

Brian H | 14 augustus 2013

One of the first to get a car in Norway reports losing only 1 km. overnight.

AlMc | 14 augustus 2013

Brian H...Thanks for the info. That would be a HUGE improvement.

RZippel | 15 augustus 2013

@stimeygee:

LI-Ion cells self discharge with rates around 3% per month, a little quicker when full. This effect is neglectable compared to the previous firmware losses. It will become a factor, especially after a range load, as soon as the firmware has the standby usage under control and the next but much less relevant thing to complain about ;-)

dtesla | 15 augustus 2013

Tesla is shipping new cars with 5.0.

5.0 has noticeably less drain. Though the MS does take time to wake up. I get into my car and see a Tesla T on the instrument cluster while the car wakes up. By time I get my seat belt on the car is ready to go (so not a problem). My MS is on a separate meter so very soon I should have some hard numbers on what the vampire drain is down to (Previous average is 3.5 KWh per day.... assuming Tesla's 89% charging efficiency is accurate)). Last night I forget to charge my MS and I only lost 1 mile of range (I don't consider 1 data point to be a significant result).

Several other changes. My favorite is 5.0 has a car up option (verses north up) on the 17" NAV display.

stevenmaifert | 15 augustus 2013

@dtesla - Do you have the option to turn off sleep mode if you don't want to use it?

dtesla | 15 augustus 2013

I just checked. I didn't see an option to control turn sleep mode on/off. I don't remember the Forum, but someone has posted screen shots of the release notes. Tesla really should post the historical notes online for reference. Sssoooo many changes/options to keep track of.

Today I ran A/C. Compressor did not run full out at beginning. A/C was much quieter, but took a little longer for cabin to cool down.

stevenmaifert | 15 augustus 2013

@dtesla - Thanks for checking. I'm still running V4.1 and have the option to turn it off.

AlMc | 15 augustus 2013

Steve...My understanding is that you do not have the option to turn off sleep option with this update. However, realize that it has been reported/confirmed by several new owners that have 5.0 that the car 'wakes up' in less than 10 seconds and that they are losing only 1 mile overnight to the vampire!

Nice trade off IMO.

AmpedRealtor | 15 augustus 2013

* dreaming *

Jarleø | 16 augustus 2013

On norwegin forums several new owners report a 1km vampire drain overnight.

Excelent

Go_Peddle_4_me | 16 augustus 2013

@dtesla

Does the sleep mode allow you to keep all your personal settings or do you have to re-apply them?

AlMc | 16 augustus 2013

wdazew; All info saved.

dtesla | 16 augustus 2013

All my personal settings are kept. The car does not turn off completely. It goes to a low power state. Though it doesn't appear the entire car goes into a low power state. The 17" console seems to stay powered up. Just so people know, this is all based on observation. I don't have any special knowledge of how Tesla choose to implement this feature.

My FOB works as before. My iPhone App still talks to my MS.

I will try to resist my desire to dive/charge my MS for 24 hours to get an accurate mile/day range loss with version 5.0.

brian4591 | 16 augustus 2013

Here's an article on 5.0 and vampire drain that I just came across.

http://insideevs.com/tesla-model-s-firmware-5-0-adds-sleep-mode-to-cut-v...

Brian H | 16 augustus 2013

45 sec. still on connectivity: " The mobile app may also take a while (~45 sec) for the secure connection to the car to be re-established when in this low-power mode."

dtesla;
That would require a lot of very strong coffee! "my desire to d[r]ive... my MS for 24 hours". ;)

Bob W | 16 augustus 2013

@dtesla, with 5.0 does the Model S go into "deep sleep" right away or after a delay? In other words, if you exit your car and return 30 seconds later, does it take another 10 seconds to power up the driver console? If not, about about 60 seconds? Or 5 minutes?

dtesla | 17 augustus 2013

There is a delay before the car goes to sleep. I have never timed it so I don't know how long it takes. I can only say it's > 15 seconds for the car to go to sleep.

The wake up is less then 10 seconds. It normally takes 2-3 seconds to wake up... once in a while it takes a second or two longer... the longest it has ever taken is 8 seconds. The wake up time is not consistent. I haven't had 5.0 long enough to notice patterns like the longer the car is asleep the longer it takes wakes up. Perhaps the MS will need coffee in the morning... just like me.

Normally sit in the drivers seat and put my seat belt on and the car is ready to go. Sometimes I must wait a second or so for the car to be ready. You might be surprised how annoying it is to wait that extra second or two, after my seat belt is on, for the car to finish waking up.

Also in the last 17 hours my MS has lost 4 miles of range.

AlMc | 17 augustus 2013

dtesla: The 4 miles lost in 17 hours is a higher rate than most are reporting. From what I have seen/heard/read it is about 1 mile for every 8-10 hours of 'sleep'.

My 4.5 firmware was costing me about 1 mile for every 3 hours of sleep.

Seems to be about 66-75% improvement in loss

What were you losing before the update?

AlMc | 17 augustus 2013

Got this from a TMC thread about vampire drain:

Update:
Parked car at 9pm last night with 185 range. This AM at 07:30, still with exactly 185 range. 0 miles of loss.
FYI: parked in garage with temp about 75 degrees.

RedShift | 17 augustus 2013

Oh goody, oh goody, oh goody!

Pages