I'm sure the car started the "charge soon" warnings a while before the shut down. It was probably even limiting, showing the dotted line, the power available. But when the charge gets that low all it takes is pressing the accelerator a little too much to drop the voltage and cause a shutdown. If you are driving like a grandma on basically flat roads you can likely use more of the charge than going uphill or trying to accelerate.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Electricfan
Active Member
I'm sorry, I don't think I quite understand you. You're not saying you bought a new (or CPO, doesn't matter) Model S from Tesla and they delivered it to you with a completely flat battery are you?
Sparrow
S105/ Roadster 189
Yes, I am saying that Tesla delivered my brand new Model S with a totally flat battery. This was a brand new Signature Series Model S and back then the Tesla Delivery Specialist would deliver the car directly to your home. I had them take it to a nearby Lowe's because the semi truck it was on could not get near my house. When they pulled it off the truck, we discovered that it had 0 miles of range left. Since I had never driven a Model S and the Delivery Specialist met me at the Lowes, I had him drive the car the 2.5 miles to my house with the totally flat battery to get a charge.
Electricfan
Active Member
I'm just dumbfounded sometimes by the things Tesla does. Its an electric car, you paid a fortune for it, and they didn't even charge the f-ing battery. Not to mention how bad it is for the battery to be in that state. Just stunning. My second Model S they delivered with the floor mats in the frunk, door gaskets hanging loose and a big plastic part broken under the frunk lid. I guess this is why people say Tesla is a tech company that makes a car to put their tech in. I'll sure be glad when they have some real competition.
They probably did charge the battery, and then sat around with vampire draw. It's not difficult to drain the battery in 2-3 weeks this way. Non-technical people tend to dismiss how much energy leaves the car regardless of whatever activity the car was involved in. Just plug in every night and the problem goes away, right?
DB 2
Member
In a much earlier version of the software (maybe 5.xx) I was able to go 20 miles past the "Charge Now". That was driving at 45MPH with the HVAC off, the steering in Sport Mode, and we barely made it to the Supercharger. I don't want to see how much farther it will go, with the newer firmware. They seem to be much more conservative.
I've driven coast to coast in my P85D and I've become comfortable reaching the super chargers with 5-10% left in the battery. I've found (at least with my car) that the Tesla range estimator is a bit conservative. That's a good thing though in my opinion. Bjorn (sp?) the youtuber did a video about the reserve left in the battery pack after the battery has reached 0%. He came up with the conclusion that the older the car gets, the less battery reserve it has.
I've been down to 3% with no ill effects on driving...but my car had less than 20K miles on it too. I would guess the cars with over 100K would be more susceptible running out of juice as they got closer to the 0% mark.
I've been down to 3% with no ill effects on driving...but my car had less than 20K miles on it too. I would guess the cars with over 100K would be more susceptible running out of juice as they got closer to the 0% mark.
Attachments
Tam
Well-Known Member
I think owners can get into trouble by driving in deep cold winter with low mileage left in the battery gauge.
For example, Tesla Supercharger is only about 30 miles away and there are still 50 miles left on the battery gauge.
So instead of going straight to charge before the hotel, the car rested uncharged at the hotel for a cold winter night.
In the morning, the 50 miles on the battery gauge could be quickly and automatically consumed by main battery heater and there won't be enough battery left to drive another 30 miles.
dgpcolorado
high altitude member
That may be but a Tesla warns the driver of this when stopping with less than 20% IIRC.
Of greater importance for cold weather trips is charging when one arrives and the battery is warm. Otherwise it can take a long time for the battery to warm up and start charging after a night spent in the cold. I've learned to do this even in just moderately cold temperatures. Saves time to charge when the battery is warm.
I routinely arrive home with less than 5%, sometimes as low as 2%. Lowest I've been was 2mi remaining last week, driving around Vallejo trying to figure out how to get into a parking lot that had charging and very bad signage (miss the turn, drive 1.5 miles to turn around). Did not think I was going to make it! But have never had any problems below 5% and trust it enough to skip the Vacaville charger on my way home as long as it says I'll make it with any percentage.
I know its an old topic but I cannot stand quite on some comments here and arguments given as reasoning, after I have experienced a sudden shut down at 4% SOC. like instant, not reduced climate etc... as 2 roads towards charger were closed (not marked). and although according to plan I should have been there with 10-12% SOC, the car shut down 2 miles before with 4% SOC. I was driving like on Egs, with on sudden powers requirements.
some arguments on algorithms...
1st, SOC of a battery is very easy to measure, down to less than 0.5% margin of error. The problem can be if Tesla uses the pack voltage (indicative of nothing) instead of lowest cell voltage, indicating remaining power, as as soon as the lowest one hits the low voltage cut off the car will shut down. a volume of fluid is much more difficult to measure to that accuracy as position in regards to sensor varies on gradient the car is sitting at.
2nd, an ICE car will newer run out of fuel at 0% indicated. thought it will stop reporting range at 60-100km (as usage without advanced navigation as the Tesla has its impossible to predict the average consumption) but even so, a SOC is same as fuel level (not the indicated range which is remaining fuel * consumption), and should always reflect the actual remaining capacity
some arguments on algorithms...
1st, SOC of a battery is very easy to measure, down to less than 0.5% margin of error. The problem can be if Tesla uses the pack voltage (indicative of nothing) instead of lowest cell voltage, indicating remaining power, as as soon as the lowest one hits the low voltage cut off the car will shut down. a volume of fluid is much more difficult to measure to that accuracy as position in regards to sensor varies on gradient the car is sitting at.
2nd, an ICE car will newer run out of fuel at 0% indicated. thought it will stop reporting range at 60-100km (as usage without advanced navigation as the Tesla has its impossible to predict the average consumption) but even so, a SOC is same as fuel level (not the indicated range which is remaining fuel * consumption), and should always reflect the actual remaining capacity
The problem isn't that Tesla can't determine capacity. It does really well adjusting for temperature and other conditions. The issue is that not all bricks/groups are the same. They age differently and drift apart over time. When the weakest one is empty the voltage drops suddenly and the car will shut down. That's what happens when it's unexpected before reaching 0%. I was able to see this happen on the CAN bus when mine shut down just before 0%. The older the battery pack the more likely it will happen. Since you have a remanufactured pack it is very likely the modules and bricks are different. Looks like the BMS in the old packs isn't able to keep track of the variations as well as it should.
The BMS, with modern firmware (anything from maybe 2016 onward), is very... VERY good at calculating usable capacity. In short, if your car shuts down before 0 miles, then it very likely has a problem.
The BMS is using some very complex algorithms, including what appear to be full simulations based on historical readings, to calculate the capacity of every single cell group (all 84 or 96). I've pitted this against bench testing equipment, and generally get to within < 1% the same data.
I'll preface this next part with the note that usable range below zero is NOT guaranteed, nor has it ever been in any firmware version.
The modern BMS's target an "empty" of roughly -3%. In most cars this is ~10 miles or so. (This is actually the reason for some visible rated range drop in early Model 3's and such when they tweaked this on those vehicles.) This "buffer" (it's not really a buffer) is adjusted based on the BMS's perception of the error in its readings, and can grow as much as needed to make sure 0 on the dash is at least a mile of rated range. In the case of problem batteries, this can be a super high value, like 70% "buffer".
The key thing is that this buffer is not necessarily usable, and the error is greater the less the car is driven below about 20-30% or less SoC. Or put different, the more the car is short cycled above 20-30% (like daily driving from 90% to 70%). The BMS mitigates this with a variety of techniques and is generally always pretty cautious. It's basically a way to move the window of range it shows as available into an area of the capacity calculation's error bars that effectively guarantees capacity is available there.
In theory, if the BMS is reporting 0 miles of range, if perfectly calibrated you'd still have a few miles left "below" zero.
In reality, 0 should be treated as 0. Don't count on any capacity below it.
Above 0, however, if the BMS decides it has to stop the vehicle it means something is too far out of whack either as estimated idle or under load. This usually happens worse at very low SoC (<15%), and is exaggerated by the wear on the battery, and other factors like short cycling adding error to the capacity calculations.
Generally today, though, if your car stops you with a positive value on the range meter... something is probably wrong with the battery pack. This is what is guaranteed to happen in the case of screwy "repairs" like mixing modules from other packs or cutting out individual cells as touted as a fix on YouTube... things guaranteed to eventually leave you stranded.
Here's some of the output from one of 057's internal diagnostic tools showing data from the BMS tracking a ton of data about the battery pack at the individual group level. This one is a working pack with one group in one module a bit weaker than the rest. Since the BMS has been aware of this, and has adapted to this over years and many tens of thousands of miles, its able to anticipate the needs of that group and work accordingly. The CAC delta is getting pretty close to problematic here, though, and results in a periodic "Maximum Charge Level Reduced" error as a result. This pack should have around 68-70 kWh usable, but the CAC delta is causing the BMS to be more and more cautious as it fights to maintain order with the growing imbalance.
(This particular one turned out to be corrosion from moisture causing a mild short on that low CAC group.)
Here's another where some mismatched modules were introduced, eventually causing a failure as they were thrown more and more out of whack with the rest of the pack. The BMS had no clue how to handle this situation and couldn't make any reasonably predictions. So instead it had to lock out something like 25% of the pack, and eventually shutdown.
Voltage balance is only a small part of the equation. The BMS strives for capacity balance, and it works in interesting ways to achieve amazing results with homogenous original unmodified packs.
Long story short, the BMS is very very good at its job, and under normal conditions, you should never have a shutdown with > 0 miles.
The BMS is using some very complex algorithms, including what appear to be full simulations based on historical readings, to calculate the capacity of every single cell group (all 84 or 96). I've pitted this against bench testing equipment, and generally get to within < 1% the same data.
I'll preface this next part with the note that usable range below zero is NOT guaranteed, nor has it ever been in any firmware version.
The modern BMS's target an "empty" of roughly -3%. In most cars this is ~10 miles or so. (This is actually the reason for some visible rated range drop in early Model 3's and such when they tweaked this on those vehicles.) This "buffer" (it's not really a buffer) is adjusted based on the BMS's perception of the error in its readings, and can grow as much as needed to make sure 0 on the dash is at least a mile of rated range. In the case of problem batteries, this can be a super high value, like 70% "buffer".
The key thing is that this buffer is not necessarily usable, and the error is greater the less the car is driven below about 20-30% or less SoC. Or put different, the more the car is short cycled above 20-30% (like daily driving from 90% to 70%). The BMS mitigates this with a variety of techniques and is generally always pretty cautious. It's basically a way to move the window of range it shows as available into an area of the capacity calculation's error bars that effectively guarantees capacity is available there.
In theory, if the BMS is reporting 0 miles of range, if perfectly calibrated you'd still have a few miles left "below" zero.
In reality, 0 should be treated as 0. Don't count on any capacity below it.
Above 0, however, if the BMS decides it has to stop the vehicle it means something is too far out of whack either as estimated idle or under load. This usually happens worse at very low SoC (<15%), and is exaggerated by the wear on the battery, and other factors like short cycling adding error to the capacity calculations.
Generally today, though, if your car stops you with a positive value on the range meter... something is probably wrong with the battery pack. This is what is guaranteed to happen in the case of screwy "repairs" like mixing modules from other packs or cutting out individual cells as touted as a fix on YouTube... things guaranteed to eventually leave you stranded.
Here's some of the output from one of 057's internal diagnostic tools showing data from the BMS tracking a ton of data about the battery pack at the individual group level. This one is a working pack with one group in one module a bit weaker than the rest. Since the BMS has been aware of this, and has adapted to this over years and many tens of thousands of miles, its able to anticipate the needs of that group and work accordingly. The CAC delta is getting pretty close to problematic here, though, and results in a periodic "Maximum Charge Level Reduced" error as a result. This pack should have around 68-70 kWh usable, but the CAC delta is causing the BMS to be more and more cautious as it fights to maintain order with the growing imbalance.
(This particular one turned out to be corrosion from moisture causing a mild short on that low CAC group.)
Here's another where some mismatched modules were introduced, eventually causing a failure as they were thrown more and more out of whack with the rest of the pack. The BMS had no clue how to handle this situation and couldn't make any reasonably predictions. So instead it had to lock out something like 25% of the pack, and eventually shutdown.
Voltage balance is only a small part of the equation. The BMS strives for capacity balance, and it works in interesting ways to achieve amazing results with homogenous original unmodified packs.
Long story short, the BMS is very very good at its job, and under normal conditions, you should never have a shutdown with > 0 miles.
Last edited:
I'm not really particularly interested in doing this experiment. I would expect that, when it says 1 mile left that there won’t be much more than 1 miles left. If I turn off the A/C or the heater, there might be 1.5 miles left. It might also do damage to the batteries
Thanks a lot for the response. I had a fear that this is probably the case, and I have it booked in for a service. will see how that goes. I just had a MCU 2 installed a month ago and got fresh updates. at the moment it is sitting on 2022.20.7. with some 2500km with that version, and 5500km since MCU 2 upgrade.
Is it possible that the latest installs of the software were still not calibrated against low SOC ? I tend to use my car 10-60% SOC but its not uncommon to be in 5% at arrival to a charger, and I installed the latest sw version with a battery at 96% just before the long road trip (where this happened)
yes, its a reman (one of the ones you have reported as possible to have "issues", sn 1088815-01-E ), and at 10% SOC I would see some of the blocks drop bellow to 2.576V on the Can/Bluetooth with "high load".
Thanks a lot for your time for this and all other posts. I have read so much (and Learned I believe)
So yesterday, my 2018 Tesla Model 3 performance with 105,000 miles on it just shutdown when it was at 3% while I was driving. But I actually had my original HV battery replaced under warranty at around the 50,000 mile mark so I only have around 50,000 miles on this current HV battery pack. I did notice that I could not drive faster than 65-70 mph though when I was at 3% to where I actually floored the accelerator pedal about 4 to 6 times to make sure. Could me flooring the throttle this many times at 3% caused it to prematurely go into its Shutdown state with everything you said about how good the BMS is? Or should I get it looked at by Tesla service?
Luckily the shutdown occurred right when I was approaching a supercharger so I was able to coast right into the site. But since I had to stop anyways to back into a stall with prevented me from shifting into reverse, I had to put it into tow mode (neutral would not work) to push it into a stall.
EstebanP212
Member
Lowest I’ve got my battery down to is 3% the rule i follow now is %5 is lowest I’ll go when trying to drive far and decide if I can make it without stops it says I’ll arrive below 5% lll supercharge for a few minutes
Similar threads
- Question
