Yeah, the BMS wont have knowledge of the true top capacity without a near full discharge and a near full charge. Inside the full voltage range it uses the current shunt and cell voltages to estimate SoC. Extrapolating out to 100% from lower capacities will generally result in gross errors in either direction (usually resulting in a lower value in cars more than a few months old or so and higher in newer cars that haven't had much calibration at all). You can still get an idea by extrapolating proportionally (range / percent), but I wouldn't take it as an accurate number and give it at least a +/- 5% error margin. The car generally tends to have a better estimate of range closer to 0% and closer to 100% overall. So extrapolating from 5% or 95% to 100% is likely more accurate than 80% to 100%, for example.
this is interesting and useful information, as well as the info about the BMS not knowing true top capacity without a near full charge/discharge - thanks. Since I was the one who earlier brought up the topic of range mode, I'll just add that every time I've turned on range mode I noticed the slight increase in rated range, and every one of those times was after shortly after charging, but before starting the car. Next time I'll start the car and look at the numbers. but this brings up a question - if "the end of charge number means very little, especially before some calibration has happened", with all this discussion of 100% rated range and calculations based on that number, doesn't that mean everybody's numbers here (and in every other thread discusing rated range) are possibly a bit off unless they started the car before recording the rated range? i.e. should you always start the car before looking at the rated range displayed to get the most meaningful number? or is this a non-issue most of the time (not sure how to know/when the calibration happens). I'm guessing I'm not the only one who looks at the rated range number on the dash after a charge but before starting the car and driving...
If you've charged to 100% recently, the number at 100% subsequent times should be pretty close, regardless of starting the car or not. If you're charging to 100% just to check your max range or otherwise not frequently, you should start the car first and have range mode off to get a more meaningful read.
Ok, I'm fine with ending the discussion about your estimated capacity numbers, as it may not be appropriate for this thread, but I believe my inquiry has brought up other issues that will affect the rated miles display that were not previously mentioned in this thread, but does have a measurable difference in results. 1) The need to hit 0%/100% for BMS calibration 2) enabling/disabling range mode effects 3) number after charge vs a few miles after: Is this capacity from low load, meaning if you do run your car accessories without engaging drive, it'll give you the equivalent of "272 miles" of usable capacity in kWh? How much is this from running it down by 3 miles vs the change in the load profile?
Thanks!!! I don't have a good reason to charge up to 100% in the foreseeable future but I'll keep this in mind when I do as it'll be interesting I think... Jeff
For mine. Rated range = 258.9 x 295 wh/mi = 76.4 kWh. CAN reports 76.3 kWh. Yep, looks good. @wk057 Does CAN 3 report an impedance measurement from the BMS that can indicate overall battery health?
So in a hypothetical charge of my S70D to a mid-point, I'm putting some serious error bars into the car. Hypothetically, the BMS reports 120 RM, while in reality there is 114RM. So from this, I take some conclusions, regarding the car's power to adversely surprise: In transit, the car will warn to slow down later, even as it tries to revise its overly optimistic RM estimates; In transit, the car will tic off RM faster as you reach your destination, than would be suggested by the instantaneous Wh/mi; (I'm guessing) instantaneous Wh/mi will still remain measured with minimal error; The chances that the car approaches '0 RM' increase, with miles left to go -- leading to a increased demand for 'emergency charging' at the doorstep of the destination SuperCharger (I think Bjorn Nyland had a video where he pushed his car the final 200M to his snow-bound SuperCharger).
All seem correct. I should make note in the OP. I'm admittedly not 100% sure why the numbers change so substantially after actually starting the car. My theory is that as the final charge trickles in something that updates the range display gets out of sync. Then when the car is turned on and it actually changes a little (even sub 1 RM) it updates to reflect the actual BMS data. This is mostly speculation, however for sure I know that the number displayed after charging but before starting the car is not always accurate. Nope. If internal resistance is actually tracked by the BMS it's not exposed anywhere. Generally if you charge to a lower SoC often the error will tend towards displaying less rated miles available as the pack will tend to underestimate capacity. However, I've seen it both ways and it will overestimate as well. Either way, as you discharge the pack and expected voltages don't match up with power usage the BMS will just adjust towards the correct value as things progress. You should never hit 1 mile left and not have any juice. You *could* hit 0 and still have a little, but it's unlikely. The BMS does do a decent job estimating range these days < ~15 miles or so.
For fun, I dug out the photo I took after that one supercharge I mentioned earlier. Memory was a little off, was 273 rated miles with ~3.2k on the odo. Range mode on. Then my next shot was before plugging at my next stop ~210 miles away. So, a look at the numbers. 273 rated miles at the end of the charge and before moving, traveled 210.1 miles using 65.6 kWh at 312 Wh/mi and arriving with 37 rated miles. Well, 273 sounds awesome. That's 81.125 kWh usable if it were true! However... consuming 65.6 kWh ticked off 236 rated miles from the 273 number. 65.6 / 236 = ~278 Wh/mi. Now, I think we can all agree that 278Wh/mi is not the value per rated mile on any Model S nor has it ever been. But, let's assume that's right... and I have 37 RM left. Well, that'd be 10.286 kWh remaining, for 75.886 kWh total. Or, if we use the real number of 295 Wh/mi, it would be 10.915 kWh remaining, for 76.515 kWh total. So, something is obviously not telling the whole story, and its the end-of-charge rated miles of range display. I clearly recall ticking off many rated miles over the first few miles of the trip, but Wh/mi was not high as I was driving very conservatively since this was my first long trip with the S. Anyway, just a personal data point showing the end-of-charge rated range number to be bogus at times. Doesn't seem to be nearly as bad in newer firmwares, but still some error there for sure.
My confidence is eroding fast. Now revising: Hypothetically, the BMS reports 120 RM, while in reality there is [[114]] 105 RM...So, now, potentially, I'll have 15ish miles disappear (while driving my S70D down from 50% to 0%) if I've biased the BMS-estimation by mostly charging and driving through the 50-100% full zone. So far, however, I've been pretty balanced in driving (and garaging) my car with equal frequency down at 20% SoC and 80% SoC. ... so maybe this strategy will moderate the degree of error? We'll see.
Also, why does BMS State of Energy vary from state of charge (and rated miles) on the low end? Basically what I care about is which number I should place greater confidence in. I've been in a situation where the IC reports 5 miles of range (2% SOC) left but CAN is showing 4% SOE and something like 3.1 kWh remaining. Which do I believe?
I don't believe this is about end of charge reading. For me in my car, this discrepancy is always true. If it was a misestimation, the miles should come back *somewhere*, yet if you sample consecutive drives, it never does. There appears to be some sort of fudging going on. Instead of doing 65.6 / 236 = 278Wh/mi, try looking at it as 236 * 0.295 = 69.62 kWh actual used. Therefore, the trip meter is probably wrong by 6% for this drive OR the trip meter is right and the energy just never existed, and the rated miles count was fudged down to compensate. Shady...
Uhh yeah, I'm now around 233 miles at 90% with very few miles on the odo. It's only a couple miles more than 85's with like 30k miles on them. 90%*5kWh/310Wh = 14.4 miles I should have on those cars, before even looking at years/tens of thousands of miles of degradation. WTF.
Cycling the pack *might* help if my suspicions are true. I definitely notice that my 100% rated range ticks up a few miles after a road trip.
Love this thread. The UI on my 90D X seemed to indicate 330 Wh/mi was the rated range, but some unscientific calculations I had done seemed to suggest it was more like 320 Wh/mi. @wk057 posting the value of 320 makes me believe even more so that is the efficiency I need to achieve in order to get the rated miles. After 16k miles and 9 months in my X I have about 2.7% degradation as a 100% charge is 250 miles for me now.
In my experience, 100% charges never tick up displayed range. I range charge often and deep dod often (it's a 60--I need to top up before a trip and am used to landing at a supercharger with 8-12 miles remaining). I was hoping there would be a more comprehensive hack that essentially wipes any 'learned' calibration....
Actually that's almost exactly the value I need to hit for rated to match actual on my early 2013 S85...
Data Point - Highest single leg in my 2012 P85 was 75.9 kWhr - 242.2 miles ( 265 rated - 0) temp ~70F @303 rated (10K miles on odo, firmware 4.4) 75.886 eh? hmm
So how does one explain people driving miles past 0%? I know it has been documented a few times over the years in different models. Is it just some rounding or other error in how it is shown, so in reality even though it is showing 0%, it is really 0.4% or something like that?
@Footbag just went about 4 uncomfortable km on 0%, but it didn't look like something I'd like to try. 4km's not much of a buffer anyhow. Not sure what all of his stats were (kW hr used, etc.) but might be in his video.