I imagine the reason they don’t do this is that you could easily find yourself in a situation where you decide to upgrade to SR+ but don’t get additional range because degradation had been hidden. I’d be pissed if I paid $1500 and got 2 more miles of range or something.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Calibrating BMS with a software locked battery (SR)
- Thread starter lateulade
- Start date
@AlanSubie4Life I have a full battery’s worth of driving today, and did a semi-scientific experiment for the first leg.
Charged to 100% (201rmi per display / 201.6rmi, per Teslafi) and took off. I was in P for 10 minutes while running an errand, but otherwise was moving. After driving exactly 100rmi (and 50% battery usage per car/TeslaFi) , I stopped and looked up my TeslaFi stats, which disagree from the car’s stats fairly significantly, and way more than 10 minutes in Park would justify.
Car says I went 93.8 miles and used 21kwh (it flipped from 20 to 21 seconds before I logged this data) at an average of 219Wh/mi.
TeslaFi says 93.29 miles driven (close enough for government work), but 22.65kwh used at 243wh/mi, which is 8% more battery used and 10% less efficient.
In an ideal world, I want to believe Tesla’s efficiency number of .219 AND TeslaFi’s usage number of 22.65, because that would mean I have 45.3kwh total (minus a ~4.3kwh buffer, right?), which means 49.6kwh, and effectively no real degradation. If I assume Tesla is right, I have only 42kwh of capacity.
Charged to 100% (201rmi per display / 201.6rmi, per Teslafi) and took off. I was in P for 10 minutes while running an errand, but otherwise was moving. After driving exactly 100rmi (and 50% battery usage per car/TeslaFi) , I stopped and looked up my TeslaFi stats, which disagree from the car’s stats fairly significantly, and way more than 10 minutes in Park would justify.
Car says I went 93.8 miles and used 21kwh (it flipped from 20 to 21 seconds before I logged this data) at an average of 219Wh/mi.
TeslaFi says 93.29 miles driven (close enough for government work), but 22.65kwh used at 243wh/mi, which is 8% more battery used and 10% less efficient.
In an ideal world, I want to believe Tesla’s efficiency number of .219 AND TeslaFi’s usage number of 22.65, because that would mean I have 45.3kwh total (minus a ~4.3kwh buffer, right?), which means 49.6kwh, and effectively no real degradation. If I assume Tesla is right, I have only 42kwh of capacity.
AlanSubie4Life
Efficiency Obsessed Member
Great! Thanks for the data! For the return trip I assume you will briefly stop to Supercharger, but try to capture the details of each leg.
This is the first trip meter detail I have seen for the SR, and it aligns well with the SR+ (SR) expectations in the table (ignore the SR line as it is incorrect - I was guessing at the constant as I did not have data - I will correct that soon). 205Wh (trip) per rated mile. From 219Wh/rmi, 95.5% reduction due to buffer, 2% due to some other loss factor (which is consistent among vehicles for the most part).
Summary of my prediction for your battery:
Full: 219Wh/rmi*201rmi = 44kWh
Buffer: 44kWh*4.5% = 2kWh
Usable Trip 201rmi -> 0rmi = 44kWh*0.955*0.98 = 41.2kWh
Recharge AC Energy to add 201rmi @ 236V, 24A:
201rmi*209Wh/rmi / 0.882 = 47.6kWh (8 hours 24 min).
As far as TeslaFi goes, it has been extensively documented that certain stats it provides are not accurate, or need some correction factors. I don’t use it so can’t provide specifics - some data is fine, others are not. You can search posts for TeslaFi with me to get some info - I have asked questions and people have clarified what to not trust in the past. But anyway you can safely ignore that data - it is not reflective of what is actually happening and will just confuse you. Not worth trying to align it.
The car is correct.
@AlanSubie4Life Thanks for your calculations! If I’m reading them correctly, it’s your theory that I have ~15% degradation (41.2khw vs an expected 47.65kwh — calculated by (SR+ capacity of 54.5*.916 (SR software lock))*.955 (buffer allowance). If so, doesn’t that seem extreme for 11 months / 30k miles?
Unfortunately, I ended up having to make a bunch of stops and errands which resulted in a 6% loss.
However, I found a setting to adjust the TeslaFi consumption calculations, and adjusted it to reflect only .901 of its consumption calculations (.243/.219), so the wh/mi and battery consumption levels are now matching my car. I’m gonna drive around a few days and see what I can find.
Unfortunately, I ended up having to make a bunch of stops and errands which resulted in a 6% loss.
However, I found a setting to adjust the TeslaFi consumption calculations, and adjusted it to reflect only .901 of its consumption calculations (.243/.219), so the wh/mi and battery consumption levels are now matching my car. I’m gonna drive around a few days and see what I can find.
AlanSubie4Life
Efficiency Obsessed Member
Approximately, maybe.
But let’s wait for the detailed time logging of the recharge event at home. Right now we are trusting the car meter, which while I have no reason to not trust, it is a possible source of error. The recharge event should be fairly definitive (though even that does assume a certain efficiency, it will be within a % or two).
I would not trust TeslaFi even after correction except for approximate data. The correlation factor may vary.
AlanSubie4Life
Efficiency Obsessed Member
Cool.
Remember, for precision, it is important to log the rated miles for precision. Not the %.
However, this will do.
You could go into the car and quickly flip distance display to miles, and log the rated miles and the exact time (so that you can log charging time and miles from that point). But if you're going to do this, make sure that the HVAC is completely off.
AlanSubie4Life
Efficiency Obsessed Member
Great. Probably you were around 16km. So to get fully charged to 323km (201miles), I would expect it to take:
201rmi-10rmi = 191rmi
(191rmi*209Wh/rmi/0.883 )/ (24A*238V) = 7 hrs 55 minutes
It's saying 7hr 35minutes, but it is subject to rounding error from various sources - I'm going to guess closer to 8 hours still. I'd check in when it gets to 40% and 70% or so to see the new estimate, and get a new capture of the current voltages and currents.
If your notifications are turned on and you think the time is accurate enough (I have no idea how accurate those are), I guess you can rely on that without having to stay up until 2 in the morning.
Honestly, even a timed charge to 80% or 90% would be reasonably definitive - there's no reason to think that there is significant nonlinearity in the gauge.
To me it looks like everything is pretty close to the predictions, unfortunately.
AlanSubie4Life
Efficiency Obsessed Member
Awesome. If you get a chance, just grab some random screen captures from your app as the charge progresses. They don't have to be at exact intervals or anything - the time on your phone on the screen capture, and the voltage and current & % are all we need. It will help us to establish linearity of the charging. I don't think we'll capture anything weird - I expect it to be basically perfectly linear, but the more data the better as long as we're doing this.
Just need three or four points, once every hour or two.
Here we have it! Started at 6:09pm with 6rmi, ended at 1:59am with 201rmi (195rmi added over 7hrs 50mins). TeslaFi indicates 43.1kwh added. I estimate 1.3kwh was left on the battery for the 6 miles of range, for a total of 44.4kwh (not counting the buffer).
Does this seem sound?
Does this seem sound?
Last edited:
AlanSubie4Life
Efficiency Obsessed Member
No. Those numbers from TeslaFi are definitely incorrect.
Your charge actually started at 6:06 as you said. We have to take TeslaFi’s word from it on the end, but likely close to 2AM.
So 7 hours 53 minutes to add 195 rated miles.
I would have predicted:
(195rmi*209Wh/rmi/0.883)/238V/24A= 8 hr 5 minutes.
So that is within 2% which is a good result and validated the method.
This formula confirms, because we know from Tesla the efficiency is about 88.3% under these conditions, that you really HAD to add only:
195rmi*209Wh/rmi = 40.76kWh
Your wall kWh (we know, with very high accuracy) was 238V*32A*7.88hr = 45.0kWh (so I would have predicted 39.9kWh = 0.883*45.0kWh added for your battery)
40.76kWh means 90.5% efficiency instead. Again, within 2% of my prediction based on Tesla’s data.
Your overall battery stats are:
40.76kWh + 6rmi*209Wh/rmi + buffer
= 44kWh (including buffer)
Or:
42kWh (not including buffer)
This is exactly as predicted above.
Realistically the method probably has about 1.5% error, so say +/-0.5kWh.
Compare to charging constant prediction of:
219Wh/rmi * 201rmi = 44kWh (including buffer).
Thanks for doing this. This confirms definitively that those rated miles are usually very accurate, and definitely are in your case. There is no chance that you have energy “hiding” anywhere.
You have 44kWh vs the original 48.2kWh, so no question: 8.8% capacity loss.
You used 45kWh from the wall at 90.5% efficiency to add 40.8kWh of energy to your battery, taking it from 6rmi to 201rmi (195rmi added).
I think this is reasonable, though somewhat on the low end of users.
Your BMS is completely calibrated, it appears.
Last edited:
Thanks! That’s a bummer, but at least I know where I stand — fortunately, you can’t throw a rock here without hitting a supercharger (we’re starting to get V3 as well), so even with my crazy high mileage lifestyle, that’s never a concern.
If I were to upgrade back to SR+, what sort of rmi do you expect I’d see? Frankly, it may not be worth it if I’m only going to get 220 for $1500, because I don’t miss/care about the rest of the SR+ options. May just spend my $2k refund check on basic AP (I declined AP at purchase to save money) and call it a day.
AlanSubie4Life
Efficiency Obsessed Member
For the 202rmi we can adjust the calcs above appropriately. Just a little bit more energy than we thought by 0.2kWh.
For the 43kWh, this is just a display thing. The car ALWAYS shows the rated miles added * the CHARGING constant rounded to nearest kWh. Not the actual kWh added to the battery (if you had SMT or whatever this would be obvious, but please trust me...).
If you calculate that:
196rmi*219Wh/rmi = 42.9kWh
AlanSubie4Life
Efficiency Obsessed Member
I would expect 220rmi. I could be wrong as there are a number of ways they could do this, but that would be my guess.
If I am wrong, I would guess that the actual answer would (initially!) be higher, not lower, than 220rmi.
Last edited:
AlanSubie4Life
Efficiency Obsessed Member
For completeness, we really need someone with an SR+ with a battery in decent shape (more than 230 rated miles) to repeat this experiment and charge from around 5rmi to about 202rmi, and show that the time is the same, and they still have quite a bit of energy to add still. Or better yet an SR with an awesome battery (anyone?).
They would need to be in a warm climate like Florida or Southern California too. Otherwise at this time of year energy could be spent warming the battery and screw up the experiment.
Set charge rate to 24A at around 240V
Record initial and final miles.
Carefully record the charging time with screen captures.
I say SR+ just because there are not that many SR vehicles out there.
This would just demonstrate that indeed the energy available and rated miles are proportional to the time spent charging it. I know it is true but it might help eliminate some doubts that people have.
They would need to be in a warm climate like Florida or Southern California too. Otherwise at this time of year energy could be spent warming the battery and screw up the experiment.
Set charge rate to 24A at around 240V
Record initial and final miles.
Carefully record the charging time with screen captures.
I say SR+ just because there are not that many SR vehicles out there.
This would just demonstrate that indeed the energy available and rated miles are proportional to the time spent charging it. I know it is true but it might help eliminate some doubts that people have.
I'm in the exact same situation. I'm still deciding if I should upgrade. If I can get at least 235 I would, but not if I get 220. (before the downgrade I had around 238) (now 202)
AlanSubie4Life
Efficiency Obsessed Member
Do you guys both have historical data showing the exact size of the rated miles step at the exact time you downgraded? It makes no sense, but maybe they are downgrading SRs more than they “should” be.
Similar threads
