Almost 15% range loss Model 3 Awd

[AlanSubie4Life]

My battery showed something closer to 300 the last time I did a 100% charge. Stats claims it would show 299 (said 303 yesterday, lol). The remaining to 90% (where it shows I need 25 kWh) is pretty much right (90% of 73.8 is 66.4, car says it thinks I have 41 kWh remaining).
I usually charge every night, if I remember I will try to check the can data in the morning.

I see. So that would calculate out to 245Wh/rmi*300rmi = 73.5kWh plus the 3.3kWh buffer. I guess the question is whether the buffer is subtracted from the 73.5kWh (the video posted above suggests that but does not make it clear). The evidence suggests that it is not subtracted from it. So that would make your battery 76.8kWh (which would have been 79.2kWh if it had 310 miles at a full charge - same as the EPA result - which is probably not a coincidence).

Anyway, it seems like it's possible it will align exactly with the charging screen data (same scaling). We'll see. To get good alignment of the data, you'd have to keep track of vampire drain losses or minimize them, since the charge is happening overnight.

 

[diamond.g]

lol, now the car (app) says I have 55% and I haven't done anything. For the record I also have 13909 miles on my car. 3.1% degradation for me.

 

[TimothyHW3]

lol, now the car (app) says I have 55% and I haven't done anything. For the record I also have 13909 miles on my car. 3.1% degradation for me.

Interesting. Can you get a hold of ScanMyTesla for Android and post a picture of the values there? That would be interesting. Especially cell imbalance and nominal full

 

[ajdelange]

Interesting. I just pulled some can data...
So 292 shows me that my screen lies to me (car says I have 54% but can bus says I have (SOCui) 58%.

I think part of the problem here is that no one here really knows what SoCUI, or SoCmin or Nominal Energy Remaining (to use the terms in the Bjorn video) really means nor does anyone have any idea about how what the car displays, and what it reports via the API relate to those. I come from an engineering background and I can remember weeks of effort to make sure that all participants in a multi billion dollar program agreed on a definition of what Signal to Noise Ratio meant and that once the standard was promulgated everyone involved adhered to it. You see SoCUI in the CANbus data and make the obvious assumption that this means the value of SoC reported by the User Interface. Obvious, right? But the UI displays 54%. So those two things clearly are not the same thing (similar discrepancy noted in the Bjorn video). You conclude that the UI in the car is deliberately deceiving you. Why would Tesla want to deceive you? Possible answer: they want you to think that you have less energy available than you do so that you'll get off the freeway and pick up some charge.

At this point I'm thoroughly confused. SoC is really only measurable in terms of pack OC voltage but we can't shut off the motors in the middle of the Harbor Freeway to measure OC voltage and so the rolling SoC estimates have to come from ∫i(t)v(t)dt. But in evaluating that integral we have to recognize that we will get different values in charging from Empty to Full than we will get discharging from Full to Empty because all of the energy put into the battery isn't available to be taken out because not all the energy put it can be taken out. Some is lost to heating of the cells internal impedance and more will be lost to balancing unless "flying capacitor" balancing is used. Now it's a simple matter to determine the internal resistance of the battery on the fly ( it's ∂V(t)/∂i(t) ) and so it would be trivial for Tesla to determine how much of the energy delivered to the battery terminals is actually delivered to the chemical system. But do they do that? I certainly don't know. In any case I'm starting to wonder if some of these discrepancies are explained by the difference in energy required to be added to bring about a given change in SoC and the amount of energy available in bringing about the same change in SoC in the opposite direction.

 

[diamond.g]

Interesting. Can you get a hold of ScanMyTesla for Android and post a picture of the values there? That would be interesting. Especially cell imbalance and nominal full

My unit doesn't have that capability. I got the CANDue before he added the ODB2 dongle to it. Not sure if I can patch ESP32 to pass ODB2 over bluetooth.

 

[TimothyHW3]

This info is all in the EPA document. The recharge event energy for the AWD is 89.6kWh.

This link isn't for the AWD (I'm too lazy to look it up but the document is at the same site and I have it on my computer):

]

You don't really need any documents to check what the rated miles in terms of capacity kWh was. You take the energy tab grey line value which is exactly 153 and multiply it by 500km, which is the rated in km. This gives you total of 76.5 nominal full including the buffer. This is also what we are seeing from the CAN on a new car - I have seen reports of 77.5-78 too. So since this is not an exact science and the BMS is kind of speculating what the capacity is, we can use 77 as a reference.
My nominal full is around 75.7 and yes, there is some degradation after 10,000 miles. I have been doing 65% DC fast charging, sometimes with 200kW, but I think it holds pretty well so far.

 

[TimothyHW3]

I think part of the problem here is that no one here really knows what SoCUI, or SoCmin or Nominal Energy Remaining (to use the terms in the Bjorn video) really means

We have a pretty good understanding what they mean. The problem is that this is BMS, it calculates this stuff and if it is uncalibrated there are deviations.

SOC is the SOC from 0-100% and is calculated by (nominal remaining - buffer)/(nominal full - buffer)*100
This is the SOC you see on the screen.

SOC UI is the full capacity, including buffer. And there are also min and Max values, because as we said, this is not the same as measuring water in a tank...You have to measure voltage across 2750 cells.
In my case SOC was 80.8% and SOC UI was 82.8% which was the exact middle of min 81.9 and Max 83.6. But as I said, these are just very close estimates.
From what I see the car uses SOC min as a reference for nominal full pack remaining or vice versa. Kind of to be on the safe side I guess.
Nominal remaining was 61.9kWh

As you can see, there is a small difference of 80.8% SOC shown onthe screen(without buffer) and SOC 81.7%, actually full remaining with buffer. This is why you can't really measure exact degradation with SOC and especially in miles and not km.

 

[AlanSubie4Life]

You take the energy tab grey line value which is exactly 153 and multiply it by 500km, which is the rated in km. This gives you total of 76.5 nominal full including the buffer.

Well, I get 75.95kWh (310rmi*245Wh/rmi), and I would say that does not include the buffer, based on the charging screen evidence.

This is also what we are seeing from the CAN on a new car - I have seen reports of 77.5-78 too. So since this is not an exact science and the BMS is kind of speculating what the capacity is, we can use 77 as a reference.

Maybe! It's just hard to say. As I said above, I think it would be useful to compare the CAN bus data to the charging screen data, so we can understand what is the relationship. We shouldn't assume the CAN bus kWh should align with the charging screen kWh, unless it is verified.

As far as the energy tab value is concerned: 245Wh/mi is the energy tab gray line value (I think). To derive the value for projected miles displayed on that screen, you take your remaining rated miles and multiply by 230Wh/mi / (current efficiency over last 5/15/30 miles (or km or whatever)).

So if it says you have 155 remaining rated miles, and your last 5 miles were done at 197Wh/mi, then it will project to 181 miles of range. The 245Wh/mi number is not used for that calculation as far as I can tell.

155rmi * 230Wh/rmi / (197Wh/mi) = 181 miles

This is easy to verify, anyway - don't take my word for it! I'm not sure what the purpose of the line is, to be honest, except to remind you of the charging constant. If you exactly match the 245Wh/mi line, then your projected range will be less than your remaining miles by the ratio 230/245.

In other words, you can use that projected range number to derive that ~230Wh/rmi constant.

rated miles/ projected miles = current efficiency / constant

 

[AlanSubie4Life]

To derive the value for projected miles displayed on that screen, you take your remaining rated miles and multiply by 230Wh/mi / (current efficiency over last 5/15/30 miles (or km or whatever)).

So if it says you have 155 remaining rated miles, and your last 5 miles were done at 197Wh/mi, then it will project to 181 miles of range. The 245Wh/mi number is not used for that calculation as far as I can tell.

155rmi * 230Wh/rmi / (197Wh/mi) = 181 miles

This is easy to verify, anyway - don't take my word for it! I'm not sure what the purpose of the line is, to be honest, except to remind you of the charging constant. If you exactly match the 245Wh/mi line, then your projected range will be less than your remaining miles by the ratio 230/245.

In other words, you can use that projected range number to derive that ~230Wh/rmi constant.

rated miles/ projected miles = current efficiency / constant

Correction...this is all wrong...was from memory. Anyway, replace 230 with 245 above.

It makes the projected range optimistic but as others have said it is probably correct if you account for the reserve. So this screen represents the distance to a dead battery, not to zero rated miles. I just got it backwards.

So at 245Wh/mi on the trip meter chosen averaging interval with 310 rated miles, it will show 310 miles of range.

But since we know at 245Wh/mi on the trip meter it takes 245/234 (the 234 I cannot get, but there is reason to believe my estimate will always be low, due to vampire) times as much energy to replace that energy, we know this means the capacity (of a new battery) is:

Energy consumed per trip meter: 245Wh/mi * 310mi = 75.95kWh (after completely exhausting the reserve)

75.95kWh (trip) * 245Wh (charge) / 234Wh (discharge) = 79.5kWh which is about 0.3% off the EPA number.

That’s how I see it working out anyway. (It’s also possible that the car would not travel 310 miles at 245 Wh/mi which would explain the slightly high estimate or why using my 230 value does not work...)

These of course are “true” kWh. What the BMS reports could be scaled. The only way to determine that is to do a charging event and compare the deltas in BMS kWh and the charging screen kWh.

 

[rtheron]

how many times have you charged the car to 100% and do you charge level 2 at home/work or superchargers only?

 

[diamond.g]

These are my 90% charge.

 

[TimothyHW3]

lol, now the car (app) says I have 55% and I haven't done anything. For the record I also have 13909 miles on my car. 3.1% degradation for me.

You mentioned on page 1 the CAC reset and linked an old Reddit post. Unfortunately Tesla deleted the file. Do you have a copy of the procedure or know if they changed the internal number? Do you know what the procedure exactly is, do they just flip a switch and reset the BMS?

 

[TimothyHW3]

Well

As far as the energy tab value is concerned: 245Wh/mi is the energy tab gray line value (I think). To derive the value for projected miles dconstant

You "think" or did you measure it? Just drive near the line and it will show the number. Plus like I said, there are conversion deviations due to miles being inaccurate. Also wether it is 76,76.5 or 77/78 doesn't matter, because the UI does not show anything above those 76-76,5

I will try it myself and switch to miles to see what the rated consumption exactly is.

 

[diamond.g]

You mentioned on page 1 the CAC reset and linked an old Reddit post. Unfortunately Tesla deleted the file. Do you have a copy of the procedure or know if they changed the internal number? Do you know what the procedure exactly is, do they just flip a switch and reset the BMS?

You had access to the toolbox? Toolbox no one has indicated in the thread that the article has changed.

 

[TimothyHW3]

Oh, I thought it was public? Do you have access, can you get the file and post it here?
Or if no access to the file- do you know what the procedure is?

 

[diamond.g]

Oh, I thought it was public? Do you have access, can you get the file and post it here?
Or if no access to the file- do you know what the procedure is?

I am not sure, I don't have access. I thought the car had to be plugged in to a computer to do the procedure.

 

[AlanSubie4Life]

These are my 90% charge

In order to figure out the scaling, we need:

1) 352 reading at a lower charge state, say 30-40% charge.

2) Charge car to 80-90%. Picture of charging screen showing kWh added (must be in % display mode to show kWh)

3) immediately get another 352 reading with all the kWh numbers.

You "think" or did you measure it? Just drive near the line and it will show the number

Initially I did this from memory (which is why I said “I think” and also got the whole thing wrong).
Then I drove home, and the line is at 245Wh/mi. (And I also noted my other error/bad recollection explained above.) If you drive at 245Wh/mi (trip meter average over the chosen interval), projected range and rated miles remaining will be equal, numerically. But of course we know that projected range must mean to absolutely zero battery, not to zero rated miles (since we know the actual discharge constant (not the line) is ~230Wh/rmi, not 245 Wh/rmi).

 

[mswlogo]

It would help us see if he's just burning a ton of power and thus the range is actually accurate based on his usage. If he's getting 250 Wh, then its a problem. If this guy is using 400 Wh then maybe not. Some people are running AC on high, driving at 95 mph on the interstate and wondering why EPA range isn't happening.

There is absolutely nothing wrong with driving 95 mph and AC on High it should have very little impact on degradation, but it could have some over the long haul (equating to more charge cycles).

wh/mi is a fair question because if he is driving it normally and the wh/mi is very high it might indicate a problem.

A better initial question might be is, how many miles does he have on it?

Miles == Charge Cycles. That matters way more than time or driving style.

If OP said 50K miles then 15% might not be very far off.

If OP said 10K miles or less, something seems very wrong.

 

[KhalHagrid]

2019 Model 3 LR RWD
9000 miles

I have owned the car for ~6 months.

For first 5 months,
80% = 261-263
90% = 290-292
100% = 323 and 325 (charged to this level twice and car was driven right away on both occasions)

Charging habits:
Charge to 80% nightly
Charge to 90% two times a week (I drive to work only twice a week, which is a ~170 mile round trip. On these days, I start at 80%, get to work and charge to 90% (free juice). I don't really need to charge, but you know..."free juice, will have". When I get back home, I plug it back in at 80%. On other days, I sometimes drive around town and plug it back in at 80% for the night).

One month ago, coincidentally after a firmware update, I noticed that the estimated range started dropping with each charge.

80% = 259 followed by 257 the next time followed by 255 the time after
90% = 287 followed by 285 the next time followed by 283 the time after

It seemed odd that every time I was plugging the car in, the estimated range dropped by 2 miles after having been consistent for 5 months.

I chatted with a Tesla support person online who said that their remote diagnostics suggested that the car was holding less charge than other cars with the same mileage/age and suggested I bring it in to a SC.

I set up an appointment with the SC. The tech from the SC texted me saying that there was nothing wrong with the battery and that the 50%-90% cycling range was probably skewing the numbers. He suggested I bring the SoC down to 10-15%, then charge the battery back to 90%, unplug the car and let it sit for several hours for the electronics to re-calibrate. He then went ahead and released the service appointment.

I tried the above 3 times. It did not work but only dropped the estimated range further (which is where it is at now).
90% = 280-282
80% = 251-252
100% = 312

I set up another appointment with the same SC. Another tech texted back saying that the EPA of the car is 310 and that the battery is performing as expected. I told him that I have the RWD and not AWD model and that I have charged to 325 before. He replied back saying that it is expected degradation (so, just in a matter of seconds, he went from EPA=310 to expected degradation). I asked him if he can confirm if the loss is indeed degradation or merely a calibration issues. He replied a few days later saying that the battery is holding 97.7% charge and released the appointment.

So, basically inconsistent answers from 3 Tesla people I have spoken with.

I have set up an appointment with a different SC. Let us see what snake oil they have in store at this location.

 

[mswlogo]

2019 Model 3 LR RWD
9000 miles

I have owned the car for ~6 months.

For first 5 months,
80% = 261-263
90% = 290-292
100% = 323 and 325 (charged to this level twice and car was driven right away on both occasions)

Charging habits:
Charge to 80% nightly
Charge to 90% two times a week (I drive to work only twice a week, which is a ~170 mile round trip. On these days, I start at 80%, get to work and charge to 90% (free juice). I don't really need to charge, but you know..."free juice, will have". When I get back home, I plug it back in at 80%. On other days, I sometimes drive around town and plug it back in at 80% for the night).

One month ago, coincidentally after a firmware update, I noticed that the estimated range started dropping with each charge.

80% = 259 followed by 257 the next time followed by 255 the time after
90% = 287 followed by 285 the next time followed by 283 the time after

It seemed odd that every time I was plugging the car in, the estimated range dropped by 2 miles after having been consistent for 5 months.

I chatted with a Tesla support person online who said that their remote diagnostics suggested that the car was holding less charge than other cars with the same mileage/age and suggested I bring it in to a SC.

I set up an appointment with the SC. The tech from the SC texted me saying that there was nothing wrong with the battery and that the 50%-90% cycling range was probably skewing the numbers. He suggested I bring the SoC down to 10-15%, then charge the battery back to 90%, unplug the car and let it sit for several hours for the electronics to re-calibrate. He then went ahead and released the service appointment.

I tried the above 3 times. It did not work but only dropped the estimated range further (which is where it is at now).
90% = 280-282
80% = 251-252
100% = 312

I set up another appointment with the same SC. Another tech texted back saying that the EPA of the car is 310 and that the battery is performing as expected. I told him that I have the RWD and not AWD model and that I have charged to 325 before. He replied back saying that it is expected degradation (so, just in a matter of seconds, he went from EPA=310 to expected degradation). I asked him if he can confirm if the loss is indeed degradation or merely a calibration issues. He replied a few days later saying that the battery is holding 97.7% charge and released the appointment.

So, basically inconsistent answers from 3 Tesla people I have spoken with.

I have set up an appointment with a different SC. Let us see what snake oil they have in store at this location.

It's possible you have a problem, but there is no way they will listen with that little a drop at this point, which very well may go away in time.

At most, ask them to run a diagnostic/monitor it.