Model 3 Range Constant Check

[SageBrush]

As most of us know, the range meter takes the usable kWh in the battery and divides it by a constant to display the range remaining. That constant was the EPA value when I bought the car, and today I attempted to check whether the constant had changed.

Method: the general idea is to take a drive and divide kWh consumed by distance as displayed by the range meter. In order to reduce rounding errors I started and ended the test just as the range meter dropped a digit, and for kWh I reset a trip meter at the start of the test and then calculated the kWh as measured distance traveled * the Wh/mile displayed. Since the trip meter has one sig digit and my test was measured at 15.5 miles, the inaccuracy is ~ +/- 1/300 in both the distance and Wh/mile numbers. All told, ~ 1%

Results:

• Range dropped 22 miles
• 15.5 miles and 316 Wh/mile per the trip meter

The calculated constant is (15.5*316)/22 = 222.6 Wh/mile

If I recall correctly EPA is 227 Wh/mile for my Model 3 LR. Presuming my memory is correct, this is good evidence that the constant is unchanged.

 

[AlanSubie4Life]

As most of us know, the range meter takes the usable kWh in the battery and divides it by a constant to display the range remaining. That constant was the EPA value when I bought the car, and today I attempted to check whether the constant had changed.

Method: the general idea is to take a drive and divide kWh consumed by distance as displayed by the range meter. In order to reduce rounding errors I started and ended the test just as the range meter dropped a digit, and for kWh I reset a trip meter at the start of the test and then calculated the kWh as measured distance traveled * the Wh/mile displayed. Since the trip meter has one sig digit and my test was measured at 15.5 miles, the inaccuracy is ~ +/- 1/300 in both the distance and Wh/mile numbers. All told, ~ 1%

Results:

• Range dropped 22 miles
• 15.5 miles and 316 Wh/mile per the trip meter

The calculated constant is (15.5*316)/22 = 222.6 Wh/mile

If I recall correctly EPA is 227 Wh/mile for my Model 3 LR. Presuming my memory is correct, this is good evidence that the constant is unchanged.

Agreed, the constant is unchanged (at least since that 310->325 range increase - no idea what the constant was before that! A constant battery capacity calculation would yield 234Wh/rmi before that change, but the battery usable capacity may not have been kept constant with that update - no idea).

That compares well to my tabulation. <DISREGARD> You still probably have two (maybe a little more!) significant figures due to the range drop being 22 miles, though your measurement method was good. Ideally, you wait for the clickover at the start (which you did), and then at the END of the test as well, wait for a corresponding click down - that way you get rid of that rounding error nearly entirely (if your consumption rate is slow). </DISREGARD> EDIT: Looks like you did this, somehow I missed that, so you have 3 sig figs, I agree. (Going to km is helpful too of course. And you also have to go to park before tabulating, for a short test anyway...which can introduce error due to the (hopefully small) regen event, I think!

List of Estimated Range Constants for Model 3 Variants

If you ever do a longer check (before any "range" update (which is apparently not happening)), do post back again - I've very sure of the constant for the AWD but the LR RWD results have been a bit elusive.

The other constant for your vehicle should be:

1.047*~223Wh/rmi = ~233Wh/rmi (maybe 234Wh/rmi? - reported by a user I trust, @Zoomit ) - this is used on the charging screen & for range estimation (Energy Consumption page). It is possible to determine this constant reasonably precisely by taking a picture of that range estimation screen at any time (ideally when using km), and back-calculating for a few different consumption values (higher consumption values and doing it with more range available will lead to better accuracy in the calculation). (You can also use the charging screen and swap between energy/distance, but it is less precise.)

Solid line on the range estimation screen (labeled as "rated") should be 5Wh/rmi higher, so 238/239Wh/rmi. (No idea why this line is where it is; it's an enduring mystery which I hope to be enlightened on (with supporting data) someday.)

 

[FactsBot]

As most of us know, the range meter takes the usable kWh in the battery and divides it by a constant to display the range remaining. That constant was the EPA value when I bought the car, and today I attempted to check whether the constant had changed.

Method: the general idea is to take a drive and divide kWh consumed by distance as displayed by the range meter. In order to reduce rounding errors I started and ended the test just as the range meter dropped a digit, and for kWh I reset a trip meter at the start of the test and then calculated the kWh as measured distance traveled * the Wh/mile displayed. Since the trip meter has one sig digit and my test was measured at 15.5 miles, the inaccuracy is ~ +/- 1/300 in both the distance and Wh/mile numbers. All told, ~ 1%

Results:

• Range dropped 22 miles
• 15.5 miles and 316 Wh/mile per the trip meter

The calculated constant is (15.5*316)/22 = 222.6 Wh/mile

If I recall correctly EPA is 227 Wh/mile for my Model 3 LR. Presuming my memory is correct, this is good evidence that the constant is unchanged.

What do you mean it didn't change compared to EPA? You just showed that it did drop from 227.7 to approximately 222.6. That is a good 2% drop. It is very similar to numbers observed by others.

On Tesla Bjorn's car the real battery capacity degradation is 5.7% but the rated range shows 3% drop only. Rest is hidden by the change of the constant above.

 

[AlanSubie4Life]

Rest is hidden by the change of the constant above.

I am not aware of any constant change in the last 9 months (for the AWD it definitely has not changed).

Any evidence of this?

 

[FactsBot]

I am not aware of any constant change in the last 9 months (for the AWD it definitely has not changed).

Any evidence of this?

I think it changes gradually.

Tesla Bjorn's Model 3 P; EPA: 74000Wh/500km = 148Wh/km

His first capacity test five months ago: 71100/485 = 146.6Wh/km
"Model 3 Performance battery capacity test"

Last capacity test: 67200/465 = 144.5Wh/km
"Calibrating battery in Model 3 part 2"

 

[AlanSubie4Life]

Tesla Bjorn's Model 3 P; EPA: 74000Wh/500km = 148Wh/km

The EPA kWh was ~79kWh FWIW. So should be about 158Wh/rkm, 253Wh/rmi

I think it changes gradually.

Maybe. I just have not seen it.

His first capacity test five months ago: 71100/485 = 146.6Wh/km
"Model 3 Performance battery capacity test"

Interesting. My test in February was on a vehicle with similar mileage. Still got 230Wh/rmi.

I’ve been stubbornly trying to get the 234Wh/rmi I expect (to be consistent with other Model 3 vehicles charge/discharge constant ratios), but have got 230Wh/rmi since February. Of course, measurements tend to be biased low but I tried to do a valid test.


I wonder about a couple factors: 1) was the battery warming over the course of the test? Maybe it contained more energy than BMS estimated at the initial temperature? It is good to pre-warm battery prior to the test. He had not been charging it for long before starting this test as far as I could tell - looked like topping charge only. Would have to be nearly 10rkm of energy though - seems like a lot.

2) He was at quite a low consumption (lower than I typically run). Maybe the meter has some nonlinearity?

3) The environment was generally a lower temperature than where I am. I wonder if it matters.

But the change from my expected value (234Wh/rmi) to the value I get in my tests (230Wh/rmi) is interesting.

It does kind of look like a couple % “hiding” of capacity loss. Maybe they do this first, before starting to reduce your rated miles! All kind of odd though.

 

[FactsBot]

The EPA kWh was ~79kWh FWIW. So should be about 158Wh/rkm, 253Wh/rmi

The EPA test starts with a fully charged battery and stops when the power output reduces and the wheels are losing their target speed due to the low battery.
This means the capacity EPA measures contains the bottom buffer as well. Bjorn's tool shows 3.3kWh bottom buffer.
74kWh nominal capacity + 3.3 = 77.3. EPA measured 78.3kWh, so the battery was able to deliver 1 more kWh from 0% SoC.

2) He was at quite a low consumption (lower than I typically run). Maybe the meter has some nonlinearity?

He intentionally goes slowly. Nominal capacity is specified at a specific temperature (usually 25C) and at a specific discharge current (often 0.2C or 0.25C but I don't know the number for this battery). BMS keeps the battery at 28C, this could add ~1% extra capacity compared to the nominal.
If the battery is rated at 0.2C, one should drive for 1/0.2 = 5 hours at a constant consumption which depletes the battery to 0% by the end of the trip to have an accurate capacity measurement.

If one drives faster and uses more accessories, the capacity will drop compared to the nominal value. See the difference on the example below. Let's say cutoff voltage (a parameter in the BMS) is at 3V. 0.2C will deliver 3280mAh while 0.5C delivers 3100mAh only (-5%). But the consumption between 0.2C and 0.5C is huge, x2.5. If your consumption is a lot higher than Bjorn's, you may measure 1-2% less capacity and lower constant as well.

Best would be if one does a capacity test with a brand new car and compares future measurement values using the same procedure to that reference.

 

[SageBrush]

(before any "range" update (which is apparently not happening)

What makes you say that ?

 

[AlanSubie4Life]

What makes you say that ?

It’s confusing, and I have not been paying close attention, but sounds like range increase is really only for the SR. No idea really.

 

[AlanSubie4Life]

Best would be if one does a capacity test with a brand new car and compares future measurement values using the same procedure to that reference.

Agreed. It would be good to get secondary confirmation. I just wish I had done it earlier on my car!

Also interesting would be to know whether it ever changes on the other cars - again, note that I've only ever seen people report data that looks like it is 209/219 or 223/234 or 226/237 (all differ by about 4.7%) for the other cars. (See @SageBrush post above for example). It always has been weird to me that the 230/245 (which I get over and over and over again now, over a decent range of consumption levels) did not match the ratio (and I've been looking for 234!)

There's plenty of confirming data for the SR that 209Wh/rmi is what you get.

The EPA test starts with a fully charged battery and stops when the power output reduces and the wheels are losing their target speed due to the low battery.
This means the capacity EPA measures contains the bottom buffer as well. Bjorn's tool shows 3.3kWh bottom buffer.
74kWh nominal capacity + 3.3 = 77.3. EPA measured 78.3kWh, so the battery was able to deliver 1 more kWh from 0% SoC

EPA was 79.2kWh for the AWD. I would just add that there's no reason the actual meters need to measure exactly accurately, so I don't really think there is a "discrepancy" to explain here. 234Wh/rmi*310rmi+3300Wh = 75.8kWh. Jury is still out on this, but I've got questions out to people to see whether the CAN bus kWh match the trip meter kWh (I don't think there is any reason they have to match - could just be scaled versions of one another).

See the difference on the example below. Let's say cutoff voltage (a parameter in the BMS) is at 3V. 0.2C will deliver 3280mAh while 0.5C delivers 3100mAh only (-5%). But the consumption between 0.2C and 0.5C is huge, x2.5. If your consumption is a lot higher than Bjorn's, you may measure 1-2% less capacity and lower constant as well.

Yes. Over a 13.5kW (225Wh/mi @65mph) average consumption to 25kW (340Wh/mi @ 73mph) average consumption, (0.18C to 0.33C), I've measured 230Wh/rmi in both cases.

In any case, it is pretty hard to get to 0.5C in normal driving!

In any case, I agree based on these videos that it does appear that the constant changes in the AWD over time (I don't think it's a software update related change) - apparently I just missed the change on my car.

 

[FactsBot]

I found some information.

The guy, who collects the data with the app claims that the nominal capacity of the brand new battery is around 78kWh. And the original buffer is 3.5kWh.
So the available capacity when new is 78-3.5 = 74.5kWh. This is very close to what Tesla Bjorn measured back then when he drove a new car in the USA.

Now after 40k km Bjorn's car degraded to 74kWh and has a buffer of 3.3kWh. So the available capacity is 70.7kWh. He measured 69.8kWh, rest must come from the difference in conditions of nominal and his driving.

The buffer gradually decreases with the nominal capacity.

Here is the table:

they put the constant in the column P, however it is just a number and not a calculation and it doesn't match the data in the previous columns.

Model 3 Battery Log

 

[SageBrush]

^^

Here is the unedited data in a table that groups by Model

 

[AlanSubie4Life]

^^

Here is the unedited data in a table that groups by Model

View attachment 470185

Weird that they averaged the results since it is a bunch of different vehicles. And we know the SR+ value in that table is way off (it is 209Wh/rmi)...

 

[FactsBot]

^^

Here is the unedited data in a table that groups by Model

View attachment 470185

I did not notice the other tabs. Now they added the Q and R column that weren't there yesterday.

SR+ has to be a typo, it shows 70kWh battery. It's around 74.9kWh based on the data on the right.

 

[SageBrush]

Weird that they averaged the results since it is a bunch of different vehicles. And we know the SR+ value in that table is way off (it is 209Wh/rmi)...

I made the pivot table. I should have removed the average but since I noticed it after I posted the table here I decided to leave it in case anybody was curious about a quasi fleet average

 

[AlanSubie4Life]

I made the pivot table. I should have removed the average but since I noticed it after I posted the table here I decided to leave it in case anybody was curious about a quasi fleet average

I see. Also interesting how no matter what I do I cannot measure 235Wh/rmi. I should know - that road trip definitely made the value of the constant on my car (currently) crystal clear - 230Wh/rmi for discharge rates from 0.13C to 0.25C.

 

[FactsBot]

I see. Also interesting how no matter what I do I cannot measure 235Wh/rmi. I should know - that road trip definitely made the value of the constant on my car (currently) crystal clear - 230Wh/rmi for discharge rates from 0.13C to 0.25C.

Yes, I don't understand that either.
Bjorn was able to discharge the battery at 1.3C in his latest high speed test.

During the test he lost 64.5% displayed SoC.
Consumed energy: 42.2kWh
Nominal energy: 64.5%*(73.5-3.3) = 45.3kWh

so that's a 6.8% drop

 

[earthwormjim]

@AlanSubie4Life.

Do you have any data on the charge constant for different line input voltages? Have you looked to see if it's the same constant for 120V input, or 208V input, or 240V or even 277V? It would honestly make some sense for it to have some sort of line input voltage relationship. Lower input voltages should have higher copper losses relative to the input power.

 

[AlanSubie4Life]

@AlanSubie4Life.

Do you have any data on the charge constant for different line input voltages? Have you looked to see if it's the same constant for 120V input, or 208V input, or 240V or even 277V? It would honestly make some sense for it to have some sort of line input voltage relationship. Lower input voltages should have higher copper losses relative to the input power.

It has no dependence on that stuff as it is only a measure of the rated miles added to the pack. And the associated Wh (245Wh/rmi for the AWD).

I have checked at Superchargers, at a ChargePoint (200V), and at home (240V).

Always the same constant.

 

[AlanSubie4Life]

@AlanSubie4Life.

Do you have any data on the charge constant for different line input voltages? Have you looked to see if it's the same constant for 120V input, or 208V input, or 240V or even 277V? It would honestly make some sense for it to have some sort of line input voltage relationship. Lower input voltages should have higher copper losses relative to the input power.

For example, for ~200V @ 30A, here was the result - 90 miles, 22kWh, which is about 245Wh/rmi (any deviation is rounding error presumably, which can obviously be significant). Note this is the charging constant, not the discharging "constant" (which is a LOT different, and seems to vary a little bit!).

It doesn't matter how you charge, it will always be the same relationship, at least that is my experience so far. Again, just because this is all energy added to the battery, excluding all other losses.