[Spoiler Alert + Mild Speculation] Tesla has created a monster!

[Troy]

Hey @Troy. I really appreciate your response! If the Model 3 is capable of up to 352 miles of range based on EPA estimates, why wouldn't this number show up in the UI when the LR Model 3 is fully charged?

Hi. Tesla cars in North America display the official EPA rated range or a few miles more. The rated range number is shown on the Monroney sticker and on the EPA website here below the small car icon. The official number is 310 miles. The car displays 314 miles. This is perfectly normal.

The problem with the 310 or 314 miles number is not that it is unrealistic. The problem is that it is incomparable to Model S numbers. 352 miles wouldn't be a good idea either because the EPA rated range favors city range and the Model 3 has a new motor that scored 10% better in city tests than the Model S 100D. However, in highway tests, the Model 3 80 scored only 0.7 miles less than the Model S 100D. Therefore, the most comparable range for the Model 3 80 would have been 334 miles.

Unfortunately, 334 miles happens to be a lot more than the Model S P100D's 315 mi EPA. Therefore, my guess is, Tesla voluntarily reduced the EPA rated range from 334 to 310 miles because they want to sell more Model S cars right now. By accident, the 310 miles EPA rated range number ended up being the most realistic EPA rated range number for any Tesla car. They might even use the exact same number for Typical range.

Typical range is a range unit Tesla uses outside of North America. It is something Tesla has created and it is more accurate than EPA rated range. For example, Bjorn used to have a P85. In North America, the EPA rated range was 265 miles (425 km in Canada). In Bjorn's car, it displayed 400 km (249 mi) Typical range when new. Later, Bjorn switched to a Model X P90D which has 250 mi EPA rated range (402 km). But in Europe and Asia, the same car shows 239 mi (384 km) Typical range.

I'm trying to make things easier and less confusing. Below is the latest version of real world range numbers I have calculated from dyno scores. I added the purple columns that show range after 5% degradation. On average, the battery capacity drops 5% at 40K miles and then degradation slows down. Check out this graph here. Therefore 5% seems reasonable.

By the way, earlier I said the dyno scores might not include air drag and the reason for the 0.7 multiplier must be air drag. I'm not sure about that now. I might have been wrong. I found an EPA spreadsheet that shows 3 coefficient numbers for all Tesla cars. Apparently, these can be entered to dyno settings. Maybe these include air drag. Therefore, the range numbers you see in the orange and purple cells might include aero covers.

 

[SageBrush]

Hey @MP3Mike. Yes, it makes sense for Tesla to match the range displayed on the UI to the EPA range reported on the Monroney sticker. But if there is, "behind the scenes," an additional 20-40 available miles, how will that difference manifest, if not in the UI? Will the driver be able to drive 20-40 miles before their displayed range of 310 ticks down to 309? Will the additional miles be a "failsafe" buffer after the car reaches 0 miles, similar to how an ICE car can hit "E" but still have a gallon of gas left? I'm trying to figure out how this is all going to play out. The extra miles are clearly there (based on @Troy's findings) - the question is, what is going to happen to them?

If there are an "extra" 30 miles of range over EPA in a full battery, that is about a 10% difference.

So when rated miles are 100, you can (theoretically!) drive 110 more
When there are 10 rated miles left, you can (theoretically!) drive 11 more
... ... ...

Just keep in mind that your driving habits and road/climate will likely be a lot more important than the EPA difference.

 

[omgwtfbyobbq]

By the way, earlier I said the dyno scores might not include air drag and the reason for the 0.7 multiplier must be air drag. I'm not sure about that now. I might have been wrong. I found an EPA spreadsheet that shows 3 coefficient numbers for all Tesla cars. Apparently, these can be entered to dyno settings. Maybe these include air drag. Therefore, the range numbers you see in the orange and purple cells might include aero covers.

The three term target road load coefficients should include rolling resistance, aero drag, and apparently driveline losses, which I didn't realize were included.

2. A dynamometer is used to simulate conditions of actual on-road operation. The dynamometer power absorber is adjusted so that the total "force" experienced by the vehicle is equivalent to the force measured on the road. EPA currently uses electric dynamometers with a three term force versus speed relationship characteristic of tire rolling resistance, driveline losses, and aerodynamic drag. This three-term equation is expressed as F = A + Bv + Cv2 where F is the road force, v is the vehicle speed.

https://iaspub.epa.gov/otaqpub/display_file.jsp?docid=34102&flag=1

The difference in actual power required compared to the three term road load power consumption should give a good estimate of any overhead, and at that point we should be able to come up with some estimates of 3 LR range at different speeds, less lighting/climate control. We could also do the same for the S and compare that to actual test results and data from owners to see how well the estimate lines up with real world driving.

 

[brianman]

Unfortunately, 334 miles happens to be a lot more than the Model S P100D's 315 mi EPA.

Nitpicker's corner: You're comparing the wrong vehicles. The "P" always has a range penalty. You should be comparing the 3LR with the S100D -- no ludicrous, no P, smallest wheels, and perhaps with Aero.

Arguably for it to be "fair" you should be comparing a single-motor S100 with the 3LR -- but alas Tesla hasn't offered the "biggest battery with single-motor" configuration for S.

 

[SageBrush]

The three term target road load coefficients should include rolling resistance, aero drag, and apparently driveline losses, which I didn't realize were included.


https://iaspub.epa.gov/otaqpub/display_file.jsp?docid=34102&flag=1

The difference in actual power required compared to the three term road load power consumption should give a good estimate of any overhead, and at that point we should be able to come up with some estimates of 3 LR range at different speeds, less lighting/climate control. We could also do the same for the S and compare that to actual test results and data from owners to see how well the estimate lines up with real world driving.

Any roll-down test includes drive train losses, but so does the dyno so they must do something to not count them twice.

I was thinking about this question of extrapolating an on-road power Vs speed curve and decided I could not do it yet. I don't know how to separate out fixed power use (e.g. electronics) or how to correct for the motor power curve. It may be that the latter is fairly flat for highway speeds but I don't know the value. My first guess would be to just take the absolute difference (rather than a ratio) of the Wh/mile EPA and dyno curves at 48 mph and tack that on to higher speeds. It would be a curve shift to the left.

 

[omgwtfbyobbq]

I think the coast down is how they (usually?) get the three term road load coefficient, and that's used so the dyno can accurately increase resistance as a function of speed. The part that I still don't get is how they include transmission losses in the coast down test, unless they don't consider the transmission to be part of the driveline.

The difference should just be the actual energy use less the power figures from the three term road load over time (energy) multiplied by the number of cycles (the total mileage from the test divided by the miles per cycle).

 

[SageBrush]

I think the coast down is how they (usually?) get the three term road load coefficient, and that's used so the dyno can accurately increase resistance as a function of speed. The part that I still don't get is how they include transmission losses in the coast down test, unless they don't consider the transmission to be part of the driveline.

The difference should just be the actual energy use less the power figures from the three term road load over time (energy) multiplied by the number of cycles (the total mileage from the test divided by the miles per cycle).

I'm guessing a bit here ... that the roll down test was developed for ICE vehicles and was performed by getting the car up to a set speed and then cutting the engine and letting the car coast to a stop. The forces acting to stop the car would be wind, road, tyres and the rotating parts of the drivetrain, including I think the engine unless it is disengaged from the transmission.

In an EV no juice should be flowing so the forces would be wind, tyres, road and rotating parts from the wheels to the transmission. These are all low friction parts and presumably are counted but make a small contribution.

On the dyno the inverter, motor and battery losses are recorded by the drop in battery SoC. The transmission, driveshaft, and differential also eat up battery energy and would seem to imply double counting unless the polynomial has a negative value to offset the double counting. And for what's it worth, the dyno itself is a smooth road or sorts, so a correction is needed there too.

At least that is as far as my reasoning takes me.

 

[internalaudit]

Once the 3 is available to mainstream buyers, maybe Tesla will leave temporarily increased ranges on the S and X for good?
As Hurricane Irma Impacts Florida, Tesla Sends Range-Extending OTA Update To Model S, X-Owning Evacuees

 

[seattlite2004]

Once the 3 is available to mainstream buyers, maybe Tesla will leave temporarily increased ranges on the S and X for good?
As Hurricane Irma Impacts Florida, Tesla Sends Range-Extending OTA Update To Model S, X-Owning Evacuees

Why would/should they? Maybe folks will pay for the increased range after the free trial ends. Kinda like free HBO/Showtime/Netfliex/Spotify etc...once the trial period ends...folks find they can't live w/o and they subscribe.

 

[omgwtfbyobbq]

I'm guessing a bit here ... that the roll down test was developed for ICE vehicles and was performed by getting the car up to a set speed and then cutting the engine and letting the car coast to a stop. The forces acting to stop the car would be wind, road, tyres and the rotating parts of the drivetrain, including I think the engine unless it is disengaged from the transmission.

In an EV no juice should be flowing so the forces would be wind, tyres, road and rotating parts from the wheels to the transmission. These are all low friction parts and presumably are counted but make a small contribution.

On the dyno the inverter, motor and battery losses are recorded by the drop in battery SoC. The transmission, driveshaft, and differential also eat up battery energy and would seem to imply double counting unless the polynomial has a negative value to offset the double counting. And for what's it worth, the dyno itself is a smooth road or sorts, so a correction is needed there too.

At least that is as far as my reasoning takes me.

I can't see leaving the car in gear in a coast down test. Driveline must just refer to the wheel bearings and transmission output shaft.

Document Display | NEPIS | US EPA

EV or ICE on the dyno should be the same, which is rolling, aero, and wheel bearing/transmission output shaft loads as a function of speed. The efficiency of everything else would be the difference between the road load power use and what the car actually uses.

 

[internalaudit]

Why would/should they? Maybe folks will pay for the increased range after the free trial ends. Kinda like free HBO/Showtime/Netfliex/Spotify etc...once the trial period ends...folks find they can't live w/o and they subscribe.

Or find another way to get a free trial lol.
=====
I actually might end up getting the long range depending on the reported SR winter driving range if only to make it more convenient for my wife to not have to charge up on those days of the week when she has to drive 130 miles roundtrip during winter.

I have to charge almost every common use device around the house anyway.

 

[M.V.P.]

Could we be missing one key aspect in this conversation, besides just the marketing aspect...the idea of a manufacturer built-in "reserve"? ICE manufacturers routinely under-publish their true gasoline tank capacities to allow for a hidden-to-the-driver mileage "reserve" for emergency use. This is usually takes the form of subtracting a 1/2 gallon or so from the true tank capacity and/or counting on the fuel already in the fuel lines.

Could Tesla be under-reporting their ranges to allow for this purpose as well? If so different models would require differing range "reserves" based upon weight (e.g. Model X would require a slightly higher reserve over the Model 3 to get the same emergency distance).

 

[Pollux]

Let me try to add more structure to the discussions:

The Facts:

1. The Model 3 80 (aka Model 3 Long Range RWD) scored 334 miles in the EPA test as confirmed by MPGe and energy consumption numbers.
2. Tesla voluntarily lowered Model 3 80's EPA rated range from 334 miles to 310 miles.
3. In the last two years, Tesla voluntarily lowered EPA rated range numbers 7 times for the Model S but never for the Model X.
4. To achieve the 334 miles EPA rated range with the Model S 80, Tesla used the 0.7 multiplier to convert dyno scores to city and highway range numbers.
5. Almost all EV manufacturers use the 0.7 multiplier but Tesla has never used it before the Model 3. They have always used higher multipliers for all Model S and X variants.
6. In EPA highway dyno tests, the Model 3 80 achieved 454.64 miles and the Model S 100D 455.37 miles.
7. Tesla used 19" wheels to achieve 253 miles EPA rated range for the Model S P85D.

I don't mind talking about these and going over the data again.

My speculation:

1. Tesla voluntarily lowered the Model S EPA rated range numbers to make the range gap between Model S and Model X look smaller.
2. Tesla voluntarily lowered the Model 3 EPA rated range numbers to avoid the Model 3 exceeding Model S' range.
3. The Model 3 80 will match the Model S 100D's highway range in actual road tests.
4. The Model 3 80D will exceed the Model S 100D's highway range in actual road tests.

Nothing useful to contribute to the discussion other than wanting to clearly state how much I love and appreciate what you're doing. This clear summary, especially divided by what you are labeling facts and what you are labeling speculation, touches the deep, stony place that is my ex-engineer's heart. This information is useful to me as I consider my previous experience with P85+, current experience with P100D and future choice between Model 3 base battery and Model 3 LR/80.

I am also grateful to the other thread contributors who are debating these issues with you.

Thank you.

Alan

 

[internalaudit]

For the 220 mile version with AWD and aero wheels, what would be the realistic winter range in a frigid climate averaging -5 to 5F with the cabin heater on, heated seats on and the upcoming heated steering wheel on?

So far Troy has already mentioned that 220 miles is underrated but then he too mentioned that the real driving conditions usually result in lower than EPA ratings so it's a wash and this M3 will likely have a real range of 220 miles outside of winter season.

 

[navguy12]

upcoming heated steering wheel

Please elaborate on any details you have on this. Thanks.

 

[SageBrush]

For the 220 mile version with AWD and aero wheels, what would be the realistic winter range in a frigid climate averaging -5 to 5F with the cabin heater on, heated seats on and the upcoming heated steering wheel on?

You forgot wind and road conditions like snow

 

[internalaudit]

You forgot wind and road conditions like snow

so about 40% drop in range when driving in slushy road conditions and 30% when the roads are dry?

 

[dgpcolorado]

so about 40% drop in range when driving in slushy road conditions and 30% when the roads are dry?

Those seem like reasonable numbers. Less of a range hit if you don't need to use the cabin heater. When driving in snow the speeds tend to be lower and the reduced aero drag offsets some of the range hit due to increased rolling resistance but snow/slush is still a big energy sink.

 

[wdolson]

For the 220 mile version with AWD and aero wheels, what would be the realistic winter range in a frigid climate averaging -5 to 5F with the cabin heater on, heated seats on and the upcoming heated steering wheel on?

So far Troy has already mentioned that 220 miles is underrated but then he too mentioned that the real driving conditions usually result in lower than EPA ratings so it's a wash and this M3 will likely have a real range of 220 miles outside of winter season.

Here in Portland I saw about a 15% increase in energy usage last winter. But most of what we're contending with is rain and temps that are cold, but usually above freezing. When I did have to drive in the snow, I think it did go up to about 25-30% more.

One thing that helps energy usage on the road is to pre-heat the car while it's plugged in. Crank the heat up to 80 before leaving, then turn it way down when you unplug the car. That will keep the heater off for the first part of your drive. Using the seat heaters more than the car heaters helps too, but I find heated seats get uncomfortable for me after a while and prefer cabin heat.

And I do get pretty close to rated range pretty consistently in the warmer months (late spring to early fall). Where I lose range the most is parking for any length of time. I can lose 2-4 miles of range in 1-2 hours parked. If I don't have long stops, I get just under rated range, but I live on a hill and you lose more energy climbing a hill than going down.

 

[Troy]

Hi. It looks like the European NEDC rated range is being replaced with the new WLTP rated range (Worldwide Harmonised Light Vehicle Test Procedure). There is a website about it here: WLTPfacts.eu. Apparently, all newly released vehicles have switched to this new unit last month. Existing cars have until September 2018.

I had a look and I think this change will be very interesting because apparently, the WLTP range is equal to EPA rated range that is calculated from dyno scores using the 70% multiplier. Almost all EVs except the Model S and Model X use the 70% multiplier for EPA rated range. One example would be the Chevy Bolt. It scored 236 mi EPA using the 70% multiplier.

In Europe, for some reason, the Chevy Bolt is called the Opel Ampera-e. The Opel website here says the car is expected to have 380 km (236 mi) WLTP range. Does this mean the Model 3 80 will have 310 miles WLTP range? No, because when you multiply the dyno scores by 70%, you get 334 miles for the Model 3 80 which is the range before Tesla voluntarily lowered it to 310 miles. Therefore, the Model 3 80 should score 334 miles WLTP.

Another interesting detail is the Model S 100D's range. Because the Bolt has the same range for EPA and WLTP, you might assume that the Model S 100D will have 335 miles WLTP range but I don't think so. This car uses the 75.4% multiplier to convert dyno scores to EPA rated range. If WLTP equals to 70%, then the Model S 100D will have 317 miles WLTP range. In other words, the advertised Model S 100D and Model 3 80 range numbers will be reversed.

This is a big problem for Tesla unless WLTP range also allows sandbagging. However, even then, this would make it possible to lower the Model 3 range numbers on paper but it doesn't help with the Model S numbers. Tesla really needs to switch the Model S and Model X motors from AC induction to permanent magnet motors before September 2018. Alternatively, they could replace the 18650 cells with 2170 and increase the range that way. If Tesla doesn't do anything, I think the advertised range numbers on Tesla's European design studios will look like this:

Estimated WLTP rated range numbers:

237 mi (381 km), Model 3 55
246 mi (396 km), Model 3 55D
334 mi (538 km), Model 3 80
344 mi (554 km), Model 3 80D
334 mi (538 km), Model 3 P80D

246 mi (396 km), Model S 75D
317 mi (510 km), Model S 100D
308 mi (496 km), Model S P100D

227 mi (365 km), Model X 75D
286 mi (460 km), Model X 100D
276 mi (444 km), Model X P100D

45% of Tesla sales are in Europe and Asia where Tesla.com currently shows the unrealistic NEDC range numbers. Within a year those will be replaced by WLTP range numbers.