EPA Certification Data

Discussion in 'Tech Talk' started by MelindaV, Aug 6, 2017.

Tags:
More threads by MelindaV
  1. Badback

    Valued Community Member
    M3OC Supporting Member
    Expand Collapse

    Joined:
    Apr 7, 2016
    Messages:
    1,111
    Location:
    Prior Lake, MN
    Tesla Owner:
    Reservation
    Country:
    Country
  2. JRP3

    Active Member
    Expand Collapse

    Joined:
    Jun 3, 2017
    Messages:
    31
    Location:
    NY
    Country:
    Country

    I think the Tesla blog post I linked and quoted above is probably a more accurate and relevant comparison. It explains why they might have gone with a PMAC instead of induction in the lower powered Model 3 application.
     
  3. Kizzy

    Top-Contributor
    Expand Collapse

    Joined:
    Jul 25, 2016
    Messages:
    424
    Location:
    California
    Tesla Owner:
    Reservation
    Country:
    Country
    Mm. That blog post also gives more reason to only use a single PMAC motor to start since cost is a factor (at least in materials). Now I wonder a bit about what the dual motor and performance setups will be.
     
  4. Troy

    Well-Known Member
    Expand Collapse

    Joined:
    Sep 18, 2017
    Messages:
    243
    Location:
    London, UK
    Country:
    Country
    #84 Troy, Sep 18, 2017
    Last edited: Sep 19, 2017
    @garsh, your numbers were correct. Also, I think your idea that the Model 3 55 should beat Bolt's 238 mi EPA was also correct. Unfortunately, EPA allows voluntary range reductions by car manufacturers. Therefore the 220 mi and 310 mi range numbers were less than expected.

    I don't know why Tesla voluntarily lowered the Model 3 55's EPA rated range from 239 mi (my estimate) to 220 and the Model 3 80's EPA rated range from 334 mi (this is what the calculation shows) to 310, but I do have many data points that show that this actually happened. The EPA doesn't try to hide voluntary reductions. In fact, there is an EPA document that shows all 2017 MS and MX range numbers before and after the voluntary reductions.

    However, the EPA wants to keep the data consistent. Therefore after the manufacturer decides to voluntarily lower the range, they change the city and highway range numbers to what they need to be to match the voluntarily lowered combined range. That means, it's even more difficult to spot voluntary reductions. However, here is the most interesting part: voluntary reductions don't affect MPGe numbers and those are also calculated from the original city and highway range.

    You already have the correct city and highway range numbers before the voluntary reductions. You can calculate the city MPGe and highway MPGe numbers from the 347 and 318 miles. For example, the same EPA document here says that the wall consumption was 89.41 kWh. MPGe means range per 33.7 kWh wall consumption (see the screenshot here).

    If city range is 346.528 miles per 89.404 kWh
    Then, city range is X miles per 33.7 kWh
    X= 33.7*346.528/ 89.404= 130.62 miles per 33.7 kWh = 131 MPGe for city

    Similarly, if highway range is 318.248 miles per 89.41 kWh
    Then, highway range is X miles per 33.7 kWh
    X= 33.7*318.248/ 89.41= 119.95 miles per 33.7 kWh = 120 MPGe for highway

    Because we can calculate the 131 and 120 MPGe numbers on Monroney sticker (link) from the 346.528 and 318.248 city and highway range numbers, this proves that the city and highway range numbers are correct and the combined range should be 334 miles before voluntary reductions (55% city and 45% highway).

    346.528*0.55 + 318.248*0.45= 190.59 + 143.21 = 333.8 mi combined range (EPA rated range)

    I want more people to understand this topic in detail, therefore I will post a calculation exercise for the Nissan Leaf shortly.
     
    • Informative Informative x 12
    • Like Like x 3
  5. Troy

    Well-Known Member
    Expand Collapse

    Joined:
    Sep 18, 2017
    Messages:
    243
    Location:
    London, UK
    Country:
    Country
    #85 Troy, Sep 18, 2017
    Last edited: Sep 19, 2017
    Hi, everybody. With the Model 3, we have an interesting case of sandbagging by Tesla. What I call sandbagging is officially called a voluntary reduction by the EPA. The EPA allows car manufacturers to voluntarily lower the combined range (the EPA rated range). However, they don't try to hide this. In fact, they have published yearly documents that say "Combined range voluntarily lowered". Below is a screenshot from an actual EPA document which you can download here. I found that file on this page under the first column called Datafile1.

    [​IMG]


    The screenshot is from the 2017 file. In the formula bar, you can see the original combined range before the voluntary reduction. What they are doing here is, they are adjusting the city and highway range retroactively to match the new voluntarily reduced range. In other words, the Model S P100D scored 324.3 mi EPA rated range (aka combined range where city range has 55% weight and the highway has 45%). When it scored 324.3 mi EPA, the city range was 305.9 mi but they are now reducing it at the rate the combined range was reduced. I don't think the EPA is trying to hide voluntary reductions. They are just trying to have consistent city, highway and combined range numbers where all of them are voluntarily reduced.

    What is interesting is that the MPGe numbers still use the original city and highway range. If this file didn't exist, we would still be able to calculate what the original EPA rated range should be. For example, it is possible to calculate the 324.3 mi Model S P100D EPA combined range from dyno scores. What this file does is, it confirms that the non-sandbagged range numbers we can calculate from dyno scores are correct.

    This 2017 file is very valuable because the EPA doesn't normally leave the original range numbers in the formula cells. Normally, this file should show 297.1 mi as static text and that's it. They were supposed to convert formula cells to static text. In Excel you do that by copying the cells and then right click to same cells > paste special > paste values. They forgot to do that in the 2017 file but they have done it in all previous years.

    Anyway, the story behind the sandbagging is quite amazing. I hope more people will understand all these details. I quite like talking about these. What is happening here is actually quite simple. They are calculating EPA rated range and MPGe numbers from two dyno scores. To understand this better, I will post a calculation exercise for the Nissan Leaf. The Nissan Leaf's scores are not sandbagged. Therefore the calculations are straightforward.

    After you do the Nissan Leaf calculations, if you apply the same calculation to the Model 3, the MPGe numbers you calculate will match the published numbers but the EPA rated range won't match because your calculation will show the pre voluntary reduction range.

    Summary:
    1. The Model 3 80 scored 334 miles EPA rated range but Tesla voluntarily lowered it to 310 miles. The reason is unknown.
    2. When the Model 3 scored 334 miles EPA, it was using the 0.7 multiplier but the Model S/X use higher multipliers. The reason is unknown.
     
    • Informative Informative x 14
    • Like Like x 6
  6. Bokonon

    Self-identified Teslaholic
    M3OC Supporting Member
    Expand Collapse

    Joined:
    Apr 12, 2017
    Messages:
    1,078
    Location:
    Newton, MA
    Tesla Owner:
    Reservation
    Country:
    Country
    Informative and intriguing analysis, as always... Welcome to M3OC!
     
    • Agree Agree x 1
  7. Troy

    Well-Known Member
    Expand Collapse

    Joined:
    Sep 18, 2017
    Messages:
    243
    Location:
    London, UK
    Country:
    Country
    #87 Troy, Sep 19, 2017
    Last edited: Sep 19, 2017
    OK. here is the Nissan Leaf calculation exercise: In EPA dyno tests, the Nissan Leaf 30 kWh scored these 4 numbers. Source.
    • 31.78 kWh wall consumption in city dyno test
    • 166.41 mi city score
    • 31.78 kWh wall consumption in highway dyno test
    • 136.4 mi highway score
    How did they calculate the following 3 numbers from those 4 numbers? Can you show the calculation?
    • 107 mi EPA rated range (combined range)
    • 124 MPGe city fuel economy
    • 101 MPGe highway fuel economy
    These 3 numbers are published on Monroney stickers and on the fueleconomy.gov website here. See the last screenshot below. 2016 and 2017 Nissan Leaf 30 kWh have the same scores. The Nissan Leaf uses the 0.7 multiplier to convert dyno scores to city/highway range like almost all EVs including the Model 3 but excluding the Model S/X. Then you take 55% of city range and 45% of highway range to calculate the combined range. MPGe means range per 33.7 kWh wall consumption.

    The calculation works the same way for the Model 3 and the results will show the following three numbers:
    334 mi combined range (334 was not published by the EPA but 310 mi was.)
    131 MPGe fuel economy (Published by the EPA)
    120 MPGe fuel economy (Published by the EPA)

    After that, if people are interested, we could go over the Model S P100D and the results will show 3 numbers:
    324 mi combined range (Both 324 mi and 315 mi was published by the EPA.)
    92 MPGe fuel economy (Published by the EPA)
    105 MPGe fuel economy (Published by the EPA)

    Here is an analogy to explain what's happening here: Imagine we are able to tell a person's birthdate by processing some data. The birthdate we can calculate is their actual birthdate and not necessarily what their ID shows. We know that a few people have ID's that show a different birthdate than the actual day they were born.

    When we do the calculation for Dolores, the date we calculate is different than what her ID shows. We ask the reason but she doesn't say anything. Some people suggest that maybe the calculation is wrong. Then we do the calculation for Maeve and it matches what her ID shows. Then we do it for Clementine but again the date doesn't match her ID. However, Clementine pulls out another document that shows her actual birthdate and it matches.

    In this analogy, Clementine represents the Model S P100D, Maeve the Nissan Leaf and Dolores the Model 3.

    [​IMG]



    [​IMG]


    [​IMG]
     
    • Informative Informative x 5
    • Like Like x 2
  8. Kizzy

    Top-Contributor
    Expand Collapse

    Joined:
    Jul 25, 2016
    Messages:
    424
    Location:
    California
    Tesla Owner:
    Reservation
    Country:
    Country
    Cool, cool. I'm on my phone, so can't check the file, but I'm curious if the Chevrolet Bolt's numbers were voluntarily lowered.
     
  9. Troy

    Well-Known Member
    Expand Collapse

    Joined:
    Sep 18, 2017
    Messages:
    243
    Location:
    London, UK
    Country:
    Country
    #89 Troy, Sep 19, 2017
    Last edited: Sep 19, 2017
    Hi, @Kizzy. No, the Bolt's range was not voluntarily lowered. Here are Bolt's EPA dyno test results. Source.
    • 67.420 kWh wall consumption in city dyno test
    • 364.40 mi city score
    • 66.508 kWh wall consumption in highway dyno test
    • 310.63 mi highway score
    From those 4 numbers, you can calculate the following 3 numbers shown on this page. Because Bolt's range was not voluntarily lowered, the calculation will be an exact match for all 3 numbers, just like the Leaf.
    • 238 mi EPA rated range (combined range)
    • 128 MPGe city fuel economy
    • 110 MPGe highway fuel economy
    By the way, here is a list of all voluntary reductions on that file:

    Pre, After voluntary reduction. Model
    68 mi, 58 mi. 2017 Mercedes Smart Fortwo Electric Drive Coupe (source for 68 mi)
    126 mi, 124 mi. 2017 Hyundai Ioniq Electric
    211 mi, 210 mi. 2017 Tesla Model S (60 kW-hr battery pack)
    223 mi, 218 mi. 2017 Tesla Model S AWD - 60D
    297 mi. 294 mi. 2017 Tesla Model S AWD - 90D
    273 mi. 270 mi. 2017 Tesla Model S AWD - P90D
    341 mi, 335 mi. 2017 Tesla Model S AWD - 100D
    324 mi, 315 mi. 2017 Tesla Model S AWD - P100D
     
    • Informative Informative x 7
  10. MelindaV

    ☰ > 3
    Moderator
    Expand Collapse

    Joined:
    Apr 2, 2016
    Messages:
    4,790
    Location:
    Vancouver, WA
    Tesla Owner:
    Reservation
    Country:
    Country
    #90 MelindaV, Sep 19, 2017
    Last edited: Sep 19, 2017
    Couple thoughts...
    my understanding was the .70 was the air resist multiplier. Over compensating, but typical air resistance.
    If the S and X have been using .90 have their numbers been real-world accurate? Seems they have been close.
    I’ve always heard Nissan’s epa estimates are overly optimistic - if they are also using .70, what is the miss there?(assuming their battery is still fresh)
    And lastly, anti-selling

    (Unless I misunderstood the .70/.90 and it is above/beyond the aerodynamic factor)
     
    • Like Like x 1
    • Agree Agree x 1
  11. Troy

    Well-Known Member
    Expand Collapse

    Joined:
    Sep 18, 2017
    Messages:
    243
    Location:
    London, UK
    Country:
    Country
    #91 Troy, Sep 19, 2017
    Last edited: Sep 19, 2017
    Hi, @MelindaV.

    The dyno test scores are unrealistically high for all EVs and they need a multiplier to convert them to realistic range numbers. Initially, I also assumed the reason they were so high must be the missing air drag but then I did more research and it appears air drag is included. The way they include the air drag is by letting the car coast on actual roads and they measure how much it drives while coasting from around 75 mph to 10 mph. Then they set the dyno settings so that it coasts the same distance on the dyno. I think the reason the dyno scores are high is that the test is performed at low speeds compared to real-world speeds. The EPA highway dyno test is performed at 48.3 mph and the city test at 21.2 mph average speed. See the test details tab here.

    Here is a list of all multipliers for all EVs since 2012. See the column "Multiplier (combined)". I extracted those from an EPA document. I don't know why the Model S has different multipliers for different variants but here are some details about this topic:

    1. Almost all EVs except Teslas have used the 70% multiplier
    2. Model 3 also uses 70%
    3. Model S or Model X was never 70%
    4. In 2012, the Model S multiplier was 79.6%
    5. In 2017, two of the Model S variants used the 75.40% multiplier while 5 used 73.80%
    6. In 2017, all 5 Model X variants used the 73.40% multiplier
    7. After the Model S 85 used the 79.6% in 2012, this multiplier stuck with the S85 and was never updated.
    8. All of the 79.6% multipliers are RWD Model S
    9. All of the 75.4% multipliers are AWD Model S
    10. Some of the 73.8% are RWD Model S and some are AWD Model S.
     
    • Informative Informative x 7
    • Like Like x 2
  12. TrevP

    M3OC Founder / Administrator
    Founder M3OC Supporting Member
    Expand Collapse

    Joined:
    Mar 1, 2016
    Messages:
    5,569
    Location:
    Aurora,Ontario
    Tesla Owner:
    Model X
    Country:
    Country

    First, I want to say a HUGE Welcome to @Troy to the forum. Troy has been data guy over at TMC for years so we trust his numbers and efforts!

    Thanks for the info, I have no much on my hands that I can't possibly track everything so we welcome your input and your calculators :)
     
    • Agree Agree x 10
    • Like Like x 1
  13. John

    Tech CEO
    M3OC Supporting Member
    Expand Collapse

    Joined:
    Apr 15, 2016
    Messages:
    1,076
    Location:
    California
    Tesla Owner:
    Model 3
    Country:
    Country
    While one theory may be that Tesla is "positioning" the cars versus one another to herd customers in a certain direction, that's a little cynical, no?

    Another theory that I'll put out there is that Tesla more and more adjusts to produce a more "real world" number. That might explain why they would reduce the P100D's stated range from 337 to 315 miles. Because honestly: who's going to buy a P100D and "drive it like it was on an EPA dyno"?

    I think Tesla has learned that people get bitter when they don't get at least the stated range most of the time. Once you're "selling every car you can produce," do you really need to claim every mile of range you can get, or do you start thinking ahead to how people will feel when their car never quite gets the range you could get on a dyno?
     
    • Agree Agree x 4
    • Like Like x 2
  14. Troy

    Well-Known Member
    Expand Collapse

    Joined:
    Sep 18, 2017
    Messages:
    243
    Location:
    London, UK
    Country:
    Country
    Hi, @TrevP. Thanks. I like the fact that I can edit my messages in this forum. That's my favorite feature here. Also, no dislikes. That's my second favorite feature.

    Hi, @John. The voluntary reductions before the Model 3 have been minimal. The largest was 9 miles with the Model S P100D when they reduced it from 324 to 315 miles. I'm not exactly sure why the voluntary reductions happened but they have only happened to the Model S and never to any Model X. Therefore my theory is that Tesla wanted to keep the gap between MS/MX reasonably small. Here is a table that shows the MS/MX range gap before and after voluntary reductions. If you look at the last two columns, the range gaps look more reasonable after voluntary reductions.

    [​IMG]
     
    • Like Like x 7
    • Informative Informative x 2
    • Love it! Love it! x 1
  15. SSonnentag

    Well-Known Member
    Expand Collapse

    Joined:
    Mar 30, 2017
    Messages:
    324
    Location:
    Peeples Valley, Arizona, USA
    Tesla Owner:
    Model S
    Country:
    Country
    Perhaps Tesla is compensating for the expected range of a used battery where the battery had degraded to roughly 93% of it's capacity when new.
     
    • Like Like x 2
    • Agree Agree x 1
  16. TSLAholic

    Member
    Expand Collapse

    Joined:
    Apr 8, 2016
    Messages:
    24
    Location:
    PA
    Country:
    Country
    @SSonnentag I was thinking along those same lines... As the battery experiences its initial degradation where the loss curve is rather steep, this could serve as a cushion to keep the general public from freaking out over their car seemingly losing range faster than expected. This would go hand in hand with the car being marketed to the masses as opposed to a more understanding crown of early adopters.
    Of course, the idea of anti-selling was the other thing that crossed my mind.
    I can't help but wonder how much range the dual motor car would gain on top of all this.
     
  17. SSonnentag

    Well-Known Member
    Expand Collapse

    Joined:
    Mar 30, 2017
    Messages:
    324
    Location:
    Peeples Valley, Arizona, USA
    Tesla Owner:
    Model S
    Country:
    Country
  18. SSonnentag

    Well-Known Member
    Expand Collapse

    Joined:
    Mar 30, 2017
    Messages:
    324
    Location:
    Peeples Valley, Arizona, USA
    Tesla Owner:
    Model S
    Country:
    Country
    #98 SSonnentag, Sep 19, 2017
    Last edited: Sep 20, 2017
    I wouldn't be surprised to see the dual motor option packaged with the performance option. This would be in line with keeping the options to a minimum. The price for dual motor/performance would then be quite large by my estimate.
     
    • Agree Agree x 1
  19. Kbm3

    Active Member
    Expand Collapse

    Joined:
    May 24, 2017
    Messages:
    137
    Location:
    Baton Rouge, La.
    Country:
    Country
    I wonder if the reason for the lowering is the EPA test was done with aero wheels?
     
  20. JRP3

    Active Member
    Expand Collapse

    Joined:
    Jun 3, 2017
    Messages:
    31
    Location:
    NY
    Country:
    Country

Share This Page

Model 3 Owners Club © 2018. All rights reserved.
Model 3 Owners Club (M3OC) is an enthusiast club and is not affiliated with Tesla Motors or the Tesla Motors Club Forum (TMC).
All Tesla logos are trademarks or registered trademarks of Tesla Motors.