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

[wdolson]

Not in the case of the Model 3 all 96 of the bricks are in series. The energy density of the pack is determined by how many cells are in parallel in each brick. Each brick has either 31 or 46 cells depending on if it is the standard or long range version. (i.e. the pack architecture is completely different than the Model S/X packs.)

Maybe I wasn't very clear, but we essentially said the same thing.

80% is the number most Tesla owners have reported on the threads I have seen.

I believe charging to 100% is a lot less efficient than charging to 90%. The charger does all sorts of conditioning things to the pack as you charge above 90 or 95% (different sources disagree on where conditioning starts) and that takes extra energy.

 

[Troy]

Hi, everybody. I want to add yet another example to demonstrate how EPA rated range numbers are calculated. In this example, I'm going to pick the Chevy Bolt because it uses the 70% multiplier that almost all EVs use. It appears the EPA tells the car makers that they can either use 70% to convert the dyno range to EPA rated range or they can calculate more precise multipliers of they want to.

Bolt's dyno scores:
UDDS city range: 364.4 miles (source)
HWFET highway range: 310.63 miles (source)

We use this formula to calculate EPA rated range from dyno range numbers:
EPA rated range= round(70.0%*(0.55*UDDS city range + 0.45*HWFET highway range))

= round(70.0%*(0.55*364.4 + 0.45*310.63))
= round(70.0%*(200.42 + 139.78))
= round(70.0%*340.2)
= round(238.14)= 238 miles. 238 mi is the actual EPA rated range of the Chevy Bolt.

However, the Model 3 does not use the 70% multiplier that all EVs are allowed to use. If it was using 70%, Model 3 80's range would be 334 miles. If it was using 73.8% like the current Model S versions, then it would be 352 miles EPA. Instead, the 310 mi number corresponds to 65%. Tesla is under reporting the Model 3's range number. However, I think once the Model S/X switch to permanent magnet motors, the EPA rated range numbers of all S/X versions will increase and Tesla will reduce the amount of sandbagging.

 

[insaneoctane]

...If it was using 73.8% like the current Model S versions, then it would be 352 miles EPA.

I've said it in other posts, and here it is again. If Tesla didn't sandbag the numbers, it would be their longest range car! Sandbag, sandbag, sandbag. This is very good news for prospective M3 buyers!

 

[Neoforce]

(Troy: Thanks for all of the amazing work you do digging up information on the model 3. )

If I understand Troy correctly, his speculation is that the main purpose is to keep the separation between the 3 and S/X until the high end Teslas get the new motor technology. I'm wondering if there could be a different reason.

With the model 3, Tesla has gone away from badging cars based on battery packs. Speculation is they feel the general population doesn't care or understand kWh. They just care about range. (Indeed I've seen that in friends who are not enthusiasts.). We also know that a) batteries degrade a small amount over time and b) with models like the 60/75, Tesla Has experience providing software limits on batteries.

Could the strategy of sandbagging be used to "eliminate" the battery degrading over time? So the long range model 3 is software limited at 310 miles, with a true capability of more. As the battery degrades, the software limit stays 310, because the sandbagging results in reserve battery that is more than the amount of degradation of the battery

Maybe Tesla is thinking about the casual customer. And this way, they can say in the long term that the advertised range of your car won't change over the life (or more likely within warentee) of the car. Software limits are then not for future sales, but to establish long term consistency of your expected range. (Makes it a lot easier for that casual customer.)

Not to mention that a full charge on a software limited model 3 will in itself help battery degradation. Since a "100%" charge will be more like 95-97%

Any thoughts if that might be part of the strategy in sandbagging the range of the Model 3?

 

[omgwtfbyobbq]

I believe the test results submitted to the EPA have to match what Tesla allows a driver to use/access. If they change how much energy a driver has access to, they have to submit a new set of test results to the EPA. They could do this, maybe with the intent of having the first employee/investor cars act as accelerated beta testers, but if they do end up doing it, they'll need to submit another CSI report to the EPA prior to selling the version with less battery capacity available for use. I don't think they're going to do that, but it's possible.

Keep in mind, the CSI figures, and fuel economy/range figures a manufacturer submits don't have to match up. My feeling is the EPA won't step in unless a manufacturer is really overstating their fuel economy estimates (>10% worse actual range/mpg).

Toyota really sand-bagged the crap out of the first gen Prius Plug-in.

The two 2010 Prius test vehicles averaged 43 mpg and 45.6 mpg over ~150k miles.

2010 Toyota Prius | Advanced Vehicle Testing Activity

The four 2013 Prius Plug-in test vehicles averaged 49 mpg to 52 mpg in charge-sustaining mode over ~150k miles.

2013 Toyota Prius PHEV | Advanced Vehicle Testing Activity

That's 10% to 20% better mileage from the plug-in, but Toyota gives both cars the same mpg rating of 50 combined.

 

[saltsman]

Official Monroney Sticker (From Zoomit via Electrek):

131 MPGe city
120 MPGe highway,
126 MPGe combined

 

[insaneoctane]

I think it is interesting to ask yourself "what would I do if I were in Elon's shoes?" You asked your design team for the moon and they over delivered. Now you have differentiation problem with your premium models. You also have a back log problem that is well into 2018. What would you do? I think I would be very quiet about my new car. I might even "anti sell" it

 

[Troy]

Hi. I have more interesting findings based on the Monroney sticker. Let's go back to the Chevy Bolt example to explain these. Skip to Step 3 if you have read my previous message.

Step 1: Dyno tests
The first step of the process is to put the car on a dyno and do 2 tests. Here are Bolt's dyno scores once again:

City dyno score: 364.4 mi (source)
Highway dyno range: 310.6 mi (source)

Step 2: City, highway and combined range
The second step is to convert dyno scores to realistic range numbers because during the dyno test there is no air drag because the car doesn't move anywhere while it's on top of a dyno. To add the air drag effect, the dyno scores are multiplied by 0.7 if the car manufacturer doesn't bother calculating more realistic multipliers. That means we get these range numbers from dyno scores:

City range: 364.4 mi * 0.7= 255.1 mi
Highway range: 310.6 mi *0.7= 217.4 mi

The combined range is calculated by taking 55% of city range and 45% of highway range.
Combined range (EPA rated range): 255.1 mi *0.55 + 217.4 mi * 0.45= 140.305 + 97.83= 238.14

Step 3: Fuel economy
To calculate fuel economy, EPA assumes that 1 gallon of gasoline equals to 33.7 kWh (see the screenshot below). That means we need to find out how many miles per 33.7 kWh the car can drive. For this calculation, the city and highway range numbers in Step 2 and the recharge energy drawn from the wall outlet during the dyno tests will be used. The recharge energy numbers are from the same EPA file that shows the dyno scores.

Recharge energy after city dyno test= 67.42 kWh (source)
City range : 364.4 mi * 0.7= 255.1 mi

Highway range: 310.6 mi *0.7= 217.4 mi
Recharge energy after highway dyno test= 66.51 kWh (source)

OK, so, we know that the city range is 255.1 miles and requires 67.42 kWh energy from the wall outlet. What would be the city range per 33.7 kWh (per gallon of gasoline equivalent)?

City fuel economy= 33.7 kWh/67.42 kWh * 255.1 mi city range= 127.5 miles per 33.7 kWh= 128 MPGe
Highway fuel economy = 33.7 kWh/66.51 kWh * 217.4 mi city range= 110.2 miles per 33.7 kWh = 110 MPGe

You can see the same two numbers on the EPA website here. The MPGe numbers we just calculated and the numbers EPA published are a match. This confirms that we correctly identified how the fuel economy numbers are calculated. Here is a screenshot:

Model 3 80 (aka Model 3 LR RWD)
The question is, what does this tell us about the Model 3? We now have the MPGe numbers from the Monroney sticker and we have the dyno scores and energy consumption numbers from the EPA document and we know how the calculation works. So, let's reverse the calculation for the Model 3.

City fuel economy= 131 MPGe (source)
Recharge energy after city dyno test= 89.4 (source)

Highway fuel economy = 120 MPGe (source)
Recharge energy after highway dyno test= 89.4 (source)

If city range is 131 mi per 33.7 kWh recharge energy,
then city range is X mi per 89.4 kWh recharge energy.

City range= 131 mi * 89.4 kWh/33.7 kWh= 347.5 mi
Highway range= 120 mi * 89.4 kWh/33.7 kWh= 318.3 mi

Combined range (EPA rated range)= 347.5 mi * 0.55 + 318.3 mi * 0.45= 191.125+143.235= 334 miles

Conclusion:
Even when using the 0.7 multiplier that all EVs use, the Model 80 achieves 334 miles EPA rated range as suspected based on MPGe numbers on the Monroney sticker. For some reason, Tesla didn't like that and decided to voluntarily lower the range number like they have done many times. My guess is they want to continue selling more Model S cars while ramping up Model 3 production but if the Model 80 achieves 334 miles EPA rated range, the Model 80D would achieve 347 miles EPA rated range. That doesn't look good for the Model S 100D with 335 mi EPA, at least until they switch to the newer motors.

However, Tesla has never used the 0.7 multiplier for any Model S or X. The lowest they ever used was 0.733. Here is a screenshot that shows the whole list. In other words, this appears to be a case of double sandbagging. First, they lowered the multiplier to 0.7 which is the default one every car maker uses. But that wasn't enough and they voluntarily lowered the EPA rated range.

This shouldn't be a surprise because Tesla has done the same thing to the Model S 7 times during the last 2 years to make the gap between S and X look smaller. Below is a screenshot from an actual EPA document which you can download here. The file shows that the Model S P100D achieved 324 miles EPA rated range but Tesla voluntarily lowered it to 315 miles. I guess they didn't like the 324 miles number because the Model X P100D has 289 mi EPA and the gap would look too large. As a result, we have inconsistent EPA rated miles in different Tesla cars. 1 EPA rated mile in a Tesla should be equal to 1 EPA rated mile in another Tesla but it isn't.

 

[SageBrush]

Step 2: City, highway and combined range
The second step is to convert dyno scores to realistic range numbers because during the dyno test there is no air drag because the car doesn't move anywhere while it's on top of a dyno. To add the air drag effect, the dyno scores are multiplied by 0.7 if the car manufacturer doesn't bother calculating more realistic multipliers. That means we get these range numbers from dyno scores:

The 0.7 fudge has to do with normalizing the dyno results to normative American speeds, braking behavior, and accessory use.

Aero and road frictions have been programmed in already with the roll-down tests.

 

[kwest2]

So what is the practical implication of the "actual range" being grated than the "EPA range"? The screen will show 0mi remaining, but you should actually have +24 (=334 - 310) miles remaining?

Do we know anything about the standard battery yet, e.g. if it will have better performance than the 220 mi range figure?

Thanks for all the work.

 

[insaneoctane]

@Troy,
Great work, thanks for showing the math. I would like to point out that your statement:
"but Tesla didn't like that because the Model X P100D has 289 mi EPA and the gap would look too large"

was opinion. I think you have a good hypothesis, but we really don't know conclusively why they did it. Your message was chock-full of facts, I just wanted to keep fact from opinion. Thanks again!

 

[scaesare]

More specifically, the voltage comes from how many cells are in series. The Model S/X modules have 6 cell units in series and then there are a bunch of these in parallel. This makes each module 25V. They put 16 modules in series on the large packs (400V) and 14 in series on the small packs (350V).

Tesla can arrange the cells any way it wants, but if they get 96 cells in series in how ever many modules, they will get a pack with 400V. The energy density of the pack is then determined by how many 96 cell strings are in parallel.

When the original small pack was designed, they probably weren't thinking of supercharging and 350V was fine. The only time it's an issue is when fast charging DC. Now that supercharging is a key component of the Tesla brand, having the small and large pack Model 3s charge at the same, high rate is a good selling point.

Overall charge rate (as long as within the voltage & current limits the supercharger is capable of supplying), is still going to be limited by the overall pack charge rate.

If both packs have the same nominal ~350V (~400 peak) and the large pack is 75kW, that means it has an effective Ah capacity of 214Ah. Charging at 120 KW would mean accepting a 1.6 C rate, which is similar to similar sized packs on the S.

If the small pack is 55KW, that implies a Ah capacity of 157Ah. Charging at the same 120KW rate would imply having to accept a charge rate of 2.2C, something not likely.

 

[insaneoctane]

So what is the practical implication of the "actual range" being grated than the "EPA range"? The screen will show 0mi remaining, but you should actually have +24 (=334 - 310) miles remaining?

Do we know anything about the standard battery yet, e.g. if it will have better performance than the 220 mi range figure?

Thanks for all the work.

We know that the standard 50 kWh Model 3 battery pack is made of 2,976 cells and the 74 kWh ‘long range’ battery pack consists of 4416 cells. 2,976/4,416 = 0.674. The standard battery model should have at least 67.4% of the range of the LR and ought to have more due only to the approximately 300 lbs of weight savings. 334 miles X 67.4% = 225 miles

 

[SageBrush]

Overall charge rate (as long as within the voltage & current limits the supercharger is capable of supplying), is still going to be limited by the overall pack charge rate.

If both packs have the same nominal ~350V (~400 peak) and the large pack is 75kW, that means it has an effective Ah capacity of 214Ah. Charging at 120 KW would mean accepting a 1.6 C rate, which is similar to similar sized packs on the S.

If the small pack is 55KW, that implies a Ah capacity of 157Ah. Charging at the same 120KW rate would imply having to accept a charge rate of 2.2C, something not likely.

Agreed.
If I can manage 70 kW throughput with the SR charging from 20-80% I'll be delighted

 

[Troy]

Hi, @SageBrush. Yesterday, I have heard the same argument on Reddit here about the 0.7 multiplier but I disagree with that argument. First, let's have a look at this quote from this EPA document.

Optional equipment that increases aerodynamic drag and which has a projected installation rate of over 33 percent on a carline in a test group must be installed on the test vehicle or accounted for if analytical methods are employed to determine the road-load force specification. This applies only to optional equipment or features which affect aerodynamic drag (e.g. roof rack). Such optional equipment that has a projected installation rate of less than 33

In other words, car manufacturers can use some analytical methods to account for air drag. Now let's look at the other quote:

When a vehicle is tested for emissions and fuel economy on a chassis dynamometer, the load from aerodynamic drag, friction, and tire losses associated with road operation must be simulated. Since the introduction of the SFTP requirements, EPA has required manufacturers to supply representative road-load forces for vehicles at speeds between 15 km/hr (9.3 mph) and 115 km/hr (71.5 mph) for emissions and fuel economy testing (reference 40 CFR §86.129-00 and § 600.111-08).

The person who mentioned this quote on Reddit understood "must be simulated" as changing dyno settings to account for air drag. I understood it as "needs to be added on top of dyno scores". If you look at the next bold section it says EPA is now asking for those numbers so they can simulate air drag more accurately in calculations.

Anyway, this discussion is a bit academic because we know that Tesla used the 0.7 multiplier because without the 0.7 you can't get to the MPGe numbers from 89.4 kWh wall consumption.

 

[McRat]

I will stick with my hunch that the 2018 M3LR with the 19" tires and no PUP will be the longest range EV when it hits the streets.
We will see.

 

[Troy]

@insaneoctane,
Thanks. I was changing the wording while you wrote your message.

@kwest2,
There is a Consumer Reports video here that shows the real-world range of a Model S 75D at 65 mph constant speed. It achieved 235 miles. We also know that the highway dyno score of the Model S 75D is 358.49 miles based on page 19 of this document. In addition, we know the dyno scores of the Model 3 80 and cell counts of 80 kWh and 55 kWh Model 3 packs. Putting all this data together, I came up with the following real-world range numbers at 65, 70 and 75 mph speeds.

Earlier, I made the same calculations using EPA rated range numbers but I believe those were less accurate because the Model 3 has much better efficiency in the city because of the new motors. For example, the Model 3 80 achieves 10% more range in the city than the Model S 100D, but on the highway, they have the same range. So, I decided to focus on highway range and completely ignore city range. Therefore I switched to dyno highway scores for more accurate calculations. The range numbers you see here are without aero wheel covers.

 

[SageBrush]

^^ I agree -- fairly academic to most.

You could probably go back in history to earlier fudge factors that the EPA has used. I think the original 2-cycle was pretty much reported as is. If I am right then the 0.7 fudge factor was not present in the beginning when the EPA was not trying to leave no American driver behind.

 

[Troy]

Hi, @SageBrush. What is interesting is that the multiplier was 79.6 in 2012 and 2013. You can see it here. But in the following years, it dropped to 73.8% most of the time. In other words, EPA rated range was becoming more and more accurate over time but unfortunately, the Model 3 has messed up that trend by achieving 10% better scores than the Model S 100D in city dyno tests. It is a shame that city range has 55% weight. It should have 0% weight and only the highway range should matter. Both the Model 3 80 and the Model S 100D achieve ~298 mi range at 65 mph. What does it matter if one of them has 10% more range at lower speeds? It shouldn't matter in real life because people typically don't drive that much in their home town.

However, interestingly the Model 3 is the perfect car to break hypermiling records. In other words, whatever hypermiling scores the Model S has achieved so far, even the RWD Model 3 80 will exceed that. Tesla needs to hurry up switching the S/X to the new motors.

 

[MP3Mike]

Both the Model 3 80 and the Model S 100D achieve ~298 mi range at 65 mph. What does it matter if one of them has 10% more range at lower speeds? It shouldn't matter in real life because people typically don't drive that much in their home town.

I don't know where you live but here I rarely get to above 45 MPH because of traffic... My average daily speed is 32 MPH. (And that accounts for ~90% of my driving.)