You wrote, and i’ve quoted, that the 353mi comes only from a scalar increase from 0,7 to 0,746. Otherwise would have been 335 mi.

That’s basically correct. For the dyno test, they got 494.54 miles on the city test, and 446.78 miles on the highway test. For a weighted average of 473 miles in the test.

Even if they had not had the heat pump, simplistically they would have got about these same basic range numbers, since they do not use climate control to get them. In any case the scalar has no impact on these numbers – they’re from the dyno! It is determined by vehicle efficiency and the battery capacity. It does not impact how far the vehicle goes.

However, they used the heat pump to get better results on the cold cycle (FTP 20F) and hot cycle (SC03). These results play into the scalar, and Tesla choses to run a complex 5-cycle calculation (not that complex, just has a lot of numbers in it) from the EPA to come up with 0.746. Rather than 0.7 (which is the default value that most manufacturers use, even if they use heat pumps).

0.746*473miles = 353miles. (0.7032*473 = 333 rated miles – close enough to 335 – not sure why I said 335 tbh. Tiny difference. 0.7032 is the old scalar.)

Here’s the scalar calculation for the SR 2020 (which ended up at 0.7055 because no heat pump), as an example. For the Model 3 AWD this year it ends up at 0.746.

So, if the 353 mi is an an artifact,

It’s not an artifact. It is calculated as explained above. It’s directly related to the vehicle efficiency, and the battery capacity (and it is a function of all 5 test cycles (via the scalar), not just two cycles).

It seems you are assuming the projected range is unrelated to the rated range, in my view, this is unjustified unless you know how Tesla calculate the Energy App numbers.

I’m not assuming projected range is unrelated to the rated range. It’s directly related, via a formula explained above (and also below).

There are some assumptions here: Namely, we are assuming it works the same as it always has for Tesla. In past years after deduction and painstaking data collection, we know how Tesla calculates the Energy App numbers. And we’ve been able to confirm the results with SMT read backs, etc. You’ll have to take my word for now it that it works this way, but you can also verify it yourself with no special tools at all. (No SMT or similar.)

For my car, 2018 Performance, the charging constant is 245 Wh/rmi. This is what you get if you are supercharging (or charging) and you divide energy added by miles added in a session. Always. That’s why it’s called the charging constant. You’ll see it too. Except you’ll get 224Wh/mi.

It turns out this value is used in the Energy App as explained already with formulas:

Projected Range = (Rated Range Remaining * Charging Constant) / Recent Efficiency

Subtle point: (Rated Range Remaining * Charging Constant) is ONLY correct to use for “Energy Remaining” at 100%. For example, at 50% on the battery gauge, you don’t have half your energy left. You have 0.955*FullPackEnergy*0.5 + 0.045*FullPackEnergy = 0.5225*FullPackEnergy left (52.25%). Which is not half of FullPackEnergy. And at 0 rated miles/km, rather than 0, you have 4.5% left.

Subtle point 2: Also implies that the charging screen “energy added” is incorrect. We know that actual available energy added is 0.955*Charging Constant per mile added. This has been confirmed with SMT. And it’s also required for conservation of energy. But it’s something to keep in mind. And no, it is not correct after you include charging losses on the DC side of the charger, either. (That’s far too high (4.5%!), and the amount of DC energy loss during a charging cycle is also called out in the EPA documents for a full charge from empty to full. And it’s ~1.5kWh, closer to 1.7%. And that has nothing to do with this discrepancy anyway.). AC to DC losses of course are much higher than this (9.9% in the EPA test condition) but that’s a totally different topic.

Aside, please ignore this:
You can see from the statements above that this calculation really isn’t right at all. But it is what it is, as they like to say. I’m not going to go into the adjustments I would make to make it right, but it doesn’t matter. No one uses this to estimate range anyway – they use the Trip Planner which works differently as far as I can tell. And it’s close enough, and only gets more pessimistic (with respect to 0 pack energy left, including the buffer) as you get lower in energy. Which is what you want.

Beyond that, the “way it works” and the buffer capacity are confirmed by SMT, and through observation of the trip meter behavior being 4.5% lower “Wh/rmi” than the charging constant. So I see 230-234 Wh/rmi when observing the trip meter. On every single decent length trip, no matter whether my consumption is 150Wh/mi or 450Wh/mi. SMT is not strictly needed – it can all be deduced from observation of the trip meter. But SMT confirms it.

So, if you wonder about what I’m saying, just observe it carefully yourself. Trip meter and the battery gauge (in km!) is all you need. You don’t have to believe me.

Afaik, in EU no hard evidence of anything different of the 535 km at 100% appeared at this date, so it’s still unjustified to think of 2 different batteries.

@TimothyHW3 has stated differently. Perhaps he could post data from an EU car with the higher rated range.