Finally got around to finishing the comparison of 2019.5.15 to 2019.8.5. I was totally wrong about what Tesla was going to do - they increased torque all the way from ~0mph to 45mph, and probably peak HP as well (though the peak HP was not as noticeable for me, explanation below).
Gathering this was such a pain that I've lost interest in ever doing this again!
Conditions:
Stock P3D+ with original 20" PS4S with significant tread remaining:
Everything done with Creep Off, no radio, no AC/fan. I weigh 160lbs. Very minimal extraneous stuff in the vehicle.
2019.5.15: 72 degrees, 90% SoC, @Sea Level, flat as I can find (but a small amount of ripple on the surface)
0-60: 3.26s (not including rollout) With rollout: ~3.54s
Acceleration from 5mph-40mph: 35mph/1.93s = 8.11m/s^2 =
0.826g
2019.8.5: 65 degrees, 90% SoC, exactly the same location as above.
0-60: 3.17s (not including rollout) With rollout: ~3.47s
Acceleration from 5mph-40mph: 8.5m/s^2 =
0.87g
I was as careful as possible to keep the weight & test conditions as close to the same as possible. The only thing that is different was the temperature, and the battery temperature was probably pretty similar - recently charged in both cases.
So,
@SleeperService's accelerometer data is correct - the torque definitely is higher all the way up to 45mph - by 5%.
However, I found that I had to engage track mode to get more speed out of 2019.8.5. It was pulling power above 45mph without track mode - such that there was actually no difference in the speed splits from 40-60mph! This was even with 92% SoC with 2019.8.5. In fact, 2019.8.5 @92% WITHOUT track mode was
SLOWER (1.05s) than 2019.5.15@90% between 45-60mph. (1.04s) This was repeatable run after run (I got measurements of at least 5).
Track mode improved times by 0.02s. But, honestly, even with track mode on, I am not seeing the 5% increase in power, at least between 45mph and 60mph, at 90% SoC (I would expect 5% lower times for that split if there were 5% HP increase). Since the 40-45mph time was slightly faster than before (0.28s vs. 0.29s), it is possible that the peak HP (which occurred around 44mph before) is briefly ~4-5% higher (no way for me to really tell). But then it seems to converge to the same power again (for me).
45-60mph, 2019.5.15: 1.04s
45-60mph, 2019.8.5: 1.04s
Not sure what is going on with that with my car. I did not expect the power to drop off as quickly as it did with 2019.8.5. I reproduced the "lack of power"
several times - also a couple days earlier at closer to 70 degrees. Does not appear to align with the dyno data I guess; I'd have to look at those curves again. But clearly faster with 2019.8.5 (more torque) all the way up to 45mph. That's where all the winning takes place. So much winning.
Anyway, here is the "raw" data. Do note there is rounding error (which can accumulate) on the 5mph splits. But it's just 0.01 seconds here and there. I removed some splits as I went on with using the VBOX for various reasons. Ignore the 0-5mph split (that has rollout subtracted). 1-ft rollout is about 0.28 to 0.3s; hard to say exactly, and finishes just before 5mph.
2019.8.5, Track Mode:
View attachment 398452
View attachment 398453
2019.8.5, Regular Mode (I was able to reproduce this exact data several times between 92 and 90% SoC):
View attachment 398457
View attachment 398458
2019.5.15:
View attachment 398460
View attachment 398461