Tesla's 85 kWh rating needs an asterisk (up to 81 kWh, with up to ~77 kWh usable)

[supratachophobia]

The 78/79 numbers are imaginary. They're the result of incomplete calibration to the real pack, combined with the fudge data in the 3 BMS. The "rapid degradation" at first isn't degradation at all.

(To be clear, the reasoning behind why the CAN-reported values don't match reality may not be, and likely isn't, nefarious... my guess, based on what I can tell from my RE work, is that they just didn't care if those values were accurate since they weren't used by the vehicle... internally, there are more accurate values that aren't reported on CAN and require active polling to retrieve.)

So if the original numbers are over-zealous, does that mean the CANbus numbers can be trusted after the electrons settle down? Say, after 5k, 10k, 20k, etc. miles?

The reason I ask, is because on a 90D, I can charge to 100% and then taking "Usable Full" from the CANbus, divided by EPA wpm of 290 (.290), go *exactly* that amount of miles in one stretch and have the dash read 0 at the end by keeping consumption at that 290 number for the whole trip.

Now, at 100%, the car (dashboard) overstates range by the 4kw difference in Nominal and Usable so I'll never get that rated number at the 290wpm consumption, but at least I'm aware of that.

 

[emir-t]

@wk057 for those of us with root access, could you elaborate where you see the more accurate numbers in the bms? Is it the BMS section of the Basics tab? or do you get shell access into the gateway and debug bms there somehow?

 

[wk057]

On the 3 the CAN-based kWh values are basically just placeholders. They'll adjust per a particular algo, but it has no bearing on the actual values used by the BMS for range calc and such. They're essentially an old code artifact on the 3 (and presumably Y).

The only way to elicit the proper values is to get the BMS itself into debug mode where it can be queried for specific information. Unfortunately the exact "how" on this will be something I'm keeping close to the vest for the time being..... since I'm pretty sure that a) Tesla will lock it down via software updates immediately upon public release, and b) to-date, I don't think anyone else has done as in-depth of RE on the BMS software, so there's a bit of value there on my end.

 

[emir-t]

On the 3 the CAN-based kWh values are basically just placeholders. They'll adjust per a particular algo, but it has no bearing on the actual values used by the BMS for range calc and such. They're essentially an old code artifact on the 3 (and presumably Y).

The only way to elicit the proper values is to get the BMS itself into debug mode where it can be queried for specific information. Unfortunately the exact "how" on this will be something I'm keeping close to the vest for the time being..... since I'm pretty sure that a) Tesla will lock it down via software updates immediately upon public release, and b) to-date, I don't think anyone else has done as in-depth of RE on the BMS software, so there's a bit of value there on my end.

How about the CAN values on S/X?

 

[trayloader]

Exactly. Can the S/X CanBus values be trusted?

 

[supratachophobia]

Exactly. Can the S/X CanBus values be trusted?

I think they can, at least if you assume you can trust the whole CAN data. The miles algorithm is 99.99% accurate if you assume EPA consumption wpm.