MASTER THREAD: V 2019.40.1.1 170kW charging, Neural Net for Auto Wipers, Auto Lane Change quicker

[Knightshade]

The S and X have the sensor right? Have the S and X train the M3...

Not for several years- no.... once they moved to AP2 cars they ditched the sensor...so no way to train AP2+ cars with sensor-installed cars.

It seems like a smarter path for training the neural wiper net would have been to include the same $2 sensor every other manufacturer uses, and have it provide feedback to the camera-based auto wiper software on scenarios it missed

Smarter? 100% for sure.

Not cheaper though. (and some have suggested the sensor costs a lot more than $2- I know "replacement" ones for other cars seem to run $50-$150 (and higher) on ebay though obviously MFG cost is less. Multiple by the million or so cars Tesla has NOT had to add that part to, it's significant.)

 

[AlanSubie4Life]

2020 SR+ Over the 240 miles mark

Posted this request for you in the other thread...would be interesting to figure out the new constant (which is presumably about 210Wh/rmi).

Take picture of energy consumption screen and battery miles gauge

 

[M3FNATIK]

Is the improvement to auto lane changes for AP or FSD? Sorry, newbie question. Just got my TM3 Stealth less than a week ago.

 

[acarney]

Is the improvement to auto lane changes for AP or FSD? Sorry, newbie question. Just got my TM3 Stealth less than a week ago.

FSD. Standard AP will not do lane changes. You manually do a lane change which deactivates AP and then you reactivate it.

 

[Knightshade]

Is the improvement to auto lane changes for AP or FSD? Sorry, newbie question. Just got my TM3 Stealth less than a week ago.

AP doesn't change lanes at all- only EAP or FSD do that.

 

[M3FNATIK]

FSD. Standard AP will not do lane changes. You manually do a lane change which deactivates AP and then you reactivate it.

Ok cool, that has been my experience so far as well so thanks for confirming.

 

[Wreckless316]

Posted this request for you in the other thread...would be interesting to figure out the new constant (which is presumably about 210Wh/rmi).

Take picture of energy consumption screen and battery miles gauge

I can actually take this picture.

But I thought we would need a long, continuous drive to figure out the discharge constant?

Or did I understand that wrong? How can we find out the discharge constant from just those three numbers that you need?

 

[AlanSubie4Life]

I can actually take this picture.

That would be awesome, as you seem to be one of the few reporting substantially more than 240 miles. (???)

But I thought we would need a long, continuous drive to figure out the discharge constant?

Or did I understand that wrong? How can we find out the discharge constant from just those three numbers that you need?

Yes, for the discharge constant, we would need a long continuous drive with a battery at steady state temperature.

However, there appears to be a pretty well-established relationship between charging constant and discharge constant. The AWD is a bit weird, but basically in all Model 3 vehicles the discharge constant is about 0.953 * ChargeConst. It seems to be a function of the buffer size in relation to full battery size: Discharge Const ~= (1-BuffkWh/NomFullkWh) * ChargeConst. The buffer size seems to be around 4.7% of the full battery charge for whatever reason on Model 3. For the AWD at least, the discharge "constant" might be as much as 6% lower than the Charge Constant...I don't know why, and I don't have access to the other types of vehicles (SR+, RWD), so no idea whether they "actually" have 6% lower discharge constants as well.

Anyway, the charging constant is most easily derived with the three numbers on that Energy Consumption screen:

ChargeConst Wh/rmi = Projected Range (mi) * RecentEfficiency (Wh/mi) / (Remaining Rated Miles (rmi))

Obviously this doesn't work if projected range is pinned at 999 miles or some corner case like that...and it's probably a bit more accurate when you have a charge of 60-90% remaining (three significant figures in all the values...).

In all the Model 3s, this formula, using the Energy Consumption screen, seems to work. The value so derived will match that seen/calculated on the Charging Screen when swapping between miles & kWh during a charging event. But it's a lot easier to capture than a charging event.

Even if the relationship to the Discharge constant is wrong, we can still compare this charge constant with what we knew we had before for the 2019 SR+, when it was 219Wh/rmi for the charge constant.

I'm expecting the new answer will be ~210Wh/rmi, but we'll see. I don't know, which is why I'd like to see the picture.

 

[Wreckless316]

That would be awesome, as you seem to be one of the few reporting substantially more than 240 miles. (???)



Yes, for the discharge constant, we would need a long continuous drive with a battery at steady state temperature.

However, there appears to be a pretty well-established relationship between charging constant and discharge constant. The AWD is a bit weird, but basically in all Model 3 vehicles the discharge constant is about 0.953 * ChargeConst. It seems to be a function of the buffer size in relation to full battery size: Discharge Const ~= (1-BuffkWh/NomFullkWh) * ChargeConst. The buffer size seems to be around 4.7% of the full battery charge for whatever reason on Model 3. For the AWD at least, the discharge "constant" might be as much as 6% lower than the Charge Constant...I don't know why, and I don't have access to the other types of vehicles (SR+, RWD), so no idea whether they "actually" have 6% lower discharge constants as well.

Anyway, the charging constant is most easily derived with the three numbers on that Energy Consumption screen:

ChargeConst Wh/rmi = Projected Range (mi) * RecentEfficiency (Wh/mi) / (Remaining Rated Miles (rmi))

Obviously this doesn't work if projected range is pinned at 999 miles or some corner case like that...and it's probably a bit more accurate when you have a charge of 60-90% remaining (three significant figures in all the values...).

In all the Model 3s, this formula, using the Energy Consumption screen, seems to work. The value so derived will match that seen/calculated on the Charging Screen when swapping between miles & kWh during a charging event. But it's a lot easier to capture than a charging event.

Even if the relationship to the Discharge constant is wrong, we can still compare this charge constant with what we knew we had before for the 2019 SR+, when it was 219Wh/rmi for the charge constant.

I'm expecting the new answer will be ~210Wh/rmi, but we'll see. I don't know, which is why I'd like to see the picture.

 

[AlanSubie4Life]

Thanks. Slightly higher than I expected but perhaps this is due to the original SR+ being derated a bit...so maybe they have reduced the buffer size for the 2020 model in addition to reducing the constant (?)...would have to be able to compare CAN bus data...but you'd expect maybe the buffer is smaller by 0.7kWh or so to have this all "line up". If that's actually the case, the discharge constants with "?" below are probably a bit wrong...would be good to get measurements at some point but not a big deal.

Looks like ~213Wh/rmi for the charging constant for your 2020 SR+, so I'll update my Lines & Constants Tracking as follows:

----

List of current constants, as of 2019.40.x, and the position of the solid line on the Energy Consumption graph:

AWD: 250Wh/rmi solid "rated" line, 245Wh/rmi charging, ~230Wh/rmi discharge (228-234Wh/mi observed), CAN bus BMS 234Wh/rmi

LR RWD: ~239Wh/rmi solid "rated" line, 234Wh/rmi charging, ~223Wh/rmi discharge (?), CAN bus BMS ? (223Wh/rmi?)

LEMR: 242Wh/rmi solid "rated" line, 237Wh/rmi charging, ~226Wh/rmi discharge (?), CAN bus BMS ? (226Wh/rmi?)

2019 SR+/SR: 224Wh/rmi solid "rated" line, 219Wh/rmi charging, ~209Wh/rmi discharge (?), CAN bus BMS ? (209Wh/rmi?)

2020 SR+: ~218Wh/rmi solid "rated" line, ~213Wh/rmi charging, ~203Wh/rmi discharge (?), CAN bus BMS ? (203Wh/rmi?)

"?" Indicates unknown/guesses. Discharge constants do have small variation due to environmental and discharge conditions and other factors, but are generally predictable enough to be useful.

Again, these constants may well CHANGE with future software updates, especially with the latest efficiency improvements, especially on 2020 models (not clear about whether prior years will change), if the "100% charge" rated range changes as a result of the software update.

My best current understanding of the formulas (subject to revision upon provided solid evidence):

Range Projection Formula on Energy Screen:

Projected Range = (Rated Miles Available * Charging Constant) / (Current Selected Efficiency Metric) (When in averaging mode)

Actual Achievable Range (to 0 rated miles, not using any reserve buffer) at a Given Consumption:

Range = Starting Rated Miles Available * Discharge Constant / (Average Wh/mi Achieved on Trip Meter)

Full Battery Capacity (CAN Bus):

Battery Full kWh = Rated Miles @ 100% * BMS Constant + Buffer Size (Wh) (Buffer Size only available via CAN read, typically 3.3-3.5kWh for AWD)

Current Battery Energy (CAN Bus):

Battery Available kWh = Rated Miles * BMS Constant + Buffer Size (Wh)

Available Energy Before 0 Rated Miles (CAN Bus):

Avail kWh Above 0 Rated Miles = Rated Miles * BMS Constant

Available Energy Before 0 Rated Miles (Trip Meter):

Trip Meter Avail kWh = Rated Miles * Discharge "Constant" (Note this may be somewhat less than battery available energy)

Energy Required to charge vehicle as shown on charging screen:

Charge Energy = Rated Miles Added * Charge Constant (Note this is considerably higher than the "used" energy)

 

[willow_hiller]

Thanks. Slightly higher than I expected but perhaps this is due to the original SR+ being derated a bit...so maybe they have reduced the buffer size for the 2020 model in addition to reducing the constant (?)...would have to be able to compare CAN bus data...but you'd expect maybe the buffer is smaller by 0.7kWh or so to have this all "line up".

Thanks for your analysis @AlanSubie4Life , at this point, we have no reason to believe that the 2020 SR+ has any more maximum kWh than the 2019, do we?

And this data still doesn't tell us whether the extra 10 miles are coming from firmware adjustments or hardware differences, right?

 

[AlanSubie4Life]

we have no reason to believe that the 2020 SR+ has any more maximum kWh than the 2019, do we?

And this data still doesn't tell us whether the extra 10 miles are coming from firmware adjustments or hardware differences, right?

Right, I doubt it has more maximum kWh. Presumably it is more efficient, but we won't really know anything until we can compare the detailed EPA test results vs. the prior year test results.

I think it will be very hard to prove long-term whether or not there are any hardware differences. I doubt Tesla will update 2018/2019 models to have more displayed range (different constant), though I could be wrong. But even if there is no change to the constant on those older vehicles, they could still improve the efficiency to match the 2020. But measuring such a change would be nearly impossible, I feel. And then there is the issue that the 2019 SR+ actually had 247 or so rated miles per the EPA test, and was voluntarily derated. So it seems like there's really been only a very small change.

 

[Wreckless316]

Right, I doubt it has more maximum kWh. Presumably it is more efficient, but we won't really know anything until we can compare the detailed EPA test results vs. the prior year test results.

I think it will be very hard to prove long-term whether or not there are any hardware differences. I doubt Tesla will update 2018/2019 models to have more displayed range (different constant), though I could be wrong. But even if there is no change to the constant on those older vehicles, they could still improve the efficiency to match the 2020. But measuring such a change would be nearly impossible, I feel. And then there is the issue that the 2019 SR+ actually had 247 or so rated miles per the EPA test, and was voluntarily derated. So it seems like there's really been only a very small change.

Just got 2019.40.2

It has increased supercharging. Guess I'm playing catch up.

 

[Apolley]

I am happy to say that my v10 update has just started to load, 25 minutes on WiFi

 

[YusufT]

Hey welcome to the forum and good to hear you got the new update! Just as a heads up for next time, you don’t have to great a new thread to contribute to existing discussions. There’s already a thread about the V10 update

 

[adrianp89]

I just got V10 as well. For some reason I didn't see a range bump.

 

[tren01t]

Got 2019.40.2 yesterday. Took 2 download attempts on WiFi. Installed in 25 minutes with no hiccups. I will report back when I can get a deeper analysis completed. 2020 SR+ White/Black 18" AP no FSD

 

[MarcG]

Got 2019.40.2 yesterday. Took 2 download attempts on WiFi. Installed in 25 minutes with no hiccups. I will report back when I can get a deeper analysis completed. 2020 SR+ White/Black 18" AP no FSD

What version were you on prior to 40.2? I just got updated from 36.2.3 to 36.2.4, which is supposedly required before jumping to 40.x without issues, but I'm not sure how true that is.

 

[Ol'Blue]

What version were you on prior to 40.2? I just got updated from 36.2.3 to 36.2.4, which is supposedly required before jumping to 40.x without issues, but I'm not sure how true that is.

pretty sure I just jumped from 36.2.2 to 40.2.1. Could be wrong however.

 

[coconutboy84]

Has anyone tried supercharging their Sr+ after the update? Really hoping to see higher than 100kw