May I ask how you pulled that chart? It looks like it's from Stats, but I don't see that graph anywhere ...?
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Battery degradation hidden in energy consumption?
- Thread starter theStig
- Start date
May I ask how you pulled that chart? It looks like it's from Stats, but I don't see that graph anywhere ...?
It is from Stats. In the upper right hand corner of the efficiency graph there’s a link to show it this way.
Here’s mine with a bit wider temp range:
Attachments
One more point, often not well-appreciated. Average speed is not the important metric. Since wind drag is proportional to speed^3, speed excursions above the average speed have an outsized impact on energy consumption. In order to compute the relevant speed average (at highway speeds, you would technically need to take the instantaneous speed, raise to power 3, take the overall average of the "cubed speed" over the trip, then take the cube root. Technically, i suppose this would be a RMC: "root mean cubed". None but the biggest nerds would attempt this, however... Maybe if Sheldon drove an EV...
AlanSubie4Life
Efficiency Obsessed Member
True that it is proportional to speed cubed, but you get there faster, so the energy required per mile only goes up with the square of velocity. So I think (not thinking about it too carefully) that if your objective is to figure out average Wh/mi, you should take the RMS value as the relevant value. I think. But in any case your point about excursions mattering a lot is definitely relevant even if it's "only" RMS. Of course it's only the aero loss component (which can be quite large) which is scaled by this velocity (squared).
SammichLover
Banned
Correct thatefficiency drops with v^2. Great explanation of why the math works that way.
The main point of simple average speed potentially being deceptive if there was highly variable speeds still stands, though. Doing a RMS I think brings it closer, but then you’d need enough data that you could just use the known table data of Wh/mi vs speed to do a proper weighted average.
I bought the car in March 2019 and lost about 5-10%, which is reasonable given the driving conditions/habits. The weather in March 2019 was colder than now, April 2019 the same as October 2019. Historical weather data shows that. I don’t buy into the cold weather range loss argument based on that. Not that large of a loss. In my searches, high mileage vehicles like Tesloop’s don’t mention anything about WH/m or range. Only capacity and total miles driven.
Back to degradation.
I’ll try to explain it more clearly.
Battery capacity can be hidden through software tricks.
Is the following possible to program? Control system engineers that I talked to say yes.
Scenario:
Full capacity: 386km
6 months later capacity drops to 261km
Full charge today shows 386
Energy monitored going into charger (in between electrical outlet and Tesla charger) shows full 55 kw or whatever the battery capacity is.
In reality, the battery either stops charging or slowly charges until it is full at 261 while the remainder worth of energy is purposely spent running auxiliary systems. Ie: The BMS heating/cooling the battery packs until the full capacity worth of energy is used.
(My home charging speed is 50km/h. When I preheat my car while charging, the rate drops to 7km/hr. The rest of the 43km/h energy being fed in is used for heating)
The degradation is then divided up into the driving average. Higher WH/km.
This prevents people from complaining about degradation. Puts the excess consumption on driver/environmental factors. No warranty issue. Small amounts of capacity loss are shown up to the allowable 30% on full charge to mimic randomness and to divide the energy spent on running the auxiliary systems.
In a closed software system, this seems likely possible. The only way to prove this would be to either A) analyze the code (which is locked down), or B) take out the battery on its own and test it, voiding the warranty and making claims invalid as it has now been decoupled from the cars software/hardware on which it ‘is optimized”.
Back to degradation.
I’ll try to explain it more clearly.
Battery capacity can be hidden through software tricks.
Is the following possible to program? Control system engineers that I talked to say yes.
Scenario:
Full capacity: 386km
6 months later capacity drops to 261km
Full charge today shows 386
Energy monitored going into charger (in between electrical outlet and Tesla charger) shows full 55 kw or whatever the battery capacity is.
In reality, the battery either stops charging or slowly charges until it is full at 261 while the remainder worth of energy is purposely spent running auxiliary systems. Ie: The BMS heating/cooling the battery packs until the full capacity worth of energy is used.
(My home charging speed is 50km/h. When I preheat my car while charging, the rate drops to 7km/hr. The rest of the 43km/h energy being fed in is used for heating)
The degradation is then divided up into the driving average. Higher WH/km.
This prevents people from complaining about degradation. Puts the excess consumption on driver/environmental factors. No warranty issue. Small amounts of capacity loss are shown up to the allowable 30% on full charge to mimic randomness and to divide the energy spent on running the auxiliary systems.
In a closed software system, this seems likely possible. The only way to prove this would be to either A) analyze the code (which is locked down), or B) take out the battery on its own and test it, voiding the warranty and making claims invalid as it has now been decoupled from the cars software/hardware on which it ‘is optimized”.
AlanSubie4Life
Efficiency Obsessed Member
This is the only way to know for sure, but do you have comparative historical data?
I do not believe there is any really funny business from Tesla to mask any losses. The losses are right there, obvious to see, in the form of a reduced number of rated miles at a given %.
However, regardless of what I think, documented charging data showing exactly how many kWh it took to add a given number of rated miles, and how this changes over time, is really all that we can look at, to see if Tesla is changing the definition of kWh. I can go back to some of my Chargepoint sessions and look I suppose. Even that is tricky because I have to be sure the session had no "shore power" use where I pre-warmed my car...
6/23/2019: 100% charge (extrapolated): 310 rated miles.
Chargepoint added 32.483kWh. Stats says I added 125 rated miles, 30.68kWh (245Wh/rmi)
Efficiency: 94.4%
Today: 100% charge (extrapolated): 304 rated miles.
Chargepoint added 24.257kWh. Stats says I added 92.5 rated miles, 22.63 kWh (245Wh/rmi)
Efficiency: 93.3%
So efficiencies are very similar (note the charging station was not the same so there may be meter variation)...I could process more data for a long-term average - but no real evidence that the definition of kWh has changed at all with V10 or any other update.
It's better for someone using exactly the same meter to do this comparison.
Furthermore: since Tesla charges for kWh at the Superchargers, they would likely get in trouble if they started changing their definition of what a kWh was (specifically, they would need to make it smaller if they were trying to trick us or hide loss of capacity). It probably has to be reasonably accurate to not get in trouble with the Bureau of Weights & Measures. Those kWh are shown on the charging screen and there is a direct scaling from that to trip meter kWh, and furthermore a direct relationship between that and your available battery capacity (based on your rated miles available at a full charge).
Last edited:
SammichLover
Banned
If Tesla is using their estimate of the battery charge level based on V measured to determine kWh used, stored, and charged then they effectively defacto DO define what a “kWh” is. I don’t know for sure they are doing this, or if they also have a current meter built into their circuity for a sanity check, but just basing it all around battery V would be by far the easiest and would be inherently internally consistent.
FYI in the last week Tesla sent me a message on my app that they are refunding me about $26 in SC overcharges. I’d seen some weird, tiny credits showing up on the CC linked to my Tesla account but not yet reaching the $26 total (though I haven’t checked my CC activity in last few days).
No word from them what the overcharging was or how it happened. $26 is quite a bit as I rarely use a SC, almost exclusively on long trips. That’s probably between 5-10% of my lifetime SC, over the last year. I’d have to look to be sure exactly what %. But don’t know what time window the overcharging being refunded happened in?
Last edited:
AlanSubie4Life
Efficiency Obsessed Member
Yeah, I don't know either exactly what they do. I can stick with the Chargepoint data in any case, as it is a measure of AC integrated power. The main point is that it does not look like anything has changed.
Regarding the Supercharger charges - no idea how they manage this, but I guess there are a number of reasons they might modify how they charge. Maybe they decided not to charge for overhead (heat use, etc.) while at a Supercharger? I have no idea how they bill at Superchargers - no experience with billing there myself. Supported by the Chargepoint data, it doesn't look like any of their definitions have changed.
Tons of people are reporting rated range loss anyway, so that's also supportive of the thesis that Tesla is not actually hiding anything.
SammichLover
Banned
They (and someone somewhere needs to have something of a clue) aren’t telling there customers wtf is going on, why suddenly this drop that cuts so much across miles driven, build config, and age that it breaks credulity that it could actually be true physical degradation, a change in the chemical/physical state of the cells. That’s by definition “hiding something”.
AlanSubie4Life
Efficiency Obsessed Member
That's true. I just meant they aren't hiding the available energy loss, specifically.
SammichLover
Banned
They don’t need to, they just blow it off as “no problem, normal” because the number doesn’t strictly trigger the warranty.
Unless you’re a grating Youtuber with a large following, apparently .
That is not the most accurate way to measure.
How much of the 32.483kwh and 24.257 kWh actually remained in your battery? Can you tell just by calculating energy going into the charger and the on screen Tesla numbers? How much of that was diverted to other car processes in disguise?
If your degradation was 20%
Chargepoint added 32.483kWh. Stats says I added 125 rated miles (subtract 20%), reality 100 miles.
Chargepoint added 24.257kWh. Stats says I added 92.5 rated miles (subtract 20%), reality 74 miles.
The extra lost 25 and 18.5 miles respectively vanished in what people say is the cold, wind, a/c, heat, heavy foot, temps, everything else one could think of which is subjective and cannot be proven.
The only way this can be measured is directly from the battery and or code itself.
It's not the first time an auto manufacturer would add tricks to boost sales and numbers and avoid costly warranty repairs.
VW diesel gate wasn't that long ago.
AlanSubie4Life
Efficiency Obsessed Member
Yes. On the charging screen, there is a fixed relationship between energy added and the rates miles added. It only deals with energy added to the battery.
You can also experiment with running the heat while charging - charge rate will reduce and overall charging efficiency will drop - but overall energy added and relationship to the kWh added to the battery will remain unchanged (219Wh/rmi).
You can verify this yourself. If you take a few moments to review my post here and the other referenced post you’ll see what I mean.
Really the mystery is limited to WHY there is less energy available, not whether there is, or how much. That is all very deterministic.
Similar threads
