Which numbers should I believe?

[maverick3n1]

When you are viewing the actual power usages/averages over 5, 15, and 30 miles, is it reporting the full power usage including all accessories running (AC, Radio etc), or is it only running an amp meter on the power draw of the motors and reporting that? From my tests, I'm guessing it's the later?

 

[AlanSubie4Life]

When you are viewing the actual power usages/averages over 5, 15, and 30 miles, is it reporting the full power usage including all accessories running (AC, Radio etc), or is it only running an amp meter on the power draw of the motors and reporting that? From my tests, I'm guessing it's the later?

I believe it reports all the usage. Of what you mentioned, the HEAT (which you did not mention ) is by far the most significant.

You should see a small difference with AC use but it might be less than 1kW (depends on the temp!), which at 50mph would only increase your energy use by 20Wh/mi (1kW/(50mi/hr)). So that could be difficult to pick out. At lower speeds it would be easier to see.

*when I say “all” the usage, I think in general all these trackers probably read a bit low relative to the “real” number. But you can obviously make relative comparisons just fine. Just don’t use them to figure out how much energy your battery has.

 

[darth_vad3r]

I believe it reports all the usage. Of what you mentioned, the HEAT (which you did not mention ) is by far the most significant.

You should see a small difference with AC use but it might be less than 1kW (depends on the temp!), which at 50mph would only increase your energy use by 20Wh/mi (1kW/(50mi/hr)). So that could be difficult to pick out. At lower speeds it would be easier to see.

*when I say “all” the usage, I think in general all these trackers probably read a bit low relative to the “real” number. But you can obviously make relative comparisons just fine. Just don’t use them to figure out how much energy your battery has.

LOL, but why not?

 

[darth_vad3r]

To me [the trip meter] seems to read energy usage about 3-4% low (distance seems to be correct).

How did you arrive at this 3-4% energy usage error estimate? Are you saying xx.x miles * yyy Wh/mi = z Wh gives you a number lower than the actual energy usage? How are you calculating actual energy usage to know this?

Are you basing this on kWh put in during charging to replace usage? How do you know there isn't just an extra 3-4% of charging losses more than you think there are?

Or are you using some other method to determine this.

Thanks

 

[AlanSubie4Life]

LOL, but why not?

I mean, you can try (and it is kind of close), but it is just a number. There is no reason to think it would be accurate in an absolute sense though. It could read 200MWh/mi and it would not make any difference to anyone! All that matters is the energy available in your battery. For that we really only have the EPA documents.

The numbers seem reasonably comparable between cars but I am not even sure about that - at lower overall usage there are claims the number deviates more (reads even lower) from the true number.

 

[AlanSubie4Life]

How did you arrive at this 3-4% energy usage error estimate? Are you saying xx.x miles * yyy Wh/mi = z Wh gives you a number lower than the actual energy usage?

I’m basing it on the EPA measured total discharge numbers (which are ~78kWh or something for AWD). So that is the available energy - but that is taken until the battery is really truly done and the vehicle shuts down perhaps 5 miles below 0 (maybe 1kWh or more below 0).

So that means that there very likely is more than 75kWh available for the 310 to 0 discharge.

But if I look at how fast my rated miles click down and compare to Wh/mi reading, I consistently get ~230-235Wh/rmi. Not 242Wh/rmi. So I conclude it is reading a few % low.

I’ve never done a full 310-0 discharge (in a single sitting, and would have to be flat or uphill, net) so hard to say because nonlinearities are possible in the SoC estimation. But it always seems linear to me over quite a wide range.

I mean, it could be the battery is not the size they say but none of the EPA documents really support that.

Are you basing this on kWh put in during charging to replace usage? How do you know there isn't just an extra 3-4% of charging losses more than you think there are?

No, I am not. That would be a way to do it but as you say the charging losses would have to be known accurately. However, that can also be determined from the EPA document (you would have to know how they charged it, what rate, and duplicate the temperature - I assume it is detailed in there somewhere how many amps they charged at - I assume 48A but no idea). You’d have to believe it didn’t vary from vehicle to vehicle. But I think it would likely be within 1% so you could use this method. Maybe.

To me it is pretty “guessy”. But I do know that I have to get ~230Wh/mi (indicated) to get my rated range. But that might not be really 230Wh/mi, obviously. Anyway my range at my typical most efficient consumption of 275Wh/mi is 260 miles max. (Directly calculated from quoted numbers.)

 

[AlanSubie4Life]

If you tried to use the energy from the wall method:

Charging rate matching (to the EPA conditions) is important because aside from charger efficiency losses there is also a fixed drain of 200W (or whatever) when the car stays in idle mode during charging. So if you take longer to charge the car then those losses add up more and overall efficiency drops.

Tesla would likely use the max charge rate since it minimizes those idle mode losses (AC-DC converter probably slightly more efficient too) since they are present for the least amount of time, and makes MPGe look awesome. Anyway, at 200W, it’s 1.5kWh over the charge duration with maximum AC charge rate. That would go up to 3kWh (about) if you used 24A charging. Introduces big error.

 

[darth_vad3r]

I’m basing it on the EPA measured total discharge numbers (which are ~78kWh or something for AWD). So that is the available energy - but that is taken until the battery is really truly done and the vehicle shuts down perhaps 5 miles below 0 (maybe 1kWh or more below 0).

So that means that there very likely is more than 75kWh available for the 310 to 0 discharge.

But if I look at how fast your rated miles click down and compare to Wh/mi reading, I consistently get ~230-235Wh/mi. Not 242Wh/mi. So I conclude it is reading a few % low.

I’ve never done a full 310-0 discharge so hard to say because nonlinearities are possible in the SoC estimation. But it always seems linear to me over quite a wide range.

I mean, it could be the battery is not the size they say but none of the EPA documents really support that.



No, I am not. That would be a way to do it but as you say the charging losses would have to be known accurately. However, that can also be determined from the EPA document (you would have to know how they charged it, what rate, and duplicate the temperature - I assume it is detailed in there somewhere how many amps they charged at - I assume 48A but no idea). You’d have to believe it didn’t vary from vehicle to vehicle. But I think it would likely be within 1% so you could use this method. Maybe.

To me it is pretty “guessy”. But I do know that I have to get ~230Wh/mi (indicated) to get my rated range. But that might not be really 230Wh/mi, obviously.

OK, gotcha. So you are basing 3-4% lower based on comparing to that (theoretical) capacity from EPA document (some possible contention I believe on how that number is 'calculated' based on other numbers?)

But you are looking at Wh/rmi ('rmi' = rate miles). Not only are they 'rmi' they are 'drmi' (displayed rated miles).

I would argue that 'drmi' are scaled from actual 'rmi' because of the buffer to allow you to drive below 0.

e.g.

1. 310 drmi = ~310 miles left til shutdown
2. 0 drmi = ~9 miles left til shutdown?
3. "~ -9 drmi" = shutdown (of course 'drmi' don't display negative numbers presumably so it would get stuck at 0 for the last 9 miles driven).

If that's the case then you need to gross up your miles measured on the displayed rated miles delta when doing calculations, no?

e.g. a trip from 290 -> 200 used 90 drmi.
Say your trip meter shows 80.0 mi @ 250 Wh/mi = 20,000 Wh

Now you'd go 20,000 Wh / 90 drmi = 222 Wh/drmi, am I right so far in your method?

My contention is that you've calculated Wh/drmi, and if you want Wh/rmi you need to adjust by the bottom buffer factor, say 3% 9mi on 310.*

90drmi * (310 drmi / 301 rmi) = 92.7 rmi.

Plug in above and 20 kWh / 92.7 rmi = 216 Wh/mi ... huh, this is ... even lower ... so unless I talked myself in circles here I've not disproved anything but ADDED another 3-4% error to your error?

Hmmm... AH, error found, CORRECTION ... thanks dimensional analysis ... this is why I love using units ...
90drmi * (301 rmi / 310 drmi) = 87.4 rmi.

Plug in above and 20 kWh / 87.4 rmi = 229 Wh/rmi

So one method gives 222 Wh/drmi which could be considered as actually equivalent to 229 Wh/rmi.


*I cherry-picked 9mi as a roughly 3% number on 310, but it fits within Elon's 5-15mi "below zero" band referred to in this tweet:

Elon Musk on Twitter

• Elon Musk
  @elonmusk
  
  A Tesla has a usable reserve of 5 to 15 miles range even after the battery reads “empty”. This will not hurt the pack.

12:19 PM · May 26, 2018​

 

[darth_vad3r]

BTW, I have still not had a recent trip longer than 30-40km to satisfactorily attempt these measurements with any greater accuracy, but the couple times I attempted to capture ticks of displayed rated miles to try and count things as 1/2 km instead of full km to get better accuracy I noticed that some times the drmi (or in my case actually drkm) ticked UP 1 and then down 1, and then up *2*, *while I was sitting in PARK* getting ready to start a test run (after a short warmup of ~3km / 5minutes from home through a drive-thru ) ... so IMO using the displayed miles is very 'guessy' because they seem susceptible to fluctuations in measurements and/or in display-to-the-user-based-on-who-knows-what-criteria logic

IMO, the trip distance and trip Wh/distance numbers I trust as accurate -- or at least as fully 'honest' let's say. Perhaps not honest, but ... undressed / raw? They aren't dressed up and made pretty, they just show what they measured, e.g.:

1. By our best measurements, we think you went (a) 74.721 miles, and we think you used (b) 19,201 Wh. We're going to show your trip meter as (a) 74.7 miles and (c) 257 Wh/mi (based on (b) 19201 / (a) 74.721).
​

Now compare this to the rated miles display, I picture it going like this, e.g.:

1. We measured 4 module voltages of (a) x mV, (b) y mV, (c) z mV, and (d) w mV (or is it even worse and per-brick measurements? )
2. Based on the temperature,
3. some other criteria,
4. and yet some other criteria,
5. we've decided we *think* this means the battery is at X kWh (has to be a 'think' because we can't count ions/electrons )
6. That's actually Y % SoC, based on some upper end we determine as usable 100%, not actual 100%, but because we also have some (possibly dynamic) bottom buffer to let you go below 0,
7. we're actually going to show you Z.z % SoC, rounded to Z % and finally,
8. map Z.z to a final M 'miles' number to show you (based on an internal constant of C Wh/mi -- well, maybe it's a constant? LOL).

*whew*

I honestly think this is roughly how it works. This is what I mean when I say I 'trust' the trip meter the most. It's unfiltered IMO. The displayed rated miles (and SoC %) is going through many measurements PLUS calculations to *estimate* SoC PLUS filters to adjust how the numbers are displayed to the user. Relying on them is risky. I don't mean to rely on them for your life or to rely on them to gauge accurate SoC, I just mean relying on using them as an input into any calculations we're trying to do just seems more prone to risk because we don't have a grasp fully on steps 1-8 that result in their display to us

I mean if I can get 3 different numbers of displayed rated km over the course of 2 minutes while I am sitting in park with only the radio on and no HVAC ... then what hope do I have to trust that the rated distance displayed 'now' vs 'later' are consistent with each other in terms of what 'mood' they were in

 

[AlanSubie4Life]

But you are looking at Wh/rmi ('rmi' = rate miles). Not only are they 'rmi' they are 'drmi' (displayed rated miles).

I would argue that 'drmi' are scaled from actual 'rmi' because of the buffer to allow you to drive below 0.

I mostly agree with this. I guess let's start with some baseline assumptions which we won't argue about (for the purposes of this). Just because I am assuming them doesn't mean I think for sure I am right about them - they're just assumptions:
1) The Tesla EPA number for battery energy content (to a dead battery) is correct, and let's say it is 78kWh (I think it might have been a little higher, but whatever). And this amount is available to any user. (I think this is required by the EPA.)
2) Let's say there really are 9 EPA miles available past 0. (More recently Elon said you can drive 5 miles past zero - whether those were EPA miles, who knows... - you can look up the Tweet, I posted the link around here somewhere though.)
3) Let's say the miles indicator is linear, meaning regardless of when a unit of energy is taken out, the miles clicks down by the same amount (except for below 0 where this behavior is hidden). Of course it might not be, I agree.

(This is for the AWD.)
78kWh/310EPA mile = 252Wh/EPA mile

What is being displayed is not EPA miles, as you say (because there are only 310 EPA miles - 9 EPA miles = 301 EPA miles left above 0). I'll go with your drmi concept. So: 301 EPA miles = 310drmi

We can also calculate energy content of a drmi, using ratios:

x / 301EPA miles = 78kWh/310EPA miles

x = 75.73kWh (Just calculating the amount of energy that is available above 0 displayed rated miles.)

So, since there are 310 displayed rated miles for a full charge:
75.73kWh/310drmi = 244Wh/drmi

As far as my observations are concerned, I do a trip, a long one, and I calculate. For a recent trip, I did:

(Let me substitute here: Whi => Watt-hours (indicated). Those are the units displayed, to avoid confusion - they may not be real Wh. )

95.7mi @ 293Whi/mi. I started at 201drmi and ended at 79drmi. Started right after a prior drive, so battery was warm, and also the temp was steady.

So: 95.7mi*293Whi/mi / (201drmi - 79drmi) = 230Whi/drmi

So, 244Wh/drmi, and 230Whi/drmi => 244Wh = 230Whi (So, a 5.8% difference for the numbers used here.)

That's how I conclude it reads low, contingent upon the assumptions. Another explanation is that one of the assumptions is wrong.

BTW, I have still not had a recent trip longer than 30-40km to satisfactorily attempt these measurements with any greater accuracy, but the couple times I attempted to capture ticks of displayed rated miles to try and count things as 1/2 km instead of full km to get better accuracy I noticed that some times the drmi (or in my case actually drkm) ticked UP 1 and then down 1, and then up *2*, *

Yeah, and that's a reason to only do these calculations for a decently long trip. Also not a good idea to do it right after charging, or when temperature changes a lot, and also good to start the trip after you've just taken a trip, to avoid any significant battery temperature changes. It just should be started and ended with a steady-state battery temperature in equilibrium with the environment.

But in the end, to me the reason why this matters is that displayed rated miles is all I have to go on. I'm not going to drive past zero. But it clicks down at a very predictable rate (from what I've seen), based on my energy use. As above: I lose about one drmi for every 230Wh (indicated) of energy used. That's just what happens. I only have 310 drmi before I have to stop (I'm not going to ever go below zero if I can help it).

So, once I know that constant, I can look at my efficiency and see my remaining range, given the current drmi reading, assuming I maintain the same efficiency:
x: remaining displayed rated miles
y: current displayed efficiency for the trip

(230Whi/drmi * x drmi ) / (y Whi/mi) = x *(230/y) mi

For example, if I'm getting 270Wh/mi, and I'm at 250 displayed rmi:

250 * (230 / 270) mi = 213 mi (remaining range)

Note in this calculation it doesn't even matter what the "actual" Wh/drmi is (the 244Wh/drmi I calculated above). It just doesn't matter. All that matters is how fast those displayed rated miles click down relative to your indicated energy consumption.

(So this is why I don't think it's great to add the "drmi" in addition to "rmi" - it just adds another unit, with limited value...kind of just makes things confusing.) And in the end it doesn't matter how big your battery is, or what your "real" efficiency is. All that matters is that Tesla gave you the amount of energy that was available for the EPA test, and that the test article they used is representative of the car you have (test car wasn't cherry-picked)...

 

[AlanSubie4Life]

If you assume no error in the meter, there are only a few obvious ways to square things up (any combination thereof):

1) My measured numbers are wrong. (I lose fewer rated miles than I say for a given energy used, or nonlinearity in the SoC indicator could throw it off.)

2) Less total battery energy available than EPA test suggests.

3) More miles than 9 miles available below zero. (Seems unlikely).

Assuming my empirical number is correct (point 1 above), in the end, you’d need to get a scenario, with some combination of above factors 2 and 3, where somewhere between 71.3kWh (230Wh/drmi) and 72.9kWh (235Wh/drmi) are available between 310drmi and 0drmi.

I think the most reasonable explanation given all the factors is simply 1) there are a couple kWh available below 0drmi and 2) the meter reads low. Seems the least complicated and most likely. I don’t think it is likely that the EPA number is that far off.

But again, none of this really matters, as long as your vehicle is 100% representative of what the EPA tested. Does not matter how big your battery is, or even whether the measured capacity number in the EPA test document is correct! It only matters how far the vehicle went in the EPA test, and how much energy they had to put into it from the wall to make it go that far...

 

[NickFie]

So, when those loads are shown with the electrical loads that make the car motivate, they can mislead the driver into thinking that 35 mph is the most efficient from a mechanical point of view. That the driver will drive farther at 35 mph rather than, say 20, 25, or 30 mph. That is simply not so. The data are skewed by the presence of all drivers in the large ABRP sample, most of whom are not aiming to take a range-maximizing trip today before their next charge, and therefore includes a lot of short trips, and medium trips, that have higher-than-average startup and other electrical loads.

The non-mechanical loads are important, particularly in winter, since heat is a big drain. Optimal total consumption speed is higher than optimal speed considering only power used for propulsion.

If you choose to travel at 30 MPH in winter, instead of 45 MPH, you'll use 50% more KWH for heat to travel the same distance. Total battery drain will be much higher. When I drive to minimize KWH/mile, optimal cold weather speed is higher - heater energy saved by a faster trip helps pay for additional energy used by higher air resistance due to higher speed.

We have S 100D and LR 3D P-. Model S dual motor is more efficient than RWD; you can set air suspension to lower car at highway speed for improved aerodynamics. In my experience, efficiency drops sharply above 72-73 MPH.

P3D- seems to become much less efficient above 65-70 MPH.

I get lowest burn rate with:
- Chill mode acceleration
- AutoPilot
- Tires inflated somewhere between max on sidewall and recommended pressure posted in driver's door jamb.
- Modest use of climate control.

Optimal speed and climate control setting depends on length of trip and Supercharger availability. Last winter we left Corolla, NC (Outer Banks) headed for Philadelphia in our S 100D with battery at nearly 100%. We wanted to skip Norfolk Supercharger and stop only at Salisbury, MD Supercharger . My wife kept her coat on, we had cabin heat set very low and speed limits supported efficiency. That leg was 219 Wh/mile.

Final leg home, we had adequate charge. Cabin was warmer, speed was higher. 282 Wh/mile.

Consider TeslaFi or similar to track your journeys. You can see all key parameters, learn how to balance the factors to meet your goals.