Trip meter energy usage incorrect

[MaxK]

I recently completed a 2,000 mile trip and logged a lot of data along the way. P90DL Sig X.

The error I am seeing is repeatable. On a non-stop leg, here is the data:
Actual miles: 139
Rated miles consumed: 155.9 (rated range at start of trip minus rated range at end of trip)
The car is rated at 333 Wh/mi
Rated miles consumed * Rated Wh/mi = 51.9 kWh
Avg Energy Consumed (51.9 kWh / 139 actual miles): 374 Wh/mi

I have attached the trip meter for this leg. It shows:
Miles driven: 139
Total Energy: 48.7 kWh
Avg Energy: 350 Wh/mi

48.7 kWh consumed to go 155.9 rated miles equates to a car rating of 313 Wh/mi. A car rating of 313 Wh/mi and a range of 250 miles would equal a battery of just 78.3 kWh.

Thoughts? Is there a flaw in my calculations?

 

[jerry33]

1. Rated range is just an estimate based on a calculation of how much energy was used. It's very hard to determine the SOC of the battery when running.
2. A car rated at 333 Wh/mi does not mean every mile will use 333 Wh/mi or that any particular trip will average 333 Wh/mi. (I think perhaps I'm not understanding what your meaning here is.)
3. The only real way to determine the actual capacity is to do a 100% charge, run the car till it no longer will drive, and then fill it up again measuring how much energy it took to fill. Of course, doing this is not great for the battery, so the only time you would actually do this is to prove something in court (like the Nissan owners did a few of years ago).

 

[MaxK]

1. Rated range is just an estimate based on a calculation of how much energy was used. It's very hard to determine the SOC of the battery when running. Are you saying the battery range meter in the car can't be trusted?
2. A car rated at 333 Wh/mi does not mean every mile will use 333 Wh/mi or that any particular trip will average 333 Wh/mi. (I think perhaps I'm not understanding what your meaning here is.) Rated Wh/mi has nothing to do with actual usage.
3. The only real way to determine the actual capacity is to do a 100% charge, run the car till it no longer will drive, and then fill it up again measuring how much energy it took to fill. Of course, doing this is not great for the battery, so the only time you would actually do this is to prove something in court (like the Nissan owners did a few of years ago).

Thanks for the reply. I am not sure if you agree or disagree with the conclusion that the energy meter is incorrect. I have Xs so I am familiar with the system and 100% charging etc.

This problem is repeatable and the energy meter total usage divided by the rated miles consumed equals 313 Wh/mi on all legs of the trip. It is too repeatable to be anything other than intentional. The rated usage is not 313 Wh/mi. If it was, the battery would have started out at 78.3 kWh when new (for a 90 kWh battery). The car is actually rated at 333 Wh/mi (333 * 250 miles range when new = 83.35 kWh battery capacity when new - leaving the buffer of 6+/- kWh).

 

[jerry33]

Thanks for the reply. I am not sure if you agree or disagree with the conclusion that the energy meter is incorrect. I have Xs so I am familiar with the system and 100% charging etc.

This problem is repeatable and the energy meter total usage divided by the rated miles consumed equals 313 Wh/mi on all legs of the trip. It is too repeatable to be anything other than intentional. The rated usage is not 313 Wh/mi. If it was, the battery would have started out at 78.3 kWh when new (for a 90 kWh battery). The car is actually rated at 333 Wh/mi (333 * 250 miles range when new = 83.35 kWh battery capacity when new - leaving the buffer of 6+/- kWh).

I don't believe that I've ever had entirely repeatable sections of a trip. Now my experience is with the S85, not the P90DL, but it now has has 94K miles on it about half of which are long distance trip miles, so I have a lot of experience with it. Different trip legs can be close, but they are never equal except by coincidence. If I understand correctly, you're saying that the Wh/mile in the "since last charge" section is always 313? That seems very weird to me.

 

[MaxK]

I don't believe that I've ever had entirely repeatable sections of a trip. Now my experience is with the S85, not the P90DL, but it now has has 94K miles on it about half of which are long distance trip miles, so I have a lot of experience with it. Different trip legs can be close, but they are never equal except by coincidence. If I understand correctly, you're saying that the Wh/mile in the "since last charge" section is always 313? That seems very weird to me.

No not all, what I am saying is that the reported kWh for Trip B (reset each leg) divided by the rated miles consumed equaled 313 on every leg. In the example provided, the total kWh consumed as reported by the Trip B meter was 48.7 kWh. The actual rated miles consumed was 155.9 (starting battery meter minus ending battery meter) and confirmed by Teslafi - note the miles driven only equaled 139. The math is pretty simple from there. The total energy consumed (as reported by Trip B divided by the rated miles consumed equals 313 (this is consistent on every leg of the trip).

So the math says that the trip meter is saying that 48.7 kWh was used to travel 155.9 rated miles. Dividing the two equates to 313 Wh/mi as the rated usage of the car.

 

[jerry33]

No not all, what I am saying is that the reported kWh for Trip B (reset each leg) divided by the rated miles consumed equaled 313 on every leg. In the example provided, the total kWh consumed as reported by the Trip B meter was 48.7 kWh. The actual rated miles consumed was 155.9 (starting battery meter minus ending battery meter) and confirmed by Teslafi - note the miles driven only equaled 139. The math is pretty simple from there. The total energy consumed (as reported by Trip B divided by the rated miles consumed equals 313 (this is consistent on every leg of the trip).

So the math says that the trip meter is saying that 48.7 kWh was used to travel 155.9 rated miles. Dividing the two equates to 313 Wh/mi as the rated usage of the car.

Are you aware that the energy used when stopped isn't counted? (at least according to some folks).
But actually, I'm very confused. You travel miles, not rated miles. Rated miles are basically EPA numbers, which are only relevant if you drive like the EPA five cycle test.

 

[animorph]

I'm not going to worry about the kWh numbers. They are kind of secondary, though maybe someone knows what the primary units (kWh, rated miles, joules) the car uses are. Rated miles are a little more relatable for me. The calculation of energy in the battery, with any units, is just an estimate.

I would think the trip calculations would be doing the same rated miles calc (start - end) and multiplying by the kWh/rated mile factor to display the energy used in kWh. And I suppose it is possible they screwed up and used the wrong factor when doing trip calculations. Or they do something else like not including climate control consumption, or when the car is not moving. That just makes it a pretty fuzzy number to pin any comparisons on. Plenty of threads have noted discrepancies.

If it shows the correct rated miles for your car at 100% charge you have the right battery.

 

[BrettS]

I think part of this confusion comes from the fact that typically we don’t use range to figure out when we need to refuel a car. In an ICE car you know how much gas you have left by looking at the percentage of gas in your tank. A lot of ICE cars have a miles to empty indicator, but I think everyone who has used one knows that the number can vary a lot based on factors such as speed, wind, and hills. I don’t think anyone with an ICE car would look at an indicator that shows 100 miles to empty and plan to stop for gas in exactly 90 or 95 miles. It’s a nice tool to use to help get an idea of how far you can go, but people know you can’t trust them to be totally accurate.

The range indicator on electric cars will also vary based on the same factors... the amount of energy a car uses (whether electric or gas) can vary considerably if you’re going uphil at 80mph or downhill at 40mph and the range indicator can only estimate based on the energy usage for the past 20 or 30 miles of travel, but if the driving conditions change significantly then it will be very wrong.

I think electric car manufacturers have made the range indicators very prominent on electric cars because they want to help lessen range anxiety, but I think it also causes more confusion as people are tracking a number that can be constantly changing. A few weeks after I got my model S I set it to display the battery level in percentage and I think this feels much more natural. Once I made that change I never looked back.

 

[jerry33]

What I've found is that I hardly ever look at the range numbers. In daily driving there's never an issue because I start with a full daily charge every day. On a trip, I just check the trip graph to see how I'm doing compared to the estimate, and that's really only if there is some weather happening (wind, rain, snow). If I remember, I check when pulling into the next SC to see how I did (I forget to do this more than half of the time). So I don't worry about whether the numbers match anything or not. Most of the time I get better than Ideal miles anyway.

 

[Bigriver]

@MaxK, this is a very old thread but I believe I have the same observation as you for my Model X 100D with an October 2017 build.

Based on my past year of driving, 315 Wh/mile is what the trip meter says when I’m exactly getting the rated range. This is when my change in rated miles equals the actual miles driven for any individual trip. It is the break even point. If the car displays a higher Wh/mile than this, it has “used” more rated miles than actual miles, and if the Wh/mile is less than this value, the change in rated miles is less than actual miles (which I love to see). It is repeatable in that it is the benchmark from which everything seems to scale.

As a specific example, the last leg of a trip I took yesterday was 111.2 miles. The change in the rated miles was 136.3 miles. This is an efficiency of 81.6%. The car’s trip meter said 386 Wh/mile and 386*.816=315 Wh/mile. This is but one example; I have dozens of trips that this calculation holds true.

What I don’t understand and have no theory on how to reconcile is the 315 Wh/mile I’ve consistently observed vs the 334 Wh/mile that I believe is pretty accepted as the baseline for my car (usable 98.4 kWh/295 mile rate range). But if my car reports 334 Wh/mile on the trip meter, I will have used 1.06 (334/315) rated miles for every mile driven. What gives? Why this 6% difference? Am I missing something obvious? Do others see this?

 

[ntense]

I have the same experience. It seems like Tesla only shows the energy to get the car going, not consumptions for electronics, heat etc. If you look at the “projected” graph, then you have a more accurate picture with the correct consumption. My X consumes approximately 2,7 KWh/10km regardless of it’s just freezing or -32c according to the energy meter which does not make any sense in reality. When it was really cold a few weeks ago (-25C to -32C) I drove 1000km with a consumption of 335KWh. According to the meter, I consumed 275Kwh. During the complete trip, the projected graph was spot on.

 

[Harvey Danger]

Another bump for this thread. The OP raised a perfectly valid question, and the first few responses are just the usual blah blah about EPA-this and vampire-that. By the time anyone actually tried to answer the question, years had passed and I'm sure the OP had long disappeared.

Here's what someone should have said:

For discussion below,
- "rm" means rated-miles of remaining electrical energy as displayed by the battery management system
- "mi" means actual-miles traveled as counted by the odometer

The OP's question can be restated as "During a reasonably long uninterrupted trip, how efficiently do you have to drive to use 1 rm of energy for every 1 mi traveled?"

Just to get a couple things straight at the outset:

1) The number of rm's (rated-miles-of-remaining-range) displayed next to the battery icon in the instrument cluster when the battery is charged 100% full is determined, by the BMS, by multiplying [the number of kWh of energy that the BMS thinks can be extracted from the cells before the pack is irretrievably bricked] times [a "constant" value hard-coded in the firmware]. In the case of the OP's P90D, this value is 333 Wh/rm.

Other constants for models other than an X P90D can be looked up in this thread and this thread . For an X75D it's 320. For a Model 3 it will be something in the mid 200's . It doesn't change, it wasn't determined by the EPA, it isn't related to your driving style - it's just a number Tesla puts in a line of code.

This is just how Teslas work. Let's just all agree if you are not sure about 1) above, this conversation isn't for you.

2) The "breakeven" consumption rate, given in Wh/mi, displayed in the trip meter, that a driver would need to achieve in order for the displayed rated-miles-of-remaining-range to decrease at the same rate as actual-odometer-miles-traveled always ends up being A DIFFERENT NUMBER from the "constant" in 1) above. The OP observed 313 Wh/mi. I have seen something similar. But anyway, the fact that it isn't 333 or even *close* to 333 is the essence of the OP's question. It is the eternal mystery of the trip counter.

I think this is all about the buffer.

I believe you can get an approximate idea of what the breakeven number might be for your car by using the method below. You'll need to know three things in advance; how many rated miles are displayed on a full charge (in rm), the anti-brick-range-calculation-fudge-value, aka 'the buffer' (in Wh , usually it's 4,000 for an older Model X), and of course the rated-miles-conversion "constant" described in 1) above (in Wh/rm).

To be able to express the calculation in one compact line, we'll need to define some terms as follows:

let E represent the mysterious driving 'breakeven' constant we are trying to estimate
let C represent the hardcoded conversion constant Tesla put in the firmware
let R represent the number of rated miles displayed on a full charge
let B represent the "buffer" used by the BMS (in Wh , not kWh)

I came up with

E = C * [ (C*R - B ) / (C*R) ]

(For you ScanMyTesla people, C*R is 'nominal full pack', C*R - B is 'usable full pack')

Real world example:
-My X P90D has a hardcoded conversion "constant" of 333.333 Wh/rm.
-It currently shows 226 rm when charged full.
-The buffer value used by its BMS is reputed to be 4 kWh
I got this from first link referred to above.
Note 4kWh=4,000Wh. We're only using Wh here.

So for me
C = 333
R = 226
B = 4000

By my formula I get a 'breakeven constant' of

E = 333 * [ (333*226 - 4000) / (333*226) ]
= 333 * [ 71333 / 75333 ]
= 333 * 0.9469 (ie it's about 95% of the original 'rated consumption' value)
= 315.6 Wh/mi

To restate this result in plain English: in order for the displayed rated-miles-of-remaining-range to decrease at the same rate as actual-odometer-miles-traveled, I would have to achieve 316 Wh/mi in the trip meter. Spelling it out even plainer, If I had to take, say, a 200 mi trip in my XP90D, and I wanted my car to use exactly 200 rm of charge to go that distance, the trip meter would have to report 316 Wh/mi for the trip.

At least in theory.

In actual real-world driving, I find I have to do even better than that, more like 312, which is pretty close to what the OP found.

This might just be due to the trip meter missing some data. It's being calculated by a computer, and a busy one at that. It has to integrate a *lot* of data to come up with that average consumption rate. Maybe it drops a few, I don't know.

If anyone knows a better way of figuring out this number, lay it on me.

For now, I'd just sum up by saying if you know the hard-coded consumption value Tesla programmed into your firmware, you'll need to limit your consumption to about 95% of that for the battery icon to show actual range.

 

[mxnym]

In the case of the OP's P90D, this value is 333 Wh/rm.

Where did you get this number from? I always assumed that Tesla was using the EPA rating in the dash, but the number is apparently lower than that. Maybe it's the EU rating? I just looked up a 2016 P90D and a 2017 100D on fueleconomy.gov and it looks like the EPA rating for the P90D is 380 Wh/m and the EPA rating for the 100D is 390 Wh/m (this is for the version that had 295 mile range). Strangely, when I first got my 100D and it still had version 8 on it, the "rated" line on the energy chart seemed to be right at 337.5 Wh/m, and now, on whatever version I current, it's closer to 340 Wh/m. I have no idea if the number by the battery in the dash would also calculate differently because I changed that to % before leaving the pick-up bay on purchase day. In any case, if I was going to guess why, I'd guess that the BMS simply can't get an accuarte read on the kWH potential stored in the battery. Potentially interesting, though, I have noticed that the average shown in the IC meters immediately after it roll to 30.0 miles seems to always be off by a small number of Wh/m when compared to the 30 mile average in the energy chart. On the ohter hand, that may be explained by the fact that the energy chart obviously doesn't use a basic average function to calculate the average (since the IC meters likely would).

 

[Harvey Danger]

Where did you get this number from?

Not the EPA.

If there's one thing I don't know anything about, it's the EPA.

But I *do* know that the BMS keeps track of a "Calculated Amp-hour Capacity" value (CAC) for the whole pack. This is multiplied by the current pack voltage and reported on the CAN bus as a value denominated in kWh, which appears in apps like ScanMyTesla labeled as 'Nominal Full Pack'. This is the max potential energy the pack is deemed to be *capable* of containing if it were to be fully 100% charged. This number obviously varies over time, trending downward inexorably over the years due to "degradation", for instance, and also fluctuating on a daily basis due to temperature, mostly, and a few other factors I'm sure. The BMS also keeps track of a value called 'Nominal Remaining' which is its estimate of how many kWh could be extracted from the pack (starting from its current state) before its voltage reaches an irretrievably low ('bricked') condition.

When the battery is charged to 100% full, ie the Nominal Remaining reaches the same value as Nominal Full Pack, the otherwise complicated formula the BMS uses to determine 'Rated Range' (shown in the dashboard Instrument Cluster next to the battery icon and also reported on the CAN bus, units are miles not km) becomes very simple: The Rated Range is simply the Nominal Full Pack multiplied by A MAGICAL FRICKING CONSTANT CONVERSION RATE that is chosen by Tesla and hard-coded into the firmware.

Units-wise, I'll just spell out the obvious: [Rated-miles] * [the Watt-hours / Rated-miles conversion constant] = [Watt-hours] (or kWh if you prefer to do the conversion).

You can rearrange the three values as necessary: when you know two, you can calculate the third.

Tesla uses a lot of secret sauce to do the CAC estimation, but the rated miles reported is merely a derived value that comes from applying this simple conversion constant.

And just to repeat that important caveat, this is only the pure truth (ie this formula is a bulletproof never-fail prediction of the activity you will see on your CAN bus) when the battery pack is fully charged. (Anything below 100% state of charge involves an additional computational constant quaintly known as the "buffer" even though it is purely mathematical, not related to any literal buffer.)

So anyway let me repeat: I don't know anything about the EPA. I don't care about the EPA. It's ultimately a political body subject to all the unscientific 'sausage-making' that would imply. I don't know why Tesla chooses the MAGICAL CONSTANT it puts in each of the firmwares that it releases (which can obviously vary by OS version level to say nothing of needing to be tailored to the model year, trim level etc of each standard vehicle build). Maybe the EPA has something to do with Tesla's choice of which number to use here. Maybe Tesla's marketing department gets involved. Maybe Elon himself is presented with a list of candidate values and throws darts. I don't know.

What I do know is this number exists and it is knowable.

You asked where I learned some real-world values for this magical constant. My previous post contains two links to important threads elsewhere on TMC. Reading those threads should suffice to answer, but I'll expand a bit.

The first thread was started buy a guy who had root and could put the BMS in debug mode which exposes ALL KINDS of things that are not normally on the CAN bus. When he says he read directly from the firmware that a XP90D uses a conversion constant of 333.33 (at least in the firmwares he was looking at in 2017), um, yeah, I believe him.

The second thread is still being actively maintained by a Model 3 owner who obsesses about this number and has hit upon several non-technical methods for estimating it, none of which require ScanMyTesla or charging your car to 100%.

The simplest makes use of the fact that the supercharging display no longer shows the user a charge rate *averaged across the entire charging session*. A few years ago they switched over to the (much simpler and less deceptive) instantaneous charge rate. While supercharging a Tesla nowadays, you will see two numbers displayed, kW and mi/hr.

To be consistent with the terms I described above, I must remind this "mi/hr" number is Rated-miles per hour. Also I guess some people (I know from your other posts you are not one of them, but I'm trying to be helpful for others that may stumble across this discussion) might appreciate being reminded that kilowatts can be thought of as kilowatt-hours per hour.

You see the obvious opportunity for determining the significant digits of the conversion constant:

Wh/rm = kWh/h divided by rm/h , with a factor of 1000 applied as necessary to account for the difference between Wh and kWh.

When I'm supercharging my XP90D, I guarantee you if the display says I'm getting 80 kW it will *also* show 240 mi/hr . 100 kw, 300 mi/hr, and so on. When I divide the small one by the big one I always get .333333 ie 333.333 Wh/mi

The method you touched on, trying to determine where the "Rated" line is drawn on the energy graph, turns out to be problematic for a couple reasons. One is having to manipulate your driving style to force your recent averages to 'bracket' this line. But far more irritating is that fact this line is in the wrong place anyway. For whatever reason, Tesla's UI team consistently places this line slightly higher on the graph than would truly be accurate. Whether you are counting pixels or trying to match your driving consumption to this line or whatever, you will arrive at a value that is higher than the one actually used in firmware.

My real world example: I took pictures when my energy graph's average consumption dotted line was just *below* the "Rated" line. My consumption at the moment was averaging 339. Then I took a picture when my average consumption dotted line was just *above* the Rated line. My consumption at that moment was averaging 341. We've got it surrounded, folks! Sure enough, when the dotted line overlaid the Rated line perfectly and disappeared my average was 340. Mystery solved, right? The "Rated" rate must be 340 Wh/mi. Except we know its not. It's 333. Every other behavior of the car confirms it's 333. Every battery hacking firmware cracking bus sniffer confirms its 333. The 340 thing is some strange cosmetic fudge Tesla persists in doing on all its cars these days. The Model 3 thread I linked to above contains similar reports. If this number is all you have, your real constant is probably a couple percent lower. Better to find a supercharger and really nail it down that way.

Well, there's a novel-length treatment of a single number. Happy to take on some of those other topics you raised, but .. maybe not til later

 

[mxnym]

So for me, the takeaway here may be that Tesla treats the word "Rated" similar to the way the treat the word "Average". I suppose it's still possible that they use the European rating agencies ratings for "Rated" everywhere, even though that rating is technically for km. It only just occurs to me that I never thought to switch to metric and look at my energy chart to see if the "rated" line positioned matched with energy charts posted from over there when I saw them posted.

 

[mxnym]

So I just came across this post. If accurate, it confirms what I thought I saw, but completely unexplains the "rated" number in my X. Oh well.

 

[jerry33]

So I just came across this post. If accurate, it confirms what I thought I saw, but completely unexplains the "rated" number in my X. Oh well.

What I do is ignore the rating and use the percent remaining at the end of the trip as a guide. Ideally I try to keep the number at 20% though I have gone down to as little as 5%. I've been using this method since they included the percentage remaining. Before that I used both the rated miles and the miles shown in the energy graph. As long as the energy graph showed more than the rated, it was good. However, the percentage remaining is better (and easier) so I no longer use the comparison method.

 

[mxnym]

What I do is ignore the rating and use the percent remaining at the end of the trip as a guide. Ideally I try to keep the number at 20% though I have gone down to as little as 5%. I've been using this method since they included the percentage remaining. Before that I used both the rated miles and the miles shown in the energy graph. As long as the energy graph showed more than the rated, it was good. However, the percentage remaining is better (and easier) so I no longer use the comparison method.

Thanks for the advice, but FWIW, I was following up from a post above where I happened to state this (emphasis newly added):

Strangely, when I first got my 100D and it still had version 8 on it, the "rated" line on the energy chart seemed to be right at 337.5 Wh/m, and now, on whatever version I current, it's closer to 340 Wh/m. I have no idea if the number by the battery in the dash would also calculate differently because I changed that to % before leaving the pick-up bay on purchase day.

 

[Harvey Danger]

I honestly believe this display setting is just that, a display setting. For state of charge, the BMS is the single source of truth. To report friendly numbers on the dash, some conversions are done at the very end of the pipeline based on then-current display preferences like metric-vs-imperial or distance-vs-energy. The real numbers being fed into the pipeline don't take display preferences into account. They are computed using Teslas proprietary algorithms based on data *measured* from various sensors. Display numbers are simple conversions, friendly numbers *derived* from the telemetry being monitored by the cars various systems. I have to believe display preferences are reversible and just let you toggle between one view or another- the underlying measurements don't change and the BMS itself is not "stateful". It doesn't need to know anything about anything that has ever happened before (like whether a display setting was changed at some previous point in time) to decide what to do next.