Display Error or Time to Rebalance?

[aesculus]

Charged to 98% the other day since we were eating lunch at a slow restaurant.

This should have resulted in 98% of my 90 kWh hour battery or 88.2 kWh.
A few miles down the road I captured a screenshot of the trip and here is what I got:

Doing the math: 88.2 kWh - 3.8 kWh = 84.4 kWh. 84.4/90 / 100 = 93.8%. Close enough since there could be rounding going on under the covers. So the image matches the calculation but it is a small amount of discharge.

Now traveling more to my next stop (91.8 miles) I used an additional 27.7 kWh of energy.

So the math: Started with 98% (88.2 kWh) and removed 27.2 kWh driving to this point. 88.2 - 27.2 = 61 kWh remaining.

By percent of total battery this should be 61 kWh/90 kWh * 100 = 68%. That is not a rounding error. That's about 11% too low.

I am highly suspect that the 61% remaining is the same as the 61kWh that should be left but that could be a coincidence.

So the question is: Did Tesla screw up the Trip display in V10 or do I need to rebalance my car or is something else a foot.

BTW this is my 90% range after charging:

In my cars context it's EPA range was 257 so 257 * .9 = 231. I am now getting 217, a loss of 14 miles or about 6%.

So even taking the 6% off the top I should have 66% remaining by my math.

 

[mxnym]

What evidence do you have that your "90 kWh battery" ever did actually measure exactly 90 kWh or more? Beyond that, what evidence do you have that your "90 kWh battery" now measures exactly 90 kWh after degradation? My understanding from reading here on these forums and elsewhere is that the "85" and/or "90" batteries degraded worse than the "75" and "100" batteries do (that's not even considering the even newer "SR" and "LR" batteries that may theoretically have different kWh capacities from one given month to the next). Further, doing equally rough math with less assumptions while working from big numbers to small ones, I get these results:
98 percent full charge - 61 percent end charge = 37%
27.2 kWh used / .37 percent used = 73.5 kWh battery
3.8 kWh used / 73.5 kWh battery = 5%
93 percent end charge + 5 percent used = 98% starting charge

Unfortunately with no experience or policy knowledge, I don't know what to tell you to do about that. You could try to learn how to check the API and see what your battery's current kWh capacity actually is so that you know if it's anywhere near that bad, but I don't know whether or not you'd be able to get Tesla to do anything about it even if it is.

 

[ajdelange]

Let's start with the title. Rebalancing goes on all the time. There is nothing you can do about it.

Second, note that percentages are quantized to one percent. Thus 36.501 and 37.499 would both be displayed as 37%. We have to be careful about drawing conclusions based on percentages - especially differences in percentages.

Third, given the above (i.e. that we know conclusions drawn from percentage differences are noisy) we should only draw conclusions from an ensemble of observations rather than single observations or pairs of observations.

Turning to your data: you used 3.8 kWh and the battery percentage dropped by 5 %. As noted above the actual drop could have been as little as 3% or as much as 7%. Accepting 5, however, this implies that the available full capacity of your battery is 3.8/.05 = 76 kwh.

Overall you used 27.2 kWh resulting in a change in batter of 37%. Thus implies that the full capacity of your battery is 27.2/0.37 = 73.5 kWh.

If you use the same reasoning with charge data you will get a bigger number. That's because it takes more energy to charge from 70 to 80% than you can get out. Some energy gets lost to battery internal heating in both directions. All of this relies on our assumptions as to what the numbers on the display really mean and that is by no means certain. % of what? Only Tesla knows (but there are a fair number of people here who think they know).

The final observation is that you went 91.8 miles on 37% charge at a drain rate of 296 Wh/mi. That's 91.8/37 = 2.5 miles/%. I get about 3 mi/% at about the same drain rate in an X100. I assume that you have an X90 and while I don't know the actual relative size of the batteries I assume it is about 90% of the X100 so you should get about 0.9*3 = 2.7 mi/%. That's only a bit off and given the accuracy considerations mentioned above I'd say it's close enough for government work. Thus, assuming that my X is performing normally yours appears to be performing normally too.

 

[Rocky_H]

This should have resulted in 98% of my 90 kWh hour battery or 88.2 kWh.

No. Everything is based on this incorrect assumption.
Let's suppose that you got to use 90 kWh out of your battery. Do you know what would happen? You would never be able to use that battery again, because you would have destroyed it. Lithium ion batteries do not tolerate going completely dead with 0 volts. They cannot be recharged again.

So, to helpfully prevent us owners from destroying our batteries, there is an "anti-bricking" reserve that is held in the bottom of the battery with some kWh still stored in it, where the cars will shut off and disconnect with that little bit of energy remaining so that they can be recovered and recharged. So the usable energy in your battery is not 90 kWh, and you can't multiply 98% times that. It is 80-some, and that would be what you could multiply by 98%.

 

[Mewtwo]

OP please read this:
Tesla’s hacked Battery Management System exposes the real usable capacity of its battery packs - Electrek
90D/P90D – ~85.8 kWh total capacity, 81.8 kWh usable

Granted this is an old article and Tesla may have tweaked its battery pack since, but you get the idea.

 

[aesculus]

Thanks for the comments. I will look into this a bit more but this is the first time I have seen such a big diversion.

So, to helpfully prevent us owners from destroying our batteries, there is an "anti-bricking" reserve that is held in the bottom of the battery with some kWh still stored in it, where the cars will shut off and disconnect with that little bit of energy remaining so that they can be recovered and recharged. So the usable energy in your battery is not 90 kWh, and you can't multiply 98% times that. It is 80-some, and that would be what you could multiply by 98%.

This is still a mystery too. Some report that they can get their display down to 0% remaining while others have run out at 5% and even one report had a user go dead at 8%.

I always assumed that the hidden energy would not display but that may be car dependant. So it really could be that in my car at least the display is actually showing my available energy which as you say is 80-some kWh hours which at one time decided that could be as low as 82 kWh.

So using 82 kWh and charging it to 98% would give me 80.36 kWh. Subtracting my 27.2 from that gives 54.8. 54.8 / 82 is 67% so I would be back to the same issue again.

But if I take the ratio of 54.8 kWh / 90 kWh that equals 61% rounding.

So that seems to be what is occurring here.

 

[aesculus]

OP please read this:
Tesla’s hacked Battery Management System exposes the real usable capacity of its battery packs - Electrek
90D/P90D – ~85.8 kWh total capacity, 81.8 kWh usable

Granted this is an old article and Tesla may have tweaked its battery pack since, but you get the idea.

Thanks for providing this article. I remember investigating this before as my last post admitted I knew that real usable capacity of my 90 was really 82% (81.8). What I guess I was missing was how Tesla was using that in the display calculations.

Even worse when I went to bookmark this article I discovered that I had already done that years ago.

 

[Rocky_H]

This is still a mystery too. Some report that they can get their display down to 0% remaining while others have run out at 5% and even one report had a user go dead at 8%.

You are including every situation in here, though, even defects. There is the normal behavior, and the broken behavior. Regular, normal, properly functioning behavior is shutting off at 0%. Please realize that plus or minus about 1 or 2% is still considered basically 0%, because measurements just aren't that precise when it is that low. The situations where people's batteries have shut off with things like 5% or higher, are almost always a defect in one of the modules in the battery and result in replacement or repair, so that is different.

I always assumed that the hidden energy would not display but that may be car dependant.

No, the anti-bricking reserve is never displayed in any of the cars ever. At 0% or whenever the car says "shutting down now", it has reached that threshold level, where the reserve is still in there, but nothing above that, or effectively the same, where there may have been just barely a little more above it, but a slight voltage dip in one of the modules caused it to briefly dip below that threshold, so it triggered the same shutdown.

 

[mxnym]

Thanks for providing this article. I remember investigating this before as my last post admitted I knew that real usable capacity of my 90 was really 82% (81.8). What I guess I was missing was how Tesla was using that in the display calculations.

Even worse when I went to bookmark this article I discovered that I had already done that years ago.

Another thing I forgot to mention and don't have time to search for is that some batteries were further capacity limited by Tesla after some battery fires. I'm not sure if the packs in question included the "90" batteries or not, but if they did, that would be more likely to explain any difference you are suddenly observing. However, that change was some time ago (probably measurable in months), so you'd have to find better information in order to determine whether or not it correlates with your observation.

 

[ajdelange]

The EPA rating for the X90D is 262.7 mi. That's 2.63 mi/%. He's now getting 2.5 mi/%. Where's the problem?

The apparent problem stems from various peoples' assumptions about battery capacity. We don't really care about that here. We know how many miles the EPA said the car could go when new per % of displayed battery (however many kWh or coulombs that may actually represent) and we know how may miles he went per percent on this trip. That number is 95% of the EPA value. So for this trip he did 5% worse than the original EPA rating for the car. I don't see that there is any reason for concern.

 

[aesculus]

That number is 95% of the EPA value. So for this trip he did 5% worse than the original EPA rating for the car. I don't see that there is any reason for concern.

Which is close to the 6% degradation I posted about up top.

 

[ajdelange]

Are you concerned about 5 - 6% degradation in a car with 38,000 miles on it?

 

[aesculus]

Are you concerned about 5 - 6% degradation in a car with 38,000 miles on it?

No. Although I would like none.

 

[animorph]

In addition to many of the reasons above, there are losses inside the battery which are not measured by the trip meter.
Car’s energy consumption (lack of) accuracy

Measure the kWh used to charge the battery back to 98%. That will include any battery losses and will be closer to total energy consumed.

 

[Hebert]

Have you ever looked at the data at //teslike.com/range?

For my X75D, it says it was supposed to start at 237, but the range estimate starts at 231. That's exactly how my car started. And it shows that if I drive 65 on the freeway, I will get 240 miles. I find I can easily get over 240 miles on road trips. But if I drive 75, I only can go about 200 miles.

I suspect the data for your 90 is close also. It shows your car started at 251.

 

[ajdelange]

In addition to many of the reasons above, there are losses inside the battery which are not measured by the trip meter.

This was mentioned in No. 3. where it was pointed out that the amount of energy required to charge the battery by x % is greater than the amount of energy withdrawn from the battery that results in a reduction of x %. Energy transferred to and from the battery is measured by integrating the product of terminal voltage and current. There is an internal impedance between the terminal and the battery (at least we model it thus). The current that charges the battery dissipates heat in the resistive part of that impedance. Thus the stored energy is less than the delivered energy. The battery's discharge current also warms that reistance and so the energy delivered by the battery is less than the energy taken from it.

Thus we can, from the displayed battery percentage, deduce two capacities: one from charging data and one from discharge data. The former will be bigger than the latter but either can be used to estimate degradation. You just can't mix the two.

The meter measures terminal energy as described. Does it correct this for internal losses? That would be simple enough to do as the internal resistance is simply the derivative of terminal voltage with current and were that observed the transferred energy could easily be corrected to stored energy. But it doesn't seem they do that.