Do P85Ds and S85Ds have inaccurate odometers that overstate distance traveled?

[m6bigdog]

I'm thinking you probably haven't read the whole thread either. If you have, and it was some time ago, I'll remind you of a couple of salient points:

--The Ds are the first models where this is any kind of an issue. Other Model S vehicles had odometers that were much more accurate, and provided the expected figures.

--With all the Ds off by about the same amount, it appeared to us that the problem could be as simple as a bad formula or calculation somewhere. There wasn't variance within the odometers, with all of them reporting wild figures somewhere within 4% of the correct figure. All the D odometers were reporting mileage about .8% longer than it actually was. This is precise and accurate--just wrong.

--Tesla has now corrected the issue on my car. For now, I have the only known car with the fix, but that's probably because I'm the only one who has noticed it. One theory is that the fix has been made in software, but needs a TPMS reset and/or a wheel change to be activated on the car. Another possibility is that the fix somehow has to be taken care of while at a Service Center, as my car was at a SC just before the odometer started reporting accurate mileage, but I don't think this is likely, as there was no documentation on my work order about this.

I think we can stop talking about whether the problem is real or not, who it matters to and why, SAE standards, etc. (since that's all been done before) and at this point just focus on whether or not other vehicles are receiving the fix, and if so, what it takes to trigger it.

FWIW, there is no such thing as an absolutely accurate (<1%) odometer/speedometer when the vehicle tire is used to measure distance or speed.
Why? because all the tire variable that cause an odometers error can be accumulative!
1) the tire wears from the max diameter (10/32" tread depth) when new to a smaller diameter when it is warn-out.
2) the tires rolling diameter is smaller from the measure/designed diameter therefore the tires revolution/mile must be measured on an SAE standard test roller and that is +/- the same tire rolling on a vehicle on the roadway. This all by it self is +3% (814 measured vs 787 calculated) with a Michelin PS2. and the warn-out PS2 calculates to be 2% less than the new tire
3) the rolling distance for any tire manufacture, specific tire size and/or type (19" vs 21", Michelin vs Continental, MXM4 vs PS2) is different even though the tire size is the same.
4) The tire constantly slips on the asphalt/concrete roadway so the actual distance traveled will vary based on constantly changing traction forces such as regen-braking, pedal braking, acceleration, speed, head/tail winds, air density etc.
5) the vehicle drive system RWD, FWD or AWD will change the accuracy because the tire slip is greater on the drive tires so the D models may require a different rolling tire factor than the non-D models.

Therefore, even if every Model S was delivered with a <1% odometer error, tire wear and/or replacement tire would bugger that up in a few thousand miles.

So can Tesla tweak/adapt the Odometer accuracy?
I anticipate they can change the tire rev/mi. factor in the software, but they would need to constantly tweak the rev/mi. factor (at every tire replacement and as the tire wears) based on some assumed more accurate actual distance traveled like the GPS.
What does maintaining a super accurate (<1% error) odometer buy you.
IMHO, just because you can measure it doesn't signify it is useful information, as +/- 1 to 3 mile in 100 is irrelevant for any real useful purpose in the big picture of vehicle dynamics and actual performance.

 

[Andyw2100]

FWIW, there is no such thing as an absolutely accurate (<1%) odometer/speedometer when the vehicle tire is used to measure distance or speed.

Great.

I don't think you'll find that I ever stated conclusively that Tesla was doing this. In fact, I don't have any information about how Tesla calculates the distance traveled in our cars. The one thing I do know, and this seems to be the point that is somehow escaping everyone who wants to come and derail this thread and tell me why I'm being silly, etc. is that, and I'm going to type this in bold and big letters, so perhaps I won't have to keep typing it:

Previous Model S vehicles had odometers that accurately reflected the distance traveled. The dual drive Model Ss don't.

Why is it difficult to understand that based on the above, we are trying to get answers, and trying to figure out if our dual drive Model Ss can have odometers that are as accurate as the Model Ss that came before ours? If this doesn't seem important to you, that's fine. No one is making it their life's mission. Just don't follow this thread, and don't worry about it. But please stop coming to this thread to post and tell us that we're stupid, that our concerns don't matter, that we have no common sense, that Tesla has more important things to worry about, that the earth is round, etc. We get it.

And we still want to know if our cars can have odometers that are as accurate as earlier cars.

 

[gpetti]

First of all, @Andyw2100, apologies for not providing additional data, I haven't been doing any longer trips lately and have been away.
Secondly I agree that there is no need for people who feel that this is irrelevant to come here and dish out insults. I've noticed that one of the members generally has an aggressive tone in many of his(?) posts. Sometimes this is usefully provocative and sometimes just plain insulting. Also said person speaks with great authority that may or may not be warranted. I think the point about speedometers always needing to show higher than actual speed (or put another way, must never read lower than the actual speed) is fairly well known, however, I've read elsewhere (no I'm not an authority either) that the speedometer and odometer are not necessarily linked as recent class action suits regarding odometer inaccuracy typically have different requirements than the speedometer. In one of the class action suits, differences were from 2% to 4%. In my case with the data reported earlier in this thread, my difference was from about 1.6 to 2% inaccuracy. As Andy has stated, I never saw any noticeable mismatch with my previous P85+ (driving the same route). There probably was some variance based on all the factors mentioned elsewhere but they were relatively unmeasurable. Thus I agree that something is slightly off with my P85D. Hopefully I can evaluate whether the difference still exists soon.

 

[Andyw2100]

First of all, @Andyw2100, apologies for not providing additional data, I haven't been doing any longer trips lately and have been away.
Secondly I agree that there is no need for people who feel that this is irrelevant to come here and dish out insults. I've noticed that one of the members generally has an aggressive tone in many of his(?) posts. Sometimes this is usefully provocative and sometimes just plain insulting. Also said person speaks with great authority that may or may not be warranted. I think the point about speedometers always needing to show higher than actual speed (or put another way, must never read lower than the actual speed) is fairly well known, however, I've read elsewhere (no I'm not an authority either) that the speedometer and odometer are not necessarily linked as recent class action suits regarding odometer inaccuracy typically have different requirements than the speedometer. In one of the class action suits, differences were from 2% to 4%. In my case with the data reported earlier in this thread, my difference was from about 1.6 to 2% inaccuracy. As Andy has stated, I never saw any noticeable mismatch with my previous P85+ (driving the same route). There probably was some variance based on all the factors mentioned elsewhere but they were relatively unmeasurable. Thus I agree that something is slightly off with my P85D. Hopefully I can evaluate whether the difference still exists soon.

Thanks, gpetti!

And I'm not sure we actually need any more data from most of us older Ds right now, unless someone resets their TPMS or changes from snows to all-seasons or the other way around, and can compare data between the two sets of tires.

The theory I held until this weekend was that the issue may have been resolved via a software update, but only implemented when wheels were changed and/or TPMS was reset. So I was hoping new Ds would report some data, so we could see if that was true. (That's why I mentioned older Ds who reset TPMS, etc.)

I kind of shot my own theory down with my data from this weekend. So my current theory is that Tesla hasn't fixed anything yet, and that with snow tires on the Ds, the discrepancy is significantly greater than with the all-seasons, and that the discrepancy is about that .8% we had come up with when people are driving on the all-seasons.

If you have really good numbers from before, when you were on winter tires, and want to compare those to some data now, assuming you are on all-seasons now, that would probably be useful. But unitl I had posted this, I hadn't actually been asking for any data from older Ds, so there's nothing to apologize for!

Thanks!

 

[WarpedOne]

You haven't explained why this error exists now, in the Ds, when it never existed in earlier Models Ss. But please go ahead and do that. I'm sure we'd all be interested. If you can't do that, how about you take your insults somewhere else?

Before any explanation, there must be some evidence. "when it never existed in earlied Model S" is pretty strong claim.
What is the evidence that supports such a claim?

 

[Andyw2100]

Before any explanation, there must be some evidence. "when it never existed in earlied Model S" is pretty strong claim.
What is the evidence that supports such a claim?

You clearly haven't read most of the thread, or you're just baiting me. I'll indulge you with just a few examples:

From upthread: Do P85Ds and S85Ds have inaccurate odometers that overstate distance traveled? - Page 3

I noticed this on my very first trip with the P85D from NJ->NC after taking delivery. The trip meter showed about 6 extra miles vs what I expected from a dozen previous trips on the exact same route.

This is a trip I've done pretty close to 100 times since I got my driver's license when I was 16. The main portion of the trip has remained the same with minor changes over the years. With the Model S, an obvious change is different fuel stops, but the distance difference is pretty negligible.

I've done the trip about a dozen times in a Model S (in both my old P85 and my fiance's P85) and once so far in the P85D at < 1k miles. In both P85's I always got within +/- 1 mile of expected distance door-to-door. I don't really deviate from my route, and +/- 1 mile seems pretty good accounting for other variables (like being in a different lane around a bend and such).

I'll be doing the trip again in the P85D soon now that the range issue is somewhat sorted out.

There's also the side by side video comparison that wk057 did when he was doing edficiency testing. I linked to that in my opening post, and he posts and provides links to all the videos in the following post.

There are also posts from others who have owned both older Ss and Ds, like this recent one:

In my case with the data reported earlier in this thread, my difference was from about 1.6 to 2% inaccuracy. As Andy has stated, I never saw any noticeable mismatch with my previous P85+ (driving the same route). There probably was some variance based on all the factors mentioned elsewhere but they were relatively unmeasurable. Thus I agree that something is slightly off with my P85D. Hopefully I can evaluate whether the difference still exists soon.

 

[WarpedOne]

I see no evidence for your claim of all old s behaving in stated way, i only see Generalization errors.
anyway, i will leave you to believe whatever you want to believe.

 

[Andyw2100]

You haven't explained why this error exists now, in the Ds, when it never existed in earlier Models Ss. But please go ahead and do that. I'm sure we'd all be interested. If you can't do that, how about you take your insults somewhere else?

i will leave you to believe whatever you want to believe.

 

[wk057]

@WarpedOne - I'll restate what others have stated: Obviously you haven't read this entire thread. There are multiple data points for non-D vehicles and D vehicles and with 100% consistency they point to the D vehicles having the discrepancy. I personally have data from two P85s and a P85D which matches this data, along with two side by side videos P85/P85D showing ~60 mile trips that exactly reflected this issue both times.

As for the speedometer showing a lower than actual speed... in the case of the Model S this is an irrelevant point. Back in the day when the mechanical sensors users to spin the gears in the dash equipment were all sync'd up maybe this made sense. Fortunately the software in the Model S can both meet such a requirement that speedometer reads some amount lower than actual speed AND have a correct odometer reading. It's kind of like magic... except with software.

In any case, this is NOT the case with the P85 vs P85D. If you note in my first side by side video (pre-TACC) the P85 was always slowly outpacing the P85D when they were set to the same cruise speed, meaning the P85D was showing a higher-than-actual speed and the P85 was showing a closer-to-actual speed. You'll also note in the second video where I used TACC to follow the P85 that the P85D's speedometer was often reading higher than the P85's, even though the distance between the vehicles was static.

So, @WarpedOne, unless you have some data that contradicts every existing data point gathered thus far, keep the insults to yourself.

 

[wk057]

Digging up this thread, I think I'm leaning towards some update/fix having happened sometime recently.

The last few trips I've made with the P85D more closely match my wife's P85 and my traded P85.

Also, today I had a P85+ loaner. I dropped my car off and just drove home, then back when it was ready. I noted the mileage as I was getting off of the highway each time and it was nearly the same for the P85D and P85+ loaner (VIN 45xxx) with only 0.1 miles more showing on the D over 51 miles.

Unrelated, P85D efficiency was better than the P85+ on all legs of this trip significantly. For the stretch above 298 Wh/mi on the P85D and 329 Wh/mi on the P85+.

 

[andrewket]

Argh! I just changed tires and was seeing significant improvement to efficiency. Now I have to wonder if it was .250 firmware that caused part of it? @$/&:$ scientific method.

 

[Andyw2100]

Unrelated, P85D efficiency was better than the P85+ on all legs of this trip significantly. For the stretch above 298 Wh/mi on the P85D and 329 Wh/mi on the P85+.

By any chance were you driving a little slower than you usually do?

I ask because in some other thread I had suggested if you were to do a side by side again, that you also do one where you keep the speed to 65 or so, just to see if that made a difference. Because one theory I have is that the strongest benefits of torque sleep aren't seen at 70 MPH-plus, and I know you often drive 70-75-plus, so I theorized that you may have just not been getting the most out of the torque sleep benefits.

Of course another possibility, if this trip was at the same speeds, is that Tesla just tweaked the torque sleep settings so that now it is kicking in even at the higher speeds.

 

[MarcG]

Argh! I just changed tires and was seeing significant improvement to efficiency. Now I have to wonder if it was .250 firmware that caused part of it? @$/&:$ scientific method.

It's most likely your new tires. I've been keeping track of my daily commute since pre-torque sleep, and haven't seen any changes in efficiency since installing .250 a week ago (relative to previous versions of 6.2).

 

[wk057]

For what it's worth my P85D has only one remaining original from-the-factory tire. Three of the four have been replaced at different points due to punctures by debris over the last 13,000 miles

Most recently (happened Friday) I removed and plugged a puncture caused by a large brand new lag bolt that I must have picked up on the interstate... *grumbles* Tesla replaced the tire yesterday to the tune of ~$500-ish.

 

[Andyw2100]

It's most likely your new tires. I've been keeping track of my daily commute since pre-torque sleep, and haven't seen any changes in efficiency since installing .250 a week ago (relative to previous versions of 6.2).

Unless the new tires were different, more efficient tires, the newer tires would actually be less efficient than the same model somewhat worn tires. Of course andrewket didn't say either way what the deal with the tires was.

 

[MarcG]

Unless the new tires were different, more efficient tires, the newer tires would actually be less efficient than the same model somewhat worn tires. Of course andrewket didn't say either way what the deal with the tires was.

Yes I know that new tires, because they're stickier, cause worse efficiency that older, less sticky tires.

The way I read Andrew's post, he was seeing better efficiency until he changed his tires and upgraded to .250. Which is why I suggested the drop in efficiency was due to his new tires.

If that's not what Andrew meant, and that he's seeing better efficiency with his new tires, then I'm at a loss as to why since .250 hasn't personally given me improvements. More details on his tire change would be needed.

 

[Andyw2100]

Yes I know that new tires, because they're stickier, cause worse efficiency that older, less sticky tires.

The way I read Andrew's post, he was seeing better efficiency until he changed his tires and upgraded to .250. Which is why I suggested the drop in efficiency was due to his new tires.

If that's not what Andrew meant, and that he's seeing better efficiency with his new tires, then I'm at a loss as to why since .250 hasn't personally given me improvements. More details on his tire change would be needed.

I know you know that the newer tires would be less efficient, Marc. I was posting to point that out to others, and to see if we could get some additional info from andrewket.

I have been away, and not driving since receiving .250, so I can't offer any efficiency comparison, but will be able to fairly soon, though with varying wind conditions and temperatures it may take a couple of weeks before I can really state anything conclusively.

One possibility, though, for why some might see an increase in efficiency now where you aren't would be my theory with torque sleep, speed, and any tweaking Tesla may have just done. If the speed you typically travel is already getting the full benefit of torque sleep, then this update wouldn't have any effect. Or if you are travelling at a speed typically above the point at which torque sleep really can kick in, and Tesla tweaked the speed up, but not up to the speed that you cruise at, you wouldn't see a difference.

Make sense?