Real World Comparison of a S85 to an 85D Efficiency

[Andyw2100]

The 85D did better this time.

I'm going to go out on a limb and guess that it was because you used range mode this time. Looking forward to the write up!

 

[spottyq]

I'm going to go out on a limb and guess that it was because you used range mode this time. Looking forward to the write up!

Nah, that would be an obvious question that someone would surely ask. I'm sure he did something more subtle. Say, opening all the windows and the pano roof on the 85. Or he conveniently 'forgot' he had 4 bikes wrapped in a thick blanket on the roof of one car.

 

[breser]

At previously mentioned we made the trek out to the Pendleton, OR supercharger from my house in North Bend, WA and back with an S85 and an 85D. This post constitutes the writeup of that trip.

Real World Test

Before I get into the details I want to point out that this comparison is a real world test. It is not a laboratory experiment. We have not carefully controlled every difference. The 85D had two passengers compared to one in the S85 (and no despite joking about this above I did not get out a scale and control the vehicle passenger and cargo weights). Tires were of course still different. If the only comment you have is to suggest we should have swapped wheels/tires or gone to much greater effort to control the differences, don't waste your time posting. Instead go out and execute your own test with your own time and $200k worth of vehicles that meets the criteria you desire.

The Vehicles

S85 constructed in February 2013. Goodyear Eagle Tires with about 4k miles on them. Software version 6.2 (2.4.239).
85D constructed in February 2015. Michelin Primacy Tires, the original tires the vehicle was sold with. Software version 6.2 (2.4.236).

Both vehicles have a Pano roof. The standard 19" wheels.

The S85 has an A pack and charges slower at the supercharger.

Test Procedure

In general we drove the vehicles back to back. With the 85D leading to Pendleton and the S85 leading on the return. There were of course instances where we got slightly separated in traffic. The vehicles were driven in range mode with roughly the same setting configuration. Climate set to 70°F (most of the time). ThortsMD mentioned lowering the temperature in his car at some point, but I don't know the exact details of which legs and how much he lowered it. We made no attempt to try and sync battery state of charge because of the differences in charging speeds. We spent around 5-10 minutes waiting for the S85 to reach a state of charge we were comfortable leaving with (due to the A pack). We ran around 3-4 miles per hour above the speed limit in most places.

The Route

We followed I-90 East to Ellensburg, WA where we supercharged and then followed I-90 east to I-82 to SR-22 (State Route) and then SR-221 and then SR-14 and finally back on I-82 to I-84 East to end up at the supercharger in Pendleton, OR. This is a shorter route than the one suggested by the car (following I-82 straight to I-84) and saves about 15 miles and 8 minutes of driving time. This accounts for a fairly significant difference in the estimated state of charge vs what we did (i.e. you can't really draw any conclusions about the estimate the trip planner gave).

The Data


S85

				SLC*	SLC*	SLC*	RR**	RR**	TP***	TP***
Location	Time	°F	Odometer	Miles	kWh	Wh/mile	Range Mode	No Range Mode	Battery %	Estimate %
Leaving breser's house	7:15	55	27,889	35.6	12.1	340	192	190	76	39
Arriving Ellensburg, WA	8:23	59	27,966	112.7	37.2	330	103	101	41	N/A
Leaving Ellensburg, WA	9:17	67	27,966	0	0	0	230	229	92	11
Arriving Pendleton, OR		74	28,141	175.2	52.8	301	44	43	17	N/A
Leaving Pendleton, OR	13:15	74	28,141	0	0	0	243	240	96	21
Arriving Ellensburg, WA	15:49	78	28,316	174.1	56.0	322	44	43	17	N/A
Leaving Ellensburg, WA	16:28	82	28,316	0	0	0	173	171	69	25
Arriving breser's house	17:33	74	28,392	76.9	24.9	324	85	83	33	N/A

85D

				SLC*	SLC*	SLC*	RR**	RR**	TP***	TP***
Location	Time	°F	Odometer	Miles	kWh	Wh/mile	Range Mode	No Range Mode	Battery %	Estimate %
Leaving breser's house	7:12	85/53****	111,365	0	0	0	243	240	90	56
Arriving Ellensburg, WA	8:23	58	11,443	77.4	26.2	339	147	145	54	N/A
Leaving Ellensburg, WA	9:15	68	11,443	0	0	0	243	240	90	11, 12, 13†
Arriving Pendleton, OR	11:56	73	11,619	175.8	54.2	308	45	43	16	N/A
Leaving Pendleton, OR	13:13	76	11,619	0	0	0	253	250	94	20
Arriving Ellensburg, WA	15:49	79	11,793	174.7	54.0	309	55	53	20	N/A
Leaving Ellensburg, WA	16:25	82	11,793	0	0	0	205	203	76	46
Arriving breser's house	17:33	75	11,871	77.2	23.9	310	118	117	43	N/A

Leg Comparisons

Leg	S85 Miles	85D Miles	S85 kWh	85D kWh	S85 Wh/mile	85D Wh/mile
breser's home -> Ellensburg, WA	77.1‡	77.4	25.1‡	26.2	324‡	339
Ellensburg, WA -> Pendleton, OR	175.2	175.8	52.8	54.2	301	308
Pendleton, OR -> Ellensburg, WA	174.1	174.7	56.0	54.0	322	309
Ellensburg, WA -> breser's home	76.9	77.2	24.9	23.9	324	310
Overall	503.3	505.1	158.8	158.3	316	313

Overall the 85D used less energy on a total kWh and Wh/mile basis. On the way to Pendleton the S85 used less energy. On the way from Pendle the 85D used less energy. The trailing car always ended up using less energy on the specific legs.

Conclusions

I don't think these results are clear enough to say that the 85D is the clear winner. But I do think it means that it's plausable that it's slightly better.

The 85D had an extra passenger and the S85 had a lower climate control setting, which I tend to think probably offset each other. On the Ellensburg, WA to Pendleton, OR leg it wasn't that far behind the S85 on energy usage even though it was the lead car.

It's interesting to note that the 85D always ended up traveling more distance. Not a lot but it still did. This might be pushing the Wh/mile numbers down for the 85D. But it still managed to win the kWh on the return trip and the overal.

You can debate about the tires since the Michelin's are supposed to be slightly more efficient. You can also debate about the mileage on the tires and if this put the S85 at a disadvantage (just as the 85D was at a disadvantage in the first test), but based on these results I don't think the problem on the first round was the newness of the 85D tires.

Rather I think the first trip to Coeur d'Alene really does come down to just not using range mode. Range mode doesn't seem to make much difference in my experience with the RWD vehicles, but it makes a world of difference with the AWD vehicles.

But while the 85D was slightly better it really wasn't that huge of a difference. Certainly not enough that I would think it should matter for a purchasing difference. The open question of course is if the extra passenger hid enough of a difference that it should still show in the rated range and EPA number. Or if the reason the rated range and EPA numbers are showing better is the issues with the odometer on the 85D (see other threads about that) and that the vehicles were actually much closer.

Now don't get me wrong. Even making an AWD car having roughly the same range and efficiency of a RWD car is a significant accomplishment. I think Elon's comment that they improved the car in every way was fair (leaving aside the disputes over if he meant both the P85D and 85D or just the 85D).


* SLC = Since Last Charge
** RR = Rated Range
*** TP = Trip Planner
**** Charging had just finished and the car's temperature read high. By the time we reached the interstate it had a lower temperature.
† It started at 11% and then changed to 12% and then to 13% while I was just sitting there waiting to pull out and wasn't even moving.
‡ Calculated after removing the Since Last Charge recorded when leaving breser's home.

 

[Andyw2100]

Thanks for taking the time to do the follow-up test, and for posting such detailed data.

It's interesting to note that the 85D always ended up traveling more distance. Not a lot but it still did.

You don't really mean that it traveled more distance, right, but rather that it reported a greater distance traveled for what was almost certainly essentially the same distance traveled.

You are familiar with the controversy over the fact that the Ds over-state distance traveled in a way that earlier Model Ss did not, right? Edit: Just realized when I went back and read more carefully that you linked to this thread yourself above. Silly me. (I had read quickly the first time, anxious to get this post written!)

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

It would be very interesting, for the purposes of the discussion going on in the thread above, to see what something like Mapquest or Google Maps gives as the actual distance for the legs in question, so that we could see how close the S85's odometer is to reality.

 

[breser]

You don't really mean that it traveled more distance, right, but rather that it reported a greater distance traveled for what was almost certainly essentially the same distance traveled.

I wasn't making a judgement.

It would be very interesting, for the purposes of the discussion going on in the thread above, to see what something like Mapquest or Google Maps gives as the actual distance for the legs in question, so that we could see how close the S85's odometer is to reality.

From my home to Ellensburg supercharger Google Maps lists 76.4 miles.
Between Ellensburg and Pendleton superchargers Google Maps lists 175.0 miles (both directions are the same mileage).
From the Ellensburg supercharger to my home Google Maps lists 76.3 miles.

However, I don't really think you can count on the online map numbers being accurate enough to be useful in this case.

The interesting bit in the data, which I didn't notice until after I posted this. Ellensburg <-> my house was a difference of exactly 0.3 miles and Ellensburg <-> Pendleton was exactly 0.6 miles. I find it hard to believe that it came out to that exact difference if there isn't some difference in how the cars are calculating distance traveled.

All that said 3-4 tenths of a percent difference doesn't seem like it's a big deal. So I'm not overly concerned about it. Probably well within the expected error difference. Could just be a matter of the difference in age of tires.

 

[Andyw2100]

From my home to Ellensburg supercharger Google Maps lists 76.4 miles.
Between Ellensburg and Pendleton superchargers Google Maps lists 175.0 miles (both directions are the same mileage).
From the Ellensburg supercharger to my home Google Maps lists 76.3 miles.

However, I don't really think you can count on the online map numbers being accurate enough to be useful in this case.

The interesting bit in the data, which I didn't notice until after I posted this. Ellensburg <-> my house was a difference of exactly 0.3 miles and Ellensburg <-> Pendleton was exactly 0.6 miles. I find it hard to believe that it came out to that exact difference if there isn't some difference in how the cars are calculating distance traveled.

All that said 3-4 tenths of a percent difference doesn't seem like it's a big deal. So I'm not overly concerned about it. Probably well within the expected error difference. Could just be a matter of the difference in age of tires.

Thanks for posting the mileages from Google Maps.

I agree that there has to be something different about the way the cars are calculating mileage. That has been the contention of those of us discussing the odometer issue from the start.

While I agree the difference is not a big deal, if it's just a question of correcting an error in a calculation being made in the Ds, I'd like to see Tesla correct it. Except for the leg from Pendleton to Ellensburg, the S85 was closer to the Google numbers on all legs. The fact that Google has the mileage being the same in both directions of that leg makes me wonder about the accuracy of it, or if somehow the precise route you both took just actually was shorter.

And as you pointed out, the S85D recorded higher mileage than the S85 on all four legs, in a very consistent manner.

While this isn't a big deal, it also shouldn't be a big deal to fix if it's just a calculation issue.

 

[spottyq]

(snip)
We made no attempt to try and sync battery state of charge because of the differences in charging speeds.
(snip)

I knew it ! The test is unfair because a fully charged battery weights more than a empty one. By as much as 3,45µg. The test is flawed !

All joking aside, thanks again for taking the time to do such a test, to collect and post that data, and to eliminate as many variables as possible. I probably wouldn't have the patience to do it like you did !

I completely agree with your conclusion.

Sadly, I can't give you rep at the moment, but I'll try to remember to give you some as soon as I have more of it.

 

[3mp_kwh]

More food for thought, from a quick visit to Tirerack, shows OE Eagle tires at 755 revs per mi, vs. the MXM4s at 751. I'd expect the cars to have fixed revolutions per mile settings, however this difference is opposite to what is observed. The Goodyear shod S85 should over-state, relative to the MXM4 85D by .5%, if I'm looking at it right.

Thanks for the real world trial!

 

[glhs272]

breser,

Minor thought... It looks like you both left from your house. Did the S85 also leave from your house cold soaked or did ThortsMD drive to your place, top off then continue on? Just thinking that maybe the early on results were higher for your S85D because it still had to warm up, where the S85 didn't.

 

[Andyw2100]

More food for thought, from a quick visit to Tirerack, shows OE Eagle tires at 755 revs per mi, vs. the MXM4s at 751. I'd expect the cars to have fixed revolutions per mile settings, however this difference is opposite to what is observed. The Goodyear shod S85 should over-state, relative to the MXM4 85D by .5%, if I'm looking at it right.

Thanks for the real world trial!

If the above is actually a factor, it would explain why the D overstatement relative to the S was less than the .8% number that others had computed previously. The S, on non-OEM tires, was also overestimating the mileage some.

 

[ThortsMD]

breser,

Minor thought... It looks like you both left from your house. Did the S85 also leave from your house cold soaked or did ThortsMD drive to your place, top off then continue on? Just thinking that maybe the early on results were higher for your S85D because it still had to warm up, where the S85 didn't.

I drove from my house in Seattle to Breser's...about 25 miles. I did not top off at his house. It was in the mid 50's and my garage was warmer, so I was not cold soaked.

 

[breser]

Minor thought... It looks like you both left from your house. Did the S85 also leave from your house cold soaked or did ThortsMD drive to your place, top off then continue on? Just thinking that maybe the early on results were higher for your S85D because it still had to warm up, where the S85 didn't.

I drove from my house in Seattle to Breser's...about 25 miles. I did not top off at his house. It was in the mid 50's and my garage was warmer, so I was not cold soaked.

I also would not characterize my 85D as cold soaked. It had just finished charging to 90% very shortly before ThortsMD arrived (I started at 5:45 and we left around 7:15). To the degree that as noted in the data the outside temperature was showing as 85 when I pulled it out of the garage to leave, despite the outside temperature being 55.

I'd be surprised if any battery heating was going on.

 

[breser]

No comparison but had some fun going to Mt Rainier this afternoon. We actually got home with 19% left on the battery. 228.1 miles traveled, 60.1 kWh used for 264 Wh/mile.

The hilarity that is the energy app after traveling down the mountain.