P85D vs. P85 Efficiency Testing, Take 2

[joer00]

If everybody else gets dramatic improvement on .139/.140 then I guess I have to consider the possibility of something hardware related/wrong with my car that is not allowing the torque sleep to activate.

Same here, it got a LITTLE better. Went from 433 wh/mile avg to around 400 wh/mile. My all time P85 avg was 351wh/mile so my D is still more than 10 % worse !

 

[AMN]

Swap cars on the way back! If your wife is lighter than you, the D will win!
Thanks for doing the test!

 

[wk057]

So, I'm sitting here eating with my fiance and we both hear someone mention "Tesla." A group of 8 folks at a table across the way are talking about Tesla manufacturing...

Of course I'm intrigued. I decide to head over, in my Tesla cap and jacket, and pry. Turns out there was a gentleman who is an engineer from the factory there meeting with a local manufacturer of a new rivit and rivit tool they're now using for the Model S and upcoming X.

Seriously, what the heck are the odds?

Long story short everyone thought it was an exciting coincidence. Made my day.

Aside from that, my dash cam isn't cooperating now for some reason, so hopefully I at least have footage for leg 1.

- - - Updated - - -

Dash pics with trip meters with 30 mile energy graphs.

- - - Updated - - -

Looking at pics, I'm honestly not sure why the temps are so different. Both cars parked in the sun right next to each other. Guess yellow vs blue. It is about 48 out.

 

[LetsGoFast]

I also note that the P85D is ahead on the 30 mile graph and has an extra half-mile on the trip odo.

 

[NOLA_Mike]

Same here, it got a LITTLE better. Went from 433 wh/mile avg to around 400 wh/mile. My all time P85 avg was 351wh/mile so my D is still more than 10 % worse !

Maybe you have to have Park Assist Disabled. I've never seen that error after all these numerous updates. That power hungry Park Assist is eating up about 50 Wh/mi...:tongue:

 

[wk057]

I also note that the P85D is ahead on the 30 mile graph and has an extra half-mile on the trip odo.

The odometer discrepancy was expected and is exactly the same as the last trip.

 

[Noone]

Looking at pics, I'm honestly not sure why the temps are so different. Both cars parked in the sun right next to each other. Guess yellow vs blue. It is about 48 out.

Sorry, slightly off topic... but... yellow​???

 

[Andyw2100]

The odometer discrepancy was expected and is exactly the same as the last trip.

I'm going to be posting some numbers shortly, and noticed a discrepancy between the odometer numbers and the estimated distance from EV Trip Planner, with the discrepancy also being that the P85D thinks it has travelled farther. In my case some of that will be due to my driveway, my wife driving around the parking garage, etc. But I'm wondering, Jason, since there are two of you, which of your odometers more closely matches EV trip planner? And if it is your fiance's P85, is this some weird little bug with P85Ds?

 

[apacheguy]

Arrived for lunch. P85... 325 Wh/mi. P85D...... drum roll...... more suspense.......... 326 Wh/mi.

So close!

The P85D seems to definitely win on the highway. The acceleration getting up to speed seems to skew against the P85D a hair, then move back in favor at constant speed.

My final crunch of the data later will help.

At this point.... too close to call, but already 15+% better than the first side by side for the P85D.

Good result.

Well given that the EPA numbers are skewed towards highway driving, I'm having difficulty believing this result. Especially after Tesla has confirmed that the P85D is, in fact, less efficient than the single motor cars. It's close, but I don't see how the argument can be made that the D is more efficient.

Driving Range for the Model S Family | Tesla Motors

 

[wk057]

Well given that the EPA numbers are skewed towards highway driving, I'm having difficulty believing this result. Especially after Tesla has confirmed that the P85D is, in fact, less efficient than the single motor cars. It's close, but I don't see how the argument can be made that the D is more efficient.

Driving Range for the Model S Family | Tesla Motors

Difficulty believing my result? I'll be posting a video of my test like I did last time if you have doubts on my info...

Final results!

P85: 111.3 miles, 35.0 kWh, 315 Wh/mi
P85D: 112.3 miles, 35.4 kWh, 316 Wh/mi

For comparison, the results of the previous test (pre-torque sleep) were:
P85: 111.3 miles, 34.1 kWh, 306 Wh/mi
P85D: 112.2 miles, 39.5 kWh, 352 Wh/mi

Today was roughly the same temperature, however we had a bit of wind today compared to the last test which was pretty calm.

So, the P85 wins by 0.4 kWh. I'm going to say that within a reasonable margin of error the cars are now equal in terms of overall efficiency. This would correlate with the window sticker's EPA combined rating being the same for both cars.

Preliminary analysis shows almost without a doubt that the P85D is slightly more efficient at highway cruising speeds than the P85.

Going to gather up the data from my videos now. Unfortunately my dash cam decided not to work on the second leg of the trip, and my GoPro mount decided it didn't want to stay in place either. Looks like the first leg has all the video footage available from the two GoPros and dash cams, though. For the last leg I held my GoPro in place until it died about 15 miles shy of home.

More soon!

 

[Andyw2100]

I am happy to see a few other P85D owners reporting that they have seen little or no increase in efficiency between .113 and .139. If I had to guess, I’d say I may be seeing some slight increase, but that’s more of a gut feeling than anything else, and the increase is certainly nothing like what wk057 is seeing and documenting in his change to .140. Perhaps there is hope for me yet, when I get to .140. (I'm not sure what to make of your situation, NOLA_Mike. And I may be right there with you when I get .140 too. Who knows.)

I am trying to do as good a job I can of recording data for the known trips we’re taking, so that when we get future firmware releases we’ll be able to make good comparisons. (MarcG’s post yesterday inspired me!)

Below are some numbers. They are not formatted nearly as well as Marc’s. They are for two trips, both of which included about 30 minutes on the highway and roughly the same amount of time on rural roads that have varying speed limits and at times stop lights, etc. I tried to note the windspeed on the highway, using the tool I mentioned in a another post. I compare the actual P85D numbers with EV trip planner numbers, but I did not try to use their wind adjusted rated mileage number. I did try to estimate payload accurately in EV Trip Planner, accounting for the fact that I weigh more than my wife, had my 50 pound dog in the car with me, etc.

I have the Tesla 19” winter tires (Pirelli Sottozero IIs) on Cyclone wheels. (The Cyclone wheels are supposed to help a bit, as per the JB Straubel blog post.)

As I mentioned upthread, the mileage is off a bit, with the P85D believing it travelled slightly further than EV Trip Planner’s route. Some of that is definitely due to the length of my driveway and my wife having to drive around her parking garage to park, but that would not account for the entire discrepancy. I checked the route, and it is correct, so I don’t know what to make of that.

Both trips were made with range mode on, in sport mode, and with the battery at least somewhat preheated.

Distance	RM	Total Energy	Avg Energy	Hwy Speed	Cabin Temp	Outdoor Temp	Wind	Elev Change	Conditions	EV Trip Planner Estimates:	Distance	RM	Total Energy	Avg Energy
58.4	78	22.7	389	65-70	65	32	11 HW	630	Cloudy		57.2	69	20.8	364
54	69	20.8	385	65-70	68	14	6 HW	(754)	Little light snow		52.6	63	18.9	360

 

[wk057]

I can't say for sure I understand why others aren't seeing an efficiency boost. Do keep in mind that the torque sleep stuff seems to be less effective in stop/go situations. At a constant speed, though, even a low speed, it seems to work wonderfully.

 

[Andyw2100]

I can't say for sure I understand why others aren't seeing an efficiency boost. Do keep in mind that the torque sleep stuff seems to be less effective in stop/go situations. At a constant speed, though, even a low speed, it seems to work wonderfully.

For me, the two trips I documented include 30 minutes of pure, 65-70 MPH TACC highway driving. The rural stuff is a little stop and go, but that too has plenty of cruising at 40-55. These trips are by no means city, stop and go driving type trips. They are predominantly constant speed, with some stop and go mixed in.

I'm just hoping .140 makes the difference for me!

 

[lolachampcar]

apache,
This is a very small sample but my D is definitely performing better at sub-highway speeds. I just saw similar numbers to my P+ on a very well know trip for me (picking my daughter up from school) where I travel at most 45 mph. There is a very distinct movement in the control firmware from an acceleration or transient mode to a steady state mode. I can feel the car "hunt" just a bit shortly after making the transition and I can also see the power usage drop. This is the first time I have seen less than 300 WHr/mile on this particular trip with my D but routinely would get between 250 and 275 with my P+. I was right at 270 today.

I believe torque sleep is used for any sustained steady state condition. The definition of sustained in this instance is establishing a speed for more than a few seconds. Any meaningful traffic requiring throttle modulation likely excludes torque sleep which is, I suspect, is why it is described as a highway feature. My trip just now has sustained periods at 45 mph and thus allowed the use of torque sleep.

I suspect that attempting to completely remove any perceivable hunting during the transition from non to torque sleep is what has delayed deployment. I've done a good bit of real time control and the work Tesla has done to allow seamless acceleration to regen in fine enough increments as to be imperceptible is nothing short of amazing. I am 100% confident they will get there with TS as well and will likely increase its usage though out all traffic patterns.


BTW
I've not touched a configuration button in the car so it is still on normal (not range) driving mode and Insane for acceleration.

 

[wk057]

Some more pics from the beginning, middle, and end of my trip:

 

[rage_777]

Just wondering since you are using two cars at the same time, if you are drafting each other? Are you leaving a big enough cushion between the cars to not have that affect the test.

 

[wk057]

Just wondering since you are using two cars at the same time, if you are drafting each other? Are you leaving a big enough cushion between the cars to not have that affect the test.

For both tests the P85D followed the P85. For this one I set the TACC at a middle value, which left a pretty good amount of space between us, as before. I seriously doubt it would be close enough to make any drafting difference given the profile of the Model S.

 

[joer00]

For both tests the P85D followed the P85. For this one I set the TACC at a middle value, which left a pretty good amount of space between us, as before. I seriously doubt it would be close enough to make any drafting difference given the profile of the Model S.

Just to make sure I can compare with my (not so good results). Your tests where done with INSANE mode and Range Mode Off ??

I tried on my way home from work to go in Sports and Range Mode and this indeed showed lower consumption at steady highway speed.

Will test over the weekend with different modes on and off.

 

[wk057]

Just to make sure I can compare with my (not so good results). Your tests where done with INSANE mode and Range Mode Off ??

I tried on my way home from work to go in Sports and Range Mode and this indeed showed lower consumption at steady highway speed.

Will test over the weekend with different modes on and off.

No, I used the same settings as the previous run, which was sport + range mode.

 

[EarlyAdopter]

I just updated from .139 to .140. I'll see if I can make a run of my test loop tonight. I didn't see any efficiency gain on .139. Hoping that .140 shows an improvement.