My P85D's efficiency is still improving with firmware updates

[AWDtsla]

I would hope that for the most part, like the other weather conditions, any fluctuations due to barometric pressure would be offset by the number of legs I recorded. I could be wrong about that.

If you record enough samples then they it will average out. Also if you have enough samples you can plot the change in efficiency, this should point out if there was a step change during a firmware update or gradual change (say due to climate or tire break-in).

 

[Andyw2100]

If you record enough samples then they it will average out. Also if you have enough samples you can plot the change in efficiency, this should point out if there was a step change during a firmware update or gradual change (say due to climate or tire break-in).

I will supply all the data I have for the summary I offered in my first post. If you or anyone else would like to manipulate the data further, have at it. (It may take me a little while to get that data into an easily postable format.)

Edit: The data is below. There is one trip under "Kim South .236" where my wife actually drove the route I ordinarily drive south, which is why that one trip is slightly longer than her others. There's another trip under Kim South .250 where she had to take a short detour to get home, due to a road closure.

	Date	Mileage	Range Miles	Total Energy	Avg Energy	Temp	Wind	Weather
Kim North .236
	6/1/15	53.5	52	15.5	289	51	2HW	Rainy
	6/2/15	53.4	49	14.8	277	51	4HW	Rainy
	6/3/15	53.4	48	14.4	269	45	1HW	Dense Fog
	6/10/15	53.4	46	13.6	254	54	3TW	Foggy
	6/15/15	53.5	47	13.9	259	70	2TW	Misty
	6/16/15	53.4	42	12.5	235	73	5TW	Clear
	6/24/15	53.5	47	13.7	256	58	0W	Clear
	6/26/15	53.5	48	13.8	259	62	4HW	Clear
	6/29/15	53.5	48	14.1	264	56	0W	Rain
	7/1/15	53.5	44	13.2	247	64	6TW	Clear
	7/7/15	53.5	45	13.2	247	70	6TW	Light Rain
Kim South .236
	6/1/15	52.9	58	17.2	324	54	2TW	Rainy
	6/2/15	52.8	52	15.4	292	60	5TW	Clear
	6/3/15	52.8	52	15.8	299	73	1TW	Clear
	6/10/15	52.8	57	16.7	317	76	11HW	Clear
	6/15/15	52.9	53	15.7	298	79	CW	Clear
	6/16/15	52.9	54	15.7	297	75	4TW	Clear
	6/24/15	52.9	50	15.6	295	71	1TW	Clear
	6/26/15	57.9	56	16.4	284	77	4TW	Clear
	6/29/15	52.9	52	15.3	289	70	2HW	Clear
	7/1/15	53.1	54	16	302	69	0W	Clear
	7/7/15	53	55	16.2	306	84	9HW	Clear
Andy North .236
	6/4/15	58.3	51	15	257	76	5TW	Clear
	6/11/15	58.3	56	16.3	280	77	5HW	Clear
	6/17/15	58.4	51	15.5	265	70	2HW	Clear
	6/19/15	58.4	54	15.9	273	69	3HW	Clear
	7/2/15	58.4	54	16	273	72	5HW	Clear
Andy South .236
	6/9/15	57.9	67	19.7	341	60	2HW	Rain
	6/12/15	57.9	64	18.8	325	74	5HW	T-Storms
	6/18/15	57.9	64	18.7	323	78	8HW	Clear
	6/23/15	57.9	64	18.6	321	82	6HW	Clear
	7/6/15	57.9	63	18.7	322	81	8HW	Clear
Kim North .250
	7/23/15	53.5	46	13.5	253	56	5HW	Clear
	7/24/15	53.5	43	12.8	239	64	2TW	Clear
	7/27/15	53.5	43	12.7	237	68	2TW	Clear
	7/28/15	53.5	43	12.7	237	68	2TW	Clear
	7/29/15	53.5	44	13	243	62	2TW	Clear
	8/4/15	53.5	46	13.4	250	60	4TW	Fog
Kim South .250
	7/23/15	53.1	52	15.4	290	74	2HW	Clear
	7/24/15	53	53	15.3	289	80	0W	Clear
	7/27/15	55.8	55	16.2	291	84	5TW	Clear
	7/28/15	53	53	15.4	291	88	3TW	Clear
	7/29/15	53	55	16	302	88	5HW	Clear
	8/4/15	53	58	16.4	309	68	0W	Rain
Andy North .250
	7/21/15	58.5	51	15.2	260	68	4TW	Clear
	7/30/15	58.4	53	15.5	265	82	2HW	Clear
Andy South .250
	7/22/15	58	62	18	311	75	3TW	Clear
	8/3/15	58	63	18.5	319	83	4HW	Clear

 

[Matteo]

The only thing this test shows is, your range changes when air density changes. Air density is kg of air in 1 cubic metre volume. That number changes all the time. When temperature increases air density drops, making it easier to cut through air and decreasing Wh/mi. If you continue the tests you will see that your Wh/mi numbers will increase each month in Sep, Oct, Nov and Dec.

Air drag = 1/2 * speed[SUP]2[/SUP] * air density * drag area

 

[Khatsalano]

Tesla is tweaking the powertrain with almost every patch. The torque sleep algorithm is adjusted constantly to try and gain more effect/less undesirable noise among other things.

I've found that my S85D has gotten more efficient as the summer's gotten hotter here in the Bay. I really just chalk it up to average weather and commute conditions. I can also see this when I commute in the early morning when it's cool, vs. in the afternoon when it's hot. Some of it has to do with wind too though since in the summer, the wind pattern in the Bay Area is zero wind in the morning, but tailwind in the afternoon for my commute. Also, morning traffic flows faster than afternoon traffic in the Bay, leading to a difference of average speed. My Wh/mi on morning and afternoon legs are quite different. Compared to the winter/spring, winds are less predictable and less strong in those seasons.

- K

 

[lolachampcar]

Thanks Andy. I need to go back and re-evaluate my average usage on the latest firmware.

 

[jbcarioca]

There are so many variables in play that i cannot even speculate reliably on my car changes in efficiency. My impression is that the last few updates have been tweaking dual motor optimization and probably BMS issues as well. Of course i might be learning how to drive more economically also and the weather has been warmer. Anyway, I sense increase in efficiency. My last trip of 36 miles had 243 w/mile, the same trip four months ago was 325 mile or more. My driving is not the only change. I am sure that if I carefully recorded all the environmental data I'd find the vast majority of that improvement is due to changes in environmental conditions. Despite that i am very pleased. I think my car is getting better!

 

[ArtInCT]

We have owned hybrids for six years and post out mileage in fuelly.com

there is a sine wave of efficiency results. High in the summer and low in the winter when cold.

it will be interesting to see your efficiency numbers when the hawk blows this winter Andy.

i predict, not as good as now.

 

[Andyw2100]

there is a sine wave of efficiency results. High in the summer and low in the winter when cold.

it will be interesting to see your efficiency numbers when the hawk blows this winter Andy.

i predict, not as good as now.

Of course. That goes without saying!

Last winter, before torque sleep was implemented, I'm pretty sure the same trips were in the 400-450 Wh/mi range. (Just did a quick check--average for Feb 4 thru March 31, which was after torque sleep was released, was 364.) The lifetime efficiency is still 341 Wh/mi, because of the winter, even with all these trips between 240 and 320, give or take.

I fully understand the effect of the cold on efficiency. I was just suggesting that the differences in temperature in the two different groups of trips I included data for were probably not the reason for the efficiency improvement, because the temperatures were pretty similar. I never would have suggested that if one of the firmware versions had run from, say, March through May, and the other from June through August, as that would have been ridiculous.

 

[vgrinshpun]

Strangely, you did not include barometric pressure, which can vary wildly over the course of the day. Looking at min/max in my area, it's as much as 5.17%, which correlates to air density directly.

The reason I did not mention barometric pressure in my post, is that I could not conclude from Andyw2100’s data a reliable trend to determine the difference of air pressure in July/August vs. June/July. Assuming highher humidity and higher temperatures in July/August vs. June/July, on another hand, is a safe bet.

SAE also excludes humidity from STP/NTP.

Somewhat OT here, but SAE J1349 does require tests to be conducted at a specific RH of 50%, it is just that these tests have sufficient accuracy for a relatively wide range of values of relative humidity: +/-20%.

9.1.6 AMBIENT CONDITIONS
The test must be conducted at the following ambient conditions:
a. Air Temperature: 25 °C ± 10 °C
b. Barometric Pressure: 90-105 kPa
c. Humidity: 50 % R.H, +/-20.
d. Wind speed: less than 7 m/s (15 mph)
e. Road Surface: A closed course, with dry, flat, level hard-paved surface.

Somehow this seems to work for rating everyday things, include microprocessor controlled forced induction control, which needs realtime measurement of temperature and pressure. That also means whatever the effect of humidity is, it requires no active compensation.

Yes, the effect of air humidity on air density and, therefore drag, is not very intuitive, and people usually dismiss it without taking enough time to understand why.

As I mentioned in my original post I’ve found this empirically, trying to understand why my energy consumption was different for my morning commute, given near equal atmospheric conditions otherwise. So here is what I found after doing some digging.

According to the Avogadro Law equal volume of all gasses, at the same temperature and pressure, have the same number of molecules. Base on this, a gas with the higher molecular weight has proportionately higher density. Since molecular weight of dry air is 28.966, while molecular weight of the water vapor is 18, the water vapor is almost 40% lighter than dry air. So each molecule of water that is replacing other gasses in dry air reduces total density of the air mixture.

As an example, using this calculator, the relative humidity at 75°F air, with dew point of 50°F is 41.44%, while dew point of 70°F at the same air temperature yields relative humidity of 84.47%. Converting 41.44% RH to absolute humidity using this calculator produces absolute humidity of 9%. Relative humidity of 84.47% at 75°F is equivalent to an absolute humidity of 18.35%.

Finally, calculating molecular weight of the air mixture at 75°F with absolute humidity of 18.35% vs. 9.0% indicates that air mixture containing 18.35% of water vapor is 3.68% less dense than air mixture containing 9% of water vapor.

 

[brucet999]

More likely that your tires are getting broken in and range is increasing steadily due to that.

Steadily is the operative word here. That OP is noticing an abrupt decrease in Wh/mile after the FW upgrade appears significant.

 

[caddieo]

That OP is noticing an abrupt decrease in Wh/mile after the FW upgrade appears significant.

Although I do not have the detailed data used by the OP, I have also noted improved efficiency since V.7(2.7.56). My routine driving pattern in my city and county involves 30-45 mph streets with a couple of 50-55 mph main roads. I am confining my observations to this and not to days when I occasionally take the Interstate. I know my driving habits have not changed - usually 3-5 mph over posted and I do not hesitate to pass people doing 40 in a 45 (or worse) if conditions are right. My daily circuit almost always uses the same roads but occasional back-and forth runs between destinations make my daily total mileage vary from 20 to 45 miles or so. Previously, my rated miles used was always greater than actual miles traveled. Example: Rated miles-35/actual miles-30-33; and my Wh/m at the end of the day would be in the 280-300 range. Now, my actual miles and rated range miles often coincide. And if my actual mileage is less than rated, it is not as much. Per above example 35 rated vs. 34 actual. And every now and then (perhaps a third of the time) I will get actual mileage greater than the rated miles used. Again, using the above rough example - 35 rated miles vs. 36-38 actual miles. Also, at the end of the day, my final Wh/m reading is in the 260-70 range with an occasional drop into the 250's. I apologize for not quantifying this the way the OP did, but will probably start doing it just in case somebody wants numbers. Has anybody else noted this efficiency improvement?

 

[Andyw2100]

Although I do not have the detailed data used by the OP, I have also noted improved efficiency since V.7(2.7.56). My routine driving pattern in my city and county involves 30-45 mph streets with a couple of 50-55 mph main roads. I am confining my observations to this and not to days when I occasionally take the Interstate. I know my driving habits have not changed - usually 3-5 mph over posted and I do not hesitate to pass people doing 40 in a 45 (or worse) if conditions are right. My daily circuit almost always uses the same roads but occasional back-and forth runs between destinations make my daily total mileage vary from 20 to 45 miles or so. Previously, my rated miles used was always greater than actual miles traveled. Example: Rated miles-35/actual miles-30-33; and my Wh/m at the end of the day would be in the 280-300 range. Now, my actual miles and rated range miles often coincide. And if my actual mileage is less than rated, it is not as much. Per above example 35 rated vs. 34 actual. And every now and then (perhaps a third of the time) I will get actual mileage greater than the rated miles used. Again, using the above rough example - 35 rated miles vs. 36-38 actual miles. Also, at the end of the day, my final Wh/m reading is in the 260-70 range with an occasional drop into the 250's. I apologize for not quantifying this the way the OP did, but will probably start doing it just in case somebody wants numbers. Has anybody else noted this efficiency improvement?

This makes sense because it was one of the V7 releases that talked of torque sleep efficiency improvements in Classic Model Ss (which it appears from your sig you have.) I apologize for not recalling the exact details, but since they didn't apply to me I didn't make the proper mental note. But I do know with 100% certainty that this improvement was recent, and in one of the firmware 7 versions.

 

[caddieo]

This makes sense because it was one of the V7 releases that talked of torque sleep efficiency improvements in Classic Model Ss (which it appears from your sig you have.) I apologize for not recalling the exact details, but since they didn't apply to me I didn't make the proper mental note. But I do know with 100% certainty that this improvement was recent, and in one of the firmware 7 versions.

Thanks for your reply and insight. I had only skimmed over the release notes when the update occurred so I went to the car and read them again. The last item was indeed about torque sleep and improved efficiency. "Problem" solved.