Stop the Press! Tesla announces REAL HP numbers for P85D and P90L

[McRat]

Well, not if the 400 HP EV is a Tesla that really only ever makes 310 HP at the battery and even less at the wheels thanks to ECE R85.

(You're coming to this conversation a bit late, McRat, so you may not understand why that is funny*, but I expect others will.)

* Feel free to substitute "pathetic" for "funny" if you prefer.

I admit I cheated and did not read the 250+ pages.

And my comments were not Tesla specific, which actually makes my posts "off-topic".

We were seeing a similar error when racing high powered turbo diesels. The numbers in the standard calculators showed more HP than dynos did due to a similar phenomena. The HP "curve" of the turbo diesel is fairly flat. Like an EV, the Torque Curve is a declining slope.

 

[Andyw2100]

I admit I cheated and did not read the 250+ pages. .

I'll give you the "Reader's Digest" version, which is sure to include my personal bias, though I will attempt at least to some extent, to be objective.

--Tesla marketed the P85D flagship product as a 691 HP car. Every article written referred to the car as a 691 HP car. Tesla made no attempt to correct this.

--It was eventually determined that the P85D could not make 691 HP anywhere in the drive train at any time ever. I believe the expectation is that the max HP is about 550, at the battery.

--Tesla explains that this is due to a battery limitation, and not limitation of the motors, and that the HP is "Motor HP." They say that given enough juice, the motors could produce 691 HP.

--There may or may not be a European standard - ECE R85 - that allows for rating the motors this way. (There is much debate on that point.)

--Many who purchased the car thinking it was a 691 HP car feel misled. It meets the 0-60 acceleration specs, but does not accelerate the way a 691 HP car would at higher speeds.

--Many believe Tesla had intended to correct this shortcoming with a software update, and they had even announced a coming software update to "improve high speed performance" on their website. That never materialized.

--Many believe that software update turned into the "Ludicrous" mode because Tesla realized hardware was required as well. That's available for an extra $10,000, or for a $5,000 upgrade charge to P85D owners. Ludicrous will get the car closer to the original 691 HP, but does not get it all the way there.

--In other countries, like Denmark and Norway, consumer advocacy groups are involved in this controversy.

--Eventually Tesla updated the figures on their website to indicate that the car was only capable of making 463 HP, with an asterisk that reads "Battery limited maximum motor shaft power." This is a great step to prevent future purchasers from being misled, but does nothing for those who purchased the P85D thinking that it was a 691 HP car and not a 463 HP car.

 

[wk057]

I'll note that the Model S power train CAN isn't like crappy OBDII CAN crap. The data is high rate with virtually no lost or delayed frames. Additionally, the data is super accurate. The speed, for example, factors in tire slip/spin. Provable by comparing speed to motor RPM during a traction control off launch.

The battery power reading is also high speed and high resolution with no delays or dropped frames.

In my logs the frame timings are generally better than +/- 1ms vs the expected interal. Tight for the highest rate data too.

 

[NOLA_Mike]

Well, not if the 400 HP EV is a Tesla that really only ever makes 310 HP at the battery and even less at the wheels thanks to ECE R85.

Too soon...

 

[McRat]

I'll note that the Model S power train CAN isn't like crappy OBDII CAN crap. The data is high rate with virtually no lost or delayed frames. Additionally, the data is super accurate. The speed, for example, factors in tire slip/spin. Provable by comparing speed to motor RPM during a traction control off launch.

The battery power reading is also high speed and high resolution with no delays or dropped frames.

In my logs the frame timings are generally better than +/- 1ms vs the expected interal. Tight for the highest rate data too.

Gotcha.

But how does it figure wheel slip when all four are spinning? Or is the AWD Tesla set up so it cannot do that, ie no LSD?

 

[lolachampcar]

McRat,
To add color to what WK said, you are looking at data gathering as if PiD requests via OBDii were being processed by the ECM. In this case, we are logging chassis CAN bus messages that are broadcast by different devices (front motor, rear motor, battery, etc.) at 100 Hz. These are the same data the ECU uses to perform its tasks (including passing it though with delay to the OBDii port on request). Going straight to the source gives very accurate real time data.

WRT TC, well now we are talking. I too want to see how things are traction limited and how AWD control determines slip. My initial impressions are that the control is using jerk to identify slip initiation and then balancing power reductions front to rear such that only one side of the car is slipping at a time. I'm currently collecting a lot of rain acceleration runs to try to better understand how Tesla is doing it. One thing is certain; Tesla was not fibbing when they said their traction control time constant was much smaller than ICE. I'm seeing incredibly fast power fluctuations which is entertaining.

Lastly, Tesla runs open diffs and uses wheel braking to control slip as well as power regulation.

 

[wk057]

Btw, I keep forgetting to post this....

I kind of want to find a good local dyno now...

 

[Andyw2100]

I kind of want to find a good local dyno now...

And an automotive journalist?

 

[lolachampcar]

Others are now working on some of the logging stuff so I started its own thread.
Chassis CAN Logging To ASCII Text Plus Graphing

 

[AWDtsla]

Btw, I keep forgetting to post this....

View attachment 109592

I kind of want to find a good local dyno now...

I am so damn jealous right now. Shows everyone how hard it would have been for Tesla to make this a true drivers car.

 

[McRat]

Just in case this was not mentioned in the 250+ pages:

This same situation came up three times that I'm personally aware of -

The 1999 Ford Mustang Cobra. The 320HP factory claim was wrong. People noticed it immediately. IIRC, Ford offered to "fix" the cars for free to increase the power, but I'm not sure about buybacks.
2003? Mazda RX-8. Mazda did buybacks on these for overstating HP. Ironically one of the magazines gave it Car of the Year.
2001? Mazda Miata. Buybacks also.

 

[whitex]

Andyw2100, good digest though you may have missed a detail or two:

--It was eventually determined that the P85D could not make 691 HP anywhere in the drive train at any time ever. I believe the expectation is that the max HP is about 550, at the battery.

As you stated later in the same post Tesla has had this number published on Model S | Tesla Motors since end of November. P85D battery limited motor power is listed at 463hp, so the expectation is not 550hp.

--Many who purchased the car thinking it was a 691 HP car feel misled. It meets the 0-60 acceleration specs, but does not accelerate the way a 691 HP car would at higher speeds.

Eh, no it doesn't meet 0-60. There was another "gotcha" footnote added by Tesla at some later date that states today that "0-60" for P models is measured with 1ft rollout, so that spec is more like 8-60mph, rather than 0-60. Since only the P models use that method, and the 85D and the rest use actual 0-60, it makes the P appear to perform better than it actually does, presumably in order to help justify the additional $20K in price differential. By similar logic I can say my P85D does 0-120mph in 1.0s, then a few months later add "with ~1/4mile rollout" to prove I was right.

--Many believe Tesla had intended to correct this shortcoming with a software update, and they had even announced a coming software update to "improve high speed performance" on their website. That never materialized.

That was not a pure "belief", but a reasonable interpretation of the note in the design studio that promised an OTA that would much improve highway speeds acceleration and provide 0-60 way faster than anyone has ever seen outside the factory.

 

[llavalle]

Btw, I keep forgetting to post this....I kind of want to find a good local dyno now...

Ok, I'm officially super jealous now. With all that snow here, disabling TC for some nice drift would be VERY nice.

 

[Andyw2100]

Andyw2100, good digest though you may have missed a detail or two:


As you stated later in the same post Tesla has had this number published on Model S | Tesla Motors since end of November. P85D battery limited motor power is listed at 463hp, so the expectation is not 550hp.

Thanks.

The 550 number was what some of the knowledgeable people here had come up with as the max number possible long before Tesla published the 463 HP number. Some also think the Tesla number is actually conservative.

Eh, no it doesn't meet 0-60. There was another "gotcha" footnote added by Tesla at some later date that states today that "0-60" for P models is measured with 1ft rollout, so that spec is more like 8-60mph, rather than 0-60. Since only the P models use that method, and the 85D and the rest use actual 0-60, it makes the P appear to perform better than it actually does, presumably in order to help justify the additional $20K in price differential. By similar logic I can say my P85D does 0-120mph in 1.0s, then a few months later add "with ~1/4mile rollout" to prove I was right.

Yes, I purposely avoided the whole 1 foot roll out controversy. But you are correct in that it was another area in which some people feel misled.

That was not a pure "belief", but a reasonable interpretation of the note in the design studio that promised an OTA that would much improve highway speeds acceleration and provide 0-60 way faster than anyone has ever seen outside the factory.

Well, whether or not Ludicrous was really the result of Tesla not being able to deliver on the promise made on the website via software alone is what I meant when I said "many believe." There was no question about what Tesla had published on the website, and really no reasonable question as to the fact that it was never delivered.

 

[wk057]

I am so damn jealous right now. Shows everyone how hard it would have been for Tesla to make this a true drivers car.

Ok, I'm officially super jealous now. With all that snow here, disabling TC for some nice drift would be VERY nice.

I'll note that the P85D in Insane mode is virtually uncontrollable with it set to "TcOffEspOff"... my poor tires...

 

[llavalle]

I'll note that the P85D in Insane mode is virtually uncontrollable with it set to "TcOffEspOff"... my poor tires...

Yeah... I can imagine...

 

[stopcrazypp]

Just in case this was not mentioned in the 250+ pages:

This same situation came up three times that I'm personally aware of -

The 1999 Ford Mustang Cobra. The 320HP factory claim was wrong. People noticed it immediately. IIRC, Ford offered to "fix" the cars for free to increase the power, but I'm not sure about buybacks.
2003? Mazda RX-8. Mazda did buybacks on these for overstating HP. Ironically one of the magazines gave it Car of the Year.
2001? Mazda Miata. Buybacks also.

All three cases had been mentioned in some respect through the threads discussing this matter.

Here's my comment on this thread, basically unlike all those cases, where power was lost vs the rating for various reasons (Cobra because parts different in production version, Mazda because of US spec vs Japanese spec emissions equipment), no power was lost in the Tesla case. This case was the difference between what some people expected and how the power was rated.

The distinction may seem subtle, but is really critical in terms of the law. If there was actually power lost in consumer cars (as was the case of examples by others of the Ford Mustang and Mazda Miata), that is a really straightforward case, because the number is literally false (the car does not make the power even under the automaker's own rating system). In this case, however, it is not so straightforward. The number is not false, but it may be misleading, and this makes a big difference. In the USA, determination of misleading would be survey based and you need a certain percentage of people misled to have a case. Of course in Denmark, it may be different.

I discussed it in more detail here:
http://www.teslamotorsclub.com/show...hread/page51?p=1175559&viewfull=1#post1175559

 

[AWDtsla]

I'll note that the P85D in Insane mode is virtually uncontrollable with it set to "TcOffEspOff"... my poor tires...

Add ~15% wheel slip before Tc and ~20 degrees of yaw angle before Esc engagement. That should do the trick.

 

[3mp_kwh]

I'll note that the P85D in Insane mode is virtually uncontrollable with it set to "TcOffEspOff"... my poor tires...

Mine is stuck at "TcOnTurnKillPower"

 

[whitex]

Here's my comment on this thread, basically unlike all those cases, where power was lost vs the rating for various reasons (Cobra because parts different in production version, Mazda because of US spec vs Japanese spec emissions equipment), no power was lost in the Tesla case. This case was the difference between what some people expected and how the power was rated.

So advertising power measured with Japanese emissions equipment was deemed misleading for people who bought it with US emissions. So how is publishing motor power measured while connected to an industrial power supply not misleading to people who bought it with a battery? Assume (Japanese emmisions = grid power) and (US emissions = battery power).