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

[stopcrazypp]

You insist on some sort of actual verification of battery capacity as part of EPA testing it simply is not a stated requirement.

The presentation you linked yourself clearly says there is a battery capacity test that applies! It is not my own personal insisting.
"On-Dyno Battery Capacity Test: Extrapolation based upon dyno test, not from standard battery test data
– 55 MPH steady-state speed, in 50min segments with 10 min rests.
– Accel and decel specifications defined"

The two figures Tesla MUST prove are MPGe (based on AC recharging input) and range. I am not saying either figures Tesla have published are incorrect.

They can report range by doing UDDS + HWFET * by 0.7 (to approximate the new 5 cycle tests) measuring DC draw and using their own knowledge of total DC capacity. (In the same way an ICE manufacturer can with an MPG figure and knowing what size tank they put in the vehicle).

Here's the problem: the EPA does not intake a usable battery capacity number (separate from test data). They may separately intake the brochure nameplate number (85kWh or 60kWh for example), but no where do they ever take in a usable battery capacity number. This number is taken from the dyno based capacity test mentioned above (note how they say "not from standard battery test data"). The test data will be passed to EPA as part of manufacturer filing.

This is different from gasoline case where the brochure nameplate tank size is all they need to get the range. Also the gasoline car range is not a legally required number on a new car sticker:

However, the EPA range number is legally required for EVs:

However after a bit more digging it would seem more likely the more recent proposal (2012) currently in J1634 is to take a fully charged car run four rounds of UDDS, two rounds of HWFET cycles (again with a approximation factor). Then just use a steady 55mph of "filler*" for the rest of the test and "official range" is simply when the car stops.
(* I can't help but think 200 miles of steady 55 mph does wonders for the reported range)

Note under either regime this doesn't tell us as consumers how much usable battery the car has, because we have no way of knowing the efficiency rates of the chargers / chemical losses during battery recharge.

(We are getting way OT here, this should be in the battery thread...)

I'm not saying the consumers will ever see this number anywhere (they won't), but that the measured usable DC Wh capacity number is reported to the EPA as part of the test data (or can easily be derived from DC wh/mi and range at that cycle). I know that they easily measure this number during the test because they necessarily have to set up a DC consumption probe on the car just to get the efficiency numbers. So if we are able to get the actual raw EPA test data, this will show up there.

In your example of using 55mph filler cycles until depletion, by the end of the test, the DC probe would have measured a total Wh consumed during test, which is where we can get the usable capacity. If Tesla disabled brick protection during test (as wk057 speculated originally) that capacity would not only be ~5% higher, but the range would also be ~5% more.

Thus if EPA were to do that test independently with a production car, that discrepancy would show up (and would be an issue EPA is obligated to address, given the range number shows up on the new car sticker).

I would be interested to read the source document for that 55mph filler procedure too if you have a link.

 

[smac]

In your example of using 55mph filler cycles until depletion, by the end of the test, the DC probe would have measured a total Wh consumed during test, which is where we can get the usable capacity. If Tesla disabled brick protection during test (as wk057 speculated originally) that capacity would not only be ~5% higher, but the range would also be ~5% more.

Thus if EPA were to do that test independently with a production car, that discrepancy would show up (and would be an issue EPA is obligated to address, given the range number shows up on the new car sticker).

I would be interested to read the source document for that 55mph filler procedure too if you have a link.

Can we clear something up. I don't believe Tesla needed to allow the bricking protection to be removed. (And I took wk057's comment to be in jest). What I am saying is under the current (and still by the standards bodies own admission slightly flawed) testing regimes for BEVs which are still in flux to ensure they meet real world consumer expectations, a Model S will indeed hit it's numbers. In fact I'm sure give 100 miles at 55mph, I would best 208 miles range of my car. I got to 190 without trying, with A/C on, heated seats on, radio on, with an out and back trip with the majority at 70 mph.

What I am saying is the EPA (who don't perform the tests) figures in no way give us is a calculable method for determining battery size.

Slide 5, manufacturers can opt to use the MCT
https://www2.unece.org/wiki/download/attachments/5801244/EVE-05-12e.ppt?api=v2

•The MCT consist of repeated sets of UDDS, HFEDS and Constant Speed Cycle (CSC))

Slide 6 says the same test may be used to determine efficiency/ consumption too.

A summarized explanation can be found here :
How the EPA determines an electric vehicle’s range - not as simple as it sounds - Torque News


I think you are completely overthinking it with DC probes and such. A much more straightforward approach that would be applicable to any vehicle and easily tested anywhere with repeatable results would look something like this:

Step 1) Charge car fully
Step 2) Drive car on a dyno using a mixture of tests, then 55 mph steady cruise (constant speed cycle), until it stops.
Step 3) Record how far the dyno's odometer went up.
Step 4) Recharge car from the mains. Record kWh drawn on the charger.
Step 5) Submit answer from step 3 as official range, and 4 as the official consumption over that distance.

Unless I'm going mad this is exactly what is being suggested here:
http://www.gaccmidwest.org/fileadmi...InfoV_Presentations/Lohse-Busch_eMobility.pdf (Page 47)


Look I wouldn't really care too much, but after 4 days of ownership Tesla decided to flash my car to a new basis for the Wh line. (and why I've taken such an interest in the whole topic). I've heard your argument on this before that it is somehow to better represent NEDC figures (though this is flawed in itself, as it should have gone down not up, under NEDC rules here).

The resultant messing around with the "numerical optics" of the dash has nasty subtle side effects, like completely screwing over the range assurance to the point it's out by 50+ miles constantly trying to send you on pointless detours every time you use sat nav. (The plus side in my case is even driving in snow I can beat the range indicated on the dash, so I have no need for the cr@ppy range assurance thing, but I can't switch it completely off and it's very annoying with the popup messages).

So OK I have a very special circumstance to be upset with Tesla's fooling around with numbers, but it's just one other data point where caveat emptor must be applied when looking at any number Tesla publish. 0-60 times, 1/4mile times, model comparative range figures, battery sizes, power figures, .... Each of these will impact a different set of buyers differently, but sooner or later it will push another group of supporters buttons.

IMHO this is just so damaging to the brand, many of whom's patrons are clever technical people that will figure out these marketeer sleight of hand sooner or later, and it feels so unnecessary.

We've set off down this path now though, and I'd happily put this behind us if we saw any change in behavior, but as far as I can see it's just more of the same. I just hope with the Model 3 the leopard can change it's spots.

 

[AWDtsla]

Note under either regime this doesn't tell us as consumers how much usable battery the car has, because we have no way of knowing the efficiency rates of the chargers / chemical losses during battery recharge.

First of all, this is easily measurable, and I've done it several times. This is no harder than a freshman physicis lab.

Not only that but the car directly reports usable battery just like it does for power output. Now if someone is claiming either of these is incorrect that's even more scandalous, but I don't think they are. People are just failing to do the units conversion.

 

[McRat]

... after 4 days of ownership Tesla decided to flash my car to a new basis for the Wh line ....

Question: Doesn't the Wh number "learn" from your driving? I'm not sure I've seen a modern vehicle that doesn't learn it's MPG average (or EV range).

<--- Not a Tesla owner yet. It's still on our potential "list". Nearly pulled the trigger, maybe later in the year.

 

[smac]

Question: Doesn't the Wh number "learn" from your driving? I'm not sure I've seen a modern vehicle that doesn't learn it's MPG average (or EV range).

<--- Not a Tesla owner yet. It's still on our potential "list". Nearly pulled the trigger, maybe later in the year.

Not as far as I can tell. It seems baked into the firmware and variant (60, 70, 85, 70D, 90D etc.) dependent.

Don't let my moaning put you off, the car is great (QC teething issues aside), and as long as you do your research you know what you are getting in to and therefore make an informed decision, everyone comes away happy. It's a shame we can take things at face value, but then that's true of most things in this world from big corporate businesses, so maybe I expect too much.

- - - Updated - - -

First of all, this is easily measurable, and I've done it several times. This is no harder than a freshman physicis lab.

Not only that but the car directly reports usable battery just like it does for power output. Now if someone is claiming either of these is incorrect that's even more scandalous, but I don't think they are. People are just failing to do the units conversion.

So the test result has 2 outputs under MCT.

1) Miles driven doing a series of standardised tests + 50 mph steady state until car stops = Mileage.
2) Total AC kWh needed to recharge the battery, = Overall efficiency

How can you derive DC kWh usable from these two facts?

Post the formula. and I'll gladly acknowledge the failings in my freshman physics classes.

 

[AWDtsla]

So the test result has 2 outputs under MCT.

1) Miles driven doing a series of standardised tests + 50 mph steady state until car stops = Mileage.
2) Total AC kWh needed to recharge the battery, = Overall efficiency

How can you derive DC kWh usable from these two facts?

Post the formula. and I'll gladly acknowledge the failings in my freshman physics classes.

You take Tesla's value for the DC usable number, which is your "rated" line in the energy analyzer. You can validate that number is the EPA number by working backwards using inputs like measured vampire loss and and measured charging efficiency.

Now the DC usable capacity is simply the product of rated miles and rated energy usage. Simply take the rated miles consumed in your test.

All of that works.

Now the real problem is that the trip meters under-report DC usage, even using Tesla's own rated miles as the reference. On a recent trip I had the trip meter report 48.8 kWh used, but the rated mile math says that 52.8 kWh was consumed. I can get that result with two different forumla's and get EXACTLY the same result. I can use either the charge percentage difference and the amount of charge each percent is, or I can take Tesla's rated Wh/mile and compare to reported rated miles. It all agrees. There is potentially a calibration problem with the BMS thus each number reported is equally wrong, however the rated mile math MUST add up eventually, you should not have accumulating errors. If every time I do this calculation it always loses kWh, then that's actually what's happening and it's not a calibration problem. You posted some achieved numbers above. You cannot trust those numbers.

 

[smac]

You take Tesla's value for the DC usable number, which is your "rated" line in the energy analyzer. You can validate that number is the EPA number by working backwards using inputs like measured vampire loss and and measured charging efficiency.

Now the DC usable capacity is simply the product of rated miles and rated energy usage. Simply take the rated miles consumed in your test.

All of that works.

I would agree if the rated line matched any semblance of reality. Mine is at 345 wH/mi, which is frankly a joke, and contradicts the official "party line" (and physics) that an S60 > S85 in terms of efficiency.

Now the real problem is that the trip meters under-report DC usage, even using Tesla's own rated miles as the reference. On a recent trip I had the trip meter report 48.8 kWh used, but the rated mile math says that 52.8 kWh was consumed. I can get that result with two different forumla's and get EXACTLY the same result. I can use either the charge percentage difference and the amount of charge each percent is, or I can take Tesla's rated Wh/mile and compare to reported rated miles. It all agrees. There is potentially a calibration problem with the BMS thus each number reported is equally wrong, however the rated mile math MUST add up eventually, you should not have accumulating errors. If every time I do this calculation it always loses kWh, then that's actually what's happening and it's not a calibration problem. You posted some achieved numbers above. You cannot trust those numbers.

I agree with this part. I'd actually like to repeat my 100%-0% trial, now armed with my CAN logger. I think this sort of rigorous testing is (unfortunately) required to try and work out what is happening behind the scenes of the IC. (Short of doing the stuff wk is doing by physically disassembling the packs)

But going back to my original assertion, for Joe Public this sort of testing is well out of scope, so purely using legislatively mandated figures we will always struggle to work out usable pack kWh.

 

[AWDtsla]

I would agree if the rated line matched any semblance of reality. Mine is at 345 wH/mi, which is frankly a joke, and contradicts the official "party line" (and physics) that an S60 > S85 in terms of efficiency.

The absolute value is not relevant to the correctness of the math. Although I do agree it's complete ********. My car is 310. Maybe if I can go downhill both ways.

 

[smac]

To give you an idea of how far out an EU 60 is:

I should note this is after a start up cycle (which really hurts range for me doing < 10 miles per day, generally with regen capped)

This is from a longer run, and after torque sleep was released for RWD (which made a big difference)

 

[stopcrazypp]

What I am saying is the EPA (who don't perform the tests) figures in no way give us is a calculable method for determining battery size.

Slide 5, manufacturers can opt to use the MCT
https://www2.unece.org/wiki/download/attachments/5801244/EVE-05-12e.ppt?api=v2

Slide 6 says the same test may be used to determine efficiency/ consumption too.

Next line in slide 5 says this suggests same DC capacity determination as noted before (confirmed in last link):

This is similar to the approach being proposed by the WLTP for shortening the duration of the BEV testing

This is what it says on slide 6:

In addition to using J1634 to determine battery electric vehicle range, the EPA also applies J1634 to determine energy consumption.
Vehicle AC Energy Consumption = AC Energy to the Charger for Recharge / Distance Traveled
Vehicle DC Energy Consumption = DC Energy from the Battery while driving / Distance Traveled

The only way they can get DC energy is to have a DC probe on the car.

A summarized explanation can be found here :
How the EPA determines an electric vehicle’s range - not as simple as it sounds - Torque News

Key point in that article:

As the vehicle is driven over several iterations of the standardized drive cycles on a chassis dynamometer, the EPA measures the energy drawn from the battery pack over each cycle. The vehicle’s total usable battery energy (which is obtained by charging the fully depleted battery following the MCT test to its fully charged state, and is not the total energy content of the battery) is then used to determine the electric range for each drive cycle type. Overall efficiency is calculated from the cycle data by accounting for the charging efficiency of converting AC electricity to DC electricity to be stored in the battery.

I think you are completely overthinking it with DC probes and such. A much more straightforward approach that would be applicable to any vehicle and easily tested anywhere with repeatable results would look something like this:

Step 1) Charge car fully
Step 2) Drive car on a dyno using a mixture of tests, then 55 mph steady cruise (constant speed cycle), until it stops.
Step 3) Record how far the dyno's odometer went up.
Step 4) Recharge car from the mains. Record kWh drawn on the charger.
Step 5) Submit answer from step 3 as official range, and 4 as the official consumption over that distance.

Yes, they measure a composite AC Wh from the grid also at the end of the test (this gives them data for determining charging efficiency), but they need the DC probe in order to measure the consumption of every cycle (the car sticker reports both city and highway numbers separately, plus they need it to do the 5-cycle math).

Also, the steady state 55mph drain does not reflect any of the 5-cycles (it is only done in order to get a battery capacity number), so they can't just use the range number from the end of the test as the sticker number. Rather they use the measured battery capacity and then use the composite 5-cycle Wh/mi to determine the final range. This is what the article says they do.

Your suggested approach that uses only AC Wh (where no DC probe exists at all) only works for the true long form test (meaning they do a 17+ hour test for city and then a test for highway cycle separately; potentially also the other 3-cycles too).

Unless I'm going mad this is exactly what is being suggested here:
http://www.gaccmidwest.org/fileadmi...InfoV_Presentations/Lohse-Busch_eMobility.pdf (Page 47)

See page 44 in your last link:

Short-Cut Method in General:
1. Find battery capacity (on-dyno)
2. Run test cycles (UDDS, HWY, US06) to find Efficiency
3. Use consumption and capacity data to find Range

So far every document you linked says the EPA measures DC Wh consumption and also determines battery capacity. So once again, if we can get our hands on the actual EPA documents, I suspect we will either see the measured DC Wh capacity directly or it can be easily derived from the final DC Wh/mi and range numbers.

 

[AWDtsla]

I don't know what you two are bickering about, it's irrelevant what the EPA did, it only matters that the EPA range rating was baked into the rated miles/consumption calculation. Could have been 100 miles, or 1000 miles with the same battery, doesn't matter. The battery still comes up short, just like the power. The car itself tells you. It just tells you something very different than teslamotors.com

 

[stopcrazypp]

I don't know what you two are bickering about, it's irrelevant what the EPA did, it only matters that the EPA range rating was baked into the rated miles/consumption calculation. Could have been 100 miles, or 1000 miles with the same battery, doesn't matter. The battery still comes up short, just like the power. The car itself tells you. It just tells you something very different than teslamotors.com

The original talk about EPA was about getting a usable Wh capacity number from the raw EPA test data to compare with wk057's 77kWh.

The rated miles/consumption calculation number has nothing to do with that. AFAIK that number is made up and does not have to match EPA (EPA test does not care at all what the display says, it tests until the car can't continue anymore). I believe that number had been changed in updates and also varies depending on car, so largely irrelevant in terms of determining usable capacity. I remember back then people trying to do all sorts of math to get it to work and I believe it was found out was non-linear near the end of the charge.

 

[ggnykk]

Looks like ALL horsepower numbers are gone in the latest update to the website.

 

[whitex]

Looks like ALL horsepower numbers are gone in the latest update to the website.

Gone from the design studio. Hp is still available in specs (Model S | Tesla Motors).

 

[Andyw2100]

Looks like ALL horsepower numbers are gone in the latest update to the website.

Gone from the design studio. Hp is still available in specs (Model S | Tesla Motors).

You two missed Groundhog Day by about a week.

Haven't stick to this topic recently. I just found out they stop showing horsepower rating for all model.

I still see it here:
Model S | Tesla Motors

 

[thegruf]

Looks like ALL horsepower numbers are gone in the latest update to the website.

Now all they have to do is use the same test method for acceleration for each model to allow the average customer to make a reasoned choice without having to resort to technical articles about drag strip racing and forums to extablish what the true delta is (and that is even assuming they spot and read the small print even then).
(Have to say that in many ways this one irks me even more than the hp BS).

 

[MikeBur]

Now all they have to do is use the same test method for acceleration for each model to allow the average customer to make a reasoned choice without having to resort to technical articles about drag strip racing and forums to extablish what the true delta is (and that is even assuming they spot and read the small print even then).
(Have to say that in many ways this one irks me even more than the hp BS).

I believe some have figured this out: http://www.teslamotorsclub.com/show...han-expected?p=1362008&viewfull=1#post1362008

"Seeing this, I feel better knowing that 20K+ would only get me a car half a second faster"

 

[lolachampcar]

#166 one of many plots used in generating the opinion below

I just have to revisit this thread.
Somewhere along the line, someone tried to convince me that putting bigger motors in the D had value because those bigger motors could deliver more torque in the low speed realm prior to the battery limit coming into play.

Well fast forward to today and we now have the car's own CAN bus data to play with (thanks to WK's work).
The D is torque limited in the first second. Period. Full stop. To boot, that torque limit does not appear to be anywhere near the specified torque capability that Tesla was promoting.

I thought the Big Torque Motor argument was without merit on the basis that the tires could not handle it. We now KNOW the Big Torque Motor argument has no merit because the people that are marketing them are not allowing them to generate anywhere near the torque they are marketing (before the other Elephant in the Room that is the battery limit comes into play).

 

[Andyw2100]

#166 one of many plots used in generating the opinion below

Just pointing out for those that may not be able to tell, the #166 is in a different color, and thus is a link to a post in a different thread.

The "plot" lola is referring to is an actual plot--a graph. Considering the debates that raged in this thread I feared that someone not realizing there was a link might not realize that and might take "plot" to mean "diabolical plan", etc.

 

[lolachampcar]

Good Point!