P3D+ Acceleration

[SleeperService]

The way I've been measuring the performance of my car (P3D Stealth, if my name didn't give it away) is to use my phone camera set at 240 frames per second and then recording the splits at various intervals frame by frame. From my calculations the acceleration from 0-7 mph is actually a bit harder than the average from 0-60, just over 1 G in fact. My data is HERE for anyone that might be interested. My "start time" for the calculations is the frame just before the speedometer clicks over from 0 to 1 mph. Using an OG Pixel and the app "SBS Viewer" if it's important. Perfect measurement method? Probably not. Pretty consistent and good enough for me? Yep. Also: long time lurker, first time poster: Red 2018 P3D Stealth. Great car...

 

[AlanSubie4Life]

From my calculations the acceleration from 0-7 mph is actually a bit harder than the average from 0-60, just over 1 G in fact. My data is HERE for anyone that might be interested. My "start time" for the calculations is the frame just before the speedometer clicks over from 0 to 1 mph

Cool. This seems fine, though obviously subject to refresh rate/update rate issues, as I'm sure you've considered.

At a minimum, though, I would modify your calculations to calculate 1-5mph acceleration rather than 0-5mph - because your method is measuring the time to accelerate from 1mph to 5mph, and you have no information about 0-1mph. You'll find the acceleration with that modification is pretty much in line with the ~0.85g number (I came up with 0.91g with your spreadsheet). (With VBOX it actually looks like slightly higher acceleration than anywhere else from 2-4mph, but slower from 0-1mph - but all of that could be simple measurement error...but interesting all the same.)

0-1mph takes about 0.1sec (100ms). Hard to say precisely, but for sure greater than 50ms.

You could put a couple feet of labeled tape on the ground next to a wheel, and have a mounted GoPro to measure the missing 0-1mph time value with precision, I suppose. I assume they have high frame rate capabilities too, though maybe not high enough. Practically it might be difficult I suppose.

 

[bhzmark]

Forget the plots then.
Just look at the times. I see no reason to think those are seriously off.

Based on my VBOX experience, that is not hard to believe. Just think the Dragy data is pretty suspect. a vs. t might be fine (all the issues with the v vs. t plots may be simply plotting issues). The integral of a with respect over time should equal v vs. t. But it does not, in any of the Dragy plots I have seen...the noise on the a vs. t data is to some extent expected, but that wouldn't have much effect on the final 0-60 results as it would be averaged out.

 

[bhzmark]

One data point of validation.

 

[AlanSubie4Life]

Forget the plots then.
Just look at the times. I see no reason to think those are seriously off.

Probably the times are pretty good. Would be interesting to see how the two boxes (and the proposed high rate camera method) compare on the same run, though.

It's just tough because people look at the plots and sometimes draw inaccurate conclusions (eg - the 0-10mph is really fast, based on velocity vs. time data).

EDIT: The easy thing would be for Dragy to fix their plotting to make it self-consistent. Would not be hard. If their data is good everything else should follow.

 

[bhzmark]

And

 

[MagnusMako]

Probably the times are pretty good.

Yes, they are definitely good. Have you ever looked inside the Dragy app? You can see the leaderboard times for dozens of Teslas and they are on par for advertised or even higher. Brooks from DragTimes has used both a Dragy and VBOX with same times, or even better times with the VBOX. You can see his comparison video and times with his Hurcan right in line with what it should be. I also snipped a ss of his time comparisons and you can see how dead on they are between each other during testing of both in same runs. Also, my runs are all very consistent with each other and in line with what I've seen reported everywhere else. I know you're highly concerned with the plots not looking right but I think it's safe to say the times themselves are pretty accurate, which I think is really what matters the most to most people.

 

[SleeperService]

Cool. This seems fine, though obviously subject to refresh rate/update rate issues, as I'm sure you've considered.

At a minimum, though, I would modify your calculations to calculate 1-5mph acceleration rather than 0-5mph - because your method is measuring the time to accelerate from 1mph to 5mph, and you have no information about 0-1mph. You'll find the acceleration with that modification is pretty much in line with the ~0.85g number (I came up with 0.91g with your spreadsheet). (With VBOX it actually looks like slightly higher acceleration than anywhere else from 2-4mph, but slower from 0-1mph - but all of that could be simple measurement error...but interesting all the same.)

0-1mph takes about 0.1sec (100ms). Hard to say precisely, but for sure greater than 50ms.

You could put a couple feet of labeled tape on the ground next to a wheel, and have a mounted GoPro to measure the missing 0-1mph time value with precision, I suppose. I assume they have high frame rate capabilities too, though maybe not high enough. Practically it might be difficult I suppose.

I've updated the spreadsheet taking your suggestion into account. I've now got "0 mph" as the frame where you can JUST see the power bar start to move in the video and my original "0 mph" is now 1 mph. Makes sense to me. Also shows that it IS pretty doggy off the line from 0-1. Probably pulling torque to stop tire slip between my MXM4s and the cold road!

 

[AlanSubie4Life]

I've updated the spreadsheet taking your suggestion into account. I've now got "0 mph" as the frame where you can JUST see the power bar start to move in the video and my original "0 mph" is now 1 mph. Makes sense to me. Also shows that it IS pretty doggy off the line from 0-1. Probably pulling torque to stop tire slip between my MXM4s and the cold road!

Cool. You have better information now. Power bar seems like a decent method. And basically that data looks pretty consistent with what I have seen with the VBOX, and matches the torque/acceleration plots from Dragy (initial ramp from low torque to a local maximum torque, followed by a slight decrease to the steady-state torque which persists to around 45mph). I wonder if a tiny amount of time with your method could be shaved off by really slamming the pedal.

There has been a lot of discussion here about why torque is initially pulled, and a couple hypotheses exist, but it is NOT due to traction limitations. The MXM4 can easily hold 0.9g, and if you go to stickier tires, you see exactly the same result.

@dfwatt
Looks like you removed your post because you answered your own question, but I was going to say “Pretty sure that’s a lambo, dude!”

 

[SleeperService]

@AlanSubie4Life
Yeah, it is interesting about the initial ramp. I wouldn't ACTUALLY want to do it unless/until I'm nearly ready to replace my tires...but I always fantasize about four wheel burnouts that are completely impossible right now. Because... #fastcar

I'm anxiously (like most of us are...) awaiting the "5% boost" just to KNOW where the curve is fattened up. Would be nice if it's kept a bit more robust from 45 to 100, but we shall see. Anyone have insight into why the P3D doesn't hold its power peak flat like the 3D and ramps it down pretty aggressively from 60 - > 100? I KNOW it's still fast in that region compared to most cars, but it's the fall off from 450 hp to ~350 hp that makes it FEEL "not fast in this speed region."

 

[SleeperService]

I'm going to go ahead and speculate an answer to my own question above: increasing back EMF as the motor RPM climbs. The 3D has enough voltage overage from the batteries to counteract it all the way up, but the P3D is voltage limited.

 

[brianman]

Anyone know the 0-60mph time is for the P85D?

Officially it was 3.2s before Ludicrous upgrade and 2.9s after. The P90D was 2.8s with Ludicrous.

 

[Oyinko]

I got the peak power increase this morning and went for a quick spin with a cold battery charged at 80%. It felt a tad faster, not at the initial pull but between 15mph and 40mph, you can feel the car pulling much more. I can't wait to see some of your dyno test to put some numbers.

 

[drewgray]

English Racing had already dynoed the software update on their P3D, so we know where the power is added.

Our Performance Model 3 Tesla

 

[stickman]

View attachment 390922 English Racing had already dynoed the software update on their P3D, so we know where the power is added.

Our Performance Model 3 Tesla

sweet. That's exactly where you'd (or at least I'd) want it. Usable power for passing etc (and out of turns on the track). Faster off the line I don't need.

 

[SleeperService]

sweet. That's exactly where you'd (or at least I'd) want it. Usable power for passing etc (and out of turns on the track). Faster off the line I don't need.

Agreed, that's where it currently feels "slow" and could best use the boost. Wonder why they didn't plot the "after" torque curve too? It can't be identical...gotta have more torque at the same motor speed to have more power...

Edited: Because it looks like they DID plot the "after" just not the "before." Oh well, should just be a directly proportional change along the range anyway.

 

[Daniel in SD]

sweet. That's exactly where you'd (or at least I'd) want it. Usable power for passing etc (and out of turns on the track). Faster off the line I don't need.

This plot shows that Tesla should have a "track junky" option to put a taller gear ratio in.

 

[SleeperService]

This plot shows that Tesla should have a "track junky" option to put a taller gear ratio in.

I heard that update 2020.4.2 is going to enable the hidden 2 speed transmission. (only on P models though)

 

[stickman]

This plot shows that Tesla should have a "track junky" option to put a taller gear ratio in.

OT: is that a miata in your profile pic? And are you like 7' tall?

 

[roguenode]

I got the update notice this morning. Did a 0-60 before applying the update on a flat stretch (Dragy verified) and got my best 0-60 yet at 3.51s for that location. Applied update and after a few more runs I did get a couple of 3.47's. The car did seem to pull a bit harder towards the end. Given that dyno plot it seems about right.