Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

P90D Performance/Power Data

This site may earn commission on affiliate links.

Pete90D

Fan of Red Lights
Jul 25, 2015
478
4
Bay Area, CA
I thought I'd go ahead and spin off a new thread for the power and performance data. I went out tonight, charged to 100% and did some testing.

Max power I was able to get was 457KW, which is 1KW higher than the other night. I can provide the CSV of the data to anyone that wants to analyze speed vs power vs time.

Regarding the "Max Battery Power" option. I kind of don't like it. It can take a long time to kick in, and even if the car has been used it won't always be "Ready". Also when you use it you will eventually have your fans kicking on like you were at the SC. It will also turn off on you. It was "Ready", I did a launch, and then it was "Preparing" again. You have plenty of power without it and I don't know that it really gives that much more as of right now. Maybe if they are still tweaking things we will see an improvement.

0-60 tests:
- Tires @ 48 psi
- 215lb driver
- Windows up
- A/C off
- Music off
- Suspension low
- SOC varies (will analyze times to find out the best times)
- Max Battery Power, not sure because I did so many runs.

Best of tonight:
- 3.255s with 0% slope, no rollout
- 2.895s with 0% slope, 1ft rollout

Here was the other night
- 3.236s with 1.689% slope, no rollout
- 2.901s with 1.689% slope, 1ft rollout


Graphs from tonight

060nr.jpg


060ro.jpg





Here's the 30-60 I got during one of the runs.

3060.jpg



So we're not at 2.8s without rollout, but we're at the top of it with rollout. We're over 1.2Gs during acceleration. I'm clearly having massive fluctuation in the G forces, but this time I didn't feel like I was losing traction so I don't know what's going on.
 
The g force fluctuation is interesting and slightly worrying.

Here's a chart produced by the same vbox with the same antenna on my P85D:

90SOC0to60PlusTorque.jpg


I wonder if there could be a drive unit issue or some other issue causing irregular output not related to traction? If there is, I wonder if it could be limiting maximum output somehow?

- - - Updated - - -

Also, for comparison, a sample run that had a 3.153 0-60 (1 ft) had a 30-60 of 2.023 vs your faster 1.784 30-60.
 
The g force fluctuation is interesting and slightly worrying.

Here's a chart produced by the same vbox with the same antenna on my P85D:

View attachment 92191

I wonder if there could be a drive unit issue or some other issue causing irregular output not related to traction? If there is, I wonder if it could be limiting maximum output somehow?

- - - Updated - - -

Also, for comparison, a sample run that had a 3.153 0-60 (1 ft) had a 30-60 of 2.023 vs your faster 1.784 30-60.

Just sent you the file for analyzing anything vboxverify can't... like HP numbers
 
Pete, we should rule out antenna noise issues by collecting some data without the car moving at all and making sure the g force reading is pretty stable. If the vbox verify web inteface won't process that, send me the file and I'll run it through vbox tools or you can install that yourself but it needs a PC.
 
Wonder if that ~3 Hz oscillation is due to some structural vibration of the mounting/location of the vbox? You can see the same frequency in the speed data, which i assume they integrate the accel signal to calculate...

It's an external antenna. There is no accelerometer in the vbox. It uses GPS Doppler to read g force. I was dubious about this at first as my past cheaper performance meters used 3-axis solid state accelerometers, but after getting an external antenna and running some baseline tests, I was convinced that a clean signal could produce very stable readings.

But in either case, since I'm able to get a clean reading it's either mounting location and interference from something or the car is not producing a smooth output.
 
Thanks for the explanation, it makes sense and may explain the signal noise/oscillations.

So the GPS provides the position and time (x,y,z,t) data from which the calculations can be made, then the result probably gets coordinate transformed into the body axes for longitudinal, lateral and vertical position, speed and accel. There is quite a bit of calculations involved and any 'noise' gets amplified by the derivatives of position and velocity to get acceleration, v=dx/dt and a=dv/dt.
 
It's an external antenna. There is no accelerometer in the vbox. It uses GPS Doppler to read g force. I was dubious about this at first as my past cheaper performance meters used 3-axis solid state accelerometers, but after getting an external antenna and running some baseline tests, I was convinced that a clean signal could produce very stable readings.

But in either case, since I'm able to get a clean reading it's either mounting location and interference from something or the car is not producing a smooth output.

I have it mounted on the windshield and it's solid and the antenna is on the top of the pano routed through the passenger window.
 
I'm skeptical of the G force readings. Fluctuations of a half G up and down in a half second would be VERY noticible to anyone inside the car and should correlate to significant curves in the speed line. Something ain't right.

Agreed. Something is amiss. The very same vbox and antenna result in rock solid signals for me as you can see from the graph I posted.

- - - Updated - - -

Pete, I looked at the resting portions of the file where you have significant time spent at 0 MPH. The g-force signal when not moving is fluctuating ± 0.05 g's. So interference can't explain the wild swings. Perplexing.
 
Thanks for the explanation, it makes sense and may explain the signal noise/oscillations.

So the GPS provides the position and time (x,y,z,t) data from which the calculations can be made, then the result probably gets coordinate transformed into the body axes for longitudinal, lateral and vertical position, speed and accel. There is quite a bit of calculations involved and any 'noise' gets amplified by the derivatives of position and velocity to get acceleration, v=dx/dt and a=dv/dt.

The g-force value is derived from the Doppler of the signal itself rather than using the timing data that is used for position. So it's actually worse because any thing that reflects the signal will mess up the Doppler reading badly while the timing data in the signal used to calculate location won't because the receiver will read the data out the strong primary signal rather than a reflection.

- - - Updated - - -

Perform a smooth braking test, from 60-0 and check the results. Doesn't need to be quick, just keep it out of the ABS. Should produce a nice g force curve. May help track down the source of the problem.

+1. Nice suggestion. Just don't get rear ended again Pete :wink:
 
To be precise, the GPS, uses Time (T) and gets T and ∆T measurements from at least 4 satellites to solve for 3-axis position and absolute time. It does the same with RF carrier frequency (F) to solve for 3-axis velocity and local frequency reference error. To measure acceleration, it has to do derivatives on velocity measurements or use the magical Kalman Filter in the GPS receiver to estimate acceleration as well.

I would trust a 3-axis accelerometer much more. There are Apps for the iPhone and Android phones that will log data from the internal phone sensors directly for comparison.
 
To be precise, the GPS, uses Time (T) and gets T and ∆T measurements from at least 4 satellites to solve for 3-axis position and absolute time. It does the same with RF carrier frequency (F) to solve for 3-axis velocity and local frequency reference error. To measure acceleration, it has to do derivatives on velocity measurements or use the magical Kalman Filter in the GPS receiver to estimate acceleration as well.

I would trust a 3-axis accelerometer much more. There are Apps for the iPhone and Android phones that will log data from the internal phone sensors directly for comparison.

This was my argument to Racelogic when I first got my vbox and had crappy noisy g-force readings. I had quite a long email exchange with Mark Kibble at Racelogic. I had assumed when I bought the product that it had an accelerometer and when he told me it didn't I immediately hit him with the derivative of the derivative argument and how the second derivative would be too noisy. PM me if you want me to forward the thread to you. Here's a quote from one of those emails:

"Some of the information will reflect more on our automotive line (anything dealing with RTK or dual antenna systems), however, the way GPS is used for calculations is the same between our Motorsport and Automotive products. As noted on this page, our equipment uses the Doppler effect to calculate speed, rather than triangulation. This is important to note as through this process, we stand behind our equipment being able to accurately measure Speed, Time, Distance, Lat/Long Acceleration, and so on. "

https://www.vboxautomotive.co.uk/index.php/en/products/data-loggers/28-how-does-it-work#velocity

So there is only a first order derivative. That said, it's still critical to have a clean signal so you need to use an antenna if you want a really clean signal.


- - - Updated - - -

Wait, I just re-read your post. I think we're actually in agreement here.

That said a 3-axis accelerometer is going to be the most accurate of course but one derivative of velocity if the signal is clean will also be very good.

Pete, I just realized there is probably one significant factor here that is different. All my runs were done out in wide open fields without any buildings in site on Snelling road just north of Merced where it's still flat. Wide open grassy fields with nothing in site. I'm worried now that if you're doing this in a setting with buildings around you in the streets that this could pollute the signal. That said, the derived g-force value is pretty clean when sitting still.
 
What is truly amazing these days is that driven by iPhone and Android volume, so-called 6-axis chips (3-axis accelerometers, and 3-axis gyroscopes) are down to single digit dollar pricing, even in small quantities. See MPU-6500, 6-Axis Gyro+Accel, InvenSense Inc., CDI


V-Box does sell an Inertial Measurement Unit. That should add to the accuracy of the measurements. Has anyone priced this option or tried to use it?
 
I have new numbers. I wasn't really trying to beat my previous ones, but I got the 1ft down by 0.085s

0-60 3.168s
0-60 2.810s (1ft)
30-60 1.726s

1ft.jpg


060.jpg


3060.jpg


- - - Updated - - -

SOC was closer to 80% on this run. I started at 85% then drove a bit, then when I got home (not far from the fastest run) I was at 80%
 
Very impressive Pete and thanks for sharing. That was close to breaking into the 2.7's. Hoping you get a chance to take your P90D drag racing and get some decent 1/4 mile times.

Thanks. I will note that this time I do have a pretty bad headache after several runs, which is new. In the past I only got them if I didn't have my head back from 30-60 and it was due to some neck discomfort.