Race tire mu, non aero car

[jtc127]

Even 10 hz is way high. I'd take your logged data and put a 1 or 2 hz low pass on it if you want to get realistic numbers. Log at high rate and crush it down with a filter rather than sample at low rate.

I will try that when I go home tonight with 2011 data at the much higher sample rate. I found that AiM has a low_pass function available in the math channel section but haven't seen documentation or and example on how to use it yet.

[gato azul]

well, according to their Manual (RTFM- )
http://www.aim-sportline.com/download/s ... 01_eng.pdf
http://www.aimsports.com/software/downl ... 10_eng.pdf

Race Studio has a built in low pass filter as a standard math function.

low_pass --> Function for low pass filter

Therefore, it should not be too difficult to implement one into your analysis.
Other then that, keep in mind what GSpeedR said about sensor mounting, that will help to reduce the unwanted noise from vibrations, and should benefit your signal quality.

Code: Select all

Low_pass(p1, p2)     p1 = channel identifier (enter your LatG channel here)         If p1(t) <= p2  ---> p1(t)
                     p2 = constant           (enter the cut off frequency here)     If p1(t) >  p2  ---> 0

[munks]

If you crack open your copy of "The Pneumatic Tire" (a quick search will net you a free copy), then look through the graphs in Chapter 11, you'll see that some compounds under ideal conditions can *approach* a coefficient of 3.0, or at least exceed 2.5. Now on the one hand, it would be very difficult to get the entire contact patch operating at ideal conditions, much less on all 4 tires at the same time, but on the other hand there's undoubtedly better compounds nowadays optimized for specific track surfaces.

[Jersey Tom]

If you crack open your copy of "The Pneumatic Tire" (a quick search will net you a free copy), then look through the graphs in Chapter 11, you'll see that some compounds under ideal conditions can *approach* a coefficient of 3.0, or at least exceed 2.5.

Tested on sand paper quite likely

[Belatti]

How much peak longitudinal G does a top fuel car mark in the "non aero" first tenths? Has someone has a G profile for the whole 5 seconds a run endures?

There are quotes of 4G, but that must be "aero", since aero forces should be noticeable in less than 0,7 seconds...

[GSpeedR]

Though I don't know a whole lot about Top-Fuel, I don't think there is a 'non-aero' section of the run, as the engine exhausts create a significant amount of downward thrust even statically (static vehicle, not engine). In addition, those are pretty ridiculous tires and perhaps they are capable of such friction coefficients(?).

[Jersey Tom]

The surface is also treated with VHT

[jtc127]

The surface is also treated with VHT

Plus massive amounts of tire rubber for the first hundred some-odd feet

[jtc127]

FYI this was what the data looks like before the filtering



So using the LPF excludes frequencies below it's cutoff, how do you determine an appropriate cutoff frequency?

[gato azul]

So using the LPF excludes frequencies below it's cutoff, how do you determine an appropriate cutoff frequency?

above the cut off frequency
if you choose 2Hz for example, all data below or equal 2Hz will pass the filter and you see them on the screen, and all data which have a frequency higher then 2Hz will be set to zero, in this case.(not pass the filter)



maybe this is worth a quick read
http://books.google.com.ec/books?id=GOh ... ta&f=false

[Tim.Wright]

So using the LPF excludes frequencies below it's cutoff, how do you determine an appropriate cutoff frequency?

5Hz works quite well for handling signals like AX, AY, yawrate.

I'd be more worried about the noisy wheel speed signal. Why is it like that? Wheel skipping off the ground?

Tim

[Jersey Tom]

So using the LPF excludes frequencies below it's cutoff, how do you determine an appropriate cutoff frequency?

That's contingent on what question you're trying to answer, and some a priori knowledge of what you're after. If for example you wanted to use your vertical and lateral accel channels to roughly determine bank angle of a corner you'd probably crush it down to a 0.5-1.0 Hz low pass. If you wanted to capture the ride response of your car with a natural suspension frequency in the 2-3 Hz range, a 5.0 Hz low pass is probably good.

I would doubt that even in a kart you are exciting handling inputs at more than 1 Hz. 1-2 is probably just fine for a LPF on Lat G if you're after sustained cornering.

[jtc127]

So using the LPF excludes frequencies below it's cutoff, how do you determine an appropriate cutoff frequency?

That's contingent on what question you're trying to answer, and some a priori knowledge of what you're after. If for example you wanted to use your vertical and lateral accel channels to roughly determine bank angle of a corner you'd probably crush it down to a 0.5-1.0 Hz low pass. If you wanted to capture the ride response of your car with a natural suspension frequency in the 2-3 Hz range, a 5.0 Hz low pass is probably good.

I would doubt that even in a kart you are exciting handling inputs at more than 1 Hz. 1-2 is probably just fine for a LPF on Lat G if you're after sustained cornering.

I'm messing around with values from 1-5. The wheelspeed is most likely accurrate. That section was under full power with a decreasing lateral load, the larger spike is the car being upset by a bump. Video here: http://youtu.be/Koq5dWhpBxA?hd=1&t=1m