Coastdown testing question: does starting speed matter for consistency?

[MetroMPG]

A question for the physics-minded among us:

I'm wondering how important is the consistency of initial speed, A, in coastdown testing where the data gathering will begin at speed B, somewhere below A.

EG: let's say I decide to measure coastdown times/distances from 90 km/h to 0 km/h.

Does a vehicle decelerate at a different rate from, say 90 to 80 km/h depending on whether the initial speed (clutch in/neutral) is 95 km/h or 110 km/h?

Just wondering for the sake of producing the most reliable results.

And no, my cruise control is currently NOT working, which is really why I'm asking . Don't want to rely on wetware (me) to accurately control speed if I don't have to.

[Daox]

I think there will be differences, but they will be statistically irrelevant.

[wyatt]

You will see a difference, but it's not because the Cd or Crr is different, it's because the Cs (coefficient of skin friction) isn't taken into accuont with coast down tests.
Instructables, how to measure the drag coefficient of your car:Measure the drag coefficient of your car
help with simplified cDA/cRR formula
help with simplified cDA/cRR formula

[MetroMPG]

Wyatt - please explain in greater detail. How does what you said affect whether the car decelerates at a different rate between speeds B & C depending on at what speed, A, the coasting begins.

Also, the instructables page says nothing about the relationship of the top/initial speed where coasting starts to the speed where data collection starts. It simply says start "above" the speed where you wish to start recording data.

[Piwoslaw]

Originally Posted by wyatt

You will see a difference, but it's not because the Cd or Crr is different, it's because the Cs (coefficient of skin friction) isn't taken into accuont with coast down tests.

Say that instead of starting at 95 km/h you start at 120 m/h. For a certain time you will be going faster than 95, so the air drag on the surface of car will be greater. This friction will cause the surface to become hotter, creating a "warmth bubble" of hot, rarified air around the car. In the right conditions, this "bubble" will have zero resistance on the car, letting you coast indefinitely.
Just kidding ;P

[KJSatz]

I don't think it should matter as long as you start coasting several km/h faster than your "start" point for all the tests. Forces (and accelerations) do not have a "lingering" sort of quality. So when you end up at your test start speed (B), the car is experiencing various "negative" forces that are dependent on current speed that should be the same for the car driving that speed at another time, regardless of if it was going slower than B a second ago, at B a second ago, or faster than B a second ago. Since you're in neutral at this point, the "positive" force is zero for all the trials as well. So it should be the same.

[MetroMPG]

Thanks - that gets to the crux of it: I wasn't sure how significant (in terms of time) that transition from a powered (constant speed) state to a coasting state was.

[aerohead]

It probably doesn't effect consistency,but the higher initial speed allows the air drag to "emerge"dramatically from the velocity/time coordinates of the coastdown curve.Since the air drag force varies as the square of the velocity,if you take the car above 100 km/h and monitor the deceleration,the air drag is really significant and Newton's F=MA relationship will shine light on aero vs RR at these higher velocities.Down around 32 km/h Reynolds Numbers are falling apart and the RR portion from the becomes dominant,helping you to identify it from the aero.SAE recommends starting from around 112 km/h,begin taking your trace at 100,and continue down to 32,back-to-back in each opposing direction for a minimum ten runs.

[wyatt]

Metro: I think I may have misunderstood (read too fast) your post. I thought you were asking if it made a difference if you started at 110 km/hr or 90 km/hr, and were thinking of starting at a variety of different speeds to do coastdown testing, and I remembered that co_driver had pointed out the viscous component of the equation that can change the results slightly. However if you look at the last post from the link above, he found his viscous drag component was about 1 Newton per Meter/second, and he was using a 90' Firefly. Sorry for the confusion, but you may be able to use a similar rule of thumb when doing your testing and obtain even more accurate results. Good luck!

Originally Posted by co_driver

As one is bound to find (from all of the discussions in this forum), there are no simple solutions. But there are rules of thumb.

In a spreadsheet that I have developed (its a beast!), I do include the viscous component - and guestimated it to be ~1 for a 2-wheel drive vehicle (where F(viscous) = cV*v (v in [m/s]). It could be ignored, but will have an influence on the standard pair of coast-down tests (high speed and low speed). Hint: (greater than) 3-speed coast down tests will show this minor factor (did this in uni many years ago w/ driveshaft torque measurement and data logger).

This is notably small when compared to the other two inputs:
eg. '90 Firefly:
@ 25 m/s:
F(friction) = 81.3N (cF = 0.010, m = 830kg)
F(viscous) = 25.0N
F(aero) = 236.3N (cD = 0.36, A = 1.75m^2)

[Disclaimer: YMMV, above numbers may or may not simulate real life, ...]

[MetroMPG]

Phil, thanks for the info. Couldn't ask for more than the SAE's instructions!

Makes sense now, Wyatt, thanks.