My new way of measuring on-road drag
 

Note: It appears that this method is problematic on cars with electronic throttles, where throttle blade position can vary, even with a fixed accelerator pedal position being held.

Introduction and background:

https://www.youtube.com/watch?v=tVAokIdaXm0


Using the technique to quantify changes in drag coefficient:

https://www.youtube.com/watch?v=Wi-TemRJ7CM

I will present the throttle-stop data here as I measure it. Unless otherwise stated, all data is for my Gen 1 Honda Insight that has the following modifications:

• electronic height-adjustable air suspension
• full ABS undertrays front and back
• rear diffuser
• rear acrylic fins
• rear GOE222 profile aluminium wing
• MoTeC dash including 10Hz GPS speedo
• No IMA – turbo, water/air intercooled and MoTeC ECU

Windows open/closed
Windows closed – 104 km/h
Windows open – 98 km/h
Calculated change in CD = 12 per cent

(Katz J, Automotive Aerodynamics, Wiley, 2016, Page 292 - gives a range of 10-16 per cent change in CD with windows open.)

Ride height
Ride height low - 105 km/h
Ride height high - 102 km/h
Calculated change in total drag - 6 per cent

(Rob Palin, former Tesla aerodynamicist, said to me of this value: I'm surprised by the small size of the ride height one. Ride height tends to be pretty linear (especially if you maintain a constant reference area!), with the deltas rising more sharply if you have very ugly wheels that you're now hiding better. The Insight had decent wheels, and the side cover at the back, so maybe it just isn't that sensitive.)

Rear wing angle
Some scatter in measurements around plus/minus 10 degrees wing reference angle, but the line of best fit shows an 8 per cent increase in car drag from a wing angle of plus 10 degrees to minus 10 degrees. That seems too much for the wing alone so I assume it is also influencing the wake and/or action of the diffuser.

At a wing angle of 80 degrees (ie air brake), car CD increases over the slipperiest position of the wing by 25 per cent.

I have now made some mock-up aero attachments to test for drag variations on the Insight.


Strakes to go in the rear diffuser:
https://i.postimg.cc/JzXTRQcD/DSC-3420.jpg


Side skirts:
https://i.postimg.cc/ncrFSXKw/DSC-3421.jpg

Rear spoiler that can be positioned at two heights:
https://i.postimg.cc/QN9sc2jR/DSC-3419.jpg

External front air curtains (and also can be trialled as propulsive wing elements for front bumper corners):
https://i.postimg.cc/134LZ64b/DSC-3418.jpg


All (except for the wings) are in plywood that can be taped into place and easily removed. (Wings are GOE222 aluminum profile.)

I actually wonder if any of these will make any measurable changes in drag (plus or minus), but we will see.

Am also going to try the Insight with the rear wheel spats (side covers) on and off.

Just waiting for the rain to stop. (Here in southern New South Wales, Australia, we've been in drought for years, so can't complain when we start getting some good days of rain.)

.

I like your method. :thumbup:

Plus, using a Honda Insight with its high mpg capabilities, expands the scale of resolution to see and discern subtle differences more easily.

Are you going to try testing smooth wheel coverings too ? They will most likely fall below your 2% threshold but, they might move the needle enough to gain a little Insight. (pun intended :p)

Looking forward to your results as you test the various aspects of automotive aerodynamics.



:)

>

What are your thought on using this method with a hand throttle and a stop vs a pedal stop?

If you haven't seen some of the users here's methods here is one. Hopefully pictures show up for you. I downloaded a plugin that lets me see photobucket pictures. https://ecomodder.com/forum/showthre...vic-30097.html

I am not sure the Insight 'expands the scale of resolution' - the lower the CD, the harder it is to make improvements (but I suppose easier to see if you're increasing drag?).

I won't be running full wheel covers. I think the Insight wheels are already very good, and I want to retain brake cooling openings.

Yes you could use a hand throttle - any method of temporarily holding a fixed throttle angle will work.

It is still raining, so no testing for the moment - the tape won't stick.

But I was thinking about the above situation, where wing angle seems to increase drag to a greater extent than would be expected. I think perhaps it's because the nose of the car is being lifted.

The wing is positioned behind the rear axle line, and we know it develops measurable downforce. The air suspension control system is not set to rapidly correct at speeds over 80 km/h (50 mph) and so it is likely that when the wing angle is set to provide lots of downforce, the rear suspension is being compressed and the front slightly lifted.

Of course, the beauty of on-road testing is that all factors are taken into account - the outcome is what it is.

I got impatient and tried some testing in very strong winds. However, the results were too inconsistent to be useful. So while you can test in some wind, really strong winds appear to be too variable, even over a 5 second or so average of top speed.

Yes. That makes sense. Real world conditions are not really an advantage because we are talking about uncontrolled variables changing willy-nilly. When I did my coast down testing of my flat decklid spoiler in spring 2019 in the desert at El Mirage heat changes as the sun rose in the sky were noticable. Your method would be an interesting way to retest my flat decklid spoiler after I finish rebuilding it. It would take less time.

My coast down method was different than the one you describe. Rather than measuring time between two speeds, I coasted from X speed to a full stop, measuring distance with GPS. But I think I need better GPS. Mine was reporting accuracy of only to within 20 to 50 feet. I was coasting 2600 to 3500 feet to stop, so that's roughly 0.75 to 1.9 % inaccuracy in my calcs from just the GPS. Not awful but not great. What GPS are you using to confirm speed?

Here is my El Mirage data: https://ecomodder.com/forum/showthre...tml#post602571