Coastdown Testing The C-Max

[SnowStorm]

Decided to try a coastdown test on The Enterprise.  I could not believe how long it takes to coast to a stop!  I was sure the motor was pulling a bit as the car went on and on.  But no, after stopping, there was no pull until shifting back to Drive.  Here's my setup for taking data.  Phone with rubber band holding it to a metal L bracket attached to some paneling.  Very stable videos.

 

 

PROCEDURE:  Find a flat road with no wind (or cars!), start the video, speed up to just over starting speed, shift to Neutral, coast for two miles (yes, it takes at least two miles just to coast!) , stop video.  Repeat in other direction.  Do it several times, etc.  (Kept passing some guy on a bicycle.  Figured he was wondering what that nutcase driver was doing but he didn't seem to pay much attention.)  Anyway, go home, play all the fool videos and use a stop watch to get all the splits every 5 mph.  Enter data, average the runs and adjust the Crr (coefficient of rolling resistance) and Cd (coefficient of drag) to get the equation below to fit the averages.  I had 4 good runs (2 each way) with tires at 49 psi cold and then did 2 more (1 each way) after letting out 11 psi.  Tire pressure had only gone up 2 psi from "cold" so actually dropped pressure to 40.  Air temperature was about 76F.  Here's the data (seconds) and resulting fits.  Odd runs in one direction, even in the other.

 

 

 

  ---------------------49 PSI Cold ----------------------- ---------- 38 PSI Cold--------   Speed Run 1 Run 2 Run 3 Run 4 Fit to Avg Run 5 Run 6 Fit to Avg 70 0 0 0 0 0 0 0 0.00   65 6.17 6.73 5.98 6.97 6.95 5.69 6.65 6.84   60 14.37 14.98 13.8 15.21 14.72 12.61 14.54 14.49   55 22.27 23.6 22.31 24.35 23.47 21.98 23.67 23.07   50 31.89 34.48 32.21 34.37 33.33 31.57 34.09 32.72   45 42.1 45.17 41.1 46 44.50 38.92 45.71 43.63   40 55.28 59.14 53.16 60.7 57.18 51.06 61.75 55.96   35 69.23 74.2 67.23 74.38 71.59 65.78 76.32 69.92   30 85.42 92.02 82.48 91.84 87.94 81.07 94.55 85.72   25 103.12 109.48 98.63 108.83 106.45 95.19 113 103.50   20 122.99 131.8 119.07 133.12 127.23 113.25 134.63 123.37   15 146.28 156.81 141.49 157.87 150.28 133.36 157.2 145.31   10 170.27 183.73 166.91 182.62 175.38 157.03 179.82 169.10   5 190.55 205.48 183.81 200.85 202.07 177.57 198.03 194.32  

 

Results for 49 psi:  Crr = 0.0082,  Cd = 0.29

Results for 38 psi:  Crr = 0.0087,  Cd = 0.29

 

Equation was:   a = -(Cd*A*0.5*rho*V²)/M - Crr*g

   where:
a = acceleration in m/s²
A = frontal area (I used a rough estimate of 2.60 sq-m but need a better number)
rho = air density = 1.22 kg/m³ (might not be quite right for actual temperature and pressure)
V = velocity in m/s

M = mass of car & "contents" in kg (I used 1740)
g = Standard gravity = 9.81 m/s²

 

(All the above subject to corrections!  Plus a better equation would take rotational friction into account which is proportional to speed - not speed squared.  However, I had all the above stuff from a spreadsheet where I did the Lexus years ago.  It had a Crr of 0.012 and Cd of 0.35.)

 

But I'm glad we don't have to worry about all this to have fun driving the awesome C-Max! :shift:

[Smiling Jack]

Snow Storm,

 

Outstanding.  Thanks so much for posting !!!!!!

 

Great idea to use the cell phone video.

 

I used to try to take coast down data on a couple of cars years ago (before cell phones, let alone cell phone video) actually reading a stop watch at just a very few speed points and and using combination of memory and jotting down while driving.  I finally gave up because I could not get the right road to be traffic free long enough for the repeats and reverses.  Traffic is even worse around here now.

 

I wasn't interested in trying to get drag coefficients,  I wanted to measure the horsepower required vs. speed  (rolling, aero and total).

 

I may have a shot at what I can do with your data.

Edited June 11, 2014 by Smiling Jack

[SnowStorm]

Here is a csv file with the data (change extension back to csv).  Above data looked OK when posting - don't know what happened.

File has 3 more runs that aren't as good.  Again, odd runs in one direction, even runs in  the other.  You will see a difference between the two groups so you can't really use Run 0 or it will through the average off.

C-Max Coast-Down Test.txt

[fbov]

Well done!

 

And no surprise that the forum SW scrogged your table... CSV come through fine!

 

How'd you find the level road? The one thing EVing has taught me is that roads aren't flat no matter how flat they appear, and nearly all roads look flat when you're driving on them! Best I've found is a nice hill, for a different kind of coast down test...

 

HAve fun,

Frank

[hybridbear]

Awesome data! I want to do something similar in the FFH to determine the Crr difference between the 17-inch tires & the 18-inch tires. Have you thought about how to account for the internal friction of the vehicle?

 

And, like Frank says, are you sure the road is truly level? Even only a few feet of elevation change matters. Google Earth has some road elevation data but I'm not sure how accurate it is.

Edited June 12, 2014 by hybridbear

[plus 3 golfer]

Snowstorm, I've done coast down on two separate occasions over the last year logging variables with DashCommand.  What I do is to minimize the square of the difference between actual speed and calculated speed.  My intent was to see if I could notice any aero difference among stock, grille covers, and gas pods.  The bottom line is that when I would add the gas pods, the data got worse. The grille covers reduced the Cd.  One potential issue is the grille shutters.  There’s no way no know with certainty what position the shutters are in when conducting the tests as I have to accelerate briskly to 75 mph (and coolant temperature does rise), then shift to neutral and coast (and coolant temperature does come down).  If the shutters aren't in the same position during the test runs, then the data is suspect. Also, any wind gusts or increase / decrease in wind speed could affect results.

 

 

Here's a curve with actual data and a curve fit to each run. The return runs are the lower set of curves. The curve that seems in error is the lower blue curve. It is significantly lower than I would have thought it should be.   BTW, in an attempt to cancel wind and elevation change, you run in one direction and record results (out), turn around, and make the run back to the starting point (return) and average the results.  There was about 10 feet of elevation change per mile.  

 

 

The scan interval of DashCommand varies but is generally every 250 milliseconds. The speed from DashCommand is recorded in integers which means there will successive scans at the same speed which shows up on the graph as there are multiple circle markers at a given speed.  I curve fit the data and used the equations to compute speed vs time for the runs and then averaged the return and outward run times to get the composite speed vs time data to go into the model.  I then compute the square of the difference between actual speed and model speed for each time increment.  I used Solver to minimize the sum of the errors.  I made multiple runs with different frontal areas but kept Cd stock = 0.30.   I found the most consistent results were when FA is 2.27 m^2 and Crr = 0.0132.  If I changed FA too much the minimized error would grow and the results made little sense.  The road surface was rough and tires at 44 psi and I would think that some of the rotation losses (axles, traction motor and so forth) are also affecting the data and shows up as a higher Crr than one might expect.

 

Here are the results: Cd stock = 0.30, Cd grille covers = 0.279 and Cd covers and pods = 0.315 with Crr = 0.0132 for all runs.

 

 

rho = 1.132 kg/m^3 

g = 9.81 m/s^2

A = 2.272 m^2

M = 1770 kg

 

 

I ran similar tests again several months later at a different location and got similar results with and without grille covers.  I ran a pods only test this time and there appeared to be no significant difference between stock.  Since I can't control the shutters and wind and since I have to drive about 20 - 30 miles to find flat, untraveled roads, I've decided not to run any more tests. :)  I would caution against using this data in determining the benefits of grille covers or gas pods. ;) 

 

 

Edited June 13, 2014 by Plus 3 Golfer

[Jus-A-CMax]

Impressive Snowstorm & +3, thanks!

Edited June 12, 2014 by Jus-A-CMax

[ptjones]

Good job Snowstorm & Plus 3 Golfer.  With the testing I did on shutter operation with out A/C on and new update they don't open until 210*F at slow speeds less than 50mph and 215*F above 55mph. Plus 3 Golfer did you monitor your temps? If they were below 210*F the shutters were closed all the time. Interesting info on GasPods. :)

 

Paul

[plus 3 golfer]

Paul, I have to check whether I kept the original log files from DashCommand which will have CT. I may also have saved the original data exports to ScanXL but I usually only export the data I need from ScanXL to spreadsheets. I think the temps were below 210 and 215 for the stock run. I know for the runs with covers the max exceeded 230F but did fall quite a bit during coasting. So, it's possible that the shutters started open and closed during the coasting. Since the covers aren't air tight the data may have been affected if the shutters changed position.

 

I'd like to run more tests but like I said, I have to drive about 35 miles to get to flat lightly traveled roads where the risk of getting a speeding ticket is virtually nil. I might have time to do more tests in late summer as I will likely be in that area then.

[ptjones]

Paul, I have to check whether I kept the original log files from DashCommand which will have CT. I may also have saved the original data exports to ScanXL but I usually only export the data I need from ScanXL to spreadsheets. I think the temps were below 210 and 215 for the stock run. I know for the runs with covers the max exceeded 230F but did fall quite a bit during coasting. So, it's possible that the shutters started open and closed during the coasting. Since the covers aren't air tight the data may have been affected if the shutters changed position.

 

I'd like to run more tests but like I said, I have to drive about 35 miles to get to flat lightly traveled roads where the risk of getting a speeding ticket is virtually nil. I might have time to do more tests in late summer as I will likely be in that area then.

I would imagine with covers on it wouldn't make much of a difference. IMO   I'm wondering if wheel covers would make much of a difference? It is getting to look like I'm close to getting 47mpg at 70mph with mods and drafting. My last trip to AZ I got 46mpg average and 62.5mph average. :)

 

Paul