Test Track
Welcome to the Test Track.  The GetBetterMPG test course  and how we chose and validated it

Steps for this project include the following. This portion of the project is complete

  1. Layout a test course of about 100km
  2. Test ride the test track - done twice with the ODO reading 63.8 miles
  3. Test ODO and GPS on a straight course to determine ODO correction factor - done on 5/23, see results below
  4. correct test track distance - 63.8 x  1.0197 =   65.059 miles
  5. Correct everything to metric.  65.059 miles = 104.70231km

I needed a consistent course over which the to test the various alterations. This seemed at first an easy thing to do, I really need to stop thinking that!  I started with 100 km as a rough estimate of what would be accurate enough. Admittedly it was a "throwing a dart at a dart board" estimation, but then this is a one garage opperation isnt it. 

Browsing around the thousands of maps that lie around my home and office I came up with something that looked promissing. I would pull out of the GetBetterMPG command center and head West on US-44 all the way to Norfolk, thats about 20 miles. This takes me past Canton Cycles where I have been known to hang around and drink coffee at times. This seemed like a good emergency bail out stop if I needed one. 

In Norfolk I would turn South on SR-272. This route wanders through the woods and past a georgeous lake where I could stop and swim if I needed to do so for some reason, hey, ya never know!  When I run into Goshen I turn West on SR-4 for a short ways then turn south on SR-63. In Litchfield I would turn East on SR-118 unless I need to first stop at one of the many coffee stops there, Litchfield has more than its share of good coffee shops for some reason.

On route 118 I go straight which eventualy merges me onto route 4 East. I follow this until I run intot he Farmington river turning left just before I get wet on SR-179 North. Here I proceed north to route 44.  I turn right on 202/44 for a few hundred yards then into my starting point. The route works out to ~63.8 miles or 102.7 km, not bad for a first stab at a course of about 100 km.

A brief look at this route through the eyes of Google Maps with the Terrain shows that this route is far from flat. In fact there is one long and very steep hill to climb between Winsted and Norfolk and another between Litchfield and Collinville.  There are country highways with 55mph limits and towns with multiple traffic lights, even one light at the top of a hill for which I must use the brake to not roll backwards. There is even a traffic circle to contend with. These should help prove that the track can approximate the real world without forcing me to putter through crowded cities. Once I get the GPS out on the track I will grab some elevation changes too. In the end I need to be sure that what I test here could also be run on the long route from LA to NY.

The results of my first stab at the ride turned up a problem though. The first option played out as 59.8 miles at www.mapsonus.com and maps.google.com.  When I rode the first route on 4/24/2008 the BMW's ODO said it was 63.8 miles, a varience of 4 miles.  Yipes!  If I am to gain any accurate results from all this testing  I will need to do better than that. The ride was done on a 2005 BMW R1200GS with 11k miles on it and brand new tires. I would think the ODO would be fairly accurate. It would seem that either my ODO, or Mapsonus and Google, needed some calibration.  I decided that telling mapsonus and Google to fix their sites might come off sounding a bit like I was a nutter.   The only thing I could do was to find out how far off my ODO was and correct the distance for my test course. 

A GPS seemed like a good way to go but the sampling rate left something to be desired.  If it sampled too slowly then any curve would input a varience in the overalll milage reading by cutting a straight line where the curve should be. A straight line with a substantial enough wait period  at both the start and end points would allow for the sample rate to play out and would give me an accurate reading of milage. The problem was the whole straight line thing.  New England is not know for long straight roads. I remember some of those in OK and TX, but not here!

Another rumage through the piles of maps turned up an interesting bit.  Interstate 91 in Massachussettes turns a bit but North of the last crossing of the Connecticut river it makes a dead straight shot for the Vermont border. Should give me more than 60 miles of straight road. The next closer straight road looks like some stretches of Pennsilvania and Ohio and nobody wants to ride their highways so MA it is.

On 05/23/2008 I started at exit 21 on Interstate 91in Northern MA. and ran non stop to exit 26.  I marked the on ramp and off ramps with orange lines so somebody else can go repeat the test rideBoth are marked on the left side of the ramp. Tire Pressure was checked at 36f and 42r before and after the calibration run and both test track rides.  Two GPSs were used, one Garmin legend  and one Garmin Legend cx. The CX does altitude to check for distance variance caused by altitude change hich is non Linear to the Legend. On the calibration run both GPS said the distance was 15.50 miles with an accuracy of +/-  18'. The ODO said 15.2.

The math works out like this.

  • (ODO miles) x (correction factor, X) = (real miles)
  • 15.2miles x X =15.50miles
  • X = 15.5/15.2
  • Correction factor = 1.0197368421052631578947368421053

To use it we do this

  • (Test track ODO reading) x (correction factor) = (real test track miles)
  • 63.8 x 1.0197368421052631578947368421053 = (real test track miles)
  • Test track is 65.059210526315789473684210526316 miles
  • We will call it 65.059 miles or 102.676 km

Once I have the bike out on the track I will need to start calibrating it's own ODO. Since the OEM speedo has no option for calibration I have opted to add an eletronic computer. The unit I choose is from a bicycle and allows for wheel OD to be entered in millimeters. I will be able to check my measurements by checkig the ODO against the known track distance.  

This will be important when I ride from LA to NY since I will have no other way to judge how far I really rode other than by the bikes ODO. I will need to calibrate Suzi-2's ODO before the trip given what I have learned here. At the end of the LA-NY run I will return to check that it is still in calibration. Then I will know the distance is correct

Pertinent facts.

  1. 63.8 miles = 102.6761472 km according to http://www.asknumbers.com/LengthConversion.aspx
  2. 1 mile = 1.6093 km
  3. 1km = 0.6213miles