ridecamp@endurance.net: LONG: Re: GPS question

LONG: Re: GPS question

Jim Mitchell (navion@lsbsdi2.lightspeed.net)
Mon, 10 Feb 1997 19:47:59 -0800

JOSEPH PETER UHLARIK JR wrote:
>
> Your test results with GPS were fascinating, especially when you
> mentioned 1 percent and 3 percent errors. I am hoping that you will
> share with us some of your future results.
> I've given enough preliminary results for now. I'll continue testing it against known
things and write a better report in the future.

> Just to make things interesting, (at least to me :-) ), I am wondering
> if the mountain bike distance was longer than the GPS distance beause as you
> ride, the front wheel does not go perfectly straight, thereby
> increasing the distance that it covers over the ground.
> I'm really into distance measurement with my hobbies and have been for years. The
truth is I haven't accurately calibrated my mountain bike for at least 1 to 2 years. I
calibrate it by measuring a course with a surveyors tape 528 feet long at least. I then
ride the course as straight as I can with the same air pressures in the tires, weights
on the bike and on the same footing as the trail I want to measure. I'm fatter now than
two years ago, my tires are older and the softness of the trail all make a difference.
But because I use a straight course and ride it at the speed I normally do I should have
eliminated my wobble error since it should be the same. (except as I tire)
Using the calibration given in the manuals for tire sizes or even a tape measure and
rolling the wheel once along it are too crude for me. They give greater errors since the
weight distribution and speed and footing are all wrong.
I also go to local straight, flat county roads that are one section long and go from
surveyors monument to monument. I can get this distance from the county surveyors office
and then I have a 1 mile course that I use to calibrate. This works great for my road
bike but is not good for trail measurement since a mountain bike tire sinks deeper into
dirt than it does on an asphalt road.

The possibilities for errors of small magnitude are almost endless but I try to
account for most of them. Now that I'm riding with the GPS on my helmet I even worry
about how much my head wobbles and if this adds a significant amount to the GPS
distance! At this point there may be greater GPS errors than my head wobble and
continued testing will tell. I've also written Garmin (the manufacturer) and I am
awaiting their response to the accuracy and tolerance of their unit. The bottom line to
me is how well it continues to match my wheel odometers. If after a month or so of
testing in open country it always agrees with 2% or less I will be really amazed. First
I need to recalibrate all my wheel odometers.
Another neat feature that will help tell it's accuracy is the track log. It stores
points along your route automatically whenever you change directions. These can then be
printed out on a pc and overlayed on a map of the area you just rode. If you rode on
known roads then the overlay should match those roads. I've ordered the software and
will be checking this also. My unit just stored 725+/- points on our 19.5 mile ride last
Saturday. This is a point on average every 142 feet. This should give me a pretty good
overlay on the 1 inch equals 1.12 mile map I'm using. This would be 42 points to every
inch on my map. Almost a continous line!

> I am also curious if you have compared the bike, the GPS and the car over
> a straight line from one know place to another.
I plan to do this on one of the county roads mentioned above.
>
> All I am saying is that you are getting accuracies where a lot of these
> "little things" that in the past could be neglected may be becoming the
> major contributors to the little bit of error (or is disagreement between
> each of the measurements a better way to phrase it) that still remains.
> Agreed

> Restating that last thought, which of the three measurements that you have
> is really the correct one? (As some people have been know to say, it is
> sure nice to have a problem like that. :-) )
> I've reached a point where there is never a correct one. Only close and always some
tolerance which I like to state. Refering to measuring a race course, I like to do it
with several different wheel odometers and when we all agree with +/- 1% I take the
average and call it good. I've never measured an endurance course yet but in the past I
measured running, triathlon and biathlon course.
But now this GPS is a whole new ballgame. The reason I like it is because I can now
measure and maybe map things from my horse instead of having to return with a mountain
bike or motorcycle to get the distances.

Jim Mitchell

Home Events Groups Rider Directory Market RideCamp Stuff