Garmin distance difference between plotted course & actual

bobbywise
bobbywise Posts: 23
edited September 2012 in Road general
I'm new to cycling, so maybe this is normal.

I plotted a course at http://connect.garmin.com/course/2052029 -- came in at 40.38 KM
Sent this course to my Garmin Edge 800, and followed it (never went off the course).
The result of following the course is at http://connect.garmin.com/activity/220104283 -- came in at 38.62 KM

That's like 1700m shorter :!: I can understand a few meters of difference here and there, but 1700 :!:

To top it off, if I then "Save as course" the activity that tracked 38.62 KM, it then shows 40.26 KM :!: :?: :!:

Anyone know why this is ?

Cheers,

Rob

Comments

  • No idea exactly why it happens, but I had something similar with a route I created using the British Cycling mapping tool - the actual ride distance, as recorded by my Garmin 500, was about 4 miles less over an 80 mile ride.

    Perhaps they use some algorithm when saving the course file that strips some of the course points out in order to keep the file size relatively small. If they are, this could certainly introduce distance errors.

    I find the routes created with BikeRouteToaster to be more accurate, but that's just my experience.
  • Thanks dee -- yeah, I trust my GPS, so I'll just use the Garmin for plotting new interesting routes, and not pay too much attention to the distance they give.
  • Pretty simple really - the Garmin takes data points roughly every six seconds so the course plotted is a series of lines - not sweeping curves.

    When you use a route plotting programme which is based upon Google Maps or similar, the interval is much smaller so you get less 'shortcuts'. So for example, if you rode a roundabout with a circumference which takes you 18 seconds to go around, the Google Maps will give you a circle of close to 18 straight lines whereas the route recorded by a Garmin would actually be a triangle with each side roughly equal - big difference.

    The second and much simpler reason is that the Garmin at best has an accuracy for every point is 20ft whereas your route plotter is pretty much spot on.

    The more corners and roundabouts you traverse on your route, the less accurate your route will be distance wise. Just zoom in on your route around Montbert and you'll see the Garmin route recorded isn't on the road at all in places.

    Its the limits of physics.
  • Gizmodo
    Gizmodo Posts: 1,928
    Out of interest, do you have a speed sensor on the wheel connected wirelessly to the Garmin?
  • MattC59
    MattC59 Posts: 5,408
    Gizmodo wrote:
    Out of interest, do you have a speed sensor on the wheel connected wirelessly to the Garmin?

    Was going to ask the same question. Depending on whether this is set up accurately or not, you could have quite a discrepancy between the distance taken from the road, the distance that the GPS thinks you've travelled and the distance plotted.
    Science adjusts it’s beliefs based on what’s observed.
    Faith is the denial of observation so that Belief can be preserved
  • I think the wheel wasn't set up correctly (had it to 2100 mm which was for my mountain bike, rather than automatic).
    I got a new bike at the start of this week (Trek Madone 4.7, and I think the wheel size is 2170 mm).
    I'm thinking maybe it's best to just put the wheel size to automatic, and let the GPS calculate the distance/speed.

    I'll do the same course a few times with "automatic" and manual set to the correct diameter and see if that makes a difference.

    I always assumed that the GPS took priority over the wheel diameter, but maybe it doesn't !
  • napoleond
    napoleond Posts: 5,992
    I had terrible trouble recently with my Garmin auto detecting wheel size so in the end I measured it myself. In fact now I've ditched the cadence sensor as I'm back with a powertap.
    The worst occasion was when I did a 118 miler and the Garmin said it was only 89!
    Insta: ATEnduranceCoaching
    ABCC Cycling Coach
  • So I set the wheel size to automatic, and did the same course this morning.
    Voila -- now rolls in at the expected 40.3 km.
    So happy I managed to fix the issue :-)
    Funny today though, went over a speed hump an the saddle angle moved. So was forced to stop and adjust it ! Glad I had the mini tool kit with me !

    Here's today's cycle : http://connect.garmin.com/activity/221958859 -- 40.25 km
    Here's yesterday 'corrected' cycle via the GPX file http://connect.garmin.com/activity/220530417 -- 40.35 km

    The difference is marginal, so im now happy with the distance from the Edge 800.
  • Just another update.
    So as per previous post, Sunday's ride was logged accurately, so I figure that setting the wheel setting to automatic sorted the issue. After riding a little on the Sunday he edge 800 beeped and said wheel size calculated. The ride was logged accurately, and I assumed it was because automatic wheel forced use of the GPS, but maybe it doesn't ?

    Today I go out to do a short 31.3 KM, and it comes in at 31 KM !!!! Again, I correct it by exporting GPS file and reimporting in garmin etc., but was frustrating. Then, I check the wheel setting out of interest, and it was back on custom with 2000mm !!!! The previous time I have it set to automatic !!!

    Does anyone know if the Edge 800s use the automatic wheel setting calculation to then cutsom set the wheel size ?
    I would have assumed that if you set automatic it would stay on automatic.

    If this carries on I'll remove the magnet from the back wheel so only cadence is registered and the Edge 800 will then be forced to use the GPS for speed/distance. Of course I prefer to keep the magnet on in case of tunnels etc.

    Any tips from existing Edge 800 users ?

    Cheers,

    Rob
  • Gizmodo
    Gizmodo Posts: 1,928
    If you want accurate (and it sounds like you do) then why not measure the actual circumference of your wheel and put that in as a custom value? Line up the inner tube valve with a line on the ground as a starting point. Sit on your bike with the tires inflated to their normal level. Roll forward (with the help of a friend is best) until the wheel has done exactly 1 revolution. Measure the distance between the 2 points.

    I'll look at my Garmin today because I changed the setting yesterday from Custom back to Automatic, so I'll see if mine has reverted and get back to you.

    EDIT: Just checked and mine is still set to Automatic this morning. Do you have the latest version of the firmware on your device?
    http://www8.garmin.com/support/download_details.jsp?id=5237
  • Hey Gizmodo,

    I have the latest firmware.
    I factory reset the Edge 800, and put wheel to auto.
    Did the same course as yesterday, and it came in at 31.28km as expected -- http://connect.garmin.com/activity/222243987

    I'll go out tomorrow and see if the Edge 800 remembers automatic.
    Out of interest I switched it on after today's ride and the wheel setting now says Automatic 2076mm.
    According to the edge manual, 700 x 23 tires are 2095 mm.
    I measured the tire running it along think masking tape, and marking the two points, yielding 2105 mm.

    So the measured value is 10 mm more than what garmin's manual quotes and 29 mm more than what he auto function calculates.

    Having said that, the activity from today where the auto value calculated to 2076 mm seemed bang on with what I expected from the course !

    I'm curious tomorrow when I go out if I should reset it to automatic to remove the 2076 mm, or if I leave it will it recalculate and update anyway ?

    Cheers,

    Rob
  • Gizmodo
    Gizmodo Posts: 1,928
    I must admit I've never been that bothered about the minor innaccuracies of what is a "consumer" grade (rather then professional) unit.

    If I were you, I would take your measured 2105 and the Garmin suggested 2095 values and pick the average of the 2 - so I would enter 2100 as a custom value. GPS is inaccurate - until we get Galileo :D
  • Just an update.
    Went out today, again on automatic (it was automatic_2076mm), and after the ride it was automatic_2081mm.
    The distance corresponded to what I plotted within 100m or so over 53.3 km, so I'm pretty happy with that.
    Seems like automatic setting is now working well :-)