Garmin 500 gradient accuracy

elderone
elderone Posts: 1,410
edited August 2013 in Road general
I like the gradient function on my 500,it,s one of the reasons I upgraded from the 200.My question is,how accurate is the real time reading,as compared to strava the gradients are miles apart.
On one hill strava has it as 27% max yet on the hill my garmin said 19%.Which if any is closest to the truth.
Dulce et decorum est Pro patria mori

Comments

  • I don't trust the accuracy of Strava at all. I can ride the same route twice and get different segments or missing info for each one. I have even ridden with a friend same time and route etc and we had different height gains, different segments and the times were miles apart despite riding side by side. As to gradient they are not in the real world.

    Either you get up easily or not, that is what I go by
    Viner Magnifica
    Fondreist TF3 1.2
    Cervelo P3
    Aeron
    Viner Icarus
  • rolf_f
    rolf_f Posts: 16,015
    Strava I assume uses a digital terrain map so it will be only as accurate as the resolution of the tiles making up the map.

    On the other hand, what the GPS records in terms of gradient is clearly bunk. GPS are accurate over distance but if you park up and leave the GPS on chances are it will record a lot of movement despite being static. If you think about it, if the GPS records a data point every few seconds, and you are climbing at, say, 5mph (assuming your gradient is sufficient that you want to brag about it!), then if there is a few metres error in position on each point (which over time cancel each other out) and a few metres (or even less) error in vertical elevation, then you will end up with a potentially vast over or underestimate of instantaneous gradient.

    So in terms of a climb on which has a particular steepest section, I wouldn't believe either Strava or the GPS. Your best bet to determine it would be to measure the horizontal distance via an online map and then sit with your GPS (properly calibrated to altitude) at the bottom of the ramp for a few minutes, recording the displayed altitude and then repeating at the top. But hopefully you don't really care that much!
    Faster than a tent.......
  • The Garmin 500 also has a barometric altimeter so I'd say it's probably going to be reasonably accurate as long as you're going up a reasonably gradient.

    http://support.garmin.com/support/searc ... 0000000%7D
  • rolf_f
    rolf_f Posts: 16,015
    The Garmin 500 also has a barometric altimeter so I'd say it's probably going to be reasonably accurate as long as you're going up a reasonably gradient.

    http://support.garmin.com/support/searc ... 0000000%7D

    Not enough for measuring gradient on the fly.
    Faster than a tent.......
  • Mikey23
    Mikey23 Posts: 5,306
    Doesn't the 200 have the same gradient function? I thought that the software and hardware were identical, the only difference being that you couldn't bolt things on to it.. I find it quite amusing when going uphill to see how the percentage gradient copes. I think it gets it right overall but the changes are pretty random from point to point
  • Rolf F wrote:
    The Garmin 500 also has a barometric altimeter so I'd say it's probably going to be reasonably accurate as long as you're going up a reasonably gradient.

    http://support.garmin.com/support/searc ... 0000000%7D

    Not enough for measuring gradient on the fly.

    Why not? Barometric altimeters can be very accurate and when the rate of change is integrated over time the result should be pretty good. It should be more accurate than GPS at any rate.
  • chris_bass
    chris_bass Posts: 4,913
    Mikey23 wrote:
    Doesn't the 200 have the same gradient function? I thought that the software and hardware were identical, the only difference being that you couldn't bolt things on to it.. I find it quite amusing when going uphill to see how the percentage gradient copes. I think it gets it right overall but the changes are pretty random from point to point

    the 200 doesnt have the barometric altimeter so it just works out the gradient using the gps signal i think, so its only as acurate as the gps signal you have.

    I find the 500 takes a little while to notice you are going up hill buit it seems fairly consistent, no idea if its right though, but similar hills seem to ge the same gradient so its at least calibrated with itslef!
    www.conjunctivitis.com - a site for sore eyes
  • diamonddog
    diamonddog Posts: 3,426
    200 gives a more accurate reading when loaded up to Garmin Connect than when riding as long as you have enabled the Elevation Correction so I would imagine the 500 would be the same as barometric pressure is not constant whether at altitude or sea level but I could be wrong. :)
  • sungod
    sungod Posts: 16,547
    barometric altimeters in a consumer gps/similar device have considerable lag

    if you imagine moving the device up/down in a sawtooth profile, the measured altitude will underestimate the maxima and overestimate the minima, a plot of altitude will also have a temporal lag, i.e. vs. the true ground position the altitude plot will be displaced

    vs. grp, barometric will be better for assessing relative altitude, but, unless you halt at peaks/troughs to allow it to stabilise, it will tend to underestimate the altitude change

    for cycling, day to day variations in atmospheric pressure do not really matter, it's usually the total metres climbed over the course of a ride that is of interest, rather than altitude per se
    my bike - faster than god's and twice as shiny
  • Schoie81
    Schoie81 Posts: 749
    A friend of mine is a surveyor who uses GPS to take levels on site surveys related to Ordnance Survey national benchmark levels. To get accurate levels, the station must be picking up a signal from at least three satellites and that usually takes a little while (30-60 seconds) to fix on those signals, and that's from a stationary location. Trees and buildings, and even hills if they're steep and you're close to the bottom of them, can also obscure the signals creating further difficulties. Whlist I'm sure GPS tracking devices don't work to the same accuracy (the levels are taken to +/- 5mm), so maybe don't need three signals, but I imagine a moving device, passing trees, buildings, lorries etc.. would struggle to be all that accurate on level/elevation data.

    I find Strava and other online activity tracking websites to be quite inaccurate in terms of elevation, but I don't know if that is the fault of the website or the GPS device providing the data.
    "I look pretty young, but I'm just back-dated"
  • hatch87
    hatch87 Posts: 352
    Schoie81 wrote:
    I find Strava and other online activity tracking websites to be quite inaccurate in terms of elevation, but I don't know if that is the fault of the website or the GPS device providing the data.

    It depends, at first it uses the data from the Garmin device, but if you click on elevation correction it uses data taken from records. The US has accuracy between 3 - 10 meters whereas everywhere else is 30 - 90 meters. Ultimately if you're cycling in steady weather the the barometer will be more accurate. But changeable weather will cause odd results.

    The other thing is Strava segments is based on the data from the person that created it so if they did it another day with different weather and a different device then you are going to get the differences you are experiencing
    http://app.strava.com/athletes/686217
    Come on! You call this a storm? Blow, you son of a bitch! Blow! It's time for a showdown! You and me! I'm right here! Come and get me!
  • chrisw12
    chrisw12 Posts: 1,246
    A couple of observations from cycling up luz Ardiden a few weeks ago.

    I had my Garmin 500 set to read altitude and gradient as I climbed this beautiful mountain. The French place markers at every km stating the average gradient and altitude. My gps would pretty much match the altitude markers, there was a small under error but it seemed to stay consistent. The gradient function seemed pretty accurate but it would drop a reading when going through heavy tree cover, this suggests to me that the gradient is calculated by GPS and not barometer variation.
  • sungod wrote:
    barometric altimeters in a consumer gps/similar device have considerable lag

    if you imagine moving the device up/down in a sawtooth profile, the measured altitude will underestimate the maxima and overestimate the minima, a plot of altitude will also have a temporal lag, i.e. vs. the true ground position the altitude plot will be displaced

    vs. grp, barometric will be better for assessing relative altitude, but, unless you halt at peaks/troughs to allow it to stabilise, it will tend to underestimate the altitude change

    for cycling, day to day variations in atmospheric pressure do not really matter, it's usually the total metres climbed over the course of a ride that is of interest, rather than altitude per se


    Yeah, there will be a little bit of hysteresis from the altimeter but I suspect the error will be insignificant when compared to the GPS positional error. In any case I've just done a quick practical test. I started up my Edge 500 and turned the GPS off - I then went up stairs and the increase in elevation was correctly displayed within a few seconds which suggests the barometric altimeter works fine.
  • sungod
    sungod Posts: 16,547
    yes, consumer gps altitude accuracy can be dire, barometric is much better

    on steep sections the lag can lop off several metres, for checking gradients it's best to wait for it to stabilise, otherwise not an issue
    my bike - faster than god's and twice as shiny
  • elderone
    elderone Posts: 1,410
    Chers lads,not sure any the wiser but I think Ill believe me garmin as that cost lots and strava is free.
    Dulce et decorum est Pro patria mori
  • plowmar
    plowmar Posts: 1,032
    As others have said there is definitely a lag of about six seconds, but that lag is constant throughout the ride on my 705; so the information you are collecting for the down/upload is accurate just not when you are looking at it.

    My problem is that it only shows 10% when I know dam (sic) well that it should be 30%. :wink: