Bike Forums

Bike Forums (http://www.bikeforums.net/forum.php)
-   Clydesdales/Athenas (200+ lb / 91+ kg) (http://www.bikeforums.net/clydesdales-athenas-200-lb-91-kg/)
-   -   Ascent accuracy on Garmin Edge 500? (http://www.bikeforums.net/clydesdales-athenas-200-lb-91-kg/828834-ascent-accuracy-garmin-edge-500-a.html)

CBuff 06-29-12 03:53 PM

Ascent accuracy on Garmin Edge 500?
 
My new Garmin Edge 500 is reporting just over 150% of elevation gain over the MapMyRide statistics for the same route.
I fully expected some discrepancy, but that is ridiculous.

Anybody figured out which is more accurate? (Is suspect that MMR is closer to the truth).
Is there something i need to do to "dial in" the Garmin to increase the accuracy?

Any insight would be greatly appreciated.

Seattle Forrest 06-29-12 04:05 PM

Upload your tracklog to Garmin Connect, click the button to enable elevation corrections, and then compare.

I don't know if the 500 uses only GPS, or also has a barometer. GPS altitude tends to be good to within about 100 to 150 feet (it's drastically better in lat/lon terms) and barometric altitude is accurate to about 10 feet, as long as it's been calibrated for the air mass you're riding in. If your Edge is only using GPS, it won't be terribly accurate. The correction GC does is to look up the elevation from USGS data, for the locations you rode through, and calculate the totals from that.

Mithrandir 06-29-12 04:11 PM

MapMyRide is less accurate than your Garmin. Their elevation algorithm ignores inclines that are less than a certain threshold that they set; usually isn't a problem with straight up-and-down hill climbing, but in an area like mine which has many more rolling hills, it's dreadfully inaccurate. RideWithGPS has a much more accurate elevation profile; compare your Garmin with that and see what the difference is.


That being said, things like clouds and humidity can affect your Garmin's accuracy.

Mithrandir 06-29-12 04:12 PM

Quote:

Originally Posted by Seattle Forrest (Post 14422653)
Upload your tracklog to Garmin Connect, click the button to enable elevation corrections, and then compare.

I don't know if the 500 uses only GPS, or also has a barometer. GPS altitude tends to be good to within about 100 to 150 feet (it's drastically better in lat/lon terms) and barometric altitude is accurate to about 10 feet, as long as it's been calibrated for the air mass you're riding in. If your Edge is only using GPS, it won't be terribly accurate. The correction GC does is to look up the elevation from USGS data, for the locations you rode through, and calculate the totals from that.

Garmins (all of them as far as I know) use barometric altitude.

CBuff 06-29-12 04:21 PM

Mith,

Thanks for the RideWithGPS reference.

Forrest, will try the elevations correction at garmin connect and see what that does.

After I check these out, I will report back.

goldfinch 06-29-12 04:24 PM

For the ride I am doing tomorrow ridewithgps shows close to twice the elevation gain as mapmyride. :(

ChrisM2097 06-29-12 04:28 PM

I have a Garmin Edge 500, and live right around the corner from you, in Milpitas (same barometric pressure - which I believe is what the Garmin uses to determine elevation). I've found it's pretty accurate.

All of the below are the same ride:
http://connect.garmin.com/activity/187860301
Garmin - Elevation Correction On: 2,286'
Garmin - Elevation Correction Off (From Device): 2,218'

http://app.strava.com/activities/10369166
Strava: 2,215'

http://www.mapmyride.com/routes/view/108068821
MapMyRide: 1,864'

http://ridewithgps.com/routes/1373785
RideWithGPS: 2,353'

It looks like MMR is the one that's underestimating the elevation. The device elevation is very close to that of the Garmin Connect site, and Strava. Why there's a 500' (and 0.4 mile) difference between MMR and RWGPS, I don't know.

When I was doing that ride, and compared certain landmarks to elevations found in Google Earth, I've found the Garmin to be very accurate.

Wolfador 06-29-12 04:30 PM

Quote:

Originally Posted by ChrisM2097 (Post 14422743)
When I was doing that ride, and compared certain landmarks to elevations found in Google Earth, I've found the Garmin to be very accurate.

I have done / found the same thing with my Garmin.

CBuff 06-29-12 05:01 PM

Thanks all.

My small sample test confirms the consensus here; that MMR tends to underestimate elevation change, that corrected Garmin seems to be pretty accurate and that RideWithGPS tends to over estimate the elevation gain.

From this point forward, I will be tracking the corrected Garmin number.

Guess I have been climbing a little more that I thought!

johnny99 06-29-12 05:56 PM

MapMyRide is notoriously inaccurate for cumulative elevation gain. Same goes for all the other free mapping web sites. I trust Garmin (units with barometric altimeter) a lot more, if you're sure you let the unit calibrate before starting your ride.

And no, MapMyRide does not consistently underestimate elevation gain. Their data is just not detailed enough (especially in areas with lots of steep hills) and their numbers can be very high or very low. For example, one of my favorite bike routes includes a bridge and a tunnel. MapMyRide does not count either, so they include the elevation gain over the top of the mountain (even though I took the level tunnel) and also the elevation down and back up the gorge (even though I took the level bridge). Even if your route doesn't have tunnels or bridges, if you ride a road that is near a ridge or canyon, they may use elevation data from the bottom of the hill or the top of the hill, even though your actual route is neither. Their errors continuously add up since elevation gain is cumulative, though they do use some smoothing algorithms to try to reduce the error accumulation.

vesteroid 06-29-12 06:03 PM

I don't trust any of them...I had one ride with garmin, training peaks, and strava all about a thousand feet apart

CraigB 06-29-12 07:46 PM

Quote:

Originally Posted by Mithrandir (Post 14422680)
Garmins (all of them as far as I know) use barometric altitude.

If you check Garmin's website, and narrow your search of bike units by feature, one of the features you can search by is a barometric altimeter. And it says only one of the three units has it. But if you look at the listed specs for each unit, it shows up on both the 500 and the 800. Only the 200 lacks it. Who knows what's correct.

TrojanHorse 06-30-12 12:01 AM

The 500 does have barometric altitude... it generally sets your altitude based on your GPS position when you turn it on, and then it just goes by the delta.

I don't think the "corrected" elevations are correct either, necessarily. i have rides with unusual spikes in the middle of a gentle grade that make me question the value of "correcting" the elevation. leaving it alone does leave me with less elevation per month but oh well. It seems to be more accurate on actual mountains rather than rolling hills. YMMV.

CBuff 06-30-12 12:14 AM

Those of us tracking elevation need to use something.
I'd prefer to use the most accurate measure, and it appears to be the corrected Garmin 500 (with barometric altimeter) number.
So until I find something better, I'm sticking with that.

rm -rf 06-30-12 02:41 AM

No elevation calculation method is perfect.

The Garmin barometer is very accurate. Holding the Garmin, it will count off 1 foot at a time as I walk up stairs. But it's affected by changes in air pressure, even over a period of minutes. I was stopped for 10 minutes in the Appalachians, and a storm cloud arrived. The Garmin elevation dropped over 200 feet while sitting still. And I think that wind pressure on fast downhills affects it, too.

I never use the "corrected" elevations. This adjustment is more useful for GPS units with no barometer, since it's difficult to calculate elevations accurately just using GPS signals. So the mapping data elevation adjustments can help there.

Even with the elevation drift from weather changes during the ride, I think the barometric elevations are more accurate. Any method probably can be off by 5% or 10%. But that's good enough for bike ride recording.

Here's an example where the mapping data is less accurate than the Garmin barometer.
This is a screenshot from the free MyTourbook software. I selected their Adjust Altitude option, using SRTM elevation data. But the software has to determine the exact elevation of each point on the route, by taking known elevation points and estimating the road's exact elevation. On a mountain slope, it can be off a few feet, and the differences can add up.

The green line is the Garmin data. The red line is the mapping data. The Blue Ridge Parkway road's grades are smooth like the Garmin, with no steep grades. The differences are small, 20 or 30 feet, but that can add up over a whole ride.

Scenic overlooks on the Parkway have an elevation marker, and the Garmin would often be within 10, 20 or 30 feet. That's very good.

https://lh4.googleusercontent.com/-5...k%2520adj1.jpg


The full route:

https://lh3.googleusercontent.com/-c...k%2520adj2.jpg

rm -rf 06-30-12 02:57 AM

I haven't used MapMyRide for a few years, but back then, they were ignoring any elevation change less than 10 meters (33 feet).

That makes some sense, since a little roller on a flat road probably shouldn't count toward the elevation gain. But it does underestimate rides with a lot of small hills, and most other calculations count every tiny bump in elevation.

I did a recent 38 mile ride on very flat roads, with three or four hills about 60 feet high. But the Garmin still reported 1200 feet of gain for the whole ride. The highest point in the ride was only 100 feet above the lowest elevation. All those 3 foot, 10 foot, or 20 foot elevation changes added up.

Mithrandir 06-30-12 07:52 AM

Quote:

Originally Posted by TrojanHorse (Post 14423979)
The 500 does have barometric altitude... it generally sets your altitude based on your GPS position when you turn it on, and then it just goes by the delta.

I don't think the "corrected" elevations are correct either, necessarily. i have rides with unusual spikes in the middle of a gentle grade that make me question the value of "correcting" the elevation. leaving it alone does leave me with less elevation per month but oh well. It seems to be more accurate on actual mountains rather than rolling hills. YMMV.

Agreed. Last year on my 2nd century the Garmin told me I climbed 2000 feet, then when I turned on elevation corrections that jumped to 5000 feet. There ain't no way I did a 5000 feet course, I'd be dead!

JakiChan 08-08-12 03:22 AM

I hate to necropost, but as a new Edge 500 owner I was wondering - is there an accurate database or something so that I can get the elevation for my driveway? I usually leave from here, so that seems like a good elevation point to set.

ChrisM2097 08-08-12 07:27 AM

Quote:

Originally Posted by JakiChan (Post 14579464)
I hate to necropost, but as a new Edge 500 owner I was wondering - is there an accurate database or something so that I can get the elevation for my driveway? I usually leave from here, so that seems like a good elevation point to set.

Google earth

volosong 08-08-12 08:22 AM

Interesting discussion.

Quote:

Originally Posted by rm -rf (Post 14424090)
...This is a screenshot from the free MyTourbook software...

I like "free". Available for Linux. I'm all over it! Thanks for the tip.

CJ C 08-08-12 10:21 AM

I have a Garmin forerunner 305 and elevation is never correct. I live in the flatlands so i really dont care what is says anyway

JakiChan 08-08-12 02:09 PM

Quote:

Originally Posted by ChrisM2097 (Post 14579978)
Google earth

And what's the accuracy of that dataset? I couldn't find any information.

ChrisM2097 08-08-12 02:17 PM

Quote:

Originally Posted by JakiChan (Post 14582046)
And what's the accuracy of that dataset? I couldn't find any information.

If you really need to know the exact elevation off a location with a foot or two, you may have to do the work yourself.

As I mentioned earlier, i compared the elevations of certain landmarks, as gathered from Google earth, to my Garmin edge 500. As I passed those landmarks, I glanced down at the screen and the elevation was pretty much exactly what I expected to be.

That's good enough for me.


All times are GMT -6. The time now is 08:20 PM.