Anyone's Garmin give way wrong elevation numbers?
#1
Thread Starter
Banned
Joined: Oct 2014
Posts: 1,297
Likes: 0
Anyone's Garmin give way wrong elevation numbers?
I rode the exact same loop a week apart. 62.1 miles both times, 1,600ft the first time and 2,900ft the second time.
I don't expect it to be perfectly accurate, but it gave almost 2x the climbing one time.
I don't expect it to be perfectly accurate, but it gave almost 2x the climbing one time.
#2
Senior Member
Joined: Sep 2013
Posts: 614
Likes: 1
Everyone's Garmin gives terrible elevation. Commercial GPS is highly inaccurate for tracking elevation changes, especially when the unit is in motion.
On Garmin Connect you can elect to have the elevation corrected when you upload your file.
On Garmin Connect you can elect to have the elevation corrected when you upload your file.
#3
Thread Starter
Banned
Joined: Oct 2014
Posts: 1,297
Likes: 0
When I started using a Garmin instead of my iPhone (Strava) I noticed it started giving me more elevation on the road biking I do in my really flat area, but gave about the same numbers for mountainous areas.
#4
Senior Member
Joined: Jan 2012
Posts: 1,051
Likes: 106
From: Kingwood, TX
Bikes: Cervelo Caledonia Rival AXS eTap, Blue Norcross AL, Lynskey Helix
I believe Garmin's use the atmospheric pressure difference to calculate altitude. I don't think they do a very good job when the pressure varies during a ride. Not sure how Connect corrects.
#5
Senior Member
Joined: May 2013
Posts: 4,400
Likes: 106
From: SF Bay Area
Bikes: Bianchi Infinito (Celeste, of course)
What Garmin unit? Only the 500 and up have barometric altimeters. Garmin connect uses topographic maps to supply altitude corrections. The barometric does a better job with small features like overpasses. Raw GPS is terrible for altitude.
#7
a77impala
Joined: Oct 2005
Posts: 1,519
Likes: 13
From: Central South Dakota
Bikes: 04=LeMond Arravee, 08 LeMond Versailles, 92 Trek 970
If it is a barometric altimeter it relies on current air conditions. Unless pressure stays constant readings will vary, I consider them ballpark figures.
#9
Thread Starter
Banned
Joined: Oct 2014
Posts: 1,297
Likes: 0
Final question--so my Garmin will read atmospheric pressures, Garmin Connect will fix it based on topographical maps, then what happens when it goes to Strava? Will Strava take the corrected number or come up with its own?
The only reason I use Garmin Connect is to interface with Strava. Garmin Connect totals always differ a little bit from Strava on the same ride. I imagine they interpret the gps download with different algorithms.
The only reason I use Garmin Connect is to interface with Strava. Garmin Connect totals always differ a little bit from Strava on the same ride. I imagine they interpret the gps download with different algorithms.
#10
Senior Member
Joined: Sep 2013
Posts: 614
Likes: 1
I don't use Strava so maybe someone else can chime in on this.
#11
Senior Member
Joined: Jun 2008
Posts: 3,247
Likes: 7
From: Northern VA
Bikes: Moots Vamoots, Colnago C60, Santa Cruz Stigmata CC, and too many other bikes I don't ride
When you upload a ride to Strava from either Garmin Connect or the device via the plug-in, it sometimes will have the text "Elevation (?)" underneath the elevation number. If you click on the (?) sign, you can have Strava correct the elevation. I assume it uses topographic data based on the GPS information from the Garmin unit.
#13
Senior Member
Joined: Nov 2008
Posts: 7,239
Likes: 8
From: Bay Area, Calif.
I've found all the common methods for determining elevation gain on bike rides to be subject to considerable error. GPS data is noisy and can have large temporary variations due to poor signal reception or reflected signals, barometric pressure is subject to weather changes, and topographic data has its own accuracy issues and also frequently reflects the topography before the road/bridges/overpasses were built. And then, partially to try to compensate for the data fluctuations, the raw numbers are filtered using various algorithms that may help reduce the effect of data glitches but can also hide multiple ups and downs actually encountered on a ride.
Looking at the detailed elevation profile can help identify large glitches in the data and I'd expect a comparison of the two profiles recorded by the OP would shed some light on the discrepancy. But a ride with lots of ups and downs will still have lots of uncertainty about the total elevation gain since it consists of adding together a large number of small climbs, each of which has an associated measurement uncertainty.
Looking at the detailed elevation profile can help identify large glitches in the data and I'd expect a comparison of the two profiles recorded by the OP would shed some light on the discrepancy. But a ride with lots of ups and downs will still have lots of uncertainty about the total elevation gain since it consists of adding together a large number of small climbs, each of which has an associated measurement uncertainty.
#14
Senior Member
Joined: Jul 2014
Posts: 386
Likes: 9
From: Oklahoma City
Bikes: 2015 Cervelo R5 Dura Ace, 2015 Cannondale Synapse 5 Disc 105, 2006 Cannondale F300
I just got a 510 and only been out with it twice since, but I noticed this too. I didn't know which one was more accurate though (I've been using Strava on my phone). I sort of hoped maybe it was Strava that had it wrong. Damn! 300 ft elevation gain on my first full century was accurate I guess (Oklahoma City = super, super flat!)
#15
Vain, But Lacking Talent
Joined: Aug 2009
Posts: 5,510
Likes: 81
From: Denton, TX
Bikes: Trek Domane 5.9 DA 9000, Trek Crockett Pink Frosting w/105 5700
What's really fun is when you first start up your device inside your warm house and then roll it outside into the cold, dry air. You're really confusing the device at that point.
#16
Vain, But Lacking Talent
Joined: Aug 2009
Posts: 5,510
Likes: 81
From: Denton, TX
Bikes: Trek Domane 5.9 DA 9000, Trek Crockett Pink Frosting w/105 5700
I just got a 510 and only been out with it twice since, but I noticed this too. I didn't know which one was more accurate though (I've been using Strava on my phone). I sort of hoped maybe it was Strava that had it wrong. Damn! 300 ft elevation gain on my first full century was accurate I guess (Oklahoma City = super, super flat!)
#17
I've found all the common methods for determining elevation gain on bike rides to be subject to considerable error. GPS data is noisy and can have large temporary variations due to poor signal reception or reflected signals, barometric pressure is subject to weather changes, and topographic data has its own accuracy issues and also frequently reflects the topography before the road/bridges/overpasses were built. And then, partially to try to compensate for the data fluctuations, the raw numbers are filtered using various algorithms that may help reduce the effect of data glitches but can also hide multiple ups and downs actually encountered on a ride.
Looking at the detailed elevation profile can help identify large glitches in the data and I'd expect a comparison of the two profiles recorded by the OP would shed some light on the discrepancy. But a ride with lots of ups and downs will still have lots of uncertainty about the total elevation gain since it consists of adding together a large number of small climbs, each of which has an associated measurement uncertainty.
Looking at the detailed elevation profile can help identify large glitches in the data and I'd expect a comparison of the two profiles recorded by the OP would shed some light on the discrepancy. But a ride with lots of ups and downs will still have lots of uncertainty about the total elevation gain since it consists of adding together a large number of small climbs, each of which has an associated measurement uncertainty.
Also there is an option to set the start point elevation in the garmin and I set my home the same elevation that my iphone say this helps a little
#18
Senior Member


Joined: Jun 2008
Posts: 9,201
Likes: 289
From: Vancouver, BC
To be slightly more specific, I think some folks like the barometric altimeters for riding because it gives decently accurate data on the fly, particularly if one is doing lots of climbing and wants to add the ascent rate data to a page. Most of us don't use this and really just want overall elevation at the end.
#19
Senior Member
Joined: Nov 2008
Posts: 7,239
Likes: 8
From: Bay Area, Calif.
But there are still plenty of cases where the "correction" function, whether by Garmin Connect or by Strava actually makes the accuracy worse rather than better. The correction is done by looking at topographic data and if that data doesn't have the information about where there's a bridge, or overpass, or where the road builders smoothed out the landscape then the 'correction' will be way off. Plus the original topo data is also subject to measurement error and I can easily find places on topographic maps where errors were made.
#21
Senior Member
Joined: Sep 2013
Posts: 614
Likes: 1
a mobile phone have gps and the help of the internet, I trust in strava and I don't have to press the option of elevation correction like the garmin connect site!!!!! I use garmin edge 510 just for distance, speed and heart rate.
Also there is an option to set the start point elevation in the garmin and I set my home the same elevation that my iphone say this helps a little
Also there is an option to set the start point elevation in the garmin and I set my home the same elevation that my iphone say this helps a little
#22
I'm doing it wrong.

Joined: Jun 2009
Posts: 4,875
Likes: 2,814
Bikes: Rivendell Appaloosa, Rivendell Frank Jones Sr., Trek Fuel EX9, Kona Jake the Snake CR, Niner Sir9
I just looked on strava at a friends ride he did this morning...cold day, rainy. It was an out and back route about 20 miles long. The elevation is jacked from the beginning to the end of the ride. The same hills he rode in the beginning of the ride are measuring two hundred feel taller towards the end of the ride. For instance, one hill measures 430 feet at the beginning of the ride, but 483 at the end
So, I would say that his garmin's elevation feature was not accurate this day. I have no idea if he has elevation correction turned on (I doubt it), but it is pretty clear the barometric is inaccurate by quite a bit.
So, I would say that his garmin's elevation feature was not accurate this day. I have no idea if he has elevation correction turned on (I doubt it), but it is pretty clear the barometric is inaccurate by quite a bit.
#24
Middle-Aged Member
Joined: Apr 2011
Posts: 2,276
Likes: 1
From: Mesa, AZ
Bikes: Bianchi Infinito CV 2014, TREK HIFI 2011, Argon18 E-116 2013
Has anyone noticed that you can set the altitude for your location (I used a USGS TOPO) for home? If you start close enough to it, it will calibrate for altitude at start, and then the rides are always more accurate. Since I set that parameter in my 800 I typically have about 10 feet delta between climbed and descended; before I set the location of "Home" it could be has many as 100' delta per 20+ mile rides.
#25
just another gosling


Joined: Feb 2007
Posts: 20,583
Likes: 2,690
From: Everett, WA
Bikes: CoMo Speedster 2003, Trek 5200, CAAD 9, Fred 2004
Huh. I set the home elevation on my 800. I use the Garmin speed/cadence sensor. My Garmin is out in the ambient for at least an hour before I start riding. Don't know what makes the difference, but my Garmin data is alwayis spot on. When I ride a loop, ascent and descent are almost exactly the same, and the numbers are in good agreement with the RWGPS numbers for that route. Last 44 mile ride I did, ascent and descent differed by 3'. Unless something really weird is going on with your weather, barometric elevation shouldn't vary by more than 50' during a normal ride. Either you're doing it wrong or you have a defective unit.



