Announcement

Collapse
No announcement yet.

Barometer/Altitude issues

Collapse
X
  • Time
  • Show
Clear All
new posts

  • Barometer/Altitude issues

    So, after my initial 21.5mi 'test' run with the 910 and my wife taking along her 310, the elevation calculation was significant, I thought, but I wanted to give it another test on a course where I pretty much KNOW the elevation gain, having run it a couple of hundred times, and it being rather straight and a path rather than a rugged trail.

    First, here is the 21.5 mile comparison
    910 - http://connect.garmin.com/activity/142125253
    310 - http://connect.garmin.com/activity/142230504
    Difference: 923ft or 30.12%

    Second, 4 miles on the Boulder Creek Path. FWIW I've run this run many, many times...it's between 90-92ft of gain depending on which curbs I jump
    910 - http://connect.garmin.com/activity/142609130
    305 - http://connect.garmin.com/activity/137296844
    Difference: 298ft or 76.41% - OUCH.

    as an added bonus, here is my first run with the watch, a 4.44mi out and back as well. that...doesn't look like an out and back.
    http://connect.garmin.com/player/141368766

    If you look at the 910 in 'player' mode (because garmin connect still hasn't fixed the bug that starts at 0ft so it makes people who live at 5kft elevation profile flat) you can see that the elevation profile for the out and back run...is not out and back. Perhaps weather, as we are getting some here in Boulder at the moment, but wow....76% is a HUGE difference.

    If I turn on elevation correction on the 910 one, it comes close to what I usually get, off by a a few feet, nothing major.

    I don't know what to think, really. I'd rather trust the GPS, which was supposed to not be as accurate.

    The only other thing I can think of, and, I know this sounds really silly, but hear me out. The barometric altimeter is SO accurate it's calculating my arm movements going up and down? I know my edge500 has a barometric altimeter, perhaps Garmin used the code from this, which is typically used on a bike, and used it for the 910, without taking that into consideration, and maybe the calculation for elevation or 'ascent' needs to be smoothed as well? Though, that doesn't explain the elevation profile in garmin connect.

    Thoughts?

  • #2
    Here's my reply from the other thread:

    I also noticed the elevation was way off. When I started it was at 30meters however within 5km it went down to -15 meters and hovered around there for most of the run even though it should have been around 2-5 meters (runnning along a river only a few k's from the beach). A cool change was forecast to come through today which may have affected it but the qNH only ever changed from 1011 to 1015 while I was running so I'm surprised there was such a large change in elevation. Has anyone else experienced this?

    Also today I was sitting outside playing with the watch and during this time I noticed the elevation went from 31meters down to (minus) -21m then to 35 meters, over a few minutes! I didnt move during this time and the weather was stable (so no sudden pressure changes)

    Comment


    • #3
      Just wanted to make sure that you let the 910 settle on the initial calibration of the altitude before you start running? It is my understanding that it calibrates using the mean of the altitude given by GPS, but this then requires keeping the 910 still for a while when at the start position.
      DISTANCE: 3000M___ - 5000M___ - 10000M__ - 21098M_ - 42195M_ --- Current VDOT: ?
      PB______: 10:50,09 - 18:40,20 - 38:39,00 - 1:25:39 - 3:10:26

      Comment


      • #4
        Yes, I let the watch sit for a while while I get ready, always.

        Here is a run from today. It's was an exact out and back with a little bit of climb, but...the elevation chart is obviously slanted. It says I ended about 100 feet higher than where I started.

        http://connect.garmin.com/player/142838740

        elevation correction corrects it, and it's off by 100 or so ft....but I'm guessing the elevation correction throws the barometric numbers out the window and just goes by GPS.

        -Rob

        Comment


        • #5
          To test the arms mouvements theory, put the watch in your back pack or breast pocket ?

          Comment


          • #6
            Originally posted by RTIMKO View Post
            Yes, I let the watch sit for a while while I get ready, always.

            Here is a run from today. It's was an exact out and back with a little bit of climb, but...the elevation chart is obviously slanted. It says I ended about 100 feet higher than where I started.

            http://connect.garmin.com/player/142838740

            elevation correction corrects it, and it's off by 100 or so ft....but I'm guessing the elevation correction throws the barometric numbers out the window and just goes by GPS.

            -Rob
            Barometric pressure changes over time during the cause of the day and that will affect barometric height readings. Not sure how long you were out for but I would not have thought that 100 ft differential would be unheard of.

            I travelled some years ago with a watch with inbuilt altimeter. You could set the barometric pressure at the beginning of the day. Without doing anything,just leaving the watch at rest, there would be elevation changes.

            Air pressure is not constant over time.

            Comment


            • #7
              In the Polar software it is possible to set the finish elevation to the same as the start elevation, and the whole activity is then "tilted" up or down depending on if the finish elevation is increased or decreased. Of course this is only useful if start and finish is indeed (almost) level.

              I would hope that it's possible to enter GPS-positions with known elevation figure and that the barometric altimeter automatically adjust when passing (near) such a point. Very useful when running in frequently used paths. At least I'd like to be able to enter the often used start points and end points, as well as high points and low points along often used courses.
              DISTANCE: 3000M___ - 5000M___ - 10000M__ - 21098M_ - 42195M_ --- Current VDOT: ?
              PB______: 10:50,09 - 18:40,20 - 38:39,00 - 1:25:39 - 3:10:26

              Comment


              • #8
                Originally posted by PHILIPSHAMBROOK View Post
                Barometric pressure changes over time during the cause of the day and that will affect barometric height readings. Not sure how long you were out for but I would not have thought that 100 ft differential would be unheard of.

                I travelled some years ago with a watch with inbuilt altimeter. You could set the barometric pressure at the beginning of the day. Without doing anything,just leaving the watch at rest, there would be elevation changes.

                Air pressure is not constant over time.
                Sure, but with a built in GPS it should be rather simple to periodically re-calibrate this on the fly. I was out for under an hour...and the difference was over 100ft. (on a 300ft run!)That's significant. Especially for something that is being touted as more accurate than the previous GPS based model.

                At least give me some configurable options to turn off the barometric altimeter. Nothing more frustrating than trying to bag a peak and it saying you're only 100ft of climb from the top only to have it be off by a few hundred feet more of climb.

                Comment


                • #9
                  I'm seeing the same thing. I ran along a trail today that varies by at most 10' along a lake and the elevation of the trail is right at 45'. I let my watch sit for better part of half an hour before starting my run which lasted 45 minutes. The high elevation that was recorded was 219'. The low was 186'. Total ascent was recorded at 245'. So much for increased accuracy. I see several problems here. First, the elevation should have adjusted to 45', even if it took a bit. 2nd, it appears there is absolutely no filtering going on with the ascent calculation. Even though the elevation doesn't vary by as much as with GPS altitude, there is still some noise in the data. They need to filter out any up's and downs by less than 20' (or similar) since those are most likely pressure variations rather than actual elevation changes.

                  They also need to add a mechanism for manually calibrating/setting the altitude. At least to the level that you can now do on the Edge 705/800 such as creating a way point with an elevation and having it use that if you press the start button near the way point. Even if it had automatically adjusted over 30 to 60 minutes, who wants to sit at a trail head for that long waiting for the their watch to adjust?

                  I know they will fix these things with updates, but those often take a long time. These are issues they have already dealt with with the Edge so I find it ridiculous that they couldn't get this right from the beginning. This is the key features that makes this better than the 310 for running.

                  Comment


                  • #10
                    One other note. Looking at my data, I see a lot of 6 to 12 foot jumps in altitude that only last a couple of seconds. This is probably why my ascent is so high for a flat trail.

                    In addition, there are a lot of even smaller bumps in the data (like 0.6' - .fit files record in 1/5 meter granularity). The odd thing here is that the data has lots of these tiny ticks but none that are slightly bigger like 2/5 or 3/5 of a meter. So it doesn't appear to really be using the 1/5 meter granularity but rather getting some rounding errors.

                    With the 310xt, it seemed like the smallest recorded interval was a full meter. In either case, it probably doesn't make sense to go to a granularity beyond 1 meter since neither GPS or the BP sensor in the 910 are accurate to beyond that level.

                    You can see this in the following workout. You need to zoom in to the altitude to see this. http://connect.garmin.com/activity/143880165

                    Comment


                    • #11
                      Originally posted by EKUTTER View Post
                      They also need to add a mechanism for manually calibrating/setting the altitude. At least to the level that you can now do on the Edge 705/800 such as creating a way point with an elevation and having it use that if you press the start button near the way point.
                      I don't have a 910 (yet) but I do have a copy of the manual, and pages 6-7 describe how to save locations with known elevations to do exactly that.

                      Not that that's any excuse for the other problems you are having...

                      Comment


                      • #12
                        I tried another run today, this time intervals arounda 400 meter oval which would be almost completely flat. With correction turned on you can see a consistent pattern in the elevation (though it is reporting a 3 meter difference which is not correct) but with correction turned off it is all over the place and thinks the height difference is as great as 12 meters at one point! (see attached)
                        Attached Files

                        Comment


                        • #13
                          Yeah, I gave up on the barometer. I just elevation correct everything now. I've done about 5 or 6 runs with the 910 and not one has been close enough to accurate that it's trustworthy. Boo.

                          Comment


                          • #14
                            Originally posted by RTIMKO View Post
                            Yeah, I gave up on the barometer. I just elevation correct everything now. I've done about 5 or 6 runs with the 910 and not one has been close enough to accurate that it's trustworthy. Boo.
                            Wow, that was the entire reason I was going to buy one of these. My Sunnto does great with its altimeter. I don't need the triathlete functions of the 910 and I am likely to get the new Suunto Ambit. It fills my needs better. Too bad, Garmin has a nice suite of software for Macs but I want a watch that will give me ascent/descent data and one I can load courses into. A baro altimeter should not be that hard to get to work correctly. Do you think there is something wrong with your unit?

                            Comment


                            • #15
                              Originally posted by MICK003 View Post
                              I tried another run today, this time intervals arounda 400 meter oval which would be almost completely flat. With correction turned on you can see a consistent pattern in the elevation (though it is reporting a 3 meter difference which is not correct) but with correction turned off it is all over the place and thinks the height difference is as great as 12 meters at one point! (see attached)
                              Don't have mine yet but have been watching the forum to see how I need to set it up (most efficiently) once it arrives. So it looks like turning ON elevation correction is a good bet?

                              Comment

                              Working...
                              X