GPS & Software data mismatch
Hoping someone can help answer what may be a very simple question.
I have a Garmin etrex Legend C whose trip computer info such as moving average and overall average is always faster than the programs I export this data to. I have been using Topofusion and just recently converted to SportTracks and they both show slightly slower times than the trip computer.
Any ideas why the discrepancy?
It depends on the software you're using. One thing to consider is that when your GPS calculates distance and speed on the trip computer, it does not take elevation change into account. Some programs will post-process your data to allow for elevation change based upon the elevation data the software itself uses. Some will base those calculations upon the elevation data your GPS saves, assuming your GPS saves elevation data.
Depending on what software you use to view your GPS data, you will see different speed, distance, and elevation readings. Sometimes the discrepancy isn't too significant, but sometimes it can be. It depends on a number of factors:
GPS signal quality
accuracy (WAAS or not, satellite geometry, and a host of other factors)
barometric altimeter or GPS altimeter
if barometric, how often do you calibrate?
if barometric, was a weather system moving through during the course of your ride?
what is the resolution of the DEM data your software is using?
does the GPS or your software 'clean up' your GPS data by removing extra points and consolidating your route?
Any or all of these can be at play for any given occasion. All things considered, you should be impressed you get what you do get.
If, however, you want the absolute most accurate readings possible, use a wired cycle computer with the sensor placed closer to the outer edge of your wheel, and be absolutely meticulous (if not OCD) with calibrating it, carefully noting tire pressure, and making sure you weight the bike appropriately so the tire compresses as it rotates.
Nate, you make me feel dumb.
I will try and process this over the next few days.
Thank you for the reply.
You don't explain enough on what numbers are off. However, the fraction of time moving and stopped often vary by 10 percent or more, depending on what thinks you're stopped based on your speed. The total time should be identical on all, as they know your start an finish time.
Originally Posted by Papa_Otter
If your moving time varies, so will your average moving speed. You have to judge for yourself which data is actually right. I find Motion Based is close enough for me, even though it differs from my bike computer and GPS odometer.
On another note, if I climb up a hill slower than 1.5 MPH, my bike computer thinks I'm dead and stopped when I'm actually moving like a snail. My GPS is better at picking up this slower motion. So many times my bike computer will show less moving time than GPS.
It's not slow, it's doing more MTB time.
Try putting the wheel sensor closer to the hub if you have problems getting speed readings at slow speeds. Putting the sensor towards the outside edge of the wheel will give better accuracy at high speeds, but closer to the hub will make the computer more sensitive to slow speeds. You should eventually find a location that works for you. BTW, you don't have to recalibrate wheel diameter if you change the sensor location because the computer knows that each time the magnet passes, the wheel has rolled x distance, you're just changing the amount of time it takes for that wheel sensor to make another pass.
Originally Posted by BigLarry
BigLarry pointed out another potential source for the error. This seems to happen when the data is post-processed, too. I did a canoe trip this weekend that lasted 9+ hours and covered nearly 21 miles. My GPS tells me the moving time was 8:45 with a 1:45 stopped time. After post-processing, Motionbased took my moving time down to 8:30. I'm assuming that the discrepancy is because there were times when I was navigating log jams that my GPS thought I was still moving because of some signal errors (probably multipath errors considering I was in a valley surrounded by canopy cover). But when Motionbased deleted a bunch of extra points, it was better able to determine that I was more or less stationary at those points.
[quote=NateHawk] BTW, you don't have to recalibrate wheel diameter if you change the sensor location because the computer knows that each time the magnet passes, the wheel has rolled x distance, you're just changing the amount of time it takes for that wheel sensor to make another pass.quote]
Not trying to be a SA, but doesn't the hub rotate at the same rpm as the rim?
If so, wouldn't the time it takes for the sensor to rotate one revolution be identical?
Since the rim has more distance to travel per revolution than the hub, wouldn't the speed of the magnet get faster as it moves towards the rim?
Does the sensor have a preference for how fast the magnet moves over it in terms of an accuracy sweetspot?
Last edited by bsieb; 06-23-2008 at 12:31 PM.
I think you caught me, bsieb. Maybe the difference does have to do with the magnet speed. Now I'm not so sure.
One thing that hasn't been mentioned yet is the following. Unless you set your tracklog to record a point every second, the GPS and the PC software are looking at different data. The GPS can look at your speed every second and decide whether you are moving or not. The software only gets whatever the GPS stored, after the fact.
Originally Posted by Papa_Otter
Generally speaking there's a speed at which software considers you "too slow" to be moving. In TopoFusion there's a setting for this. 0.1 mph or less might be appropriate for a super hike-a-bike ride or a mountaineering expedition, but in general I find something like 0.3 mph to be closer to "reality" for MTB use.
Hope that made some sense.