Elevation change errors and their effect on power
Posted: Sun Jul 01, 2012 3:51 pm
Quite awhile back I mentioned that I was seeing up to a 5% scatter in the elevation change on a long climb that I do fairly often. Since the iBike uses barometric elevation readings to determine slope, this can result in a pretty big power error on long climbs (or descents). I thought I'd revisit this to see if there was any way to improve the results.
The clearest website I found regarding barometric altimeter measurements is here: http://www.hills-database.co.uk/altim.html
If I understand correctly, there are significant sources of error even if the barometric sensor is working perfectly. Temperature looks like the most important variable, with humidity also having an effect. The drift in barometric pressure is another potential factor on long rides, but since this climb takes me ~42 minutes, the drift should be quite small.
I tallied some of my trips up this climb below. I also looked to see if there was any correlation between the measured elevation change and two other variables: average iBike temperature during the climb, and the iBike barometric reading. Surprisingly there was a correlation with both... meaning (I think) that there is a chance the iBike could be made more accurate in this respect, using the existing sensors.
Anyone else interested in posting their numbers? I wonder if this phenomena is fairly consistent with different units and different elevations. It should be a climb that is fairly long (10 minutes minimum), that you have ridden numerous times, and can identify the "bottom" and "top" accurately.
The clearest website I found regarding barometric altimeter measurements is here: http://www.hills-database.co.uk/altim.html
If I understand correctly, there are significant sources of error even if the barometric sensor is working perfectly. Temperature looks like the most important variable, with humidity also having an effect. The drift in barometric pressure is another potential factor on long rides, but since this climb takes me ~42 minutes, the drift should be quite small.
I tallied some of my trips up this climb below. I also looked to see if there was any correlation between the measured elevation change and two other variables: average iBike temperature during the climb, and the iBike barometric reading. Surprisingly there was a correlation with both... meaning (I think) that there is a chance the iBike could be made more accurate in this respect, using the existing sensors.
Anyone else interested in posting their numbers? I wonder if this phenomena is fairly consistent with different units and different elevations. It should be a climb that is fairly long (10 minutes minimum), that you have ridden numerous times, and can identify the "bottom" and "top" accurately.