Garmin 60CSx altimeter testing
Hello there!
Inspired by the posts about the Garmin 60CSx altimeter I decided to investigate its behavior a little further and compare it with the GPS altitude data. Thanks to this forum, I first installed the 2.70 software update and got rid of the altitude bugs reported earlier.
First an short introduction to pressure altimeters. A traditional aircraft altimeter works by converting the ambient pressure to an altitude value using the fact that ambient pressure decreases around 1 mb (millibar, sorry for my metric units) for each 8 metres (27 ft). Before each flight it must always be reset according to the current barometer setting, in order to show correctly the altitude above sea level. As the barometer reading, i.e. the ambient pressure at sea level, is constantly changing with the weather, this resetting must be redone frequently (every hour or so).
In the 60CSx, this resetting is according to an internet posting done automatically by using the GPS altitude as an reference and making the adjustment every 15 minutes.
In this first test the altimeter was in auto-calibration mode and in the variable elevation mode. The GPS was stationary near a window with a view to the northeast. Satellite reception was medium (see screenshot). The real altitude of this location is around 25-30 metres (75-90 ft) above sea level.
The test was run for around ten hours during the night (22:16-08:58). GPS altitude was recorded via the NMEA interface and OziExplorer. Ozi was configured to record a track point every time the position changed 5 metres . This resulted in a track with 4591 track points (and a travel distance of 44 kilometres). The 60CSx "Barometric" altitude was recorded by the unit's own track log. This track contained 1120 track points. As a reference I used the 48-hour ambient pressure plot available from the website of Vaisala Inc., a manufacturer of industrial weather sensors, located about 10 km (6 miles) away.
Initial conclusions
Let’s start with the ambient pressure plot and compare it with the Vaisala data (at the end of their 48h plot).
To my eyes they are quite similar, with the exception of the Garmin’s -2 mb dip at 22:59 and 1 mb peak at 23:40 respectively. I have no explanation for this peak but it is possible that it is caused by a draught from an open window or similar local pressure variation.
The Vaisala data shows a pressure variation of around 2.5 mb during the test. This means that a traditional pressure altimeter would have shown an altitude variation of around 20 metres (61 ft).
Next comparison is the altitude. On the following Excel graph I have plotted the GPS altitude in dark blue. As you can see the data points are quite scattered, so I also plotted the rolling 50-point average (violet) which looks somewhat better. The yellow line is the altitude data as recorded by the Garmin altimeter.
It seems quite clear that the Garmin uses the GPS data in smoothed form to adjust its altimeter reading. In this case the GPS altitude varies quite much, so the benefit of the altimeter is mainly to dampen the variations from a 200-metre (610 ft) range to a 50-metre (150 ft) range. The altimeter plot also managed to suppress the erroneous GPS altitudes seen at around 05:30, which went as high as 18000 feet. The GPS intermittently lost satellite contact during that time.
This Garmin’s screen capture of the elevation (altitude) is a little misleading since it plots elevation over distance, not time. So the nice long flat part of the graph actually only represents 1.5 hours out of the 10-hour measurement period. I subsequently found out that the Garmin can also plot elevation over time, I will use that next time.
Finally we have the Garmin’s barometer plot. I’ll have to retest this as I have problems making sense of it. I theory a higher pressure should correspond to a lower altitude, but the comparison with the Excel altitude plot is not simple.
So far the conclusion about 60 CSx altimeter errors is that the altimeter with automatic calibration showed up to 50 metres (150 ft) variation during the test. If the auto-calibration mode would have been off, the variation would have been slightly better, around 30 metres (90 ft) (the main error due to the mysterious dip and peak in the ambient pressure plot) but in that case an initial manual calibration would have been necessary. In this mode also the natural pressure variation will introduce errors. For example, the Vaisala archives showed several recent days with a sustained pressure change gradient of 0.5 mb/hour, which during e.g. an 8-hour period would introduce a 32 m (105 ft) altitude error.
Comments are welcome, both on the results and if there is something to refine in the methodology. My plan is to test next time with auto-calibration off, and also try to see if a better GPS satellite view would improve the GPS altitude data, which to me seems less accurate than expected.
This post has been edited by 60csxuser: 26 May 2006 - 04:16 AM
Help
I have the Map76CSx that had this error in the first week and no repeat failures for the last 6 weeks. On the way into work this morning I looked at the GPS and saw the elevation at _______. I did a normal reset and still no valid altitude. I then powered the unit off and back on and calibrarted the altimeter to the current GPS elevation. That corrected the problem for now.
Posted 09 May 2006 - 07:53 PM
