Exaggerated total ascent figure

I bought this unit in October 2021. I used it on a trial walk in an area I knew. It soon became apparent that the unit was recording ascent incorrectly. The displayed figure was nearly twice the actual ascent. The unit was straight out through box with factory settings. Suspecting a faulty unit I contacted my supplier with a view to a warranty replacement. The supplier asked me to contact Garmin support for advice before returning the unit. I contacted Garmin support at the end of October 2021, who asked for various files to be sent to them. When compiling these files I noticed an anomaly between the total ascent figure and the elevation chart. Intriguingly the chart showed the correct total ascent whilst the total ascent figure was wrong. Since submitting the requested files I have not had any meaningful response from Garmin support. They won't unequivocally say there is a problem with my unit, so I can return it under warranty. I notice this problem has been reported earlier by RosscoP over a year ago.

Is there a generic problem with the GPSMAP 66s or is there a fault on my unit?

0 jlg2 over 2 years ago

Hi, Legnonefan1.

Two possibilities other than a defective device:

Could the exaggerated ascent you observed be the result of a not being aware that the 66i continues to record ascent/decent when track recording has been stopped? This is a counterintuitive design that I've complained about. If you stop the track recording the trip computer odometer stops but the ascent keeps on changing. So if you finish a hike, stop the recording, drive up and down hills, nthen look at your Trip Computer, you are bound to get an exaggerated total ascent.

Were you in Expedition Mode? In my experience, the track record elevation is wildly incorrect in Expedition Mode. I believe this is (or was) true of all 66i's. I haven't checked lately. So that experience may now be outdated. Judging from the timing of the errors, I suspect the errors are introduced during the resume process, but that's just a guess. In my experiences those errors can be more than 1000 feet and may be cumulative. I've complained about that too but not too loudly because I don't want Garmin to fix the ascent inaccuracy in Expedition Mode with a fix that increases power consumption reducing the time between charging.

0 Legnonefan1 over 2 years ago in reply to jlg2

Hi jlg2,

The error accumulated as the track was recording. After crossing 3 contour lines on a UK map the device aa reporting 50m of ascent, instead of 30m. At the top of this particular slope the device reported 300m of ascent when it should have been 160m. Interestingly I can say this particular device did not register ascent when track recording was paused or stopped! But it it did record ascent whilst descending (steeply). I'm happy to accept a 10-15% error in ascent figures but 100%? No.

There has been a development since I posted. The device started shutting down after a few minutes use. It's now on it's way back to Garmin.

I'm hoping the replacement device performs better.

0 jlg2 over 2 years ago in reply to Legnonefan1

In a perfect world, Garmin would have loaner benchmark devices, so users could compare unexpected contemporaneous, side-by-side behavior to distinguish design problems from individual unit problems.

0 twolpert over 2 years ago

Some of this may be due to settings on the device, logging interval, and environmental conditions. Or not.

Note that I have no idea what the default settings are. So I don't know what settings were in use.

Ascent/descent is based on the (barometric) altimeter. Since we are not concerned with absolute altitude (only changes), we don't need to worry too too much about calibrating the altimeter. What we do need to worry about is interference due to auto-calibration based on GPS-derived altitude or due to weather-induced pressure fluctuations.

With regard to GPS-derived altitude, check this setting:

Settings > Altimeter > Auto Calibration

It should NOT be set to continuous. Setting it that way results in continuously adjusting the altitude based on GPS readings. Vertical error in GPS location is significant. Something on the order of 1.7 to 2 times the horizontal error. Making these adjustments continuously is bad for your accuracy. This also exposes you to particularly bad readings in challenging conditions. It only takes a few of those to throw the cumulative values way off.

Especially when looking at ascent/descent, the difference between Off and Once is not significant.

If the weather is changing (for example, pressure is rising or dropping ahead of a front), this will induce errors in the ascent/descent. The unit has no way to distinguish changes in pressure due to true changes in altitude vs. changes due to atmospheric conditions. This is likely to be a (very) secondary consideration because (presumably) the time between readings is small compared to the time it takes for a significant weather-related pressure change to occur.

Finally, time between readings (typically, between logged points) matters. As an extreme example, take a round trip up and down a steep hill. If your timing is such that you record points only at the beginning and end of the round trip, the accumulated ascent/descent will be very close to zero. You will have missed all of the actual ascent/descent ON the hill. Within reason, shorter intervals will produce better results than longer intervals.

I can't speak to the contention that the unit accumulates ascent/descent when not recording an activity. Never had a reason to care about what it does when not recording.

I vaguely recall some discussion of differences in ascent/descent as recorded in the .fit files for activities vs. what showed up in the .gpx files (if you enable both .fit and .gpx logging). That sort of discrepancy would definitely be a firmware bug.

0 jlg2 over 2 years ago

I hope someone will correct me if this is wrong since I'm not sure. I believe that during a normal frontal passage, pressure altitude can be expected to change around 50 feet per hour for several hours before and after frontal passage. I think pressure altitude increases before passage and decreases after passage.

0 CowboySlim over 2 years ago

I have made a statistical analysis of my device over a period of time. All statistically significant number of points were recorded readings were collected at the same location in my backyard over a period of one week with successive points being at least 3 hours apart. Assuming that the errors are consistent with the Normal Distribution:

1. The standard deviation of the errors in the GPS derived elevations are 4 times greater than that of the horizontal values.

2. The standard deviation of the errors in the barometrically derived elevations are 12 times greater than than that of GPS derived elevations.

3. The sum of a statistically significant number of errors characterized as a Normal Distribution is zero.

en.wikipedia.org/.../Normal_distribution