GPSMAP 66s/st: Total ascent way too low (by 20% - 30%) [Bug 148]

Discussion related to the Garmin GPSMAP 66s/st GPSr
steerage250
Posts: 98
Joined: Fri Jan 04, 2019 2:48 am

Re: GPSMAP 66s/st: Total ascent way too low (by 20% - 30%) [Bug 148]

Unread post by steerage250 »

Spiney wrote: Mon Aug 12, 2019 12:11 pm When you upload the GPX track to a track analyser such as Trakan, it is recalculating the data based on the recorded track points and ignores the summary statistics. In your Trackan screen shot, you have the vertical filter set to 15m. If I'm not mistaken, this setting further filters the data from the track file (eg try resetting it to 1m and you will see the total ascent figure increase).
Spot-on @Spiney! Dropping the "Vertical" figure in Trackan to 0.1m brought the total ascent & descent close (5% less) to the totals displayed on the GPSr.

So, given the accuracy of altitude measurements, and difference between the rate that the GPSr is summing compared to the track data recorded at 1 sec intervals and summed later, I think that's pretty acceptable.

@Themistockles
I think the solution to your problem is to go slower - walk it :)
Themistokles
Posts: 45
Joined: Sun Jun 30, 2019 7:47 am

Re: GPSMAP 66s/st: Total ascent way too low (by 20% - 30%) [Bug 148]

Unread post by Themistokles »

steerage250 wrote: Sat Aug 17, 2019 12:15 pm @Themistockles
I think the solution to your problem is to go slower - walk it :)
I see. Perhaps I should keep that device until my age forces me to do so. Hopefully I have left 15 years until then ...

Joking apart, I have found two guys over here which were willing to conduct the test I mentioned. I don't trust their results completely, but now there is a slight tendency that it's just my device that is defective. I'll try to return it and to get a replacement.
Themistokles
Posts: 45
Joined: Sun Jun 30, 2019 7:47 am

Re: GPSMAP 66s/st: Total ascent way too low (by 20% - 30%) [Bug 148]

Unread post by Themistokles »

Final test result: The misbehavior I have described is due to a serial fault.

I finally wanted to know what was going on, ordered a second GPSmap 66s and at the same time returned my first one due to the misbehavior. Then I began to test the new device.

The new device showed exactly the same misbehavior as the first one. The elevation reading was following the actual elevation extremely slowly, the total ascent was calculated way too low. The figures were nearly literally the same as with the first device on all routes I have tested.

So Garmin's (probably outsourced) "development" team even is too stupid to correctly implement a simple software filter ... not that they couldn't have taken that part of the software from the Oregon 700 or any other device which does it right.

Sorry for the rant, but I am extremely upset due to this issue. Is there really absolutely no quality assurance team over there? Obviously, they are even too sloppy and too lazy to compare total ascent results between the different types of their devices for different kinds of motion (including mountain biking), which really is the very least you could expect.

I'll return the second device as well due to that defect and recommend the same to those of my friends who are also affected by the issue. Maybe later I'll consider buying another one when I read that the problem has been fixed in a new firmware version. Hopefully we'll have found a reasonable alternative from another manufacturer until then.
User avatar
GPSrChive
Site Admin
Posts: 3881
Joined: Fri Jan 20, 2017 11:29 pm

Re: GPSMAP 66s/st: Total ascent way too low (by 20% - 30%) [Bug 148]

Unread post by GPSrChive »

Themistokles wrote: Sat Aug 31, 2019 10:20 pm Hopefully we'll have found a reasonable alternative from another manufacturer until then.
I'm afraid that is simply just not an option.

I am confident this issue, and others, will be resolved in future firmware updates.

8^)
CowboySlim
Posts: 5
Joined: Sun Sep 08, 2019 12:22 am

Re: GPSMAP 66s/st: Total ascent way too low (by 20% - 30%) [Bug 148]

Unread post by CowboySlim »

gpsrchive wrote: Sat Jul 06, 2019 3:06 pm I left my GPSMAP 66 v310 (gps 270) recording a track log yesterday, all day long, for a round trip drive of nearly 150 miles. When I arrived back where I started, I paused the track log recording and saved it. I also captured screens from the Trip Computer and Recording Controls pages for later comparison.
WRT to Round Trip recorded elevations, I would like to provide a highly simplified, hypothetical example:
A 2 hour round trip of 100 miles at essentially a constant speed of 50 mph with elevations recorded once a minute on two devices, a GPSMAP 66 recording barometrically determined elevations and a Gamin inReach Mini recording GPS derived elevations. At conclusion I would export the track logs from both in gpx format. I would then extract the elevations and time values in to a MS Excel spreadsheet and calculate the net ascent for both (net ascent = total ascent – total descent).

Barometric results: With the leaving and arriving identical ambient pressures of 29.92 inHg, the net will be 0.00 feet.

GPS results: The net will also be 0.00 ft. (Prior accuracy tests of mine have shown the elevation errors to be normally distributed; consequently, the sum of the 120 errors will be 0.00)

Repeating the test: With the leaving ambient pressure of 29.9213 inHg and a different arriving pressure, due to weather change, of 29.8133, a net ascent of 100.0 feet would be calculated, However, the GPS net ascent result will be the same as those of the first test.

Atmospheric pressure vs. altitude: https://www.digitaldutch.com/atmoscalc/

Bottom line: GPS data will give appropriate elevation results while barometric data will provide likewise as long as there are no weather changes affecting ambient pressures.
Themistokles
Posts: 45
Joined: Sun Jun 30, 2019 7:47 am

Re: GPSMAP 66s/st: Total ascent way too low (by 20% - 30%) [Bug 148]

Unread post by Themistokles »

CowboySlim wrote: Sun Sep 08, 2019 4:12 pm At conclusion I would export the track logs from both in gpx format. I would then extract the elevations and time values in to a MS Excel spreadsheet and calculate the net ascent for both (net ascent = total ascent – total descent).
This is exactly the thing which I (the OP) am not interested in. Usually, I don't care a Laptop with me on my MTB tours. As I have stated at the beginning, I am only interested in the "Total Ascent" data field which the GPS device provides.

CowboySlim wrote: Sun Sep 08, 2019 4:12 pm Bottom line: GPS data will give appropriate elevation results while barometric data will provide likewise as long as there are no weather changes affecting ambient pressures.
No. GPS data is far too inaccurate and unreliable to provide a reasonable basis to compute the total ascent. The best elevation accuracy you can get with usual GPS devices is about between 10 m and 20 m. In contrast, my Oregon 700 (for example) measures elevation differences of 1 m (yes that's right, 1 m) reliably and reproducibly.

Hence, if you ride a tour with one single ascent of 1000 height meters, the difference between GPS total ascent and barometric total ascent may be marginal (if the weather stays the same). But if you ride a tour over many small hills, GPS total ascent and barometric total ascent will inevitably differ, the latter one being correct and precise and the first one being total nonsense. Furthermore, under non-optimal reception, the GPS elevation measurement may fail at all, making the GPS-based total ascent calculation totally unreliable in addition.

The inaccuracy and unreliability of the GPS elevation for sure has been the main motivation to incorporate barometric elevation measurement into GPS devices.
CowboySlim
Posts: 5
Joined: Sun Sep 08, 2019 12:22 am

Re: GPSMAP 66s/st: Total ascent way too low (by 20% - 30%) [Bug 148]

Unread post by CowboySlim »

Actually, I am not the least bit surprised with the lack of agreement with my opinions. Yes, I have had many similar disagreements on this very subject previously on the geocaching technical forum.
https://forums.geocaching.com/GC/index. ... um/11-gps/

As a result, I performed a technically correct study gathering and analyzing both GPS and atmospheric pressure data of a statistically significant quantity. Yes, I gathered the data at the exact same position in my backyard once a day for 40 consecutive days. For these tests, the results include averages, ranges and standard deviations.

My analyses are in accordance with the definitions and principles described here:
https://en.wikipedia.org/wiki/Normal_distribution

I would welcome a quantitative comparison with any data gathered similarly by any forum participant.
Themistokles
Posts: 45
Joined: Sun Jun 30, 2019 7:47 am

Re: GPSMAP 66s/st: Total ascent way too low (by 20% - 30%) [Bug 148]

Unread post by Themistokles »

CowboySlim wrote: Sat Sep 21, 2019 10:48 pm As a result, I performed a technically correct study gathering and analyzing both GPS and atmospheric pressure data of a statistically significant quantity. Yes, I gathered the data at the exact same position in my backyard once a day for 40 consecutive days. For these tests, the results include averages, ranges and standard deviations.
While the result of your study for sure will be interesting and while I don't doubt that you did your tests correctly in a mathematical sense, your test situation is not the situation I'm after. I usually don't stay at the same place during my tours :-)

While my GPSmap 66s doesn't have problems on routes where the height of each ascent (hill) is significant and there are only a few such hills, it does not compute the total ascent correctly when there are many small hills. You would get similar results (yet for other reasons) with a device which calculates the total ascent based on GPS elevation.

I think you should try yourself to see what I'm talking about. If you would like to do that:

Take both devices (i.e. one device which calculates elevation and total ascent based on GPS and another one which calculates elevation and total ascent based on air pressure) and go over a small hill which is 5 meters high one hundred times. After that, the first device will probably show a total ascent of 0 (or some other nonsense), while the second device will show a total ascent somewhere between 400 meters and 500 meters.

If you don't have such a hill near you, you can use every other place which provides a height difference of a few meters. If you have only big hills near you, you can just go up a few height meters, return to the starting point and repeat that process. For that test, it really only matters that you go up and down only a few height meters each time, but do so many times, each time returning to the same starting point.
CowboySlim
Posts: 5
Joined: Sun Sep 08, 2019 12:22 am

Re: GPSMAP 66s/st: Total ascent way too low (by 20% - 30%) [Bug 148]

Unread post by CowboySlim »

I appreciate the suggestion above that I perform further testing in order to corroborate the poster’s conclusions. However, it is not additional testing that is needed on my part, it is more complete description of my tests and more definitive characterization of the results and their implementation.
Sometimes, technical discussions are only of partial values as the participants are not in concurrence with the technical definitions of some of the parameters, In this subject, I refer to the term “accuracy” as the difference between an observed value and an accepted, standard value and the term “precision” as the differences between observed values and an accepted, standard value. https://en.wikipedia.org/wiki/Accuracy_and_precision

1. I performed a test to determine the precision of the barometric function of my GPSr in the absence of weather affects. Typically, in the early daylight hours of my locale, the winds, air movement are non-existent as observed by total lack of tree leaf movement. I recorded the 20 barometrically derived elevations, one every 5 seconds. For that elapsed time of 1 minute and 35 seconds, I assumed that the ambient pressure was constant and any differences in the 20 values of elevation were due to the GPS device’s barometric performance characteristics..

2. I performed another test to determine the precision (more precision = less data scatter) of both the device’s GPS derived elevations and those barometrically derived. I recorded both elevations once daily for 40 days at the exact same location in my backyard. As expected, the barometrically varied due to weather induced ambient pressure changes. https://en.wikipedia.org/wiki/Normal_distribution

3. Test results: I assumed that the precision of the 3 recorded elevation sets are best defined by their standard deviation values, the lesser being the more precise. That of the first, 1:35 second barometric test, was 1/6 of the 40 day GPS derived value and the 40 day GPS value was 1/12 of the 40 day barometric elevation value. The barometric only (no weather) precision was 6 times better than the GPS derived values which were 12 times more precise than those of the barometric unit values during weather changes. Or, the weather effects degraded the precision barometric elevations by a factor of 72.

When to use barometric: I was asked on another forum which method to use to measure a berm of approximately 50 feet which could be walked down quickly. I know of a nearby hill that professional surveyors had left a mark at the top noting 49 feet and one at the bottom of -2 feet, as measured with their Trimble device. I recorded both GPS and barometric at the top and the bottom, 2 minutes later, during quiescent weather conditions. Subtracting the top elevation from the bottom for both methods resulted in a far better, more precise result for the barometric method after comparing both heights.

When to use GPS methodology; Two possibilities, when there is ongoing weather issues or when many data points are to be recorded. Firstly, if there are weather changes underway, it is likely that all barometric evaluations will be negatively compromised. Secondly, even in quiescent weather conditions, when a number of sequential readings will be recorded for atrack, such as multiple points for total ascent or decent determinations, GPS will be preferred even in quiescent weather conditions. As a manifestation of a normal distribution of the data set, the number of negative value errors will be essentially equal to the number of positive errors. Consequently, the sum of the errors for each recorded elevation will cancel out and the total value of ascent or descent will be as good or better than that of barometrically derived. (Note that there are some GPSr models that cannot automatically record a sequence of GPS derived elevations while traveling.)
CowboySlim
Posts: 5
Joined: Sun Sep 08, 2019 12:22 am

Re: GPSMAP 66s/st: Total ascent way too low (by 20% - 30%) [Bug 148]

Unread post by CowboySlim »

Themistokles wrote: Sun Sep 22, 2019 10:05 am
I think you should try yourself to see what I'm talking about. If you would like to do that:
I appreciate the suggestion; however, as you can now see from my post above, I have completed more tests than I had reported in earlier posts. Additionally, I have recorded trips up and down thousands of feet, vertically, in the Sierra Nevada Mountains in California.

One thing I will do, is use two devices to record both GPS derived and barometrically derived elevations climbing and descending 3,000+ feet in nearby mountains. Accepting the errors of the GPS elevations to cancel out, differences between the two methods would be due to atmospheric changes during the trips. However, as the trips would be about 15 minutes in duration, I would not expect a substantial variation.
Post Reply

Return to “GPSMAP 66s/st”