Version 2.70 Bugs

[TracknQ]

I'm wondering whether I've discovered some bugs in Version 2.70.

 

TOTAL ASCENT BUG?

 

Yesterday, when I was walking downhill, I noticed that the total ascent reading on my 60CSx's screen was moving upward.

 

Today, I ran two tests with my 60CSx sitting stationary. Before each test, I recalibrated the altimeter to a known certain elevation, 76 feet; and I reset all data fields to zero. The first test was done with the GPS receiver and track log enabled. The second test was done with the GPS receiver and track log disabled.

 

Without the 60CSx moving horizontally or vertically, the elevation readings fluctuated upward and downward constantly, causing the total ascent and maximum elevation readings to climb over time.

 

I took a number of screen shots of elevation plots, barometer plots, and ambient pressure plots. I saw some abnormal patterns. For example, the barometer plots fluctuated rapidly. I made Acrobat Reader files to show all screen shots. Beneath each screen shot, I noted the type and time of the subject plot.

 

Some of the plots showed large feet-per-minute up or down changes in the lower left-hand corner. I highlighted in yellow the notations beneath the screen shots that showed these fluctuations.

 

First Total Ascent Test (GPS Receiver and Track Log ON)

The first elevation plot of the first test looked like this:

 

 

(Notice the rapid upward feet-per-minute change in the lower left-hand corner.)

 

The last elevation plot of the first test looked like this:

 

 

(The last screen shot of the first test was an Ambient Pressure Plot. That plot showed a total ascent of 123 feet. So, my 60CSx ascended 123 feet while sitting still.)

 

The track log profile at the end of the first test looked like this:

 

 

Notice the total ascent discepancy:

 

Profile max. elev. = 148 feet

Profile total ascent = 72 feet (148 minus 76)

Plot max. elev. = 151 feet

Plot total ascent = 123 feet

 

(All this vertical variation occurred with the 60CSx sitting still.)

 

Here's the link to the .pdf files that shows all screen shots taken in the first test:

051006_1115-1256_GPS-Track_On.pdf

 

Second Total Ascent Test (GPS Receiver and Track Log OFF)

The first elevation plot of the second test looked like this:

 

 

Within the span of a minute, two elevation plots showed these downward and upward fluctuations:

 

 

 

The last elevation plot of the second test looked like this:

 

 

(The 60CSx ascended 357 feet without moving!)

 

Here's the link to the .pdf file that shows all screen shots taken in the second test:

051006_1259-1642_GPS-Track_Off.pdf

 

ANOTHER VERSION 2.70 BUG?

 

I live in Honolulu, Hawaii. With 2.70 loaded in my 60CSx, I took several 6.5-mile walks. Although my 60CSx had strong signals from multiple SVs, low EPEs, and constant correction from WAAS #47, the elevation readings constantly crept upward over time. Despite recalibrations of the altimeter to known, certain altitudes and fair weather, my 60CSx showed elevation readings that were much higher than control benchmarks and MapSource topo lines.

 

QUESTION ABOUT VERSION 2.70?

Have any of you experienced total ascent readings that are higher than your actual total ascents, or creeping increases in your elevation readings, or anything else that seems like a Version 2.70 bug?

 

TracknQ

Edited May 11, 2006 by TracknQ

[peter]

Notice the total ascent discepancy:

 

Profile max. elev. = 148 feet

Profile total ascent = 72 feet (148 minus 76)

Plot max. elev. = 151 feet

Plot total ascent = 123 feet

I'm not sure what discrepancy you're concerned about. The 'total ascent' will generally be higher than the simple difference between the initial elevaton and the highest since you also have several ups and downs along the way and you're determining the sum of all the rises. I.e. if you start at 100', climb to 300', drop to 200', climb to 250', drop to 50', and climb back to 100', then your total ascent would be 200+50+50 = 300'.

Some fluctuation in measured altitude is normal so it's not surprising that over time you get an ever increasing value for your total ascent.

 

Frankly I'd be much more concerned about the fact that the profile chart shows the horizontal movement as about 0.6 miles although you said the unit was stationary. That's a discrepancy of thousands of feet, not just a hundred or two that you're seeing in false vertical movement.

ANOTHER VERSION 2.70 BUG?

[/u]

I live in Honolulu, Hawaii. With 2.70 loaded in my 60CSx, I took several 6.5-mile walks. Although my 60CSx had strong signals from multiple SVs, low EPEs, and constant correction from WAAS #47, the elevation readings constantly crept upward over time. Despite recalibrations of the altimeter to known, certain altitudes and fair weather, my 60CSx showed elevation readings that were much higher than control benchmarks and MapSource topo lines.

Was automatic recalibration turned on or off during these walks? Even though the weather remained fair there could still have been significant changes in normalized barometric pressure.

[+dumketu]

That is the problem with using a barometer for elevation; not that the GPS altitude doesn't vary a lot too. I have set my gps to plot pressure, and the pressure plot varies quite a bit, just as the altitude does. If the wind blows, pressure changes; if the sun goes behind a cloud, pressure changes; and if I squeeze the gps with my hand, pressure changes. At the end of the day I would just prefer to be able to turn off the altimeter altogether and rely on the gps altitude for all its flaws.

 

Glad they fixed the other things though. Two steps forward, one step back.

 

Brian

[TracknQ]

My 60CSx has continued to sit, motionless. The GPS receiver has remained disabled. I therefore assume that only the barometer has been controlling altitude readings.

 

A few minutes ago, I ended my second test with this screen shot:

 

 

Looking back in time, I started the second test at 12:59 p.m., when I recalibrated the altimeter to 76 feet and reset all data fields to zero.

 

Despite remaining totally stationary, my 60CSx purportedly ascended 614 feet.

 

As I watched the elevation plot, I saw repeated, rapid, one-second-interval fluctuations of elevation. For example, the elevation readings rapidly moved one or two or three feet upwards, upwards, downwards, upwards, downwards, downwards, etc, in a seeming random fashion. For example, 87, 84, 86, 87, 86, 89, 87, 85, 84, and so on.

 

Sometimes, when the elevation readings moved from 85 to 86 feet or from 84 to 85 feet, the total ascent reading moved up a foot.

 

I highlighted a few large upward or downward feet-per-minute ("FPM") fluctuations in the .pdf file. I didn't see any clear connection between the size of the upward or downward FPM changes and the number of feet that elevation readings would move correspondingly up or down.

 

I agree with the suggestions of others that actual changes in barometric pressure will cause a barometric altimeter's readings to change. However, the barometer plots that I captured showed rapid, short-duration ups and downs of barometric pressure. These ups and downs didn't seem to match the more gradual curves of the ambient pressure plots. (All day, the weather in Honolulu was fair and balmy.)

 

I'm no expert. That's for sure. However, based on the evidence that I've seen, I think that Version 2.70 has a bug that causes elevation readings to jump up and down too frequently and too rapidly. I believe that the resulting fluctuations cause total ascent to increase abnormally and, possibly, other altimeter malfunctions.

 

I'm hoping to hear whether or not other 60CSx users have witnessed altimeter anomalies, like the one I saw during my walk yesterday, viz. total ascent climbing while I was going downhill.

 

Thanks to all for your comments. We can work on this issue together.

 

TracknQ

[60csxuser]

Hi,

I looked at your .pdf file and specifically the ambient pressure plots. There was a 1.4 mb pressure drop in around three hours. (A windy day?) Assuming the readings were not influenced by local errors (for example changes in the air streams from the airconditioning), science says this will cause a ~36 ft increase in your measured altitude.

[TracknQ]

Thanks, 60csxuser.

 

Is there a formula I can use to convert pressure changes to altitude changes?

 

Since total ascent readings have been completely unreliable, I stopped using the total ascent data field on my plots page. I still use the max. elev. data field. In place of total ascent, I alternately use the barometer and ambient pressure data fields. (I'm uncertain whether barometer or ambient pressure readings are more useful.)

 

Aloha,

 

TracknQ

[60csxuser]

Hi,

 

The standard formula is that the ambient pressure decreases 1 millibar for each 8 metres (27 feet) of altitude increase.

 

This also means that if the ambient pressure varies due to the weather patterns, the altitude reading will change. See this link for which pressure changes are typical for different types of weather.

 

What this means, as far as I understand so far, is that a stationary GPS receiver can work quite well as a barometer, which is of great interest for weather watchers. But according to my experiments so far (see other thread) if GPS auto-calibration is enabled, the GPS-altitude deviations will typically cause more altitude errors (especially in the total ascent sum, because of the up/down jumping) than the ambient pressure variation.

Either way, obtaining a really accurate altitude reading is difficult.

With GPS auto-calibration disabled, you need to find a sufficiently accurate reference barometer. Fortunately the nearest airport always has one. Or then you take a trip to the seashore and calibrate the device there.

With GPS auto-calibration on, a very long averaging period is probably needed to smooth out the momentary errors in GPS altitude. Unfortunately the 60CSx does not offer an in-built averaging function. In addition, even if you would get the "real" GPS altitude, there is the question of the accuracy of the map datum parameters. The map datum parameters define the Earth as an ellipsoid with a certain shape, which is only a more or less accurate approximation of the actual shape.