... sometimes, anyway.
One week ago today, I rode over and met the club run at Croome Park. I recorded the ride on my Garmin. When I switch on the unit and look at 'History', the details of the ride are there. However, when I upload data from the unit to Ride with GPS, or to Garmin Connect, that ride doesn't appear. Rides before and after, yes, but not that one.
So I've copied the TCX file out of the unit's History folder to my computer and tried opening it on BikeHike. I get an error message, something about "> required at line 1", I think.
I realise that I am fallible: I do daft, absent-minded things at least as often as the average person. But what can I have done? All I do is switch the blessed thing on and off and press Start and Stop. What can even a dozy bugger like me do to cause one isolated ride to not be recorded properly? I know from experience that even if you forget to switch off and/or the battery runs flat mid-ride that a readable record is still saved in the unit's history.
After nearly 2 years using my Garmin, it still puzzles me
Moderators: Philip Whiteman, Andy Terry
George, it is possibly that the /tcx file has been corrupted, try this link for some info;
http://www.photoscene.com/kimandsteve/4 ... -tcx-files
If it is corrupt then it might be posssible to fix it using First Object XML Editor.
http://www.photoscene.com/kimandsteve/4 ... -tcx-files
If it is corrupt then it might be posssible to fix it using First Object XML Editor.
That is a neat tool, Si: makes the files much easier to follow.
I opened the file with FOXE, which seemed to have no trouble reading it, but reported something along the lines of "parsing error: unfinished tag at position xyz". Unfortunately, that message disappeared from the status bar at the bottom before I had time to consider its implications or note the precise location of the error.
Could the parsing error be the cause of the file's invisibility to Garmin Connect etc? If so, can I fix it without first learning all about XML so that I know what the tag should be?
And what might the cause be? Operator error? Unit malfunction? I notice that the article you linked says "This is all because the new firmware for the Garmin Edge 705 is not writing it’s [sic] data files correctly in some circumstances." Is this linked to my aberration? (My unit is a 605.) And, if the firmware is malfunctioning in a way that enough people are affected and writing blogs about it, how come Garmin haven't fixed it?
Throughout the life of my Garmin, it has occasionally done weird things that I don't understand. Sometimes, maybe the cause is operator error (e.g. the time when recording stopped mid-ride; I had no recollection of going near the Stop button, which isn't easy to press accidentally with your hands in a normal riding position, but maybe somehow I had). Other times, it is hard to see how that can be the case (e.g. when it has frozen midway through the start-up procedure, or when the map display has gone waywire midway through a ride). However, almost all these incidents have been one-offs, which doesn't seem consistent with a specific fault in the unit or its firmware, which one would expect to result in a consistent and persistent pattern of errors. Most of the time, the thing functions just fine.
I opened the file with FOXE, which seemed to have no trouble reading it, but reported something along the lines of "parsing error: unfinished tag at position xyz". Unfortunately, that message disappeared from the status bar at the bottom before I had time to consider its implications or note the precise location of the error.
Could the parsing error be the cause of the file's invisibility to Garmin Connect etc? If so, can I fix it without first learning all about XML so that I know what the tag should be?
And what might the cause be? Operator error? Unit malfunction? I notice that the article you linked says "This is all because the new firmware for the Garmin Edge 705 is not writing it’s [sic] data files correctly in some circumstances." Is this linked to my aberration? (My unit is a 605.) And, if the firmware is malfunctioning in a way that enough people are affected and writing blogs about it, how come Garmin haven't fixed it?
Throughout the life of my Garmin, it has occasionally done weird things that I don't understand. Sometimes, maybe the cause is operator error (e.g. the time when recording stopped mid-ride; I had no recollection of going near the Stop button, which isn't easy to press accidentally with your hands in a normal riding position, but maybe somehow I had). Other times, it is hard to see how that can be the case (e.g. when it has frozen midway through the start-up procedure, or when the map display has gone waywire midway through a ride). However, almost all these incidents have been one-offs, which doesn't seem consistent with a specific fault in the unit or its firmware, which one would expect to result in a consistent and persistent pattern of errors. Most of the time, the thing functions just fine.
Away from the context of this forum, Si has been doing some very useful detective work on my behalf. This is probably of very little interest to most, but just in case it may intrigue anyone, here's what he found:
The History file for the ride that wouldn't upload included a corrupted trackpoint:
</Trackpoint><usms>uu>z<vm><lati>53>7<qty>->nw{t}|oLoowmuw~muow{}}w><o>132.??xi|}o~nemeters><d>51>>559<t><vis>6017/57?34w8=?ty{vq~o]m}e~~osition><LatitudeDegrees>52.044011</LatitudeDegrees><LongitudeDegrees>-2.413222</LongitudeDegrees></Position><AltitudeMeters>133.314</AltitudeMeters><DistanceMeters>22130.727</DistanceMeters></Trackpoint>
Simon's discovery made me feel rather sheepish about not having noticed it myself, because it is pretty obviously not as it should be. I must confess that I had simply opened the file, taken one look at the long list of latitudes and longitudes and decided that trying to spot one that was wrong would be like looking for a needle in a haystack.
The spot to which the trackpoint relates is a crossroads in the lanes, where I made a sharp left turn ( http://goo.gl/maps/rLso ). I've been asking myself whether anything associated with that spot might have triggered the error:
Human intervention?
I wondered whether accidentally pressing a button or deliberately doing something such as changing the display on the unit at the precise moment that it is writing a trackpoint could cause an error. I can't recall for certain whether I pressed any buttons at that point. However, I think it's unlikely. Accidental pressing would be difficult when braking or pulling away, because your hands are well away from the unit. I do sometimes fiddle with the display while riding, but rarely, and I don't think I'd do it there, where I've got three difficult junctions in very quick succession. Besides, a functional unit ought to be able to cope with that, surely?
Mechanical interference?
I suppose that a violent shock or other stress at the precise moment that a trackpoint is being written could theoretically cause corruption. However, the units must be built to withstand a lot of bumping and g-forces, and nothing exceptional happened at that point.
EM interference?
There are no overhead power lines at that point, but there is a radio mast quite nearby. I have an open mind on that as a possible cause. However, I quite often ride close to the masts at Wychbold without a problem.
Location is a red herring?
I don't know enough about exactly how the unit works to know whether we should assume that corruption occurred at the time/place to which the trackpoint relates. I've noticed that the unit has a file called something like Current; does it initially record everything in that and then write a History file when you press Stop? If there is any post-processing of data, the corruption might have occurred much later.
Although I'm barely any closer to knowing what (if anything) I might do to prevent a recurrence, I am at least reassured that I am not guilty of some sort of crass techno-illiteracy; there was some sort of data recording/processing error beyond my control.
The History file for the ride that wouldn't upload included a corrupted trackpoint:
</Trackpoint><usms>uu>z<vm><lati>53>7<qty>->nw{t}|oLoowmuw~muow{}}w><o>132.??xi|}o~nemeters><d>51>>559<t><vis>6017/57?34w8=?ty{vq~o]m}e~~osition><LatitudeDegrees>52.044011</LatitudeDegrees><LongitudeDegrees>-2.413222</LongitudeDegrees></Position><AltitudeMeters>133.314</AltitudeMeters><DistanceMeters>22130.727</DistanceMeters></Trackpoint>
Simon's discovery made me feel rather sheepish about not having noticed it myself, because it is pretty obviously not as it should be. I must confess that I had simply opened the file, taken one look at the long list of latitudes and longitudes and decided that trying to spot one that was wrong would be like looking for a needle in a haystack.
The spot to which the trackpoint relates is a crossroads in the lanes, where I made a sharp left turn ( http://goo.gl/maps/rLso ). I've been asking myself whether anything associated with that spot might have triggered the error:
Human intervention?
I wondered whether accidentally pressing a button or deliberately doing something such as changing the display on the unit at the precise moment that it is writing a trackpoint could cause an error. I can't recall for certain whether I pressed any buttons at that point. However, I think it's unlikely. Accidental pressing would be difficult when braking or pulling away, because your hands are well away from the unit. I do sometimes fiddle with the display while riding, but rarely, and I don't think I'd do it there, where I've got three difficult junctions in very quick succession. Besides, a functional unit ought to be able to cope with that, surely?
Mechanical interference?
I suppose that a violent shock or other stress at the precise moment that a trackpoint is being written could theoretically cause corruption. However, the units must be built to withstand a lot of bumping and g-forces, and nothing exceptional happened at that point.
EM interference?
There are no overhead power lines at that point, but there is a radio mast quite nearby. I have an open mind on that as a possible cause. However, I quite often ride close to the masts at Wychbold without a problem.
Location is a red herring?
I don't know enough about exactly how the unit works to know whether we should assume that corruption occurred at the time/place to which the trackpoint relates. I've noticed that the unit has a file called something like Current; does it initially record everything in that and then write a History file when you press Stop? If there is any post-processing of data, the corruption might have occurred much later.
Although I'm barely any closer to knowing what (if anything) I might do to prevent a recurrence, I am at least reassured that I am not guilty of some sort of crass techno-illiteracy; there was some sort of data recording/processing error beyond my control.
-
Andy Terry
- Posts: 904
- Joined: 20 Nov 2006 14:27
- Real Name:
- Location: Bromsgrove, Worcs
Or, as we IT professionals say, s**t happens.George wrote: I am at least reassured that I am not guilty of some sort of crass techno-illiteracy; there was some sort of data recording/processing error beyond my control.
Just as a matter of interest - on your Garmin can you get tracks as GPX instead of TCX?
-
Philip Whiteman
- Posts: 2046
- Joined: 19 Nov 2006 16:17
- Real Name:
- Location: Drayton, Worcestershire
The History files are certainly TCX by default (I've never chosen to save them as TCX). I don't think you can change that in the settings, but I haven't actually looked. I will try and make time to do so.Andy Terry wrote:Just as a matter of interest - on your Garmin can you get tracks as GPX instead of TCX?
Yes, I have one of those too, Phil. It is at least as annoying as the Garmin and doesn't tell me how late I'm going to be at the café.Philip Whiteman wrote:I have terrible problems with my £19.95 hardware. Sometimes the map does not fit easily between the plastic and I have terrible difficulties when charging the battery.
Says he feeling very smug.