My bet is that the session was very close to 60 minutes long, and that the GT-31 was on a little longer. The 1-hour average for the GT-31 (13.72 knots) comes out to exactly the distance traveled (25.4 km). The Canmore track is probably just a few seconds shy of 60 minutes.
My bet is that the session was very close to 60 minutes long, and that the GT-31 was on a little longer. The 1-hour average for the GT-31 (13.72 knots) comes out to exactly the distance traveled (25.4 km). The Canmore track is probably just a few seconds shy of 60 minutes.
I have my GT-31 set to record at 0knots (in lieu of a 1kn or 2kn minimum) so it even gives me an hour average if only sailed for say 40minutes and sitting still for 20minutes. If you have it set to a numerical value greater than 0 it won't record an hour in the above scenario. Maybe the settings in the Canmore split the hour if you standstill???? I doubt it'll be anything major.
Great work Sailquik.
Snides, any idea why the canmore didn't register an hour result in latest session?
If you can see more than one hour of data in the software speed graph, it is possibly a setting in the analysis program, and may be related to the filter settings rejecting part of the track.
The programs should calculate an average one hour though, even if you only sailed for 45 minutes. It will just be slower.
Are you both using the same program? If you send the file to me I can run it in the three main programs to check.
Great work Sailquik.
Snides, any idea why the canmore didn't register an hour result in latest session?
If you can see more than one hour of data in the software speed graph, it is possibly a setting in the analysis program, and may be related to the filter settings rejecting part of the track.
The programs should calculate an average one hour though, even if you only sailed for 45 minutes. It will just be slower.
Are you both using the same program? If you send the file to me I can run it in the three main programs to check.
Problem solved- just a filter setting in my case
This is just an addon to the .SBN:
When I understood it correctly, then the .SBN format of the GT-31 has 6 bytes more in Message ID 41, than the SirfStar Binary Protocoll specs say. The spec say the payload length is 91 bytes, but the payload length of Message ID 41 is 97 bytes long in the .SBN files produced by the GT-31.
They add:
uint16_t ufSog; /* Speed Over Ground in m/sec with resolution 0.01 ,unfiltered*/
uint16_t ufCog; /* Course Over Ground [0 to 360] in degrees with resolution 0.01, unfiltered */
uint8_t SDOP;
uint8_t VSDOP;
The last two bytes of the .SBP are:
uint8_t SDOP;
uint8_t VSDOP;
No guarantee though :).
Great work Sailquik.
Snides, any idea why the canmore didn't register an hour result in latest session?
If you can see more than one hour of data in the software speed graph, it is possibly a setting in the analysis program, and may be related to the filter settings rejecting part of the track.
The programs should calculate an average one hour though, even if you only sailed for 45 minutes. It will just be slower.
Are you both using the same program? If you send the file to me I can run it in the three main programs to check.
Problem solved- just a filter setting in my case
it appears i turned mine (canmore) off 87 secs short of a full hour..
i have checked yesterdays file in ka72,gps results and action replay pro and none gave an hr result.
More testing needed.
My understanding is, to get an hour you need an hour's worth of data. I don't think either real speed or gpsarpro will average out less than 1 hr's worth of data.
My understanding is, to get an hour you need an hour's worth of data. I don't think either real speed or gpsarpro will average out less than 1 hr's worth of data.
I think GPS-Results will. I know I had a discussion with Manfred about the 24 hours settings when he assumed someone would have to sail for at least 24 hours to get the results. I assured him this was not the intention and he modified the settings. I assume the same applies to the 1hour. Send me the file and I will check it.
sailquik (at) Hotmail
Received the Canmore a few days ago and yesterday I brought it to the water first time.
I think there's a lot to say about this device.
First of all: it's stupid small. Actually an advantage, I think it will even fit into some wetsuit keypocket; but the screen is so small and digits so tiny and low contrast that it's a problem to read anything if one is 45+yo. Hopefully the software might be updgraded and this issue solved.
Second: the interface is simple, after all nicer than in the GT-31. One can still setup the logger using pc connection.
Third: in the end one gets a nice set of Doppler datas, which after all is what one uses a datalogger for by 3/4 (the missing 1/4 is the ability to check run results when on the water - small screen).
This is the main issues I think. In my view it would be easy to develope it into a watertight wristwatch-type unit with a larger screen. I am still very fond of the Garmin Foretrx format. Having larger bolder digits on it I think is only a matter of updating the software.
My understanding is, to get an hour you need an hour's worth of data. I don't think either real speed or gpsarpro will average out less than 1 hr's worth of data.
I think GPS-Results will. I know I had a discussion with Manfred about the 24 hours settings when he assumed someone would have to sail for at least 24 hours to get the results. I assured him this was not the intention and he modified the settings. I assume the same applies to the 1hour. Send me the file and I will check it.
sailquik (at) Hotmail
You may need to talk to Manfred again about 1 hour. I made a little test file that's 60 min 15 seconds long. I got 1 hour results from GPS Action Replay, but not from GPSResults v. 5.36.
GPS Results reports the length of this session as only 54 minutes. Looking at the track points, I noticed that the GP102 often apparently did not log data points when I was in the water (data were from a freestyle session, so I was in the water a lot). Some breaks were about 20-60 seconds long; the total number of points was 3249.
I went back to a session where I used both the GT-31 and the GP102. The GT-31 also sometimes dropped points when it had no reception, but breaks were fewer and shorter. Both units sometimes showed non-zero doppler speeds when the number of satellites was 0 (GT-31) or the time was > 1 sec (GP102).
The total session time reported by GPSResults indicates that the program either ignores the time since the previous data points, or assumes it is always 1 second. GPSResults will not report a 1 hour average if the time it so calculates is less than one hour, even if the total elapsed time is more than one hour. GPSAR will report a 1-hour result if the elapsed time is at least one hour, but not below.
I don't see a problem with the GP102 not recording data when it has no reception. I could plug the data into spreadsheets to see if these dropped points are the reason by GP102 distances are typically slightly lower than GT31 distances, but it's warm and sunny here, and the wind is picking up .
My understanding is, to get an hour you need an hour's worth of data. I don't think either real speed or gpsarpro will average out less than 1 hr's worth of data.
I think GPS-Results will. I know I had a discussion with Manfred about the 24 hours settings when he assumed someone would have to sail for at least 24 hours to get the results. I assured him this was not the intention and he modified the settings. I assume the same applies to the 1hour. Send me the file and I will check it.
sailquik (at) Hotmail
You may need to talk to Manfred again about 1 hour. I made a little test file that's 60 min 15 seconds long. I got 1 hour results from GPS Action Replay, but not from GPSResults v. 5.36.
GPS Results reports the length of this session as only 54 minutes. Looking at the track points, I noticed that the GP102 often apparently did not log data points when I was in the water (data were from a freestyle session, so I was in the water a lot). Some breaks were about 20-60 seconds long; the total number of points was 3249.
I went back to a session where I used both the GT-31 and the GP102. The GT-31 also sometimes dropped points when it had no reception, but breaks were fewer and shorter. Both units sometimes showed non-zero doppler speeds when the number of satellites was 0 (GT-31) or the time was > 1 sec (GP102).
The total session time reported by GPSResults indicates that the program either ignores the time since the previous data points, or assumes it is always 1 second. GPSResults will not report a 1 hour average if the time it so calculates is less than one hour, even if the total elapsed time is more than one hour. GPSAR will report a 1-hour result if the elapsed time is at least one hour, but not below.
I don't see a problem with the GP102 not recording data when it has no reception. I could plug the data into spreadsheets to see if these dropped points are the reason by GP102 distances are typically slightly lower than GT31 distances, but it's warm and sunny here, and the wind is picking up .
Yep. You're right. GPS-R needs at least 1 hour of data to calculate a 1 Hour average. I guess you should leave your GPS on for a bit longer sitting on the beach to make the extra time up to an Hour. I don't think this is ideal. For 24 hour distance we decided it was fine to span a few missed points. I think this should happen in 1 hour as well but will need to look into it a bit more.
For the GT-31, we recommend that users set to record every second, even when the GPS does not have a fix. We need this to give a better result for one hour and distance so we are not cutting time out. It is also useful to have more data when trying to analyse errors in speed tracks.
This is all good to be reminded of so if we have the chance to get firmware changes for the Canmore, we can know what to ask for.
For Snides track I got a results for 57 mins, but not for 58 mins in GPS Results. Three mins short on 1 hour data!
n also be problems with the GPS not having a good 'view' of the sky.
+1 on wearing on helmet. I do this and when I reviewed my Canmore data and it cannot see any dropouts over my sessions and I spend plenty of time in the water!
I think that's the big advantage of the Canmore, It's going to be much easier to helmet mount because of it's size.
Taking the helmet of to read the display could be a small pain.
But until the Canmore gets firmware changes, I'd be tempted to use a Gt on the arm for reading on the water, and the Canmore on the head for accurate data.
Thank you Sailquick for the many useful advices!
Didn't know wet neoprene harms GPS signal. Right now I am using the GP-102+ in a comfortable soft non-waterproof armband pocket that I bought bundled with my GT-31 and never used. The small size of the GP-102+ lets me put it into another sealed plastic bag and put both into the armband. The armband pocket provides protection from crashes/pressure peaks, and the plastic bag (sandwich bag) provides protecton against leakage. Hopefully.
I will check the paqua, at first time (a few posts above) I thought it wa a misspelling for Aquapac.
All the best with the firmware improvement! That would make a super device.
Seen the Paqua lineup, the GP-102+ is so tiny that probably would fit even in their PAQUAkey bag! Actually the GP-102+ is smaller than both my Volvo and Transporter keys!
+1 on wearing on helmet. I do this and when I reviewed my Canmore data and it cannot see any dropouts over my sessions and I spend plenty of time in the water!
Another advantage with a hat mounted GPS, the readings seem to be higher.
I always have one on the arm (in a Pacqua) for reading while sailing and my back up is in my helmet lining in a ziplok bag.
The hat mount is normally 0.15kts faster and I'll take that.
I always have one on the arm (in a Paqua) for reading while sailing and my back up is in my helmet lining in a ziplok bag.
The hat mount is normally 0.15kts faster and I'll take that.
Same here. One on the arm, two it the helmet.
However, I have not seen a consistent pattern of the helmet mounted GPS giving higher speeds. Definitely cleaner tracks though!
Thanks snides, looks very similar doesn't it.
my final comparison
can more gt31
2- 31.47 31.31
10- 28.3 28.14
hr- 9.45 9.34
a- 18.76 18.46
nm-21.03 20.74
d- 17.5 17.29
To put these numbers a bit into perspective, here's a comparison of the Flysight GPS to the GT-31 that I did in November 2013:
GT-31:
2s 29.6 +-1.6
10s 27.3 +- 1.8
5x10 25.35
Flysight:
2s 31.15 +- 0.29
10s 29.15 +- 0.14
5x10 27.13
Note the big differences, and that the accuracy on the Flysight (5 Hz UBX-chip based) was much higher, according to GPSResults. The accuracy estimates for individual data points were a bit closer (about +- 1.1 knots for the GT-31, +- 0.4 knots for the Flysight), so part of the accuracy gain seems to be from the 5-fold higher data rate. But the number of satellites tracked also contributed: the GT-31 was tracking about 8 satellites, the Flysight 12.
The wind in this session was extremely up and down, and I got just one very fast run. I had the Flysight announcing speed while sailing, and the time when I hit 31 knots was indeed the fastest segment, when I caught a big gust while going deep downwind on a super-flat stretch of water. In almost all other sessions where I used the Flysight and the GT-31, the numbers were closer.
Bottom line is that the top speeds between GT-31 and GP102 are always well within the estimated accuracy. We don't get the accuracy data for the GP102, but it seems safe to assume that the estimated error is at least as good as for the GT-31; most likely, GP102 error estimates would be lower, since the GP102 tracks more satellites than the GT-31.
As nice as it would be to have the accuracy estimates in the data file, I'd put display changes higher on my priority list for firmware changes. With contacts in, I have a really hard time reading the tiny numbers for top speed.
my final comparison
can more gt31
2- 31.47 31.31
10- 28.3 28.14
hr- 9.45 9.34
a- 18.76 18.46
nm-21.03 20.74
d- 17.5 17.29
While the short-distance stuff is easy to vary, the 1hr value is a bit concerning - bigger distance/time will usually results in negating minor differences between hardware. In this case the variation is > 1% difference... which is significantly higher (by about 1-order) than the GT-31's error rate.
To measure the error of the canmore, sit it outside for 24hrs (obviously put the GT-31 with it), then analyse the data. The device with the least accumulated distance, is likely to have the best accuracy.
I agree with boardsurfr. The differences are within the error margins we sometimes see from two GT-31's, although head mounted, side by side GT-31's are often extremely close.
The differences between GPS are commonly greater with Alphas as this is such a difficult thing for the GPS to deal with, so I would not put quite so much weight on Alpha differences.
Another source of difference could be rounding due to the conversion to .GPX format. It will be interesting to see if the differences are the same if you run the same file in the latest version GPS-Results, within which Manfred tells me he he has addeed reading of the .FIT file format.
One of the first firmware changes we must try to get is the direct writing of the files in the .SBN or .SBP file format. This is because Manfred tells me the .FIT format does not have provision for satellite position or SDOP messages.
I am hearing everyone on the small screen font. That is right on top of the list of requests!
One of the issues is that memory (and battery capacity) is limited and according to Canmore, this is not easy to change. Additional data will eat up more memory, as will recording at 5hz, and this will also use more power. We will see what can be done though.
"Another source of difference could be rounding due to the conversion to .GPX format. It will be interesting to see if the differences are the same if you run the same file in the latest version GPS-Results, within which Manfred tells me he he has added reading of the .FIT file format..."
The latest PC version of GPS Results does read the .fit files directly from the Canmore unit and this is very convenient as you no longer need to use the Canmore software. The Mac version is still a work in progress so watch for new version....
1) Can someone email me some .fit files (and any other formats you are interested in.) I'm cracking open the GPS library on ka72 and adding support for more devices, but I need some sample data. (dylan@ka72.com) (If you've emailed me files in the past, feel free to send them again, as it will take me ages to locate them.)
2) Is there serious interest in me modifying ka72 to give a 1hr result even if you haven't sailed an hour? I can do it, just wonder what the demand is?
Cheers,
Dylan.
>>>
2) Is there serious interest in me modifying ka72 to give a 1hr result even if you haven't sailed an hour? I can do it, just wonder what the demand is?
Cheers,
Dylan.
It does seem like a good idea.
Not calculating an hour because there isn't an hour's worth of data, is a bit strange.
I guess the idea could be, that you need to sail an hour to get a result, but all you have to do is leave the gps on for an hour, and you'll get the result.
So the only person penalised is the one who turned their unit off too early.
1) Can someone email me some .fit files (and any other formats you are interested in.) I'm cracking open the GPS library on ka72 and adding support for more devices, but I need some sample data. (dylan@ka72.com) (If you've emailed me files in the past, feel free to send them again, as it will take me ages to locate them.)
2) Is there serious interest in me modifying ka72 to give a 1hr result even if you haven't sailed an hour? I can do it, just wonder what the demand is?
Cheers,
Dylan.
Email with .FIT file format sent.
1) Can someone email me some .fit files (and any other formats you are interested in.) I'm cracking open the GPS library on ka72 and adding support for more devices, but I need some sample data. (dylan@ka72.com) (If you've emailed me files in the past, feel free to send them again, as it will take me ages to locate them.)
2) Is there serious interest in me modifying ka72 to give a 1hr result even if you haven't sailed an hour? I can do it, just wonder what the demand is?
Cheers,
Dylan.
Hey Dylan, I have emailed you one of my .fit files. Please let me know if there are any issues etc. :-)