Results 15 to 18 of 18
Thread: Navi GT-31
9th January 2014, 08:09 PM #15
Today, I had a session in winds from 7 to 18 knots. According to the watch readout, my maximum speed was 21.7 knots. I took the data into GPSAR. It gives me:
5 best 1 second (at least) average = 40.81km/h [22.04Knots]
1 second run n°1 = 41.89km/h [22.62Knots] (20.9 m. in 1.800 s.) -
1 second run n°2 = 40.76km/h [22.01Knots] (11.3 m. in 1.000 s.) -
1 second run n°3 = 40.60km/h [21.92Knots] (11.3 m. in 1.000 s.) -
1 second run n°4 = 40.47km/h [21.85Knots] (11.2 m. in 1.000 s.) -
1 second run n°5 = 40.34km/h [21.78Knots] (11.2 m. in 1.000 s.) -
5 best 2 second (at least) average = 40.14km/h [21.67Knots]
2 second run n°1 = 40.68km/h [21.97Knots] (22.6 m. in 2.000 s.) -
2 second run n°2 = 40.09km/h [21.65Knots] (31.2 m. in 2.800 s.) -
2 second run n°3 = 40.05km/h [21.62Knots] (22.2 m. in 2.000 s.) -
2 second run n°4 = 39.98km/h [21.58Knots] (22.2 m. in 2.000 s.) -
2 second run n°5 = 39.89km/h [21.54Knots] (22.2 m. in 2.000 s.) -
5 best 10 second (at least) average = 38.42km/h [20.75Knots]
10 second run n°1 = 39.60km/h [21.38Knots] (110.0 m. in 10.000 s.) -
10 second run n°2 = 38.34km/h [20.70Knots] (106.5 m. in 10.000 s.) -
10 second run n°3 = 38.22km/h [20.64Knots] (106.2 m. in 10.000 s.) -
10 second run n°4 = 38.09km/h [20.57Knots] (115.4 m. in 10.910 s.) -
10 second run n°5 = 37.86km/h [20.44Knots] (105.2 m. in 10.000 s.) -
5 best 100 meter (at least) average = 38.44km/h [20.75Knots]
100 meter run n°1 = 39.60km/h [21.38Knots] (110.0 m. in 10.000 s.) -
100 meter run n°2 = 38.34km/h [20.70Knots] (106.5 m. in 10.000 s.) -
100 meter run n°3 = 38.22km/h [20.64Knots] (106.2 m. in 10.000 s.) -
100 meter run n°4 = 38.15km/h [20.60Knots] (105.0 m. in 9.910 s.) -
100 meter run n°5 = 37.86km/h [20.44Knots] (105.2 m. in 10.000 s.) -
5 best 500 meter (at least) average = 35.46km/h [19.15Knots]
500 meter run n°1 = 37.63km/h [20.32Knots] (501.8 m. in 48.000 s.) -
500 meter run n°2 = 35.88km/h [19.38Knots] (508.3 m. in 51.000 s.) -
500 meter run n°3 = 35.39km/h [19.11Knots] (502.3 m. in 51.100 s.) -
500 meter run n°4 = 34.68km/h [18.72Knots] (500.9 m. in 52.000 s.) -
500 meter run n°5 = 33.73km/h [18.21Knots] (506.8 m. in 54.090 s.) -
Looking at the track data in more detail, I noticed that the watch increases the gap between data points, the slower you go. As I carry the board to the bank, it's recording in 5 second increments. Then, as I stand waiting for a gust, it waits 10.910 seconds before recording another position. When I get onto the board and start moving, the sample goes to 2 seconds and then I go from 14km/h to 18km/h and it switches to 1 every second.
When I fall and lie around trying to waterstart, the interval goes right up to 15 seconds. I know that they use the accelerometer to smooth the spikes in the speed and it seems that they also use it to save storage space (and battery?) by not recording data unnecessarily. It seems it switches to 1 second increments when it predicts that the next point will exceed 15km/h. By way of an example of a big gap in the data, this is my first tack:
Speed goes (in km/h) 25.09, 22.84, 18.61, 12.32 (1 second gaps), 6.16 (2 second gap), 2.95 (5.910 second gap), 9.47 (2 second gap), 9.86 (2 second gaps until I hit 14.48 and it goes back to 1 second),18km/h.
The track contains 1338 points in the 1,986 second session, so 11 minutes are missing. Movescount tells me that I spent 4'54" going less than 8km/h and 6'26 going less than 15.9km/h. That's 11 minutes which could explain the missing data points (the difference probably explained by the sample going up to 15.9km/h and the watch switching to 1 second increments sometime before that as well as the fact that some data points are recorded even at slow speeds).
If you are going to have a speed competition, it makes no sense to allow different hardware and software to produce the data you submit. Justice must not only be done, it must be seen to be done and having different units would always cause trouble even if each one was 100% accurate. I certainly would never argue that you should allow more than one unit onto the rankings. That's why the GT-31 is useful. I do not, however, buy the argument that the Ambit2 is significantly less accurate than the GT-31. I think the GT-31 was a good choice at the time and you have to stick with that choice but it is not the bee's knees technically. In my view, anyone who is not interested in participating in the rankings would be far better served using the phone they already have or buying a GPS watch that is less bulky, more useful, sturdier and easier to use than the GT-31. These devices will give you very accurate data that you can certainly use to improve your technique and have fun with and that any reasonable person would trust to be accurate.
Last edited by Capie; 9th January 2014 at 08:15 PM.My Boards: 2016 Fanatic Falcon TE 129, 2014 Patrik Slalom 115 vII, 2014 Patrik Slalom 92l, 1992 Windsurfer One Design, 2012 Fanatic Freewave 85l
My Sails: North Sails Warp f2016 , North Hero, North Volt
9th January 2014, 09:41 PM #16
- Join Date
- Feb 2013
- Cape Cod, USA
You cannot prove that a device is accurate by posting data from one session. Even if you would compare your data to data from a GT-31 and would get the same results in one session, that would only show that the device was accurate in that one session. The open question is how the accuracy would be over many sessions. Even old Garmins and "bad" GPS units can record very accurately during some (and perhaps most) sessions.
Your observation about the speed-dependency of the recording rate is interesting. Recording at lower intervals at low speeds is actually one way of making the data appear more accurate. Note that the speed does not seem to be the entire explanation, though - your top 1 second speed shows a recording interval of 1.8 seconds, and other lines in your table also show irregular recording intervals (2.8 s, 10.91 s).
If you want to look into the speed differences you saw, compare the speed results in GPSAR with the "Doppler" box checked and unchecked. These tend to be different even if the unit does not record Doppler speeds, for example due to smoothing (which may use doppler or accelerometer data, at least in some modes). GPSAR assumes that the speed data in input files are doppler data, and computes non-doppler speeds from the coordinates. You'd have to ask Suuonto where the differences come from - good luck with that!
If you'd want to have Suuonto watches accepted for the GPS web sites, here's what you'd have to do:
1. Get Suuonto to modify the firmware to (a) record in fixed 1 second intervals, and (b) output doppler speeds and precision estimates (which the Sirf4 chip in the watches provides) in a binary file format.
2. Do a direct comparison of the Suuonto to a GT-31 over multiple sessions, providing the actual data files. Obviously, the comparison would need to show that the Suuonto tracks are as accurate as the GT-31 tracks.
3. Convince the decision makers at the site you're interested in that your analysis is adequate.
Good luck with that - I think you biggest hurdle would be step 1. Whether or not you'd have a chance in step 3 would depend on how well you do step 2. That would certainly have to be a lot more involved than what you've done so far.
10th January 2014, 07:33 AM #17
I think there's a misunderstanding between us on this one. I am not interested in being ranked on a website and agree entirely that if this is what someone wants to do with their GPS, then they should get whatever device is standard for the site they choose (probably the GT-31). Ranking sites will be stuck with this device for some time because the choice of hardware for them is as much (if not primarily) a decision to standardise as it is to find the most accurate device.
What I responded to was the suggestion that other devices are intrinsically handicapped by the technology choices made (not to record doppler) and are as such useless. I also responded to your specific statements about how Suunto watches work. I am not sure what Suunto watch you looked at, but clearly you did not look at the Suunto Ambit that I have and possibly the one you have was broken. You said Suunto watches were not capable of recording position data every second and then that it could not sustain this for more than 20 minutes. I checked and both of those statements are incorrect. Either you looked at a dud watch or this was a long time ago. The fact is that there isn't a test I can see on the Internet where someone tests a GT-31 against a Suunto Ambit or an iPhone 5 or a Garmin Quatix and yet the speed sailors dismiss these as inaccurate.
Yes, I agree that the longer intervals are not always explained by speed. The vast majority are. There are, however, a couple of blips where a single interval is 1.9 seconds.There are also a few occasions where the watch uses a smaller interval (0.9 seconds). As you say though, without a GT-31 in exactly the same circumstances, we can't say that it would have recorded any more intervals. And I also don't really buy the argument that this results in a more accurate reading. Maybe if we were turning when the measurement was taken but if you're dividing a bigger distance by 1.9 seconds rather than 1 second, it should be just as accurate if not more accurate.
Ultimately, if someone asks the question what is the best GPS for windsurfing, the answer in my view is not the GT-31. That is the only choice you have if you have a desire to compete via a website. If you want to race, I would imagine the Garmin Quatix is the best GPS for windsurfing. If you want something very accurate and are happy with the bulk, sounds like your Flysight is a good bet. The Suunto is a good option for people who also SUP, bike, swim, surf, run, hike and do other things. It's also the one that Wind Magazine in France use for all of their testing, by the way. And if you are prepared to put it at risk, the other option is to use your smart phone.
Last edited by Capie; 10th January 2014 at 09:31 AM.My Boards: 2016 Fanatic Falcon TE 129, 2014 Patrik Slalom 115 vII, 2014 Patrik Slalom 92l, 1992 Windsurfer One Design, 2012 Fanatic Freewave 85l
My Sails: North Sails Warp f2016 , North Hero, North Volt
20th July 2015, 11:50 AM #18
- Join Date
- Jul 2015
- Kyiv, Ukraine
Im Ukrainian windsurer, and Im developd Android application for windsurf track analisyz. It support common GPS files - GPX, Garmin's FIT and TCX , GT-31 SBN files
You can find this app there - https://play.google.com/store/apps/d...d=ua.khirh.wta