OK, seems like the bugs are slowly getting sorted.

Select to expand quote

elmo said..
Any news as to when the comercial versions will be released? as my GT11 has just about had the Richard

Or

is it a better option going the "android" mobile phone option?

No news yet about release dates yet.
The Android version looks great, but won't have the error checking of the GW. I've also got reservations about reading a back lit display in bright sunshine in a pouch.
The Canmore's are on special at the moment, I think Yoyo may have a comment about them.

Select to expand quote

sailquik said..

decrepit said..

sailquik said..
>>>>

What were the SDOP values and range?

Strange Andrew, thought I had the latest version of real speed, but mine stops at usog and ucog, no sdop or vspod.

I will send you my version by email, but I thought they were the same in that respect. Mine is slight different in another area.

OK I've got v1.929 installed.
And cleaned up the file a bit, removed the tracks where I was setting the unit up, changing from 5hz to 1hz, and then turning off and came up with these figures.

5hz for 76min. Total distance 45.68m. Total dopler distance 15.03m Total doppler distance 3d 15.15m typical sdop of .058, number of satelites 9.
1hz for 117min. Total distance 52.63m. Total dop dist 13.64m. Total dop dist 3d 14.53, typical sdop of .019. sky must have moved a bit because satelites have dripped to 8.

so that's about 12m/hr at 5hz for doppler distance and about 7m/hr for 1hz doppler distance.

Select to expand quote

decrepit said..

sailquik said..

decrepit said..

sailquik said..
>>>>

What were the SDOP values and range?

Strange Andrew, thought I had the latest version of real speed, but mine stops at usog and ucog, no sdop or vspod.

I will send you my version by email, but I thought they were the same in that respect. Mine is slight different in another area.

OK I've got v1.929 installed.
And cleaned up the file a bit, removed the tracks where I was setting the unit up, changing from 5hz to 1hz, and then turning off and came up with these figures.

5hz for 76min. Total distance 45.68m. Total dopler distance 15.03m Total doppler distance 3d 15.15m typical sdop of .058, number of satelites 9.
1hz for 117min. Total distance 52.63m. Total dop dist 13.64m. Total dop dist 3d 14.53, typical sdop of .019. sky must have moved a bit because satelites have dripped to 8.

so that's about 12m/hr at 5hz for doppler distance and about 7m/hr for 1hz doppler distance.

I had a brief email exchange with Manfred about this. It appears that 5hz files will naturally have more 'noise' in them. We see this in the 5hz speed graph as a more jagged appearance. It also shows up in each data point with slightly higher SDOP error. But Manfred pointed out to me that the error is averaged over the second so it actually cancels out and becomes much smaller. So it appears that the distance recorded is exaggerated by this 'noise'. It also seems to mean that you can't really compare these sort of results from different hz.

Select to expand quote

sailquik said..
I had a brief email exchange with Manfred about this. It appears that 5hz files will naturally have more 'noise' in them. We see this in the 5hz speed graph as a more jagged appearance. It also shows up in each data point with slightly higher SDOP error. But Manfred pointed out to me that the error is averaged over the second so it actually cancels out and becomes much smaller. So it appears that the distance recorded is exaggerated by this 'noise'. It also seems to mean that you can't really compare these sort of results from different hz.

For error that is completely random, measuring at a 5-fold higher rate would increase the total error 5-fold. So the accumulated doppler distance would be expected to go from 7 m/h to 35 m/h. So it's actually good that the distance only went up to 12 m/h. There are different possible reasons for this, from systematic error sources to data filters in the GPS firmware to how exactly the distance is calculated (simple sum vs. spline, etc.).

The "more jagged appearance" of the speed graph is due to simply seeing 5-fold more points, each with it's own error. On the other hand, the track graphs tend to look smoother with 5-Hz data. May well be that most of the error is in the direction of travel (as opposed to perpendicular to it). If so, that would indeed cancel out nicely when looking at longer intervals, even just one second or two.

For speedsurfing, all the observed numbers when stationary are perfectly fine. 40 knots is about 20 meters per second, or several thousand times faster than the "drift". Of course, the actual error when traveling at speed will also increase.

Has anyone been able to run the GW-52 utility on Windows 8 I downloaded it but the utility is incompatible. I asked Locosys they said ( The GW52 USB driver/Utility has been tested on Win8 successfully.

Select to expand quote

boardsurfr said..
(Clipped) >>>>>>>
For speedsurfing, all the observed numbers when stationary are perfectly fine. 40 knots is about 20 meters per second, or several thousand times faster than the "drift". Of course, the actual error when traveling at speed will also increase.

That last sentence is an interesting point. Tom Chalko has insisted for quite some time that actual error decreases at speeds above 30 knots but I could never quite follow his explanations.

Have you seen any theories to support this?

Tried out the latest firmware today. The distance no longer dissapears when you adjust the speed genie setting, and there's a selection of 1s, 1.5s, 2s and 2.5s for the speed genie scroll rate. I've set mine to 2s and find it much easier to read in the pouch.

RealSpeed only gave me a 16kt alpha, but gpsarpro was a 20 something, the GT31 and logit also gave a 20, whether processed in RealSpeed or GPSarPRO.

Select to expand quote

sailquik said..

boardsurfr said..
(Clipped) >>>>>>>
For speedsurfing, all the observed numbers when stationary are perfectly fine. 40 knots is about 20 meters per second, or several thousand times faster than the "drift". Of course, the actual error when traveling at speed will also increase.

That last sentence is an interesting point. Tom Chalko has insisted for quite some time that actual error decreases at speeds above 30 knots but I could never quite follow his explanations.

Have you seen any theories to support this?

This is based on what I see. The doppler speeds when stationary, which are the actual error, are typically below 0.1 knots on the GT-31. The +- column in GPSResults typically shows values of 0.5 knots.

It is also a common expectation whenever measuring anything. Some error sources will be proportional to the signal, so a larger signal (in our case speed) will have a larger absolute error. However, other error sources will be signal-independent, so the relative error will typically go down.

Regardless of whether or not the absolute speed error increases with speed when windsurfing, all signs indicate that the relative speed error decreases. A 0.1 knot error when almost-stationary at 1 knot is a 10% error; a 0.5 knot error at 25 knots is a 2% error. A GT-31 typically error of 0.3 knots at speeds above 30 knots is less than 1% error.

Gents, this forum discussion and the others of similar vein is fascinating, but I'm finding it hard to keep up.

The question that needs to be answered for IT challenged sailors like me is - Now that the GT31 is almost impossible to purchase, what should we purchase, when as a result of malfunction, we have to replace it with a device that is compliant on GPS Team Challenge?

Will we have an answer to this question before spring?

Select to expand quote

powersloshin said..
So the GW52 produces .sbp files and not .sbn like the GT-31?

Yes, but the GT31 and GT11 also produce .sbp files from their internal memories, it's only their SD cards that stores the .sbn files.
So the GW52 without an SD card does the same as the GTs.

I've finished all the things I want to check until there's another firmware version. So if anybody wants to try it out, see me at the beach and you can borrow it for a session.
I'll email the resulting file back to you.

Select to expand quote

Glynn Merrit said..
Has anyone been able to run the GW-52 utility on Windows 8 I downloaded it but the utility is incompatible. I asked Locosys they said ( The GW52 USB driver/Utility has been tested on Win8 successfully.

Did you select the correct COM port in the utility?

Manfred has solved the problem I was having with gybes in GPSResults.
I was using "Smart logger" set at 10knts, so any gybes that went under that changed form 5hz to 1hz.
This upsets the filters and no alpha or NM including a gybe, result is produced unless filters are deselected.
So I think I'll stick to the 5hz setting unless going for distance.

I'll do some experiments and see if this is also the realspeed problem with alphas.

Select to expand quote

decrepit said..
Manfred has solved the problem I was having with gybes in GPSResults.
I was using "Smart logger" set at 10knts, so any gybes that went under that changed form 5hz to 1hz.
This upsets the filters and no alpha or NM including a gybe, result is produced unless filters are deselected.
So I think I'll stick to the 5hz setting unless going for distance.

I'll do some experiments and see if this is also the realspeed problem with alphas.

If the software fails, then it is probably assuming a sequential-time-series of points... ie: not taking into account the timestamp of the point....

It probably should take it into account... aka: I wonder how many alpha tracks there are, where there was a two-second gap in the data...?

Select to expand quote

John340 said..
Gents, this forum discussion and the others of similar vein is fascinating, but I'm finding it hard to keep up.

The question that needs to be answered for IT challenged sailors like me is - Now that the GT31 is almost impossible to purchase, what should we purchase, when as a result of malfunction, we have to replace it with a device that is compliant on GPS Team Challenge?

Will we have an answer to this question before spring?

To answer the first question John, you should purchase one ASAP so you have two GPS's to wear at all times!

You will also have a back up when the first one fails. That way you will not have downtime waiting to get a new one.

The second question has just been answered above.

I purchased one yesterday. Just have to wait till September now!
Hope its worth the wait.

Select to expand quote

sailquik said..

John340 said..
Gents, this forum discussion and the others of similar vein is fascinating, but I'm finding it hard to keep up.

The question that needs to be answered for IT challenged sailors like me is - Now that the GT31 is almost impossible to purchase, what should we purchase, when as a result of malfunction, we have to replace it with a device that is compliant on GPS Team Challenge?

Will we have an answer to this question before spring?

To answer the first question John, you should purchase one ASAP so you have two GPS's to wear at all times!

You will also have a back up when the first one fails. That way you will not have downtime waiting to get a new one.

The second question has just been answered above.

Thanks Andrew, with 2 GPS I'm bound to go twice as fast

I tested the GW-52 against the GT-31 and Sony/GPS Logit app today. See the GPS-Logit thread.

Here are some comparisons of Alphas between the GW-52, GT-31 and Sony/GPS Logit App.

The images are is just the one quickest Alpha I did in yesterdays session. I have not directly compared the the track maps of the others but did compare them in terms of speeds.

These images suggested that the GW-52 is a lot better for Alphas and that the Sony/GPS-Logit is a lot worse. So I did some investigation below.

GW-52:



GT-31:








Sont Experia E1/GPS-Logit App:





SDOP = Speed Dilution of Precision. i.e. The error range values for the calculation as reported by the GPS system within which we can be 99.9% certain that you did the reported speed.
ie. If you subtract the SDOP from your reported speed we can be 99.9% certain you did at least that calculated speed.

The 10 GT-31 Alphas had an average SDOP of +/- 0.289 with a high of 0.308 and low of 0.188

Taking off the high and low give an average of +/- 0.215

GT-31 Alpha 500/50m:
time m [knots]
16:05:10 480.0 25.219+/-0.190
16:06:56 489.8 24.411+/-0.244
16:03:42 488.0 24.321+/-0.195
16:08:43 479.2 23.885+/-0.188
16:33:46 493.4 23.391+/-0.247
16:32:45 499.7 23.128+/-0.308
16:31:50 495.4 22.929+/-0.261
15:53:13 477.5 22.638+/-0.200
16:07:53 489.2 21.612+/-0.190
16:06:02 488.9 21.598+/-0.194

The GW-52 Alphas had an average SDOP of +/- 0.227 with a high of 0.291 and a low of 0.195

Taking off the high and low give an average of +/- 0.223

GW-52 Alpha 500/50m:
time m [knots]
16:05:09 495.9 25.368+/-0.234
16:06:55 496.6 24.626+/-0.232
16:03:42 497.8 24.436+/-0.206
16:08:42 480.7 23.960+/-0.253
16:33:47 499.4 23.448+/-0.207
16:32:44 499.2 23.216+/-0.198
16:31:50 498.2 22.946+/-0.195
15:53:12 497.8 22.931+/-0.206
16:07:52 499.3 21.761+/-0.291
16:06:01 498.4 21.724+/-0.251

The Sony/GPS Logit does not record the SDOP values in the files as the GPS chips in the phone do not output that data.

Below I have added two columns showing the worst case error added (speed plus SDOP) for the same runs in the GT-31.(This is the best comparison as both are at 1hz)

SonyExperia E1/GPS Logit App:
Alpha 500/50m:
time m [knots] GT-31
16:05:09 474.2 25.607 (25.409)
16:06:56 491.7 24.510 (24.655)
16:03:42 489.0 24.373 (24.516)
16:08:42 472.8 24.183 (24.073)
16:33:47 489.5 23.208 (23,638)
16:32:44 488.2 23.148 (23.346)
16:31:50 498.2 23.060 (23.190)
15:53:12 477.7 22.651 (22.838)
16:06:01 493.0 21.780 (21.792)
16:07:52 481.0 21.744 (21.802)


So there are two Alphas that are higher than they should be according to the error values from the GT-31 (bold).

They are not a long way out but they still could make a difference in a list of rankings in a competition.

You will also notice that the last two runs from the GPS-Logit are reversed in ranking from the other two GPS (bold).

This highlights the importance of error margin value and how it can affect competition results and rankings.

In this case, the margin is very small and in fact if you add the max error in the GX-52 results you get the same reversal.

It highlights that accuracy and error margin matters, and in rankings where the difference between rankings is lees than around 0.25 knots, you must assume they are random rankings and give the riders a draw.

For 'fun' rankings it is probably of little consequence, but for records and the results of top level competition it is to be taken very seriously.

This is also exactly why the GPSTC and the GPS-SS insist the top ranking riders use the GT-31 (or the GW-52) for their postings as the results from those GPS can be verified.

The SDOP analysis is just one way of examining the Alpha results and comparing them. Looking at the map view and speed graph I am still inclined to deduce that the Sony/GPS-Logit Alpha are a bit more suspect, so more study will be needed.

Another factor to consider is that all the different Android phones may well have different GPS chips and implementations in them so to get a very clear picture, they would all need to be individually tested. Some may well turn out to give much better results than others.

Recommendation:

The GPS Logit App is a wonderful advance in almost every way and still has more potential. I highly recommend it for those who like to have lots of feedback and instant results. Depending on the phone it can also be quite inexpensive. (remember the GPS Logit App will not be free - nor should it be).

But for those serious about their speeds, competition and rankings in the GPSTC, I recommend using your GT-31 or GW-52 for your
posting data, preferably work on or in your helmet, and using the GPS-Logit app on your arm for a competitive tactical advantage (the value of the speed talking and instant and running averages feedback is huge!)

Hope that helps clear up some of the confusion.

Thanks Daffy, now I know how to use the sdop values!
Maybe for records the + & - values need to be posted as well.

And I've also noticed the reversal of alpha results between the units.
I also endorse a GW52 on the head, I find it easier to wear on the head than the GT31 as it's smaller and lighter.
If the conditions look like a PB is possible I'll also wear the GT31 on the shoulder of my camel pack.

Select to expand quote

decrepit said..
Thanks Daffy, now I know how to use the sdop values!
Maybe for records the + & - values need to be posted as well.

And I've also noticed the reversal of alpha results between the units.
I also endorse a GW52 on the head, I find it easier to wear on the head than the GT31 as it's smaller and lighter.
If the conditions look like a PB is possible I'll also wear the GT31 on the shoulder of my camel pack.

When we announce WGPSSRC records on the GPS-SS forum we give the recorded speed and the SDOP correction for the Claimed Speed (Record).

Select to expand quote

Glynn Merrit said..
Has anyone been able to run the GW-52 utility on Windows 8 I downloaded it but the utility is incompatible. I asked Locosys they said ( The GW52 USB driver/Utility has been tested on Win8 successfully.

I've been running that version on Windows 8.1 without too much drama.

Prior to updating the firmware of the GPS I had some trouble with corrupt data ont he GPS crashing my computer when the utility tried to access it. That was in both Windows 7 and 8.1

Here is some good news for Mac users.

Up until a few days ago I had been unable to get a connection with the GW-52 with Windows (XP) running in VMware in my Macbook Pro.

By good fortune, a clever IT support person was staying in our ski lodge last weekend. He heard my complaint and offered to take a look. Within a minute to so of flicking in and out of various venues I didn't know even existed, and after installing a driver update, he had the Locosys utility talking to the GW-52 perfectly.

I am now one very happy chappy.

So it seems that the use of emulation software to run windows and talk to the GW-52 is possible if set up correctly and should not be a blocker to Mac users benefitting from using a GW-52.

Seems KA72 won't read my sbp file anymore using my GW-52 tried asking Dylan but got no reply anyone had this problem?
Use to work but now when I click summit it just comes up saying ( Full Results from your Upload 155800021_GW52USER_20150919_145447.sbp File Visibility:) but I can't get any results for posting. Yes am signed in done everything as usual.

Glynn, It sounds like more of a KA72 website issue to me. Try here for contacts: www.seabreeze.com.au/forums/Windsurfing/Gps/New-KA72-Website/

Mark at Surf Sail Australia reports he now has them in stock, so if you've pre-ordered, expect it soon.

Thingsee1 is available for ?300 "shipping in 6-8 weeks"...
http://shop.gps-speedsurfing.com/p/67/thingsee-one-gp3s-edition-v1

If you are thinking of getting a Thingsee One, buy it in US$, it works out a little bit cheaper because of exchange rates. (US$299)

The GPS-SS guys are negotiating with Thingsee for a lower cost version. Don't know the time scale if they come to some agreement.