5hz versus 1hz

Took the gt31 and proto 5hz for a drive yesterday, had them both in a paqua (not because I expected to flood the car, just to keep them next to each other), on the dash as close to the windscreen as I could get.




Here's a classic example of a simulated gybe.





blue line is the GT31, pink is the 5hz unit.

But there are some examples that don't look as good, the thoughts are, that the satellite has only a clear view in one direction, and changing direction can cause funny things to happen as the gps adjusts to a new set of satellites.
I had thought I should attach the pacqua to the roof racks, but was worried about something nasty happening.

looks like the pink line was a beautiful drift!

Thanks for conducting trials Decrepit, the GPS windsurfing community stands a far better chance of getting a suitable product if we can be involved in the product development.

road test results.

GT31 blue prototype red,
2sec 47.64 47.48

5x10 43.89 43.84

1hr 17.36 16.95

NM 43.68 43.65

alpha 30.03 29.87
26.79 27.05
26.22 26.04
25.65 26.02
25.59 25.12

Unfortunately looks like more accuracy doesn't give you more speed, but I'm not sure how valid these results are with most of the car obscuring the satellites.

Here's the worst track to show my concern.





It's an old google earth image, there's now a roundabout there, which I did go around! Honest.
There was no cross over of my tracks.
There are no buildings in the vicinity to do strange things with refections. But it is the tightest turn, and it did similar things with both gps's. So my conclusion is the dashboard isn't the ideal place.

And I also can assure everybody, I stayed on the road the whole time on all these pics. Google earth isn't always spot on, that can account for some positional error, but when my tracks go over both sides of the lane, then that's a gps error.

When I get the chance, I'll repeat the course with the GT31 on a roofrack and see if there's any difference.

As it is hard to say whether the difference in readings is from the different type of unit or just different units, it would be useful to have two GT 31s and the prototye.

The closeness of the 5 X 10 and the nm would suggest the differences average out over time. Not sure about the hr though...

Select to expand quote

Macroscien said..
Common new standard and popular chipset are 10 Hz, so this "new" device will not be interchangeable and compatible with all other software around.

What are you talking about? a data-point is timestamped, so "more data" is already correctly handled by all the existing tools.

New standard for GPS racing going to be 10 Hz or even 20 Hz , so suddenly we are the only group stack with 5Hz. Any specific reason for that ? Dollar or two saving on manufacturing costs ? I doubt because usually standard 10Hz chipest and parts are cheaper. From memory storage point of view doesn't really matter 5 or 10 as capacity today is big enough and cheap to keep lifetime history on device.
http://www.ebay.com.au/itm/Qstarz-BT-Q1000eX-Racing-GPS-10Hz-Logger-Aus-Distributor-/271090349961
or 20 Hz
www.race-technology.com/dl1_+_20hz_gps_8_7585.html
see 20 Hz chipset $39 available now
www.sparkfun.com/products/10919

5Hz vs 10 Hz

at 50 knots at Luderitz sailor progress is recorded every 5.14 meters at 5 HZ resolution

at 10 Hz it will be 2.57 meter

at 20 Hz only 1.28 meter

Maybe that is not a Melbourne Cup - where 5 meters resolution at the finish line means 2 horse length. If could be quite a mess to dispute who won Melbourne cup if it comes to 2 horse and everybody at this range at the finish is potential winner. For sprinter 60 or 100 meters that possibly means that all arrived at same time and share the gold medal.

In our competitions - at slalom racing - also two boards length at the finish line accuracy could lead to hot dispute - specifically if you won by two length by visual and video recording and according to GPS you lost.

Apply this to rugby and 5 meters before the line or behind makes no difference if you score a point.
Get a basketball and your basket suddenly have a 5 meter diameter or Golf and 5 meters wide hole in the ground is all you need to aim at.

The first point is that current analysis software (GPSAR_PRO, GPS-Results and RealSpeed - and I think KA72) are all able to cope with 5hz and 10hz data in NMEA, Sirf Binary and UBX raw binary formats. We have been testing various 5 and 10hz loggers for a few years now.

The second point is that yes indeed, we will have 10hz capability loggers. There are a number of projects working on that, but they are quite a long time from commercial production. The most promising are based on a UBlox GPS engine that is capable of 10hz binary output, and possibly up to 14hz.

Of the examples you linked to:

We have tested the Quartz loggers over 4 years ago. They were 5hz models. The thing that disqualifies them is that they use the MTK GPS chip which will not output raw binary format (NMEA only) so they don't provide the accuracy data we require (we call it SDOP - Speed Dilution Of Precision).

We also tested some GT-31's that had an extra 5hz MTK GPS engine in them (Dubbed GT-35 by us). It recorded the Sirf .sbp data to the internal logger and the 5hz NMEA data to the SD card. It was very nice for Alphas showing smoother curves in the gybes (and lovely tracks when I tested it skiing ). Locosys decided not to proceed with this configuration for various reasons.

The Race Technology type of GPS has been around for a while as well. There are other models available (VBox?) that can record at 50hz and 100hz.
They are very expensive and bulky and really only suitable for racing car teams with budgets in the 100's of thousands and Millions! They are usually based on Survey Grade type GPS engines which are also available in things like the Trimble surveying GPS that has been used by Macquarrie Innovation, Sailrocket and Hyroptere. These things usually use differential correction, require a lot of technical post processing of their files to get the results and are $30-50,000 a pop last time I looked.

The SkyTraq Technology, Venus638FLPx is very interesting. Thanks for bringing it to our attention.
It is clearly a consumer grade GPS and quite inexpensive. If you read the data sheet on it here:
dlnmh9ip6v2uc.cloudfront.net/datasheets/Sensors/GPS/Venus638FLPx.pdf
it says it can output in a format they call SkyTraq Binary (as well as the common NMEA). This could provide us with the accuracy data we need and enable us to record the very compact sentences which are required for high hz logging. (Compare the size of NMEA files to SBP files!)

This one is worth following up to see if ant GPS manufactures are using it in there commercial loggers. Please Google away or ask SkyTraq who uses it!

Hi, today I was out with my Ambit2 S and my own 5hz logger. It is using the NEO-6m from www.hobbyking.com, 20 USD and a Fubarino SD from www.seeedstudio.com, 32 USD and some cheap OLED display. Unfortunately the antenna supports just up to 5hz. So I should have read the specs better :).

Anyway with 5hz already it looks like a lot of data:

The track has 22305 points as compared to the Ambit track that has just 3850 points. The Ambit doesn't always save points every second for some reason, even if the settings say 1s.

The .ubx file has quite a nice precision: best was 1.42 meters and average was 2.05 meters. But it is over 6MB for a 74 minute session.

I also noticed that only GPSResults can handle those .ubx files properly. GPSActionReplay can handle them only, when I set the logger to 1hz. Ka72.com cannot handle .ubx files, but that would be a nice upgrade.

Here are the files, just in case someone with more knowledge wants to compare them:

www.chumba.ch/windsurf/05071737-Mellsten-117-8.5.UBX
www.chumba.ch/windsurf/Move_2014_05_07_17_31_50_Mellsten.gpx

Greetings,
Mete

PS: SkyTraq: I order 2 of these: https://www.indiegogo.com/projects/navspark-arduino-compatible-with-gps-gnss-receiver, it looks promising, but from the specs I couldn't find any speed accuracy data yet.

I also took the proto for a run yesterday on the water with my GT 31. Wind very light was on a 7.3 and 102 ltre board. Will get Dylan from KA72.com to analyze the files and give his opinion. great that the unit was so light and small, although needed to have the wrist band in almost tightest position so might need to have a smaller wrist band for women ?

Select to expand quote

Kazza said...
They need to make womens sized Aquapacs too, mine is always slipping down to my elbow, pain in butt!

..but your elbow is no where near your butt!

Select to expand quote

sboardcrazy said...

Kazza said...
They need to make womens sized Aquapacs too, mine is always slipping down to my elbow, pain in butt!

yep.. I bought a replacement aqua pack and it allows a bit more adjustment. The older ones I've had to swim for the board a few times and almost had the GPs fly off my arm!I wasn't game to stop to shove it up higher as I wouldn't have caught the board..25kts + in big chop..

with a little help from a sewing machine, just shorten the strap...
I found mine too annoying on the arm, and hard to check the numbers too.
I attach mine to the shoulder strap of my flotation vest. Doesn't get in the way, and easier to check the numbers.