Forums > Windsurfing   Gps and Speed talk

KA72 data analysis

Reply
Created by Te Hau > 9 months ago, 27 Aug 2009
Te Hau
493 posts
27 Aug 2009 6:19AM
Thumbs Up

Hi Dylan,
am I right in thinking your system is reading only positional data, ie no doppler and no trapezoidal integration?

Dylan72
QLD, 660 posts
27 Aug 2009 10:38AM
Thumbs Up

If you upload a .sbn file or .sbp file from a GT-31 then the data is doppler. If you upload a file from a GT-11, then the data is doppler except for the Alpha-500 result, which is trackpoints.

.gpx files are treated as trackpoints.

Dylan.

decrepit
WA, 12767 posts
27 Aug 2009 7:54PM
Thumbs Up

Te Hau said...

Hi Dylan,
>>>>>>> no trapezoidal integration?


By this do you mean, using 3 measurements for the 2sec result, like realspeed does????

Te Hau
493 posts
28 Aug 2009 7:14AM
Thumbs Up

It's the integration method which GPSAR is using now, although on my older version it is an option, not sure if its still optional on updates........has the effect of dropping your speeds ..depressing, that accuracy !

decrepit
WA, 12767 posts
28 Aug 2009 11:35AM
Thumbs Up

Te Hau said...

It's the integration method which GPSAR is using now, although on my older version it is an option, not sure if its still optional on updates........has the effect of dropping your speeds ..depressing, that accuracy !


Yep, GPSAR trapezoidal, gives the same results as real speed, before it was slightly greater.

There was a discussion earlier, about how Mal calculated the 2 sec, as the result was always lower than manually averaging the best 2 adjacent results.
Then it was explained that the doppler 1 sec is not the average of 1 sec, it's an instantaneous reading, to get a 1 sec average you need 2 readings, to get 2sec average you need 3 readings, that's why it's less.
I'm sure Dylan has said that his results match realspeed's.

mathew
QLD, 2133 posts
28 Aug 2009 4:19PM
Thumbs Up

Do you have a link specifying that the doppler is 1sec instantaneous? The Sirf spec sheet suggests that the doppler is the average of 10 x 0.1 sec samples.

ka72
QLD, 581 posts
28 Aug 2009 8:37PM
Thumbs Up

(Dylan posting via ka72"s acct)

Yes the KA72 site uses trapezoidal interpolation for speed calcs. 2sec is the most affected. 2sec peak is the average of three seconds data. This is deliberate, and a bit conservative. You can probably get better results from other software by tweaking the settings. KA72 tries to be fair by applying the same rules to everyone.

Dylan

decrepit
WA, 12767 posts
28 Aug 2009 8:40PM
Thumbs Up

mathew said...

Do you have a link specifying that the doppler is 1sec instantaneous? The Sirf spec sheet suggests that the doppler is the average of 10 x 0.1 sec samples.


No Mathew, I'm just going by what was posted in a previous thread.
Could there be a difference between the GT11 and GT31, in this, cause I think they have different sirf chips?

If the GTs use the average of 10 X .1sec samples for the 1sec doppler figure then I fail to see why there should be 3 readings involved in a 2sec average???????

ka72
QLD, 581 posts
30 Aug 2009 9:47AM
Thumbs Up

(Dylan72 again...)
I've had a bit of a dig this morning, but I can't find anything that would lead me to believe that the GT-11 or GT-31 would use 0.1sec samples. Could you provide a reference?

Unfortunately, the most likely sources seem to be missing or very expensive to access (eg NMEA0183)



Subscribe
Reply

Forums > Windsurfing   Gps and Speed talk


"KA72 data analysis" started by Te Hau