Ka72 postings

My offer to Dylan to have a look at his oao importer/filter stands.

Select to expand quote

tbwonder said..
Well if there is no plan to get to the bottom of this quickly then I suggest that a note is put on GPSTC to suggest to people not to post using Motions with KA72. There are plenty of alternatives.

OK done, good idea

Select to expand quote

Xbraun54 said..

JulienLe said..
"Julien and everybody else says, the motion files work fine with all other software."

Ignore XBraun54 / Jan Hendrik de Bruin from GP3S up here. He's moody for other reasons.

Thank you for confirming my post "The easiest way to solve a problem is to deny it exist..."

This is Bull**** Xbroun54.

All the other analysis software programs, including GPS-Results which GPS-SS uses, get the correct results. It is only KA72 which is getting it wrong.

It is very clear that the problem exists within KA72 - Period. You clearly have an ulterior motive for this misinformation attack. It's beneath you.

Select to expand quote

sailquik said..

Xbraun54 said..

JulienLe said..
"Julien and everybody else says, the motion files work fine with all other software."

Ignore XBraun54 / Jan Hendrik de Bruin from GP3S up here. He's moody for other reasons.

Thank you for confirming my post "The easiest way to solve a problem is to deny it exist..."

This is Bull**** Xbroun54.

All the other analysis software programs, including GPS-Results which GPS-SS uses, get the correct results. It is only KA72 which is getting it wrong.

It is very clear that the problem exists within KA72 - Period. You clearly have an ulterior motive for this misinformation attack. It's beneath you.

No need to get offensive @sailquick.... we have some interesting examples of motion data with errors, but based on the very open and constructive feedback regarding questions posted here (and also earlier) not so keen to share them :) But hey, you got to believe what you believe i guess... don't get me wrong, the Motion is a very good device (if you can get one....) , but it's always possible to have room for improvements.... and yes also with GPS-Results we have seen issues, so it's not fair to put all the blame on KA72 it think....

Select to expand quote

Xbraun54 said..
A bit confused to read this, the fact that this seems to be a known error since september and Julien is already working on a solution for this (that's good ...i think ??) is a bit dissapointing.

You are barking up the wrong tree. There's no indication that the problem is due to the Motion. He has also looked into this issue, but he cannot fix it because the problem in NOT in the Motion firmware. Besides GPSResults and GPS Speedreader, GPS Action Replay Pro also gives the correct results with this file. But the absolute indication that this is a ka72 problem is that ka72 reports a (correct) 100 m speed that is higher than the 2 second speeds. That is not possible in windsurfing (it would require a 1 second speed above 50 m/s, or ~ 100 knots). That proves that ka72.com can read the file just fine, but can not calculate the top speed correctly for this file.

It is quite common for software bugs to show up with only certain data. When that happens, a software developer should thank the person reporting the problem, and use the data to fix the issue. In this case, the software developer is Dylan, not Julien.

Select to expand quote

Xbraun54 said..
We have seen previous "glitches" with Motions before and the only strategy around them seems to be to deny issues like hell , no transparency and after a software inprovement (why on earth????) present it like a non issue.....don"t get me wrong, there is no device on earth that's 100% but just be open about it .....

Every software will have bugs. What matters is how a software developer reacts to bug reports. The only thing I have seen when there were problems with the Motion firmware was that the bugs were fixed very quickly.

It is quite obvious that you hold grudges against Julien. Perhaps you have good reasons, but that is no excuse to attack him on a matter that is clearly not Julien's problem. I may not agree with Julien on some things, but in this case, he's gone above and beyond just by looking into this issue, even though all indications are that this is a ka72 problem. But he can't fix problems in the code that runs ka72.com.

I, like many I am sure, use KA72 with their Motion files as it is the easiest method as all is handled simply on the phone. As I am not 'playing for sheep stations' with my results, I will probably just keep uploading to GPSTC through KA72, unless of course, I see the possibility of a PB or an out of the expected result. Then I would usually forward that file to a team member for scrutiny. In view of the above discussion, does this sound reasonable?

Select to expand quote

tbwonder said..
Mikey,
Personally I would stop using KA72 until this issue is fully understood. Although the reported cases so far have shown to be lower speeds in KA72, who is to say that some speeds may be over reported.
Best to play safe.

I find it hard to see how the peak 2sec speeds can be wrong, it is the easiest calculation of all.

Ok Andrew, thanks. Looks like I better fire up the laptop, download Speed Reader, and learn how to use it. my brain hurts already.

Select to expand quote

mikey100 said..

tbwonder said..
Mikey,
Personally I would stop using KA72 until this issue is fully understood. Although the reported cases so far have shown to be lower speeds in KA72, who is to say that some speeds may be over reported.
Best to play safe.

I find it hard to see how the peak 2sec speeds can be wrong, it is the easiest calculation of all.

Ok Andrew, thanks. Looks like I better fire up the laptop, download Speed Reader, and learn how to use it. my brain hurts already.

Shock horror. I did it. And posted to GPSTC.
WELL DONE ME!

Here's a session from today where ka72.com shows wrong results: www.ka72.com/Track/t/483326
The giveaway is a 100 m speed that's higher than the 2 second peak. A 100 m run at 32 knots takes about 6 seconds, and therefore must contain a 2 second segment that is at least as fast (and, in reality, always faster).

This time, the 2 second error is less than 1 knot, but the ka72 understates 5x10 by more than 5 knots:
GPS Action Replay gives the same 5x10. I have not checked with GPSResults, but I bet it agrees with GPSAR and Speedreader.

Select to expand quote

boardsurfr said..
Here's a session from today where ka72.com shows wrong results: www.ka72.com/Track/t/483326
The giveaway is a 100 m speed that's higher than the 2 second peak. A 100 m run at 32 knots takes about 6 seconds, and therefore must contain a 2 second segment that is at least as fast (and, in reality, always faster).

This time, the 2 second error is less than 1 knot, but the ka72 understates 5x10 by more than 5 knots:
GPS Action Replay gives the same 5x10. I have not checked with GPSResults, but I bet it agrees with GPSAR and Speedreader.

It looks to me like your mini-motion track has only 10 points whereas mine has 20.

Select to expand quote

tbwonder said..
Could it be that KA72 is looking for 20 data points and averaging them rather than only using 10, when logging is set to 5hz?
Although if this was the case the 2sec speeds would never be correct.

ka72 supports all kinds of devices and rates, including watches that record at rates below 1 Hz. In the past, it had no problems with my 5 hz Motion data, and still handles at least some Motion data without problems (I checked this track from today: www.ka72.com/Track/t/483374, and ka72 gets the same results as Speedreader).

Select to expand quote

tbwonder said..

remery said..

It looks to me like your mini-motion track has only 10 points whereas mine has 20.

He has it set to log five times a second rather than 10 times a second
So a 2 sec peak is 10 points rather than 20.

Which does raise an interesting point. The errors I am seeing are when using a mini motion at 5 Hz. Is this the issue?
Are other people seeing the error at 5hz logging or 10hz logging?

Could it be that KA72 is looking for 20 data points and averaging them rather than only using 10, when logging is set to 5hz?
Although if this was the case the 2sec speeds would never be correct.

Sounds like you guys are making some progress on this issue,
With the Motion can the user set 5 or 10hz or is this set within the device firmware ?
I know with GW60 user selects to set to 5hz

10Hz (with Galileo) is available again in firmwares >= #3038. It's a redo of firmware #3007's core which was more accurate than recent 10Hz (without Galileo) firmwares. Please don't shout on top of roofs about it yet, I'd like its adoption to be progressive. The plan is to ask everyone to upgrade mid-January, along with the other upgrades, if no issues are revealed. So far so good.

Select to expand quote

tbwonder said..

remery said..

It looks to me like your mini-motion track has only 10 points whereas mine has 20.

He has it set to log five times a second rather than 10 times a second
So a 2 sec peak is 10 points rather than 20.

Which does raise an interesting point. The errors I am seeing are when using a mini motion at 5 Hz. Is this the issue?
Are other people seeing the error at 5hz logging or 10hz logging?

Could it be that KA72 is looking for 20 data points and averaging them rather than only using 10, when logging is set to 5hz?
Although if this was the case the 2sec speeds would never be correct.

I suppose if KA72 throws a couple of points out for some reason, the impact on the 10 points in the 2sec will be greater than the 31 points in the 100m, which might leave the latter higher. I don't have GPSResults at work but looking at the KML it seems the 100m overlaps about half of the 2sec run.