OSTC3, cR, sport and 2 :  heinrichs weikamp forum The fastest message board... ever.
OSTC's running hwOS 
Compass unstable
Posted by: karip ()
Date: April 28, 2019 06:30PM

Hi.
I update firmware V3.01 and hereafter compass is unstable? Only OSTC2 new version. Anyone same problem? Yes, I have calibrate many times.

Attachments: Compass.mov (381.2 KB)  
Options: ReplyQuote
Re: Compass unstable
Posted by: Titus ()
Date: May 19, 2019 12:39PM

Hello karip

I received my new OSTC2 this week and performed many calibrations (also running on v3.01).

I have to admit that the compass is not showing the same direction as a "standard" compass. I have between 20 and 40 degrees difference ... which is huge when you dive in bad visibility conditions (as we have here in my country).

Any idea from HW ?

Find attached a picture comparing a standard compass (showing direction between 200 and 210 deg (SSW)) and OSTC (showing 245 deg (SW)).



Thanks

Titus



Edited 1 time(s). Last edit at 05/19/2019 12:52PM by Titus.

Attachments: OSTC_Compass.jpeg (163.4 KB)  
Options: ReplyQuote
Re: Compass unstable
Posted by: Ralph ()
Date: May 19, 2019 06:12PM

Hi Titus,

the issue with your test setup is: the OSTC thinks the north pole is somewhere in direction of the good old suunto compass... In earnest: the small magnet in the Suunto's needle produces a magnetic field that is strong enough to cause magnetic deviation in its vincity. So as you can't have two traditional compass next to each other, you can't have a traditional one and a computer one on the same arm. Take a north bearing with the Suunto, mark it somehow on your table and then move the Suunto away before you probe the OSTC on the marked bearing.

Still (answer to the first post), with some hardware versions of the OSTC the compass sensor gain setting needs to be fine-tuned to make the readout more stable.

BR
Ralph

PS: an OSTC causes much less influence on a traditional compass as with the reverse way...

Options: ReplyQuote
Re: Compass unstable
Posted by: Titus ()
Date: May 19, 2019 07:18PM

Hello Ralph,

Thanks for the advise.

I exactly did what you said. I put the old SUUNTO compass on a place where there is no iron around at least 1m.

I draw on a paper the North and set the direction based on the N indicated by the SUUNTO. I then remove the SUUNTO and put the OSTC on the paper (taking care to not move the paper). And the result is not positive. The difference is always there (about 40° as said in my post).

Thanks

Titus

Attachments: OSTC Compass North.jpg (920.7 KB)  
Options: ReplyQuote
Re: Compass unstable
Posted by: Ralph ()
Date: May 19, 2019 08:15PM

OK, step 1 successfully completed! :-)

For the calibration, there are numerous posts already out on how to tumble the OSTC during the 1 minute calibration phase. Are you aware of the procedure? In short, the unit must be turned more or less evenly often around all 3(!) axes. Can you take a note of the calibration factors that are shown for a short time after the calibration?

Options: ReplyQuote
Re: Compass unstable
Posted by: Titus ()
Date: May 19, 2019 08:38PM

Hello Ralph


I did a calibration again. the result is even worse.

Now I have a difference between 40 and 50 degrees.

the calibration settings I noted are :

x = - 206
Y = - 993
Z = 1103

what does that mean ?

Thanks

Titus

Options: ReplyQuote
Re: Compass unstable
Posted by: karip ()
Date: May 19, 2019 09:30PM

Sorry. Problem with attached file.

Maybe now?

[www.youtube.com]

Options: ReplyQuote
Re: Compass unstable
Posted by: Ralph ()
Date: June 06, 2019 07:22AM

Hello Titus,

sorry for the delay in answering. The Cal-X/Y/Z values shown are the coordinates of the pointer that represents the OSTCs own-induced magetic field. They are generated during the calibration and then later on subtracted from the measured field to derive the direction of the true magetic field.

If any of these values would be zero, that would be a strong hint on a hardware issue. With your OSTC it's not, so i suspect some gain setting mismatch. Over the course of time, the compass chips have changed a bit every now and then, all variants need to be handled a little bit differently in software. As this is low level programming, i'll task that to Matthias... :-) If not done already, can you pls. send the serial number of your OSTC so we can nail down the exact hardware configuration?

Thanks,
Ralph

Options: ReplyQuote
Re: Compass unstable
Posted by: heinrichsweikamp ()
Date: June 06, 2019 12:42PM

Besides the LCD screen, the compass chips are those being discontinued from the manufacturers all the time.
Which is why the hwOS code now supports three different compass chip sets. After the low-level readout of the raw values all OSTC use the same piece of code to compute the heading. With the raw data displayed in the calibration screen you can check different issues: Defective axis (no movement of the value(s)) or very little variations of the values. Technically, you'll need at least 180 counts of difference in the X and Y axis and ~90 on the z-axis for a 2° resolution in the heading.

I'll look into this (again) but I'm pretty sure the low-level routines are working fine for any compass chip variant. There is still room for improvement (extra filtering) tough.

Matthias

Options: ReplyQuote
Re: Compass unstable
Posted by: Titus ()
Date: June 10, 2019 03:18PM

Hello Ralph,

The serial number is 15592

hope this help (because beside the compass not showing the right direction, the computer is also not giving the right depth !!!)

Titus

Titus

Options: ReplyQuote
Re: Compass unstable
Posted by: Ralph ()
Date: June 10, 2019 04:51PM

If that's true, pls. email HW also directly, 'cause that'll probably be a hardware issue then.

Options: ReplyQuote
Re: Compass unstable
Posted by: Titus ()
Date: June 10, 2019 05:22PM

done already but looks that everybody seems to think that it's not true (see my post on depth difference in the forum)

Titus

Options: ReplyQuote
Re: Compass unstable
Posted by: VK ()
Date: June 12, 2019 09:26PM

I have faced the exact same problem. The compass is unstable and jumps up and down constantly even if I keep it absolutely stable. Sometimes the wobbling is +-10 degrees, sometimes +-5 degrees, sometimes something different.

The compass does get north more or less right (well, hard to say exactly, because it is so unstable), but it has an internal systematic error of almost +-30 degrees. If I first align the compass with north (ie. so that the reading wobbles between ~355 and ~005) and then turn the compass clockwise exactly 90 degrees, the compass then wobbles between ~115 degrees and ~125 degrees. If I turn it another exact 90 degrees, it then wobbles between 187 and 190 (for some reason more stable now). Another 90 degrees: 242-248.

I have calibrated the compass tens of times, sometimes following the video, sometimes trying different techniques. The result is always more or less the same. The compass is very unstable and exact 90 degree turns result in the reading changing by 60-130 degrees.

I have given up regarding the compass, because it is useless in any kind of underwater orientation. I use the compass of my back-up Suunto because it is way more reliable.

Options: ReplyQuote
Re: Compass unstable
Posted by: Ralph ()
Date: June 12, 2019 09:35PM

Within my set of test computers, some perform quite stable and others do dance around. It's a known issue that tracks down to the different compass chips as written before. It's on the todo list...

Ralph

Options: ReplyQuote
Re: Compass unstable
Posted by: VK ()
Date: June 12, 2019 09:47PM

I'd like to add that the instability is not the only problem. Another problem is that a 90 degree turn results in a 120 degree change in the reading. And then in a 60-70 degree change. When I turn the computer by 90 degrees, the reading should change by 90 degrees.

Options: ReplyQuote
Re: Compass unstable
Posted by: Ralph ()
Date: June 13, 2019 06:13AM

Yeah, it's all connected, these compass chips require quite a lot of software postprocessing...

Options: ReplyQuote


Sorry, only registered users may post in this forum.
This forum powered by Phorum.