Replies to This Discussion

3D RoboticsPermalink Reply by Chris Anderson on January 2, 2011 at 8:04pm

Luis, the mags are user-configurable between 3v and 5v (via a solder jumper). Check your module: if the solder jumper is in the 3v position, you need to move it to the 5v position.

• ▶ Reply

Permalink Reply by Luis Mauro on January 2, 2011 at 8:27pm

Chris,

I already double checked the solder jumper, It´s connecting the center pad with the pad labeled 5V.

I´ll record the values supplied by the board and show in a time-stamp graph for a better understanding of the problem.

I´m thinking to make a non-sense trying: set-up the 5843 board to 3V and connect to the Arduino board through a logic level converter (Sparkfun)!

 

 

• ▶ Reply

3D RoboticsPermalink Reply by Chris Anderson on January 2, 2011 at 8:44pm

We don't see the problem on the APM board. You can see sample data here

• ▶ Reply

Permalink Reply by Luis Mauro on January 6, 2011 at 12:35pm

Hello

A question about the data from http://code.google.com/p/ardupilot-mega/wiki/MagnetometerTest
Does the magnetometer was moving or static (no movement)?

Please take a look in details the problem that I´m having (powerpoint attached file)

Attachments:

• HCM 5843.ppt, 3.3 MB

• ▶ Reply

Permalink Reply by Fabio Varesano on January 3, 2011 at 12:51am

Hi there, the problem I reported at http://www.arduino.cc/cgi-bin/yabb2/YaBB.pl?num=1291403700 was mostly imputable to the ITG3200 rather than the magnetometer. See this for more details.

 

You can give a try to my Arduino library for the HMC5843 based on Wire.

Also if you are into sensor fusion algorithms and 9DOM/DOF boards check my 9DOF/DOM orientation filter algorithm code and my FreeIMU 9DOM/DOF board.

• ▶ Reply

Permalink Reply by Luis Mauro on January 3, 2011 at 8:16am

Hello Fabio

 

First of all, thank you for your help. I already downloaded your library and I´ll try!!

IMPORTANT: I´m not using the SparFun board (3.3V) that needs a level converter to be connect to 5V devices, I´m using DIYdrones board that can deal with 3V3 and 5V devices. In this case it´s setted-up to 5V (no level converter required).

• ▶ Reply

Permalink Reply by Uwe Gartmann on January 3, 2011 at 2:13pm

Hi Fabio

 

Great stuff. What do you use to display the orientation on screen?

 

I enhanced the original APM code for the HMC5843 for my purposes: Link (goto libraries)

 

• ▶ Reply

Permalink Reply by Fabio Varesano on January 4, 2011 at 1:15am

Hi Uwe, I use Processing to display the cubes on the screen. It's a very simple programming language which perfectly fits the need of rapid graphics prototyping.

 

If you were talking about the 9DOF/DOM orientation filter above the Processing code is available on the attachments of my 9DOF/DOM blog post.

• ▶ Reply

Permalink Reply by Uwe Gartmann on January 4, 2011 at 2:18pm

very cool, thanks for sharing.

• ▶ Reply

Permalink Reply by Uwe Gartmann on January 6, 2011 at 1:49pm

Hi Fabio

 

Can you give me some support regarding your code?

How ist the syntax/format of the string send to the processing prg?

• ▶ Reply

Permalink Reply by Fabio Varesano on January 6, 2011 at 4:09pm

Sure, the sensor fusion algorithm computes a quaternion which is a representation of the orientation of the body in 3d. A quaternion is stored in memory as four floating variables.

 

These variables are then sent over the wire using an HEX representation.The format should be:

f:HEX_q1,f:HEX_q2,f:HEX_q3,f:HEX_q4,\r\n

Above, HEX_q1 means the hexadecimal representation of quaternion first component.

That's how I managed to send floats over serial without loss of precision. I wrote a blog post on this topic.

 

The Processing app just need to read the data from the serial interface, decode that back into a float and use the orientation data for displaying the rotated cube.

 

For more examples on Arduino, Serial and Processing check my blog post: Serial communication with Arduino and Processing: simple examples a...

 

Feel free to ask if you need more help.

• ▶ Reply

Permalink Reply by Uwe Gartmann on January 8, 2011 at 2:16pm

Hi Fabio

I try now since hours to run your code on my APM board. But it does not work.

 

I do not know why, but when the calculated values are written back to q0..q3, this values remain zero at all.

 

May the compiler is wrong. I found that A = A + 1 is not working correctly until I write it like this: A =+ 1

Very strange all this things. At this point I miss professional development tools.

• ▶ Reply

• ‹ Previous
• 1
• 2
• 3
• 4
• Next ›
• Page