Question about the compass_tutorial01.c

[thn_tpl]

Hello,

I have a couple questions about the compass_tutorial01.c:

1/ If I mount the BerryGPS-IMU board on top of the Pi4 board as in the attached photo, which way is "North" for the compass? Is it the red arrow direction?

2/ I have compiled and run the code but the result heading is only moved somewhere between 300 to 360 degree. Is there anything else should I check? I know this code is just a basic without calibration or compensation for tilt but I would expect it to be somewhat close to 0 to 360 when I move the unit full circle.

Thanks for any advises.

Attachments:

1. 

   Screen-Shot-2022-07-21-at-10.36.13-AM.png
   

[Mark Williams]

Hi, rotate the red arrow 90 degrees anti-clockwise, this is north.

When the BerryGPS-IMU is sitting on the Pi, you need to calibrate as the Pi interferes with compass.

If you had the BerryGPS-IMU connected via a cable and some distance from the Pi, you may not need to calibrate the compass.

Mark --OzzMaker.com --

[thn_tpl]

Hello,

I have followed the instruction how to calibrate the compass here https://ozzmaker.com/compass3/

After that I recompile the test 03 program with the result of Min/Max values. However when I did a test with my unit positioned in one direction and then 180 turn around, I got these heading result:

First test results:

heading 354.670          magRaw X 589   magRaw Y 571    MagRaw Z -220

heading 356.553          magRaw X 506   magRaw Y 527    MagRaw Z -268

heading 354.433          magRaw X 577   magRaw Y 538    MagRaw Z -238

heading 355.347          magRaw X 504   magRaw Y 528    MagRaw Z -218

heading 354.452          magRaw X 563   magRaw Y 521    MagRaw Z -213

 

Second test result:

 

heading 303.325          magRaw X -685  magRaw Y -543   MagRaw Z -282

heading 302.978          magRaw X -554  magRaw Y -514   MagRaw Z -304

heading 306.539          magRaw X -697  magRaw Y -540   MagRaw Z -285

heading 302.722          magRaw X -633  magRaw Y -504   MagRaw Z -295

heading 304.783          magRaw X -621  magRaw Y -472   MagRaw Z -321

So I expected the result to be around 180 degrees apart but it was not. Please help me figure out what did I do wrong?

Thanks,

[Mark Williams]

when you did the calibration, did you spin the unit around a few times in all directions?

Mark --OzzMaker.com --

[thn_tpl]

Yes and I did until the max/min value no longer changing before I hit Ctrl-C

[Mark Williams]

can you please share the entire code in compass_tutorial03.c

Mark --OzzMaker.com --

[thn_tpl]

Here is the code exactly from your github, all I did was to change the "defines" and recompile.

For some reason, this forum won't let me upload C code or txt.

 

[thn_tpl]

Update:

We ordered two GPS-IMU boards and after seeing the strange compass readout with the first one, I took out the second one and plugged it in. Ran the same compass03 calibrate and then compass03 with new max/min value. Then the second board compass readout is very much as expected.

Is there a possibility that the first unit has a bad compass?

Regards,

[Mark Williams]

Hi.... I am so sorry you are having a trouble.  this would be the first I have heard of a compass not working.  Ill have to ask my colleagues to see if they have seen this before.

 

In the mean time, can you do a test?
Can you remove the current print statement from compass_tutorial03.c
and put in the below print statement.

        while(1){
                readMAG(magRaw);
                printf("Raw,%i,%i,%i,",magRaw[0],magRaw[1],magRaw[2]);
                //Apply hard iron calibration
                magRaw[0] -= (magXmin + magXmax) /2 ;
                magRaw[1] -= (magYmin + magYmax) /2 ;
                magRaw[2] -= (magZmin + magZmax) /2 ;
                printf("Cal,%i,%i,%i,",magRaw[0],magRaw[1],magRaw[2]);
                //Apply soft iron calibration
                scaledMag[0]  = (float)(magRaw[0] - magXmin) / (magXmax - magXmin) * 2 - 1;
                scaledMag[1]  = (float)(magRaw[1] - magYmin) / (magYmax - magYmin) * 2 - 1;
                scaledMag[2]  = (float)(magRaw[2] - magZmin) / (magZmax - magZmin) * 2 - 1;
                printf("Rscaled,%f,%f,%f\n",scaledMag[0],scaledMag[1],scaledMag[2]);

 
recompile and then start the program again, rotate the compass 180 degrees and send the output to sales@ozzmaker.com. please also include you calibration values
 

 

 

Mark --OzzMaker.com --

[thn_tpl]

Here is the test results I sent to sales address

====Result from calibration
#define magXmax -147
#define magYmax 2372
#define magZmax 1055
#define magXmin -2983
#define magYmin -553
#define magZmin -2709

Result after compiling with the above

heading 356.282 Raw,-880,1210,-1297,Cal,685,301,-470,Rscaled,1.586742,-0.416068,0.189692
heading 356.440 Raw,-876,1223,-1334,Cal,689,314,-507,Rscaled,1.589563,-0.407179,0.170032
heading 356.765 Raw,-874,1229,-1188,Cal,691,320,-361,Rscaled,1.590973,-0.403077,0.247609
heading 356.916 Raw,-876,1205,-1287,Cal,689,296,-460,Rscaled,1.589563,-0.419487,0.195005
heading 356.349 Raw,-878,1197,-1300,Cal,687,288,-473,Rscaled,1.588152,-0.424957,0.188098
heading 356.152 Raw,-866,1215,-1295,Cal,699,306,-468,Rscaled,1.596615,-0.412650,0.190755
heading 356.641 Raw,-887,1204,-1266,Cal,678,295,-439,Rscaled,1.581805,-0.420171,0.206164
heading 356.257 Raw,-888,1203,-1315,Cal,677,294,-488,Rscaled,1.581100,-0.420855,0.180128
heading 356.227 Raw,-870,1208,-1280,Cal,695,299,-453,Rscaled,1.593794,-0.417436,0.198725
heading 356.456 Raw,-868,1210,-1256,Cal,697,301,-429,Rscaled,1.595205,-0.416068,0.211477

Turning the unit 180 degree ================

heading 312.402 Raw,-2210,283,-1261,Cal,-645,-626,-434,Rscaled,0.648801,-1.049914,0.208820
heading 312.847 Raw,-2231,288,-1288,Cal,-666,-621,-461,Rscaled,0.633991,-1.046496,0.194474
heading 312.341 Raw,-2224,287,-1225,Cal,-659,-622,-398,Rscaled,0.638928,-1.047179,0.227949
heading 312.522 Raw,-2226,281,-1288,Cal,-661,-628,-461,Rscaled,0.637518,-1.051282,0.194474
heading 312.366 Raw,-2224,301,-1293,Cal,-659,-608,-466,Rscaled,0.638928,-1.037607,0.191817
heading 312.756 Raw,-2210,296,-1262,Cal,-645,-613,-435,Rscaled,0.648801,-1.041026,0.208289
heading 313.065 Raw,-2225,303,-1239,Cal,-660,-606,-412,Rscaled,0.638223,-1.036239,0.220510
heading 312.762 Raw,-2217,287,-1270,Cal,-652,-622,-443,Rscaled,0.643865,-1.047179,0.204038
heading 312.718 Raw,-2220,308,-1258,Cal,-655,-601,-431,Rscaled,0.641749,-1.032820,0.210414
heading 312.988 Raw,-2224,299,-1257,Cal,-659,-610,-430,Rscaled,0.638928,-1.038974,0.210946

[thn_tpl]

Hello,

It's been a week and I haven't heard anything from sales@ozzmaker.com regarding the compass performance.

Any helps to guide me through with getting a good compass reading would be very much appreciated.

Cheers,

[richardp]

You may have missed Peter's email. I just forwarded it again.

We created the test program.
Please download this file into the same directory as the current code you are working with.
You can use a wget
wget https://raw.githubusercontent.com/ozzmaker/BerryIMU/master/selfTest-LIS3MDL/selfTest-LIS3MDL.c

Compile with
gcc -o selfTest-LIS3MDL selfTest-LIS3MDL.c

Can you please run it and share the output

[thn_tpl]

Sorry for the delay, I was on travel.

Here is the test result after compiled the code and ran:

---------------------------RESULTS------------------------------
| Axis | ST min [gauss] | ST max [gauss] | Self test results |
-----------------------------------------------------------------
| X | 1.0 | 3.0 | 2.976940 |
| Y | 1.0 | 3.0 | 3.130469 |
| Z | 0.1 | 1.0 | 0.526260 |
-----------------------------------------------------------------
The test results in the last column need to be between the min and max values for the that axis

[thn_tpl]

And here is the test result for second board.

---------------------------RESULTS------------------------------
| Axis | ST min [gauss] | ST max [gauss] | Self test results |
-----------------------------------------------------------------
| X | 1.0 | 3.0 | 3.191758 |
| Y | 1.0 | 3.0 | 4.354581 |
| Z | 0.1 | 1.0 | 0.570539 |
-----------------------------------------------------------------
The test results in the last column need to be between the min and max values for the that axis

[Mark Williams]

is the second board above the one with the problem?

Mark --OzzMaker.com --

[Author]

Posts