Examples pull raw data in incorrectly

[schwghrt]

All of the examples pull in the raw data as an 'unsigned int'. This is an 18-bit device. The data needs to be pulled in as 'unsigned long'.

This bug causes the first two significant digits to be dropped. This makes the readings useless. You can see it in the examples because the normal Earth field is showing up as -4 and -7 Gauss. Also, the 'heading' example is also useless with this bug.

[PaulZC]

Hi @schwghrt ,

Thanks for reporting this. I agree, it does look like we're doing the wrong thing there.

I will be upgrading the library tomorrow, and will update the examples at the same time.

Best wishes,
Paul

[PaulZC]

Hi @schwghrt ,

Thanks again for reporting this issue.

I've updated the examples to use uint32_t. v1.0.3 will be released shortly.

(Just FYI, on some platforms - like ESP32 - unsigned int is 32 bit. The examples worked just fine on those boards! ;-) )

Best wishes,
Paul

[schwghrt]

Yeah, those pesky 8-bit controllers. :-) Thanks for fixing it! Now if I can only get the Set/Reset feature to work.... Sam

[PaulZC]

Hi Sam (@schwghrt ),

Please give v1.1.0 a try. It should resolve the set-reset issue.

Best wishes,
Paul

[schwghrt]

It did resolve it! Thank you. Have you thought about compensating for the internal offset? Now that manual Set/Reset works, you can get that offset value, and then subtract it from all future measurements? (Page 18 of the Rev A datasheet.) I had manually figured out what the offsets were for the sensor (by rotating the magnetic sensor). The Set/Reset method gave me similar results. Sam

[PaulZC]

Hi Sam (@schwghrt ),

Sure. That should be easy enough to do - now that Auto-SR is disabled...! I've opened a Feature Request for it: #11

I don't have time to write it right now. I'll try and get to it in a few days.

Best,
Paul

[PaulZC]

Hi Sam (@schwghrt ),

Turned out to be quicker and easier than I thought. Please test the code in #11 and let me know if it does what you need.

You could log the offset periodically - with temperature ( getTemperature() ). Fit a line / polynomial to the data and determine the offset vs. temperature coefficient(s)? ;-)

Enjoy!
Paul