I'm having some problems getting my mbed to read in data from a CMPS03 compass chip, and save it to a text (or other) file for later use - any help would be greatly appreciated, even if it turns out to be a simple mistake! The code I'm trying to use is below:

(The various LEDs are so I could try to visually see where the program was up to)

#include "mbed.h"

I2C compass(p9, p10);
DigitalOut led1(LED1);
DigitalOut led2(LED2);
DigitalOut led3(LED3);
DigitalOut led4(LED4);

LocalFileSystem local("local");

int main() {
        
        int address = 0xC0;
        char read_cmd[2], read_result[3] = {0}, prnt_result[3] = {0};
        read_cmd[0] = 2;      
        unsigned short data;

         led1 = 1;
         wait(0.5);
        
         compass.write(address, read_cmd, 1);
         wait(0.05);
         compass.read(address, read_result, 2);
         led2 = 1;
         read_result[0] = prnt_result[2];
         read_result[1] = prnt_result[1];
         read_result[2] = prnt_result[0];

         memcpy(&data, prnt_result, 3);
         led3 = 1;
        
         FILE *fp = fopen("/local/out.txt", "w");
         fprintf(fp, "Begin Compass Log...");
                 
         led4 = 1;
    
         fprintf(fp,"The compass data is: ");
         fputs(prnt_result, fp);
         fclose(fp);
         led4 = 0;
         
        }

I'm planning to try something similar with a mindsensors I2C compass (CMP2X), hence the interest in this topic.

I think it might indeed be a simple mistake...

Aren't you overwriting the read_result data from your compass with prnt_result values (three lines afer the led2 = 1 statement)?

Perhaps if you swap those statements to something like this you might have better luck:

prnt_result[0] = read_result[0];

Hi Torben,

Would it be possible for you to explain what problems you are having explicitly? Does the file not get written to or does it give a non-sensical value? What LED does the program get up to?

Some of the things I notice are:

• After reading the contents of registers 2 and 3 into read_result[0] and read_result[1] you are then setting read_result[0] = 0 and read_result[1] = 0 due to the fact all the elements of the prnt_result array were initialised to 0. As Andrew pointed out, you'll need to swap the statements.
• The bearing is contained as a 16-bit number in registers 2 and 3, and you only read in 2 bytes to read_result, so your read_result array only needs to contain 2 elements.
• I believe the size of an unsigned short on the compiler is 2 bytes; I don't know what sort of behaviour memcpy'ing 3 bytes to a 2 byte location will exhibit, but I'm sure it can't be good.
• You're memcpy'ing zeroes to the data variable as the prnt_result array holds zeroes. I assume the prnt_result array should hold the value of the bearing in "human readable" form, but at the moment it isn't doing anything because it doesn't contain the right values.
• If you're going to use fputs to put the result into the file, then prnt_result needs to be a null-terminated string, meaning the last element of the array needs to be the null character '\0' - assuming the data variable held the correct value you could simply do: fprintf(fp, "The compass data is: %i\n", data); instead.

I have written a class for the CMPS03 which is successfully logging values to the local file system.

I hope this helps, let me know if it doesn't.

Cheers,

Aaron

Thanks for your swift responses to my somewhat amateur mistakes! I've changed the code round to eliminate the problems pointed out by you both, and then written a new program based mostly on Aaron's library, so thanks a thousandfold for that!

Cheers!