Large Math Problems?

[Tre]

I am working with the ADXL210 example code. I actually have an ADXL202E so I am modifying the code slightly. I have however, run into some problems with the actual calculations.

I left ADXL210_SCALING_FACTOR alone defined at 100,000

signed long accelerometer_read(void) { // Returns hundreths of a g
signed int32 value = (signed int32)ADXL210_SCALING_FACTOR;

printf("Value0: 0x%X%X%X%X\r\n", make8(value, 3), make8(value, 2), make8(value, 1), make8(value, 0));
printf("t1: %lu, t2: %lu, zg: 0x%X%X%X%X, ppg: %lu\r\n", adxl_t1, adxl_t2, make8(adxl_zerog, 3), make8(adxl_zerog, 2), make8(adxl_zerog, 1), make8(adxl_zerog, 0), adxl_ppg);
value *= (signed int32)adxl_t1;
printf("Value1: 0x%X%X%X%X\r\n", make8(value, 3), make8(value, 2), make8(value, 1), make8(value, 0));
value /= adxl_t2;
printf("Value2: 0x%X%X%X%X\r\n", make8(value, 3), make8(value, 2), make8(value, 1), make8(value, 0));
value -= adxl_zerog;
printf("Value3: 0x%X%X%X%X\r\n", make8(value, 3), make8(value, 2), make8(value, 1), make8(value, 0));
value /= adxl_ppg;
printf("Value4: 0x%X%X%X%X\r\n", make8(value, 3), make8(value, 2), make8(value, 1), make8(value, 0));

delay_ms(1);

#asm nop #endasm
return(value);
}

This is the output that I get from the function
Value0: 0x000186A0
t1: 4683, t2: 8431, zg: 0x0000C350, ppg: 125
Value1: 0x1BDD7BE0
Value2: 0x0000D858
Value3: 0x00001508
Value4: 0x0000002B

As you can see, Value0 is correct, 0x000186A0 is indeed 100,000d. However, as soon as that is multiplied by 4683, the PIC is returning 0x1BDD7BE0 when it should actually be 1BE9B0E0. This multiplication results in an error of 800000d.

I have tried all of the casts that I can possibly think to add. I am using a PIC18F2550.

Thanks,
Drew

[PCM programmer]

I got rid of all your make8() stuff, and just used %LX. I made a test
program and compiled it with PCH vs. 3.188 and also PCM 3.225 and
it worked OK. Results:

t1: 4683, t2: 8431, zg: 0000C350, ppg: 125
Value1: 1BE9B0E0