converting int into float dsPIC33

Question

I am trying to convert int16_t to float

code is

int16_t ADC_value;

float voltage = (float)ADC_value/1000.0f;

printf("%f\r\n",voltage);

result

when the voltage exceeds 2.0V floating point values become (-)values. but before it converts to floating point values it is normal.

code:-

int16_t ADC_value;

printf("%u\r\n",ADC_Value);  //this is correct

is that int16_t to float conversion wrong?

On an unrelated comment, the `%u` format is for *unsigned* `int`. You pass an `int16_t` (which will be promoted to `int`). Mismatching `printf` format specifier and argument type leads to *undefined behavior*. — Some programmer dude, Aug 02 '22 at 08:52
As for your problem, what values can `ADC_value` have? What do those values represent? What are the actual values of `ADC_value` that turns out wrong? — Some programmer dude, Aug 02 '22 at 08:52
But what do the values *mean*? What does the value `2000` mean? What does the value `-2000` mean? Is the value perhaps supposed to be *unsigned* instead of signed? — Some programmer dude, Aug 02 '22 at 08:57
Range of `uint16_t` is 0 to 65535. Perhaps you should be using `uint16_t`? — Ian Abbott, Aug 02 '22 at 09:03
You do know that a signed 16-bit value can only reach up to `32767`? (And down to `-32768` assuming two's complement) You should definitely use the unsigned `uint16_t` instead, which has a range from `0` to `65535`. — Some programmer dude, Aug 02 '22 at 09:05
You mention that the problem starts with a voltage of `2.0` volt. What value of `ADC_value` does that correspond to? `2000`? — Some programmer dude, Aug 02 '22 at 09:08
there was an int16_t used place in the ADC conversion program thanks for the help i didn't see that. thank you @Some programmer dude — madhawa polkotuwa, Aug 02 '22 at 09:38
@madhawpolkotuwa "when the voltage exceeds 2.0V" --> what is the value of `ADC_value` when voltage == 2.0V? — chux - Reinstate Monica, Aug 02 '22 at 09:39

FooAnon · Accepted Answer · 2022-08-02T15:33:32.807

For ADCs, I find this to be the prevailing pattern. Since you are using LTC1867, you need to determine whether the IC is operated in bipolar or uni-polar mode. The FSR is 4.096. If unipolar, you should use a uint16_t as the ADC value. If bipolar mode, you should use int16_t. In bipolar mode, the ADC is a signed twos complement, which can vary from +/- 32767, so you would need to alter the FSR variable I use below.

const float fsr = 4.096;        // Full-scale range, Volts, depends on the ADC
const float max_val = 65535.0f; // For 16-bit unsigned ADC, 2^16 -1 
float voltage = fsr*((float)ADC_value / max_val);

converting int into float dsPIC33

1 Answers1