float to char conversion with CCS 5.5

Thanks llmars for your answer.

There is no problem to compile this code but the function doesn't work ?!

I only have an information:

Recommend moving them to RAM during run time or not using
     as these are processing/power intensive

on the sprintf line

Dr.No said:

There is no problem to compile this code but the function doesn't work ?!

Sprintf function is precompiled already, resides in C run-time library. Thing is that compilers have different versions of libraries - as I already said. You have to find how to change them so you have one with sprintf that supports float. Look for project "generic" options, runtime libraries or something like that. Sorry, I have no experience with CCS, I can tell you how to do it in IAR.

Something like: Properties->CCS General->Advanced Settings and verify that the Runtime support library field

As illustration here's excerpt from SLAU132H:

--printf_support={full| Enables support for smaller, limited versions of the printf function
nofloat|minimal} family (sprintf, fprintf, etc.) and the scanf function family (sscanf,
fscanf, etc.) run-time-support functions. The valid values are:
• full: Supports all format specifiers. This is the default.
• nofloat: Excludes support for printing and scanning floating-point
values. Supports all format specifiers except %a, %A, %f, %F, %g,
%G, %e, and %E.
• minimal: Supports the printing and scanning of integer, char, or
string values without width or precision flags. Specifically, only
the %%, %d, %o, %c, %s, and %x format specifiers are supported
There is no run-time error checking to detect if a format specifier is
used for which support is not included. The --printf_support option
precedes the --run_linker option, and must be used when performing
the final link.

I would multiply the floating point number by 10000 (for 4 decimal places. i.e. 99.xxxx) and convert to long integer like so:

#include <msp430g2553.h>
#include <stdio.h>

volatile float temperature;
volatile long int temp;
volatile  unsigned int hundreds, tens, units, tenths, hundredths, thousandths, tenthousandths;
volatile long int remainder;


unsigned char string[30];
main ()
{
	


        temperature=25.0625;
        temperature *= 10000;//250625.
        temp = (long int)temperature;//250625

        tens = temp/100000;//2
        remainder = temp - tens*100000;//50625
        units = remainder/10000;//5
        remainder = remainder - units*10000;//625
        tenths = remainder/1000;//0
        remainder = remainder - tenths*1000;//625
        hundredths = remainder/100;//6



        sprintf(string, "Degrees =  %d%d.%d%d", tens, units, tenths, hundredths);
        while (1);


}

Continue along the same lines to get the remaining decimals. And if the temperature is greater than 99 start with a 100000 multiplier to get 3 digits.

@ llmars & Joseph Raslavsky

Many thanks for your very detailed answers. I'll try to found the good library and I'll also try your solution Joseph.

Happy New Year 2014 all my best greetings from France to you and yours

Just a small note...

Enabling float type to be printable using sprintf function will increase your code size for quite an amount... so be careful with that. G2553 has small flash memory which sometimes not sufficient... (happened to me once)

Also, happy new year to you too :)

Ilmars said:

• full: Supports all format specifiers. This is the default.

So the support level has to be changed in the project settings, from minimal to full.

And indeed, it adds significant amount of code to the binary. (converting float numbers to ASCII is quite a complex task)

If you want build yourself library you can use these information as reference:

1. Build an "Number" ASCII (0 - 9) array, include ".";

2. According the IEEE 754 standard analysis the single precision floating point. 

3. Suggest you use assembly language improve the efficiency. 

4. If you want build the PC side software, I suggest you use the char array and UNION data structure.

Hi Joseph,

I am working on similar code. I am working on F28069 processor (C2000 application) and doing coding using CCS

I also want to convert float to char and I tried the logic which you posted but the sprintf function doesn't seem working.

Here I attached the screenshot of CCS Debug window. Please suggest suitable solution for this problem asap

Viswa said:

I am working on F28069 processor

Then why you write in msp430 microcontroller forum? :)

Hi llmars...

As I Saw relavent question, I posted my query I didnt checked the Forum... Sorry

Viswa said:

Hi Joseph,

 but the sprintf function doesn't seem working.

  You don't need sprintf. Just use an ascii table to convert to a character. Adding 0x30 does it for the digits 0 thru 9

        valueNew[0] = hundreds + 0x30;
        valueNew[1] = tens + 0x30;
        valueNew[2] = units + 0x30;
        valueNew[3] = tenths + 0x30;
        valueNew[4] = hundredths + 0x30;
        valueNew[5] = etc
        valueNew[6] = etc
        valueNew[7] = '\0'; //string terminator