[SOLVED] Conversion from 2s complement to decimal

[north2012]

hello,

Does anybody can help me with writing function for conversion from 2scomplement (16-bit) to decimal value? I am using micro C compiler, and here is my code which does not works... I received wrong results (-30 dec) on LCD for all calculations.

// LCD module connections
sbit LCD_RS at PORTC3_bit;
sbit LCD_EN at PORTC2_bit;
sbit LCD_D4 at PORTC4_bit;
sbit LCD_D5 at PORTC5_bit;
sbit LCD_D6 at PORTC6_bit;
sbit LCD_D7 at PORTC7_bit;

sbit LCD_RS_Direction at DDC3_bit;
sbit LCD_EN_Direction at DDC2_bit;
sbit LCD_D4_Direction at DDC4_bit;
sbit LCD_D5_Direction at DDC5_bit;
sbit LCD_D6_Direction at DDC6_bit;
sbit LCD_D7_Direction at DDC7_bit;
// End LCD module connections

void main(void)
{
signed short int p;
unsigned short mask;
char *txt;
char txtp[6];
//start of program
Lcd_Init();                        // Initialize LCD
delay_ms(100);
Lcd_Cmd(_LCD_CLEAR);               // Clear display
Lcd_Cmd(_LCD_CURSOR_OFF);          // Cursor off
lcd_out(4,1,"INIT DONE...");
delay_ms(1000);
lcd_out(4,1,"CALCULATION 1");
  p = 0x038B;
   mask = 0x8000;
   while(mask)
   {
   if(p & mask)
       *txt = '1';
   else
       *txt = '0';
   txt++;
   mask >>= 1;
   }
  *txt = 0;
   ShortToStr(txt, txtp);
   Lcd_out(1,2,txtp);
  delay_ms(3000);
  Lcd_Cmd(_LCD_CLEAR);               // Clear display
  lcd_out(4,1,"CALCULATION 2");
   p = 0x0204;
   mask = 0x8000;
   while(mask)
   {
   if(p & mask)
       *txt = '1';
   else
       *txt = '0';
   txt++;
   mask >>= 1;
   }
  *txt = 0;
   ShortToStr(txt, txtp);
   Lcd_out(1,2,txtp);
 
  Lcd_Cmd(_LCD_CLEAR);               // Clear display
  delay_ms(1000);
  lcd_out(4,1,"CALCULATIONS DONE");

  while(1)
  {
  }
}

 

[jayanth.devarayanadurga]

Explain what you are trying to do.

 

[north2012]

Hello,

I am trying to connect AD5933 with ATmega128. AD5933 is impedance converter and measurement results are stored in 4 8-bit registers (Real high&low byte and Imag high&low byte). Results are in 2s complement data format and for impedance calculation Z=sqrt(R*R+I*I) I need to convert 2s complement to decimal value (see example from data sheet in attachment). Because of that I want to first test function for conversion, that is code above. I want to convert three different numbers (0x038b, 0x0204 and 0xfa3f) to decimal to verify algorithm for conversion.

- - - Updated - - -

 

[jayanth.devarayanadurga]

The datasheet doesn't mention that the data is in 2's Complement and why would they design a device which will give result in 2's Complement. The datasheet tells that the data in the registers at address 0x94 and 0x95 will be 0x038B (say) = decimal 907 and value at addresses 0x96 and 0x97 will be 0x0204 (say) = decimal 516. You have to take the squares of the 2 numbers and add them and then take the square root of the result. Where is the 2's Complement?

unsigned int R, I;
float Z;

---

Code C - [expand]
1
2
3
4
R = 0x038B;
I = 0x0204
 
Z = sqrt(R^2 + I^2);

---

 

[Venkadesh_M]

its already given in the correct format...... (they shown in the Hex)

 

[north2012]

Many thanks for your replies. In data sheet page 26 of 40 there is explanation (I attached picture).

- - - Updated - - -

here is one solution...

signed int convert (unsigned int p)
{
   unsigned short mask;
   signed int res = 0x0000;
   mask = 0x8000;
   if((p & mask) == 0x0000) //positive
   {
   res = p;
   }
   else //negative
   {
   res = p;
   res = ~res;
   res = res+1;
   res = 0-res;
   }
   return res;
}

 

[jayanth.devarayanadurga]

Ok. What is the problem? 2's complement is nothing but 1's Complement + 1. To reverse subtract 1 from 2's complement and then invert the bits of the result.

 

[Jayaraja]

Hai all,
I am trying to interface AD5933 with PIC16F886. Does anyone got a working proto board (only for AD5933)?
I am from Bangalore, India.
Regards,
Jayaraj.A