problem with matrixes

[sdherzo]

hi !
I am having a (probably ) simple problem which i am not able to solve

For an interpolation function i have to fill my matrix A & B . I have declared them & then tried to allot them values.

The problem now is that i just wanted to see ( before the values get into the interpolation function ) if the values of the matrix elements are right ... i printed out Sys_[0][0] but am getting a "junk" value. I then printed out the individual termas which make up Sys_[0][0] , i.e (double)s[2].s1 & (double) s[1].s1 ... these values are printed right. But the "calculated" value is "junk" ...

WIll be great if anyone notices the mistake :-)
Thx

    struct Sensor 
           
	{
		int nr,n;
		float d;
		double s1,s2,s3,s4,a;

	};
....
static struct Sensor s[100];
...

double Sys_[4][4]; 

double b_[4];    

double z_[4];   

....
...
...

Sys_[0][0]=  (double)s[2].s1 -(double) s[1].s1;
Sys_[0][1]=  (double)s[3].s1 - (double)s[1].s1;           
Sys_[0][2]= (double) s[4].s1 - (double) s[1].s1;
Sys_[0][3]= (double)s[5].s1 - (double)s[1].s1;


Sys_[1][0]= (double)s[2].s2 - (double)s[1].s2;
Sys_[1][1]= (double)s[3].s2 - (double)s[1].s2;
Sys_[1][2]= (double)s[4].s2 - (double)s[1].s2;
Sys_[1][3]= (double)s[5].s2 - (double)s[1].s2;


Sys_[2][0]= (double)s[2].s3 - (double)s[1].s3;
Sys_[2][1]= (double)s[3].s3 - (double)s[1].s3;
Sys_[2][2]= (double)s[4].s3 - (double)s[1].s3;
Sys_[2][3]= (double)s[5].s3 - (double)s[1].s3;


Sys_[3][0]= (double)s[2].s4 - (double)s[1].s4;
Sys_[3][1]= (double)s[3].s4 - (double)s[1].s4;
Sys_[3][2]= (double)s[4].s4 - (double)s[1].s4;
Sys_[3][3]= (double)s[5].s4 - (double)s[1].s4;




b_[0] = (double)DataVal0  - (double)s[1].s1;
b_[1]=  (double)DataVal10 - (double)s[1].s2; 
b_[2]=  (double)DataVal20 - (double)s[1].s3; 
b_[3]=  (double)DataVal30 - (double)s[1].s4;


  gj4(Sys_, z_, b_);

[Salem]

> gj4(Sys_, z_, b_);
So what does the prototype / implementation of this function look like?

[sdherzo]

hm ... whwn i am printing sys_[][] before the func i get the right answer , so i think i know where the prob was ... thbaks ;-)


I do have another problem now tho ... i wated to calculate s[i ].a for 72 values ( as 72 is the number of elemnts in the structure ... have i created the for loop correct ?? i am just getting the same s[i].a values fpr all values ... for eg. s[1].a = s[34].a ... but i know it cant be the same ;-)

 for ( i = 1 ; i <= 72 ;i++ ) 
  {

  s[i].a =    ( ( s[i].s1 - DataVal0) * (s[i].s1 - DataVal0 )) + ((s[i].s2 - DataVal10 )*( s[i].s2 - DataVal10 )) + ((s[i].s3 - DataVal20) * (s[i].s3-DataVal20)) + (( s[i].s4 - DataVal30 ) * ( s[i].s4 - DataVal30 ));	
   
   }

[vart]

in C/C++ array indexes start at zero
so probably
for ( i = 0 ; i < ndata ;i++ )

[sdherzo]

in C/C++ array indexes start at zero
so probably
for ( i = 0 ; i < ndata ;i++ )

does not seem to be the problem :-(

[sdherzo]

i think the problem is that DataVal is declared as static unsigned long & s[i].s1 as double . i broke up the formula & tried ( s[1].s1 - DataVal0) & ( s[23].s1 - DataVal0) i am just getting "-DataVal0" as the answer ... putting a (double) before the DataValo0 doenst help ...

any suggestions ???

[Salem]

Yes, post more code.
Because your attempts at guessing which lines of code are important just isn't working.

[sdherzo]

struct Sensor 
           
	{
		int nr,n;
		float d;
		double s1,s2,s3,s4,a;

	};

...

static struct Sensor s[100];
static unsigned short DataVal, DataVal1, DataVal2, DataVal3;
    static unsigned long DataVal0, DataVal10, DataVal20, DataVal30;DataVal0 = (double)mu *  ( (double)DataVal - (double)DataVal0 ) + DataVal0;
			DataVal10 = (double)mu * ( (double)DataVal1 - (double)DataVal10 ) + DataVal10;
			DataVal20 = (double)mu * ( (double)DataVal2 - (double)DataVal20 ) + DataVal20;
			DataVal30 = (double)mu * ( (double)DataVal3 - (double)DataVal30 ) + DataVal30;

for ( i = 1 ; i <= 72 ;i++ ) 
  {

  s[i].a =    ( ( s[i].s1 - DataVal0) * (s[i].s1 - DataVal0 )) + ((s[i].s2 - DataVal10 )*( s[i].s2 - DataVal10 )) + ((s[i].s3 - DataVal20) * (s[i].s3-DataVal20)) + (( s[i].s4 - DataVal30 ) * ( s[i].s4 - DataVal30 ));	
   
   }

& like i said the problem is with the subtraction as for eac block it just gives me DataVal0, ... back.

[sdherzo]

on the other hand if i perform ( s[1].s1 * DataVal0) & give the answer out as &#37;ld i get the correct answer out of the multiplication ... so why doesnt ( s[1].s1 - DataVal0) work ...

[sdherzo]

ok ... also solved ;-) declarde the s1,... as float ... Thanks anyway :-)