brainwagon / angst Goto Github PK

I was looking at the output of your code vs some other code, and noticed that compared to the original G3RUH sources, you apply a different formula to the M2 variable than he does. I'm curious if you corrected an error in his code, or perhaps overlooked what he had done, or perhaps something obvious is eluding me.

In line 241, you have the following code: M2 = RADIANS(getdouble(l1, 33, 43)) ;

However, in G3RUH's code, at line 1370, he uses M2 = M2*2*PI

G3RUH's original code is at http://www.amsat.org/articles/g3ruh/111.html

Another issue that I found while in there is where a float is used instead of a double. Changing it to a double improves the accuracy.

For all those, who want to use the implementation after year 2015, the sidereal and solar data should be updated (see http://www.amsat.org/articles/g3ruh/111.html end of code).
Change the lines 41..49 in P13.h to

static const double YG = 2014.f;
static const double G0 = 99.5828f;
static const double MAS0 = 356.4105f;
static const double MASD = 0.98560028f;
static const double EQC1 = 0.03340;
static const double EQC2 = 0.00035;
static const double INS = radians(23.4375f);

Solar data is valid until ~2030

Could it be, that in line 71 of P13.h (DateTime destructor) a semicolon is missing?

I would suggest to rename the classes to something like P13...
Especially the class DateTime could cause potential conflicts with other libraries like e.g. RTC libs which often also use DateTime in their context.

I wondered what the intention is at the constants in P13.h to declare a constant as "double" and then use the "f" literal for singleprecision at the value?
Is it to save storage space and not have the need to cast them to double when calculating/working with them?

The calculations for the sun prediction seems to differ from the original code.

Original BASIC code:

:
C = COS(-GHAA): S = SIN(-GHAA)
:

Code in P13.cpp Sun::predict lines 420/421:

:
C = cos(-GHAE) ; 
S = sin(-GHAE) ; 
:

Maybe this could be the reason, that the altaz()-method is broken.
GHAA seems to be calculated in Satellite::predict, but not used in Sun::predict (I am not really sure, if it is the same value, but it seems that is is some kind of time which should be the same for Satellite as well as for Sun prediction at the same time).
So in my opinion, the method should look like this:

void
Sun::predict(const DateTime &dt)
{
    long DN = dt.DN ;
    double TN = dt.TN ;

    double T = (double) (DN - fnday(YG, 1, 0)) + TN ;
    double GHAE = RADIANS(G0) + T * WE ;
    double GHAA = GHAE + T * WE;
    double MRSE = RADIANS(G0) + T * WW + M_PI ;
    double MASE = RADIANS(MAS0 + T * MASD) ;
    double TAS = MRSE + EQC1*sin(MASE) + EQC2*sin(2.*MASE) ;
    double C, S ;

    C = cos(TAS) ;
    S = sin(TAS) ;
    SUN[0]=C ;
    SUN[1]=S*CNS ;
    SUN[2]=S*SNS ;
    C = cos(-GHAA) ;
    S = sin(-GHAA) ;
    H[0]=SUN[0]*C - SUN[1]*S ;
    H[1]=SUN[0]*S + SUN[1]*C ;
    H[2]=SUN[2] ;
}

Regards, Thorsten dl9sec