Program listing for RPN Sunrise/Sunset

[Craig Bladow]

In the Articles section there is a listing for Program to calculate Sunrise/Sunset times. I came across a character in the program listing on line 68 (and others) that is puzzling. It looks like it should be an operator, but the section of listing below includes the other common arithmetic operators. Can someone explain what this is?

66 18
67 RCL 07
68 •
69 24
70 /
71 RCL 05
72 +
73 STO 08
74 .9856
75 x
76 3.289
77 •
78 STO 09


Link to the article: http://www.hpmuseum.org/cgi-sys/cgiwrap/...?read=1131

[Dieter]

(10-14-2017 04:10 PM)Craig Bladow Wrote:  In the Articles section there is a listing for Program to calculate Sunrise/Sunset times. I came across a character in the program listing on line 68 (and others) that is puzzling. It looks like it should be an operator, but the section of listing below includes the other common arithmetic operators. Can someone explain what this is?

If you look at the DOY routine it seems that this "•" must be a minus. This also matches line 68 where the difference between 6:00 resp. 18:00 h and the longitude divided by 15° (R07) is calculated. The same is true for the code around line 145 where, depending on flag 02, either a longitude or 360° minus this is calculated.

BTW there seems to be a similar problem with the 35s version. There are several lines with an "1." which also seems to stand for a "–" minus.

However I wonder why this "•" or "1." lines appear in the listing while elsewhere a simple "–" seems to work fine.

Dieter

[Craig Bladow]

Thanks Dieter, I'll give "minus" a try.

[SlideRule]

Craig
from the 33s listing for the equivalent section of program code:
C0029 FS? 0 If set use 18 for sunset calculation
C0030 18
C0031 RCL- G
C0032 24
C0033 /
so I'ld say Dieter nailed it.

BEST!
SlideRule

[Craig Bladow]

Thanks!

I found this also in the DOY program:

11 FS? 01 DMY mode

Should be

11 FS? 31 DMY mode

Still get a DATA ERROR, so need to do some more debugging.

[Dieter]

(10-14-2017 09:20 PM)Craig Bladow Wrote:  Still get a DATA ERROR, so need to do some more debugging.

FTR: I tried the HP41 version in V41 and it works just fine. The times for today exactly match those from a good online calculator.

However, I think the program can be improved. I also wonder if the FLOOR routine is really required. But I don't get any errors. At least not for my location and timezone. What values did you try?

Edit: The program calculates floor(R10/90) and floor(R11/90). Both R10 and R11 are the result of a mod 360 operation which always returns a value >= 0. So I think the complete FLOOR routine is obsolete and can be replaced with a simple INT command. And maybe the whole thing can be done even more elegantly with MOD.

Edit 2: I now have replaced the following lines of the original program...

Code:

 99    ATAN
100    360
101    MOD
102    STO  11
103    RCL  10
104    90
105    /
106    XEQ 'FLOOR    Floor function.  External program
107    90
108    x
109    +
110    RCL  11
111    90 
112    / 
113    XEQ 'FLOOR    Floor function.  External program
114    90
115    x 
116    -     
117    15
118    /     
119    STO  11

...with this:

Code:

 99    ATAN
100    90
101    MOD
102    RCL  10
103    90
104    /     
105    INT
106    90
107    x 
108    +
109    15
110    /
111    STO  11

And it seems to work. But please do your own tests and report here if you find any cases where errors occur.

Dieter

[Craig Bladow]

Thanks for validating the program Dieter! It turns out I misplaced a decimal point in one of the constants so I am now getting good results. Once I save this to mag card(s) I will try your modifications. Since my longitude is > 90 degrees West it will be a good test of your changes.

[Craig Bladow]

Dieter, yes it works with your changes for my location. Thank you!

[Dieter]

(10-14-2017 09:20 PM)Craig Bladow Wrote:  I found this also in the DOY program:

11 FS? 01 DMY mode

Should be

11 FS? 31 DMY mode

BTW, there is no need to check flag 31 at all. Instead of Dec 31 simply count the DOY from 1 Jan, which is 1.01yyyy in either mode. ;-)
Here is a way to do so:

Code:

01  LBL"DOY"
02  100
03  SIGN
04  LastX
05  RCL Z
06  x
07  FRC
08  101
09  +
10  %
11  R^
12  DDAYS
13  +
14  END

Edit: The original program uses data registers up to R15. Since some values are only stored and recalled once and can be held on the stack, the program can be modified so that it requires only registers up to R10. Maybe even less. And I got it down to less than 330 bytes (including the DOY routine) so that it will fit on three card tracks.

I don't know if the author of the original program, M. Joury, is still reading this forum, so I'm not sure if it's OK for him if I would post my "streamlined" version of his program. The implemented algorithm looks very accurate.

Dieter

[Craig Bladow]

Dieter,
I think the updated version, with corrections of errors in the posting on the HP Articles portion of the old website, would be a valuable contribution to the HP-41c software library.