HP Forums
HP33S HMS Conversion Bug - Printable Version

+- HP Forums (https://archived.hpcalc.org/museumforum)
+-- Forum: HP Museum Forums (https://archived.hpcalc.org/museumforum/forum-1.html)
+--- Forum: Old HP Forum Archives (https://archived.hpcalc.org/museumforum/forum-2.html)
+--- Thread: HP33S HMS Conversion Bug (/thread-61431.html)



HP33S HMS Conversion Bug - Dave A - 08-02-2004

I have noticed comments on HP33S bugs in your thread.
Has this one come up? It occurs on mine and others.

Key in -14.2900
Press HR HMS
See -14.2820

Key in 14.2900
Press HR HMS
See 14.2900

The error occurs when the angle is negative but not when positive. Does it occur on all HP33S units?

Edited: 3 Aug 2004, 6:32 a.m. after one or more responses were posted


Re: HP33S HMS Conversion Bug - bill platt - 08-02-2004

Hi Dave,

Thanks for posting this.

It seems that between this bug, and the polar conversion bug, we have a real problem to deal with. Not to mention the strange behavior with solve not finding the negative real roots of a quadratic while solving in the equaton mode, and the RAD annunciator bug that Ben Salinas found....and on an on...

we have quite a list growing here.

I will test on both of mine (I bought a 1st release WalMart version along with a newer one) to see if it is consistent(ly bad).

regards,

Bill


Re: HP33S HMS Conversion Bug - Fubar Vikinghelmet - 08-02-2004

I have this bug.


Re: HP33S HMS Conversion Bug - Karl Schneider - 08-03-2004

Dave --

Yup, that's another one!

I get the following on my 33S (CNA4134082)

Key in -14.2900 Press HR HMS See -14.2820 (INCORRECT! -- should be -14.2900)

Key in 14.2900 Press HR HMS See 14.2900 (CORRECT)

I tried a few others. The consistent factor seems to be incorrect reconversion back to H.MMSS from H.DDDD, of a negative H.MMSS value where,

MM = 3n+2 (integer n = 0 through 19); and
SS = 0.000

In other words, for any negative H.MMSS with |MMSS| exactly equal to

.02, .05, .08, .11, .14, .17, .20, .23, .26, .29, .32, .35, .38, .41, .44, .47, .50, .53, .56, or .59

->HR ->HMS will yield |MMSS| = .0120, .0420, etc.

Weird! (And unacceptable -- none of the HP-designed calc's make this mistake.)

It seems that certain internal representations of numbers are not being handled in a consistently correct fashion.

Edited: 3 Aug 2004, 12:24 a.m.


Re: HP33S HMS Conversion Bug - Dave A - 08-03-2004

Thanks for the responses and specially your analysis, Karl. A class of 40 navigation trainees apparently all hit the problem here in South Africa. We have not checked out serial numbers, if this has any significance. Mine is CNA41206382. Meantime best for general safety to avoid all negative angle conversions by inspection, changing the sign before and again after the calculation. Ever heard of a calculator recall? - I have not in almost 30 years of dealing in HP calculators. Perhaps warranty, or exchange scheme claims met for goodwill? Guess we would pay in something for a few more labels too!


Re: HP33S HMS Conversion Bug - John Smitherman - 08-03-2004

Bill, when you get a chance can you check the conversion of trig functions with negative angles, such as asin(sin(-45)) on the HP33s?

Thanks,

John


Re: HP33S HMS Conversion Bug - bill platt - 08-03-2004

Hi Johm,

Tested at -30,-20,-45,120, (degrees) all works correctly.

Have not tried RADIANS or GRADS.

Regards,

Bill


Re: HP33S HMS Conversion Bug - bill platt - 08-03-2004

Well, this bug has some interesting properties.

the bad result seems to be caused somewherein the digits past what you can get on the screen.

                            {display ten 3's}
Instead of doing -0.29 -->HR {.483333333333} -->HMS
with reuslt = -0.282

Try typing in:

.48333333333 (that's 9 "3's"}
and doing
-->HMS
you get
-0.286 which is correct--as the "6" is worth one hour.

(Don't get too upset with this "60 minute" quirk--even
the 15c does this in certain rounding circumstances--
including this one--if you type in .483...to your limit, the 11c/15c will -->H.ms return ".286" . However, if you do 29/60 --.H.ms, they will return ".29")

Now, even if you get to the .483.... using your own computations--say take -29/60---you will still get the buggy result---so it is not just the conversion program, but perhaps something more fundamental?

BTW, one of Karl's exapmles actually works ok: .08-->hr-->H.MS goes to .076---which is sort of ok).

Replacement Programs for the HR and H.MS:

Not really optimized!

LBL U    (type in your H.MS and covert it to HR}
sto A
IP
sto B
RCL A
FP
100
*
IP
STO C
RCL A
100
*
FP
100
*
STO D
60
x^2
/
RCL C
60
/
+
RCL B
+
RTN

LBL V {type in decimal HRs and get back H.MS}
STO Z
FP
60
*
ENTER
IP
STO x
x<>y
FP
60
*
sto Y
RCL Z
IP
RCL X
100
/
+
RCL Y
10,000
/
+
RTN

Of course, this routine is going to give you the "60 minute" thing, but at least it is numerically correct.

Note that you could also do this with an Equation---and then you would only need one variable:

H.MS-->HR:
((FP(100*A)/0.6+IP((FP(A)*100)))/60+IP(A)

ck=FBDC
LN=41

Bill

Edited: 3 Aug 2004, 8:19 p.m.