HP50g won't compute integrals using infinity
in [HP48]
|
Prev: HP50g -> computing whether or not (mathematical) series converge, and, if so, which value to?
Next: How many colors can be used for the HP 48 Emu emulator skins?
From: Merciadri Luca on 29 May 2010 14:23 -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi, For some weeks, my HP50g has decided not to compute integrals with `infinity' boundaries. When trying, I receive something like `Excess overflow' (it is in Spanish). Why? I looked carefully through the CAS flags, but I have deactivated every option (rigorous mode, etc.) which could lead to such problems. I really need it to compute such integrals, to verify my results. How can I? It worked, some weeks ago... Thanks. - -- Merciadri Luca See http://www.student.montefiore.ulg.ac.be/~merciadri/ - -- A closed mouth catches no flies. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Processed by Mailcrypt 3.5.8 <http://mailcrypt.sourceforge.net/> iEYEARECAAYFAkwBW6kACgkQM0LLzLt8MhxRKgCZAfjudFPqP+LctyGFP26711u8 8n8An3VShwGBRqWJI35ay5Vr6Ga538cb =jjHj -----END PGP SIGNATURE-----
From: Bart on 31 May 2010 11:52 On May 29, 7:23 pm, Merciadri Luca <Luca.Mercia...(a)student.ulg.ac.be> wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > Hi, > > For some weeks, my HP50g has decided not to compute integrals with > `infinity' boundaries. When trying, I receive something like `Excess > overflow' (it is in Spanish). Why? I looked carefully through the CAS > flags, but I have deactivated every option (rigorous mode, etc.) which > could lead to such problems. > > I really need it to compute such integrals, to verify my results. How > can I? It worked, some weeks ago... > > Thanks. > - -- > Merciadri Luca > Seehttp://www.student.montefiore.ulg.ac.be/~merciadri/ > - -- > > A closed mouth catches no flies. > -----BEGIN PGP SIGNATURE----- > Version: GnuPG v1.4.9 (GNU/Linux) > Comment: Processed by Mailcrypt 3.5.8 <http://mailcrypt.sourceforge.net/> > > iEYEARECAAYFAkwBW6kACgkQM0LLzLt8MhxRKgCZAfjudFPqP+LctyGFP26711u8 > 8n8An3VShwGBRqWJI35ay5Vr6Ga538cb > =jjHj > -----END PGP SIGNATURE----- Hi, Try resetting the CAS to the defaults: To return the CAS to its default state, press "right shift" followed by SYMB (CAT), ALPHA F3(C), DOWN cursor. With CASCFG selected, press ENTER. Or by typing CASCFG (double press "ALPHA" to lock ALPHA mode) on the command line followed by ENTER. Bart
From: Merciadri Luca on 1 Jun 2010 12:05 -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Bart <bart(a)barend.plus.com> writes: > On May 29, 7:23 pm, Merciadri Luca <Luca.Mercia...(a)student.ulg.ac.be> > wrote: >> -----BEGIN PGP SIGNED MESSAGE----- >> Hash: SHA1 >> >> Hi, >> >> For some weeks, my HP50g has decided not to compute integrals with >> `infinity' boundaries. When trying, I receive something like `Excess >> overflow' (it is in Spanish). Why? I looked carefully through the CAS >> flags, but I have deactivated every option (rigorous mode, etc.) which >> could lead to such problems. >> >> I really need it to compute such integrals, to verify my results. How >> can I? It worked, some weeks ago... >> >> Thanks. >> - -- >> Merciadri Luca >> Seehttp://www.student.montefiore.ulg.ac.be/~merciadri/ >> - -- >> >> A closed mouth catches no flies. >> -----BEGIN PGP SIGNATURE----- >> Version: GnuPG v1.4.9 (GNU/Linux) >> Comment: Processed by Mailcrypt 3.5.8 <http://mailcrypt.sourceforge.net/> >> >> iEYEARECAAYFAkwBW6kACgkQM0LLzLt8MhxRKgCZAfjudFPqP+LctyGFP26711u8 >> 8n8An3VShwGBRqWJI35ay5Vr6Ga538cb >> =jjHj >> -----END PGP SIGNATURE----- > > Hi, > > Try resetting the CAS to the defaults: > > To return the CAS to its default state, press "right shift" followed > by SYMB (CAT), ALPHA F3(C), > DOWN cursor. With CASCFG selected, press ENTER. > > Or by typing CASCFG (double press "ALPHA" to lock ALPHA mode) on the > command line followed by ENTER. Thanks. Without doing anything, everything returned to the normal, i.e. integrating 1/x^2 over [1, +\infty[ did not work yesterday, but worked today. But thanks for the tips, I'll keep them in mind. Do you have any idea of what could have caused the trouble? Thanks. - -- Merciadri Luca See http://www.student.montefiore.ulg.ac.be/~merciadri/ - -- To accomplish great things, we must not only act, but also dream; not only plan but also believe. (Anatole France) -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Processed by Mailcrypt 3.5.8 <http://mailcrypt.sourceforge.net/> iEYEARECAAYFAkwFL7IACgkQM0LLzLt8MhzHYQCfWnrN31Pyqu+ZSGxE6GRgWc3g ta4AnjxE6TYTr0FFfaDqQH9ifeWsuqr6 =FAZf -----END PGP SIGNATURE-----
From: John H Meyers on 1 Jun 2010 12:40 On 6/1/2010 11:05 AM, Merciadri Luca wrote: > integrating 1/x^2 over [1, +\infinity] did not work yesterday While calculating numerically, X^2 would overflow the standard floating-point exponent range within that argument range. See flags -20 and -21 Numeric integration is also somewhat inefficient for integrals whose range interval is infinite, particularly for a range of -\oo to +\oo for a function whose largest values are nearer to zero. At any rate, other than the weather, which rarely affects calculations, a stored value (e.g. for the variable in the formula, if by any chance different than the variable you are integrating over), a flag setting, or the problem entry itself (e.g. "1" vs. "1." or the lower limit perhaps being 0 rather than 1) would be more likely to vary between one day and another. [r->] [OFF]
From: Merciadri Luca on 1 Jun 2010 12:48
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 John H Meyers <jhmeyers(a)nomail.invalid> writes: > On 6/1/2010 11:05 AM, Merciadri Luca wrote: > >> integrating 1/x^2 over [1, +\infinity] did not work yesterday > > While calculating numerically, > X^2 would overflow the standard floating-point exponent range > within that argument range. > > See flags -20 and -21 > > Numeric integration is also somewhat inefficient > for integrals whose range interval is infinite, > particularly for a range of -\oo to +\oo > for a function whose largest values are nearer to zero. > > At any rate, other than the weather, which rarely affects calculations, > a stored value (e.g. for the variable in the formula, > if by any chance different than the variable you are integrating over), > a flag setting, or the problem entry itself (e.g. "1" vs. "1." > or the lower limit perhaps being 0 rather than 1) > would be more likely to vary between one day and another. > > [r->] [OFF] Thanks for this recommendation. That is, do you also mean that numeric integration is the only technique which is used by the HP50g? It would make me learn something! - -- Merciadri Luca See http://www.student.montefiore.ulg.ac.be/~merciadri/ - -- Man is still the most extraordinary computer of all. (John F. Kennedy) -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Processed by Mailcrypt 3.5.8 <http://mailcrypt.sourceforge.net/> iEYEARECAAYFAkwFOeQACgkQM0LLzLt8Mhwy4ACgm8b5b94f/GS17yu7Z1irwB1d GiQAoJQSw+LaQOT6qpJZPg/j3EqlE8ME =Kwoq -----END PGP SIGNATURE----- |