RE: HP48-HP50G : Error functions ERF(z), FRESNEL integrals, PROBIT, etc.

[Gil]

New version 16

The subroutine with the calculation of FRESNEL Integrals for very large real numbers
('fresnelFASTxLARGE') contained severe errors that were duly fixed up in this version.

My apologises for the inconvenience.

Now the Fresnel results given by the HP50G are very similar with the ones computed by Wolfram.

Besides, in this new version, you can always compute both the normalized and Not normalized Fresnel Integrais F_C & F_S
(just adding, when calculating the Non normalized Fresnel Integrals, the letter N [N for Non normalized] in stack level 2, before the argument x or z always in stack level 1).

Regarding always the Fresnel integrals:
when argument x is a real number : any value from -infinity to +infinity;
when argument z is a complex number a+ib, b≠0: abs(a) & abs(b) less or equal to 5