01-15-2021, 07:56 PM

(01-02-2021 04:09 PM)Nihotte(lma) Wrote: [ -> ]Code:

1) at 07.09 PM 2021/01/02

2) at 10.31 PM 2021/01/02

2) at 08.06 PM 2021/01/06

I push this post to come back on several details that I have taken the time to explore a bit since.

First point :

It seems to me that ultimately the best way to get an error 8 by wanting to calculate an integral is still to run an f SOLVE on a range of a function where there is no 0. This time, I therefore suppose that the anomaly was about 12 inches (30 centimeters) above the calculator. I think that it is indeed preferable to launch an integral calculation by f INTEGR because I never reproduced the error thus!

Second point :

I finally managed to calculate the integral brilliantly using the HP35s. It is indeed simple as long as you use the calculator correctly. The technique consists of not simply reproducing a program in the spirit of the HP15C.

I mean: the HP15C don't need an intermediate storage register to calculate a function on which we can then evaluate an integral. On the other hand, this technique is perhaps less easy with the help of the HP35s which asks to know the derivative variable.

In this case, avoid reproducing my way of proceeding and avoid using STO A followed by one or more RCL A. Only RCL A is required, in fact!

This is all the more correct, since it is easy to enter the equation via EQU and to launch the calculation of the integral from the equation present on the screen. You just need to provide the name of the derivation variable (i.e. A) when the HP35s asks for it and everything is fine!

On the other hand, it seems that the computation of integrals remains more efficient by taking into account the limit of the functions evaluated. It is therefore more judicious, according to my observations, to evaluate 0 -> 90, then 180 -> 90 (or, as Albert Chan pointed out to me 0 -> -90) rather than 0 -> 180 directly because the calculation runs up against the limits around 90.

Last point :

Eventually the HP35s gave me a clean result using RAD directly with π based values instead of DEG and starting values in degrees.

Thanks again to Valentin Albillo who offered us this challenge which turned out to be rich in lessons and questions for all those who took part!