symbolic constants in CAS

[jgreenb2]

I'm a self-confessed CAS newbie -- I never used it on my 48 or 50. But now I want to explore it a bit and I'm confused.

If I want to find the Laplace transform of exp(x) I execute:

laplace(exp(x),x,t) and correctly get 1/(t-1) as a result.

Now if I want to symbolically evaluate exp(a*x) for an undetermined constant 'a' I should get 1/(t-a) as the answer. But if I use 'A' as the constant, it's evaluated and I get 1/(t-3) as the answer (assuming 3 is stored in A).

If I try using 'a' as the constant:

laplace(exp(a*x),x,t) I get an error because 'a' is undefined.

How do I enter laplace(exp(a*x),x,t) and get back 1/(t-a) ???

[Han]

Have you tried to use this command using the CAS view? Press the [CAS] button (it's black, toward the top right of the keyboard).

[jgreenb2]

All the examples above were executed in CAS.

[jgreenb2]

A few learnings but no solution. First, CAS is based is giac. This must be common knowledge but not to me! Which means, of course, that Parisse is actually Bernard Parisse -- Hello! So that explains how to find the "missing" documentation on CAS (like what are the legitimate arguments to assume(), etc.).

So I downloaded a version of XCAS and tried to see what to do. There, it works exactly as I want:

laplace(exp(a*x),x,t) returns -((1/(a-t))) with no complaints about the symbol 'a'.

That gives me hope it should be possible on the prime, but again, how???

[Han]

On the HP Prime, there is a major difference between 'A' and 'a' besides one being uppercase and the other being lowercase. The variable 'A' is always a real-valued variable and can never be undefined. On the other hand, 'a' may be treated as an undefined variable. What you likely experienced was that using 'A' in your calculations will actually convert 'A' to the value stored in this variable.

Try using the laplace command but with undefined variables such as 'a'. If for some reason 'a' seems to not work, then it may be the case that you had previously stored a value into 'a'. To make matters worse, the 'a' as a variable created in Home view is different from the 'a' variable created in CAS view. To ensure that all versions of the variable 'a' are "deleted" you can use the memory browser to delete the Home view version, and purge(a) to delete the CAS version.

[Mark Hardman]

(12-22-2013 07:42 PM)jgreenb2 Wrote:  That gives me hope it should be possible on the prime, but again, how???

This looks like a definite bug in a very fundamental transform.

If I try to help the CAS along by rewriting the expression as an indefinite integral, I get the correct result:



But, when the definite integral is supplied:



You get the same error as from the Laplace command:



Mark Hardman

[parisse]

Just checked, it works with giac/Xcas. On the Prime, there is an additional limitation for checking sign of expressions (if there are 2 or more variables in an expression, here it's for the sign of a-t) because error handling is more primitive, and this is the reason why you get this message. I hope I have found a way to get the sign also on the Prime.

[Rich]

Hi,

Has anyone found a resolution to the issue in this thread? I tried searching on this forum and a few others, but have not found a solution.

I am taking Modeling and Control this semester at college, I would like to be able to use Laplace transforms from my HP Prime.

Entering this: laplace(e^(-a*t),t,s)
Gives me this: "Undef/Unsigned Inf encountered in limit"

Entering this: assume(s>0)
then checking if the CAS s variable exist in var menu, it does

Entering this: laplace(e^(-a*t),t,s)
Gives me this: "Undef/Unsigned Inf encountered in limit"
I then go back to check if the s variable exist and it no longer does. I then did a factory reset in the FLS utility menu and the same issue happens.

Any help would be great. Thank you

[Unlockcell]

what is the correct process for laplace transform? I have the same problem as previously mentioned friend! "

thank you,

Unlockcell