odesolve: found a number with a magnitude greater than 10^307

[Cornel]

Hi,

[Werner_E · Accepted solution]

Hmmm, lets see:

For example

Primes numeric uses IEEE numbers, so any calculation which yields a result of more than about 1.7*10^308 throws an error - Primes numerics can't handle numbers that large. Thats something you have to accept and live with.

[Cornel]

But Is normally is small/very small..so, its not good to put such value like 3A for Is...

[Werner_E]

---

@Cornel wrote:

But Is normally is small/very small..so, its not good to put such value like 3A for Is...

---

Sure! Chnging the value does not solve your initial problem, its just avoids the error thrown by the tool you use. It may mean that you either have to use a different tool or find a different approach to set up your problem which avoids numbers that large. You can't handle Dirac impulses in Prime either, don't you?

You may give it a try but I don't think that Primes new symbolic way to solve ODEs would be capable enough to solve your system, though. If you want to give it a try, your first attempt should be without units.

[Werner_E]

Additional comment:

Your solve block (using the Is=1nA) works OK up to about 8.8 microseconds.

Its now up to you to find out which expression would throws a result too large for Prime to handle. Maybe(!?) then you would be able to rewrite/reformulate the expression so it avoids values too large. Good luck!

[Cornel]

But there is a way to limit values which are calculated inside the odesolve block to not exceed 10^307?
Or the only way is as you said to try to find out which expression would throws a result too large for Prime to handle and then to rewrite/reformulate the expression so it avoids values too large? (this can be not so easy)

[Werner_E]

---

@Cornel wrote:

But there is a way to limit values which are calculated inside the odesolve block to not exceed 10^307?

---

Maybe - if you can find out which calculation throws the error it may be possible to rewrite it in a way so that the numbers calculated remain within the allowed limits.  This can turn into quite tedious detective work.
A simple push-button solution like you always keep wishing for, where the software takes care of everything, is unlikely to exist. But then I'd be happy to be proven wrong.

A very simple and most obvious example:

Written that way the function f throws an error for larger arguments:

Rewriting the function, this equivalent function has no problems

Of course its not always that easy to spot the root of the problem and cure it 😞