ACCEPTED SOLUTION
Accepted Solutions
Valery, what did you do to get Mathcad throwing that error?
My best guess is, that there is some kind of redundance, that the five equations are not independent.
The error message thrown was new to me in conjunction with odesolve. I saw the same message only in the form "....but has the form: unitless". So I tried to reproduce that very error with a much simpler system of equations and this is what I came up with
So maybe its a good idea to recheck the equations for consistency.
Its always the last vector element which is flagged red - it does not mean, that there is something wrong with that specific function.
You told us you had a mistake in one of your equations in your Prime sheet. When you corrected that, was the error message the same? The first message Prime throws, the one about not being able to create a table of derivatives, could correspond to an dependent system, I guess. The second, unknown error about an expression being too large lead me to think that it could have to do with the complexity of your system and an unknown intern Prime limit.
BTW, you can insert the infinity sign as index in a variable name in Mathcad 15, but its a little bit tricky:
1) insert the infinity sign anywhere on the sheet, select it and copy it to the clipboard (Ctrl-C)
2) Begin typing your definition, e.g. "u."
3) Press Ctrl-Shift-K and watch the cursor getting red. This is a special tricky mode which allows you to insert characters in a variable or function name, which are normally not allowed.
4) Now insert the infinity symbol from the clipboard (Ctrl-V)
5) Press Ctrl-Shift-K again to leave the special mode and keep on typing ":3"
Good Example, Alan.
In Prime2 my example gives an error about Prime needing more equations (so Prime's Odesolver is a bit cleverer than the one in MC15 according to more suitable error messages). Your example throws the same error which is seen the first time in Tim's file:
In contrary to Tim's solve block, recalculating the sheet does not change the error (In Tim's file we get an "unknown error - expression too large" then.
So it seems the problem with Tim's system is not due to an error in the ODE system but its simply that Mathcad and Prime are not capable enough to solve it. Is there a workaround? While I could create an example in MC15 where rearranging the terms made it work, I was not able to do so with your example. At least not with simple rearrangements. Substitution derivatives and rearrangement can do the job, but would be difficult with the 5 eq system. Next pic shows how sensible Mathcad is:
Does this prove, that I can work with derivatives on the right hand side?
It is possible to have derivatives on the RHS (see Alans first example), but Mathcad seems not to like them, as Alan wrote, and ever so often will refuse to do its job. The exact conditions when MC will accepts that kind of equations and when it fails are unclear to me.
Find attached a file which follows Alans idea of solving your equations for the derivatives first. Of course I haven't done it by hand but let MC do the work.
I included Alans example in front of the file, which failed with the same error as yours and the method worked fine for it.
When I try to apply it to your system, Mathcad crashed on my machine. It could be because of lack of memory, I could imagine. Also the solving for the derivatives is very slow, I haven't timed it but I guess it took about an hour on my machine (which is rather slow and has only 1 G of mem, so you may have more luck).
I also ran into a strange error of Mathcad not being willing to make substitutions with a function name consisting of a Greel character - fortunately there is a workaround (other than renaming the function):
