Skip to main content
R
Rik19823-Newcomer
3-Newcomer
October 14, 2024
Solved

Solve block: bad solve body

  • October 14, 2024
  • 1 reply
  • 1751 views

My students and I use Mathcad to solve problems of dynamics. My students this year (prime 10.0.0.0) come sometimes across an unknown error when using a solve block. The error mentions "bad solve block"; refer also to the screenshot. I have never come across it before and I have no clue what the reason could be for this error.

 

After retyping the equtions and the "find" command, the solve block executes correctly. But I don't see a difference between the original solve block and the one after retyping it. So it stays a mistery why one time it executes and the other time it doesn't.

 

Schermafbeelding 2024-10-14 223314.png

Best answer by Werner_E

Are you sure that the equations are correct?

I notice that you defined t.0 and a.2 but they are never used.

 

Solving the system symbolically shows that you run into troubles (division by zero) if a.0 = a.1

Werner_E_0-1728942756603.png

Here I changed a.1 from 1.5 to 2.5 m/s^2 and the symbolics seem to provide a valid solution:

Werner_E_1-1728943130794.png

Unfortunately the numeric solve block still fails 😞

 

EDIT:

Ahhh! Seems you found a bug in Prime 10, Feel free to report it to official PTC support.

The problem is the inline evaluation in the definition of the guess value for v.1.

Delete it (and evaluate in a separate region if necessary) and the solve block works OK - at least if a.0 and a.1 are different!

Werner_E_2-1728943603424.png

As we have seen from the symbolic solution, there are two sets of solutions stemming from the underlying quadratic equations. In the solve block above we got the (presumably undesired) solution which includes negative values.

To get the other set of solutions you may either provide better guess values (even t.1:=9s will do the job with the values given) or you add additional constraints, for example t.1>0

Werner_E_3-1728943744258.png

 

1 reply

W
Werner_E25-Diamond IAnswer
25-Diamond I
October 14, 2024

Are you sure that the equations are correct?

I notice that you defined t.0 and a.2 but they are never used.

 

Solving the system symbolically shows that you run into troubles (division by zero) if a.0 = a.1

Werner_E_0-1728942756603.png

Here I changed a.1 from 1.5 to 2.5 m/s^2 and the symbolics seem to provide a valid solution:

Werner_E_1-1728943130794.png

Unfortunately the numeric solve block still fails 😞

 

EDIT:

Ahhh! Seems you found a bug in Prime 10, Feel free to report it to official PTC support.

The problem is the inline evaluation in the definition of the guess value for v.1.

Delete it (and evaluate in a separate region if necessary) and the solve block works OK - at least if a.0 and a.1 are different!

Werner_E_2-1728943603424.png

As we have seen from the symbolic solution, there are two sets of solutions stemming from the underlying quadratic equations. In the solve block above we got the (presumably undesired) solution which includes negative values.

To get the other set of solutions you may either provide better guess values (even t.1:=9s will do the job with the values given) or you add additional constraints, for example t.1>0

Werner_E_3-1728943744258.png

 

    R
    Rik19823-NewcomerAuthor
    3-Newcomer
    October 15, 2024

    Hi Werner

     

    Thanks a lot for your comments!

    • you are right, the value for a.1 should be -2 m/s² (mistake in the file of student); but that still doesn't solve the bad solve body
    • indeed deleting the inline evaluation of v.1 makes the difference and the solve block executes
    • interesting is that after executing the solve block and then adding the inline evaluation of v.1 is no longer a problem for Mathcad. I understand that Mathcad has some "rules" of how to enter values and commands, but this is then something I wouldn't expect. Moreover, teaching/encouraging students to use Mathcad (oh yes, they are reluctant!) is more easy if you can "explain everything"... 🙂

     

    But many thanks! I can tell my students not to do the inline evaluation in a Solve block.

     

    For your information/interest, I include my copy of the complete dynamics problem.

     

    Best regards

    Rik

    W
    Werner_E25-Diamond I
    25-Diamond I
    October 15, 2024

    I tried to track down the error using a simpler example, but I failed in reproducing the error. Inline evaluation of a guess variable worked fine with and without units and also by defining the variable in a non-default unit first.
    Not sure what exactly triggered that error in your sheet.

    When I again loaded your sheet now to correct the value of a.1 I deleted the inline eval of v.1 and the solve block worked OK. Just out of curiosity I readded the very same inline evaluation and expected the solve block to fail again .... but it didn't!!???

    Werner_E_2-1728992298820.png

    So the reason for this bug remains a mystery and cannot be explained (at least not by me). But it sure should be fixed by PTC.

     

    Inline evaluations were the cause of some unexpected problems already back in the times of real Mathcad (Mathcad up to version 15), Most of the time calculating a value and in the same region also displaying the resukt would work perfectly fine but sometimes it failed with a strange unexplainable error. So I got used to avoid inline evaluations altogether (although I never was 100% consequent in doing so).

    That said, I would not use

    Werner_E_0-1728991799091.png

    even though it works pretty fine, but rather use something like

    Werner_E_1-1728991850881.png

    just to be on the safe side.

     

    Terms & ConditionsAccessibility statement