Skip to main content
A
adar1-Visitor
1-Visitor
April 8, 2013
Question

Eigen value resbustitution error

  • April 8, 2013
  • 2 replies
  • 1110 views

There appears to be some bug in Mathcad 15. In the attached file I used the function getvals for the eigen value problem. I get the same answer by calculating the determinant of the matrix. However when I substitute the eigen values in the main equation, it leads to a very big number. It should give me zero or close to zero. However, when I solve the problem with fraction numbers as results it gives me the exact answer.

The problem is that the solution given in the first case should lead to a small number close to zero not a huge number with 10power 9. see attahed sheet.

2 replies

A
AlanStevens19-Tanzanite
19-Tanzanite
April 8, 2013

The float command affects all intermediate calculations, not just the final display. Avoid it as much as possible when using symbolic maths in Mathcad. Better as:

AvoidFloat.PNG

The limited precision of the numerical calculation in getvals leads to inaccuracies here.

Alan

W
Werner_E25-Diamond I
25-Diamond I
April 8, 2013

As Alan already wrote its not an error but rather a disadvantageous behaviour of the symbolic "float" modifier.

Your calculations seem to be very sensitive with respect to numerical accuracy and so you see this big discrepancy.

See below how the float precision affects the outcome from float,5 -> 10^18 down to float,250-->10^-227 (which is close enough to zero, I guess).

If the intend of using float was just to see the numbers, assign the symbolic eval (without float) to a variable sols, as Alan had done, and in a separate region type sols=. Don't do it in the same region after the symbolic eval or you will loose precision again.

floatprec.png

Terms & ConditionsAccessibility statement