Levenberg - Marquardt chi^2 Error

[newvi]

In the Documentation (labview 6.1) on the Levenberg Marquardt.vi It shows that the chi squared term (which is minimized) is the sum of (y-f(x,ai)).

But should this be sum of (y-f(x,ai))^2.

The original form would allow large errors in the opposite direction to cancel. The squaring is need to prevent this problem and is implied by the term chi^2 and not just chi.

I tried fitting some real data with the Levenberg module and it gives that same answer as if the square was not taking place.

Can anyone with more knowledge of labview code confirm that the code is also in error and provide what should be a simple fix to this vi.

[altenbach]

You can verify from "Lev-Mar xx.vi" from inside the "nonliner Lev-Mar fit.vi" (here the help is correct) as well as from Levenberg-Marquardt.vi (here the help is incorrect as you noticed), the code correctly uses the square in both cases.

See for example the attached image.