Bug in coefficients output of General LS Linear Fit LabView 8.5.1?

[Fil_]

Hi

I compiled my application from LabView 7.1.1 to LabView 8.5.1. This application use the "General LS Linear Fit.vi". I noticed that there differences in the calculation of the coefficients.

There are attached two files, one comiled in LV7.1.1 and the other with LV 8.5.1, with two different results. Please make sure you open this files with the correct version.

If I change in "General LS Linear Fit.vi" the link to the lvanlys.dll of the LV7.1.1 version, the calculations are correct.

Any help would be appreciated, thanks.

Fil

[Ray.R]

Your attachments are missing.

R

[Fil_]

Sorry

Thank you for the notice.

[Andre_Saller]

The code of the lvanlys.dll changed in the Version 8.5 because of additional functionality and bugfixes. One of these bugfixes concerned the linear fit functions. Becuase of that, you see different results depending on which version of LabVIEW you are using. I recommend using the later Version, as the algorithm used in the new version is more robust and exact.

André

[Fil_]

Hi André

I think the results in the 8.5 Version are completely wrong. If the result would deviate a few percent one could say that the new version could be more exact. But not like that.

Fil

[dan_u]

Fil,

in your data you have 3 columns with constant values (1, 0.5, 0). These coefficients can't be estimated correctly, because the fit quality is exactly the same if you change 2 coefficients in opposite direction. If these parameters are really constant you'll need to combine them into one. If these parameters can vary, you need better data for the fit to be reliable.

Daniel