| T O P I C R E V I E W |
| Joshua Wilson |
Posted - 02/20/2018 : 06:19:45 AM
OriginPro 2015 Academic
Operating System: Win 7
I have some data from device simulations that I'd like a heat map of. The data is not regularly spaced in the y-direction and the first z-column is
1.99E+10
3.56E+10
7.03E+10
1.56E+11
4.00E+11
1.21E+12
4.42E+12
2.03E+13
1.22E+14
7.33E+14
3.38E+15
1.26E+16
4.01E+16
1.17E+17
3.47E+17
1.24E+18
7.29E+18
When I convert to a matrix using XYZ Gridding the first column is converted to:
1.9923E10
5.798E10
1.78555E11
5.13573E11
1.56319E12
4.77659E12
1.01348E13
2.39127E13
1.22018E14
-2.55965E13
1.86239E15
3.50662E15
5.9944E15
4.26972E16
7.61601E16
2.53708E17
7.2864E18
The data now contains negative values. This occurs regardless of which random gridding method I use. Am I doing something wrong? Or is it a bug?
I know a similar problem was discussed here: https://www.originlab.com/forum/topic.asp?TOPIC_ID=9859, but I need a heat map with a spectrum bar for the z-values and it doesn't seem like that can be done with the contour plots.
Thanks in advance for your help.
Joshua Wilson |
| 1 L A T E S T R E P L I E S (Newest First) |
| Drbobshepherd |
Posted - 02/20/2018 : 09:11:04 AM
Joshua,
Gridding routines typically make use of cubic polynomials to maintain smooth continuance in the interpolated curves and their first derivatives. If there are large gaps, the powerful cubic term can drive the interpolated results far from the data's neighborhood.
You could try using linear interpolation if you do not require smooth first-derivative curves.
Or, based on the large range of your data, I would recommend gridding in the log domain. This will guarantee no negative values.
DrBob |
|
|
