ConvertToMDMinMax (local/global) on WISH

[Owen Arnold]

Not sure if this is best done by Alex or Andrei.

These issues have been raised by users of WISH, they may be best done as separate tickets.

1) ConvertToMDHelperMinMaxGlobal does not produce the expected results for wish. These are the limits according to Pascal:

For q_lab, with a full instrument (10 banks, for one side, q_x would be halved), we should get :
q_x : -8 to +8
q_y -1.5 to +1.5
q_z : 0 to 16

Running

#Load 10 panel wish file
limits = ConvertToMDMinMaxGlobal(WISH00028328, QDimensions='Q3D', dEAnalysisMode='Elastic')
print limits 

gives (array([-12000., -12000., -12000.]), array([ 12000., 12000., 12000.]))

2) ConvertToMinMaxLocal has properties that should not affect the limits calculated. For example, Lorentz Correction, Ignore Zeros

3) Documentation does not effectively describe the difference between ConvertToMDHelperMinMaxGlobal and ConvertToMDHelperMinMaxLocal. For example why is one local and one global?

[Owen Arnold]

I'm ccing Pascal as he may have more questions and would probably like to be updated with the progress.

[Andrei Savici]

Quick comments:

• For 3: the global routine allows one to move detectors up to twotheta=180 degrees, and to rotate sample in any way. The first part is to be able to combine runs on HYSPEC (moving detectors using motors) or TOPAZ (move detectors by hand in different cycles). The local calculates limits for Q in whatever detector/goniometer configuration you have in the matrix workspace. So if you rotate the sample, the limits will change. It is then hard to add two MD workspaces at different orientations.
• For 1: in Qsample, all limits are the same, since you can in principle rotate the sample in any way you want. The big number comes from (my guess) a bin with low TOF. The routine does not know about the flux, so even if you have just background there, a low TOF will give a huge momentum. I suggest that the workspace be converted to wavelength or momentum, and rebinned to reasonable limits

[Owen Arnold]

Moved owen's tickets to 3.5

[Stuart Campbell]

This ticket has been transferred to github issue ​10434