SANS: Problem with horizontal integral in the ISIS SANS diagnose tab

[Steve Williams]

Enter SANS2D2500 as the run number and click "Display Horizontal". The plot shows the maximum numbers of counts towards the edges of the plot (x=0) instead of near the centre(x~100)

[Steve Williams]

Generally confirm in instrument view that the spectra used in the rear and front-detector sums include all and only those actually in the detector

[Steve Williams]

Ticket #2435 should help things

[Steve Williams]

(In #2435) There was also a change to the instrument configuration in cycle 4/9 (they used their working detector as the front instead of the rear)

[Nick Draper]

Is this already resolved?

[Nick Draper]

Bulk move of tickets at the end of iteration 28

[Nick Draper]

Accepted and assigned tickets moved at iteration 29 code freeze

[Nick Draper]

Bulk move of tickets to iteration 31 at the iteration 30 code freeze

[Anders Markvardsen]

There is no 'display horizontal' in the ISIS SANS Diagnostics tab

[Nick Draper]

The name on the button was changed - it means Horizontal Integral

[Nick Draper]

Moved at end of release 2.1

[Nick Draper]

Moved at the end of release 2.2

[Gesner Passos]

Reproduce the error

[Gesner Passos]

Replying to Steve Williams:

Enter SANS2D2500 as the run number and click "Display Horizontal". The plot shows the maximum numbers of counts towards the edges of the plot (x=0) instead of near the centre(x~100)

The error is related to the definition of what are the Spectrums that belong to a single DetectorBank. For example, the DetectorBank='front' is a square Bank whose dectors IDs goes to 2,000,000 to 2,191,191. But, what you want is to get the SpectraIDs that were acquired by those detectorsIDs.

The current implementation inside SANSDiagnostics.cpp part from the following assumption:

"The minimum number (2,000,000) will be related to the minimum SpectrumID, analogous to the maximum"""

This assumption seems not to be valid for this file. And usually, in the moment of Croping the workspace to get only the data related to one DetectorBank it is sending without one row, but the SumRowColumn knows only how to execute integration on a square dataset. So, you get the wrong image.

That is why the SumRowColumn issues this warning:

The input workspace has 36482 spectra.This is not 128*128 or 192*192 - did you make a mistake?

I attached a python script that shows what is happening. Just execute the python code

[Gesner Passos]

Solve getSpectraList wrong assumption

Change the calculation of the spectra ids limits from SANSDiagnosis that used to return a wrong range due to a false assumption. This patch solves the issue, but advise is given in order to improve the performance. re #2470

Changeset: e47d661bb1f18889046242cb8c8a65e0373ae922

[Karl Palmen]

Unable to find ISIS SANS diagnose tab, but do find a SANS ISIS diagnostic tab. It has no "Display" horizontal. Other features are disabled. I don't seem to know enough about this GUI to do the test.

[Nick Draper]

This specific bug fixed, but two more created for the next iteration #6042 and #6043 for the next release

[Gesner Passos]

Solve getSpectraList wrong assumption

Change the calculation of the spectra ids limits from SANSDiagnosis that used to return a wrong range due to a false assumption. This patch solves the issue, but advise is given in order to improve the performance. re #2470

Changeset: e47d661bb1f18889046242cb8c8a65e0373ae922

[Stuart Campbell]

This ticket has been transferred to github issue ​3317