The can and sample workspaces aren't filling the same Q range

[Steve Williams]

As Richard says below after a reduction the can and sample workspaces that are produced should have the same range in Q:

I am going mad again. I am changing the minimum radius in the gui to compare results.

Depending on the order I do things in I can get less bins at the minimum Q in the can that in the sample or (sample-minus-can) result, as per attached plot. (Was 1.28.11783 of Weds 18 May) “Show data” on the workspaces confirms there are 2 less points at the start of the can_tmp workspace

Is Mantid “caching” the can results from a previous round and not noticing I have changed the radius ???? – if so then what background is it actually subtracting off the first few points ?

Richard

[Steve Williams]

A set of test runs for this is 7174 7098 7097 7176 7096 7097 I'm looking for an effect in the geometry as suggested here:

Just tried going back to 2-12 in the user file L/WAV and TRANS/FIT (rather than 1.75 to 14 as before)

The TRANS/FIT box seems not to pick up what is in the user file, my gui “remembered” a 14 ang max from previous session?

And it still give different Q range for the “can”

PREVIOUS COMMENTS could be a red herring as 14/12 = 1.166, is also very close to 47/41 = 1.146 the different beam stop radii !

Have now SWAPPED the sample and can, the weaker scattering run 7176-add has the missing Q bins at small Q, so perhaps, As Steve suggested this morning, perhaps it is an “empty bin” causing truncation somewhere ??? or something odd in the geometry ??

Richard

[Steve Williams]

(In [12018]) Correct a bug in SANS display mask that was stopped viewing workspaces with few counts re #3078

[Steve Williams]

(In [12029]) Correct a bug in SANS display mask that was stopped viewing workspaces with few counts re #3078

[Nick Draper]

Accepted and assigned tickets moved at iteration 29 code freeze

[Steve Williams]

Perhaps this is not a problem since ticket #3105. We could check again what Richard Heenan thinks.

[Steve Williams]

There are still problems! Richard has replyed with another test case. Place these runs into the SANS GUI

          Tr     db
Sam 7187  7185  7182

Can 7179   7183  7182

In that case the workspaces:

7178rear_1D_1.75_14.0

7178rear_1D_1.75_14.0_sam_tmp

7178rear_1D_1.75_14.0_can_tmp

must all have common bins.

[Steve Williams]

In [13107]:

SANS plots don't need the end zeros removed with the new Q1D, center finder remains to do re #3078

[Steve Williams]

In [13108]:

Revert a file I shouldn't have sent re #3078

[Steve Williams]

In [13574]:

The test case should work now fixes #3078

[Stuart Campbell]

This ticket has been transferred to github issue ​3925