Using 4m TRANS to reduce 12m SANS

[Anders Markvardsen]

Answer to Steve's question below

I had some Users last week on SANS2D who needed to run at 12m, so the count rates were very low. In addition, the divergence of the scattered neutrons means that transmissions measured on the beamstop M4 at 12m are unreliable, so current practice is to do separate M3 transmissions at 12m. This eats up valuable measuring time.

I wondered whether we could quickly blast the transmissions out at 4m using M4 and then use those transmissions to reduce the 12m data.

In principle this should be ok, but I think something funny is going on in Mantid. See attached.

When I calculate a 4m M4 transmission when the sample scattering run is at 12m I get a slightly different answer to when I calculate the SAME 4m M4 transmission when the sample scattering run is also 4m.

[Gesner Passos]

Summarizing the issue. If you select any of the masks presented by Stephen and apply the reduction for the run (Sample = 18617; Trans = 18617; Direct = 18618). After, you can just change the Sample to 18783. The transmission workspace changes its value.

It should not, because, the transmission depends only on Trans and Direct, and not on the Sample.

It was found that the reason for this was the assumption that the Z position of the Detector is always the same for the Sample, Trans and Direct (The value is the Rear_Det_Z inside the file).

This caused the transmission calculation to depend on the position of the Rear Detector on the Sample experiment. So, the sample was first at 12m but the Trans was at 4m. So, the monitor position is set differently from the two calculations which caused the difference on the calculated transmission.

As it is implemented right now, there are some values:

• Front_Det_Z
• Front_Det_X
• Front_Det_Rot
• Rear_Det_Z
• Rear_Det_X

which are read just once for the Sample, and it is kept inside the SANS2D(ISISInstrument) as its attribute. When the move_components (method) is called, it will use those values to define where to move the components.

In order to solve this issue, it is necessary to allow these values to be read from the workspace itself, before moving the components.

[Gesner Passos]

Allow the move_components to read the values from the workspace info

Before moving the components, the SANS2D(ISISInstrument) will read some values from the workspace info. This will ensure that the positions are moved according to the real position when the data was acquired.

re #7314

Changeset: 1e13f2747a3bb415b69370741393eb1f0ee23a7d

[Gesner Passos]

After the given commit, the position now is done based on the current workspace information. This solved the issue of calculating the transmission workspace as can be seen at the figure:

Tester: (easier from someone here at ISIS)

In order to reproduce this picture do:

• Open ISIS Sans interface
• Select SANS2D instrument
• Set the searching directories to cycle_13_1
• Open the Mask at: ndxsans2d/user/Masks/MASKSANS2D_131M_12m_8mm_Adams_using_4m_TRANS.txt
• Fill the Scattering Sample=18617; Transmission Sample=18617; Transmission Direct=18618
• Press Load/Reduce1D
• Rename Workspace::18617_trans_sample_1.75_12.5_unfitted to sample_4m_transmission
• Change the Scattering Sample=18783
• Press Reduce1D
• Rename Workspace::18617_trans_sample_1.75_12.5_unfitted to sample_12m_transmission
• Select both workspaces and plot with errors.

[Anders Markvardsen]

Steve writes:

Excellent!

I have just tested this fix and it does the job! Thank you.

[Stuart Campbell]

This ticket has been transferred to github issue ​8160