Optionally Apply Background Correction

[PeterParker]

sans_feature This issue was originally TRAC 9350

The ISIS SANS interface needs to be expanded to allow background corrections using so-called "dark runs".

My understanding is that there are two kinds of dark run; "natural background" (i.e. source off) and "dark current" (i.e. source on, but shutter closed).

It is not clear whether or not this would be best implemented as a User File command, as checkboxes in the GUI, or both.

Implementing this seems trivial enough, though we'd need to be careful about scaling a 10-hour background run before applying it to a 1-hour data run, run example. It may be useful to see how this has been done at the SNS.

We are to account for periods where they have been used.

---

Keywords: Bilby

[PeterParker]

@PeterParker (2014-07-31T12:53:26): More from Richard:

• We're to normalise to the number of good frames x width of a time bin, assuming that the bins are the same width for the dark current run. We need to put in a check for this, of course.
• We need to make sure the data is in floats (not in integers) when we do the subtraction, since the correction will be very small.
• The GUI is cluttered enough, and this will be a niche enough feature so that adding a field to the interface in not warranted in this case. Just have a User File command.

Richard is going to send an email to the SANS team to ask about whether or not the subtraction should be on a time-channel-by-time-channel basis, or whether it should be integrated. More to follow.

[AntonPiccardoSelg]

From Richard:

In our case “background” could be a beam on “Cd run”, “fast shutter closed” or “main shutter closed”. These could be normalised to either microamphours or (if you think cosmic & electronic noise is dominant) to time. For electronic noise and/or “cosmic ray background” we might do a long “beam off” run in a shutdown.

Thus I think we need to specify: run number (or file name); normalise to data collection time (=number of good frames x collection time in each frame) or to good uamps; whether for monitors or for which particular detector bank.

Since we already use “BACK” for tof monitor electronic noise, http://www.mantidproject.org/SANS_User_File_Commands#BACK we could use e,g.

BACK/REAR/UAMP/TOF/RUN=12345 BACK/FRONT/UAMP/TOF/RUN=12346 BACK/MON/RUN=23456/TIME/MEAN

/TOF means apply subtraction in tof /MEAN is to average the RUN over the whole spectrum and subtract the same count rate from all tof bins in each spectrum, this is appropriate if electronic noise or cosmic background dominate.

From Richard:

Either /UAMP or /TIME must be present for the normalisation, else flag an error.

Either /MEAN or /TOF must be present for the subtraction method, else flag an error. ( /TOF to flag error if range of time channels does not match actual data. Possibly flag warning if rebin of time channels is needed)

/REAR, /FRONT or /MON must be present for SANS2d /MAIN, /HAB or /MON must be present for LOQ

From Stephen:

We can’t use /TIME as the parser already looks for /TIMES. J

Let’s use /DURATION instead of /TIME.

for the online help: all four cases (/UAMP/TOF, /DURATION/MEAN, /UAMP/MEAN, and /DURATION/TOF) are potentially valid, but we would only expect the first two to be used routinely…