MuSR Spectrum Types

Now that we have observed a raw MuSR Time Histogram plot in the Java GUI it is time to start implementing other types of TD-MuSR Spectra we will want the MuSR_Server to be able to generate and download as Tables of Columns for plotting by the GUI. The GUI will include a Spectrum Type menu and dialogue methods where one specifies packing and other custom options for Spectrum generation by the Server. (This will require a different dialogue for each Spectrum type.) "Overrides" of HistHdr parameters is handled earlier. Since the Column names must be short enough to use in algebraic expressions (eventually) I am partial to using an extremely compact form for the menu items, echoing the Column names that will result; a Help button could provide a key to more verbose descriptions until people catch on - after that they will probably prefer the short form anyway. Anyway, I need all you µSR folks to look carefully at the list below and think about both the specifications and the nomenclature. Is it clear? Is it unique? Is it enough?

Spectrum Type Menu Item name
(e.g.)		 "label"
(e.g.)		 [units]
(e.g.)		 Ref.
Time-Integral mode (x, y1, y2, . . . ) TI TI "Sclr 1" [counts per bin] N
Raw Histogram Hi H1 "Hist 1" [counts per bin] 1
t<0 Background-Subtracted Raw Histogram Hib H1b "Hist 1 - B" [counts per bin] 1,b
Single-Histogram Asymmetry Ai A1 "Asymmetry" [Hist 1] 1,a
Background-Subtracted Single-Histogram Asymmetry Aib A1b "Asymmetry" [Hist 1 - B] 1,a,b
Two-Histogram Background-Subtracted Asymmetry Aijb A12b "Asymmetry" [(H1-H2)/(H1+H2)] 2,A,b
All-Histogram Background-Subtracted Asymmetry SAb SAb "Asymmetry" [(F-B)/(F+B)] N,S,A,b
All-Histogram Complex Asymmetry [RRF optional] CA CA "{Real/Imag.} Asymmetry" [RRF=___MHz] N,C,A[,R]
All-Histogram Background-Subtracted Complex Asymmetry [RRF optional] CAb CAb "{Real/Imag.} Asymmetry" [RRF=___MHz] N,C,A,b[,R]

References:

1:
Requires parameter input for only one Histogram.

2:
" " " " two Histograms.

N:
" " " " all Histograms!

b:
Measure time-indep. background from t<0 bins and subtract from N(t) first.

a:
Fit No and B numerically, subtract B, divide out No exp(-t/µ ).

A:
Subtract t<0 backgrounds, then form A = (F-B)/(F+B). Unknown , .

S:
F = i ( Hi b Ci ), B = i ( Hi b [1-Ci ] ) where Ci = "Forward" direction cosine for the ith Hist. (Errors from Poisson statistics after the sums.) Untested method. Conventionally we designate some Histograms "F" and others "B" rather than double-counting as indicated here. May need revision.

C:
Re A = i ( Ai cos i ), Im A = i ( Ai sin i ), where i = Phase Angle for the ith Hist. (Errors from Poisson statistics after the sums.) [With "b" the t<0 backgrounds are subtracted before forming 1-Hist Asymm's.] Actually each Histogram will also need an "on/off" switch to tell whether it should be included or not, since i does not provide this. Untested at TRIUMF, but something like this has been used at ISIS.

R:
Rotating Reference Frame transformation (and any post-packing) applied after formation of CA[b].

The inclusion of multi-Histogram versions allows elegant solutions to the RRF alignment problem and also implements ISIS- or BOOM-type runs with many Histograms; but there will need to be some very clever interfaces for gathering the required info (like angles and so on) for every single Histogram; here is where the "like last time" button will be absolutely crucial! (Also a "remember this for later" button!)

Suggestions/critiques/comments/requests, please! Remember, what we do now will be what we are locked into later, maybe for all time....

Jess H. Brewer
Last modified: Sat 06 Nov 1999