next Fit of a single histogram
up How to create an initialfile
previous Spectrum commands
index Index

Fit of an asymmetry spectrum

  In this case the SPECTRUM commands section looks like the one here. The first line is 

SPECTRUM 1,    ,    ,    ,    , ,,,,
!    alpha,beta,rpha,rasy,l/t0, 5xindirect
with

alpha ratio of detection efficiency of the forward and backward telescope ( alpha = NB,0 / NF,0 ),  
beta ratio of asymmetries in forward and backward telescopes ( beta = AB,0 / AF,0 ).  
Comment: The exact equation for the asymmetry and the meanings of alpha and beta you can look up in Phys. Rev. B 35 (1987) 6597.
rpha, rasy Used if histogrammes 1,2 and 3,4 are fitted together. You then need two spectrum statements: one for 1 and 2, the other one for 3 and 4. In the latter one you have to define an other alpha (beta) and the ratios.    
l/t0 Either a lifetime (e.g. µ-SR) or t0-channel if phase is fixed.  

                    
                    
In the line 

sys$data:[musr.dat_cecunisn]012182.tri
you call for the data (here 012182.tri) that shall be fitted. The other specifications determine where the file is stored in the computer network. You have to write the full runnumber plus extension. The spectra files must have the format *.tri, *.dat or *.gps. If you use a gps-file (which you can get e.g. from RESERL) you have to say

GPS=012182.gps    (GPS= has to be written with capital letters!).  

Comment: MSRFIT has severe problems with files containing more than 12MEv. Then the gps-format is the only data format that will work.
The command 
SP 2
in the following line specifies that the asymmetry B+F/B-F should be formed. Further available commands are SP 3 which simply forms the ratio B/F and SP 4 which is the rotating frame analysis*  . This is only of use for transverse fields B > 1kG. 
H 1, 2, 0, 0
specifies histograms for which the asymmetry ratio has to be formed. Only two histograms are allowed (an alternative would be H 0, 0, 3, 4 for left/right histograms). 
B 70, 240, 70, 240, 0, 0, 0, 0
specifies the channels for determining background counts. You can find out these channels e.g. by using the program SCAN. For TRIUMF data the background channels are usually below the t=0 channel, for RAL data use late channels.
Comment: When you want to fit left/right histograms you have to fill the slots the following way: B 0, 0, 0, 0, 70, 240, 70, 240.
Notice: After this background definition two more SPECTRUM commands are possible 
M(ARKERS) T0, FGCH, LGCH
and (or) 
RAW_BINS  Tbin T0 ns A =   0.00000,   0.00000,  B =   0.00000,   0.00000,  C = 0.00000,   0.00000,  D =   0.00000,   0.00000
In the line MARKERS   you can determine time-zero, first and last good channels if different from the datafile.
              
The line 
T 0, 5., 50
determines start   and stop time (µs)   as well as the binwidth (ns)   for fit and plot. At TRIUMF the binwidth is 2.5ns, at RAL 16ns.
Notice: You can not plot a range which is not fitted and you can not plot with more points/unit time than used in the fit.
You can change   the number of points/unit time for any part of the spectrum by defining a second SPECTRUM command (see Using different binwidth for one spectrum).