EFLECH

Site map
Up EFLECH EFLECH variables INDEX
INDEX variables Index example Download
NTHREADS
Thread count as used in parallel, should be equal or less the number of CPU's in your PC. The parallel efficiency of EFLECH is quite good. You may put an assignment to NTHREADS into the eflech.cfg file and it will be valid for all EFLECH runs.
VAL[i,j]
Measuring file(s) (j=1,...) for i-th intervall. Normally, this entry is generated by TEIL automatically.
WMIN[i],WMAX[i]
Angular range for measuring values of the i-th intervall. Normally, this entry is generated by TEIL automatically.
WMIN2[i],WMAX2[i]
Angular range for peak search of the i-th intervall. Normally, this entry is generated by TEIL automatically.
OUTPUT[i]
*.par-file, peak list of i-th interval in internal form. If this file exists at starting time, is will be used as input for starting parameters. Normally, this entry is generated by TEIL automatically.
STARTPARAM[i]
This *.par-file will be used for starting parameters, OUTPUT[i] will not be used.
PROTOKOLL=Y
Switches on the full iteration protokoll.
TEST
Concatenation of all possible changes, which will performed automatically to reach the output model. Default is TEST=ND+-234U, which switches all changes to possible. The letters mean:
N
Search for new peaks.
D
Search for possible deletion of peaks.
+
Search for possible improvement of peaks RP=2->3->4
-
Search for possible simplification of peaks RP=4->3->2
2
Peaks may change to RP=2 (caused by N or -)
3
Peaks may change to RP=3 (caused by N or + or -)
4
Peaks may change to RP=4 (caused by N or +)
U
Search for simplification of background
RU
(Starting) degree of background polynomial. Otherwise, EFLECH selects a sufficient degree.
EPSILON
Maximum accuracy. Means relative error of any measured angular point. Defaults are: Therefore, in principle, counting much more than 4000 impulses for any measuring point makes no sense (in reflection geometry). But, for bad signal-to-background-ratio, the allowed accuracy per point is enlarged.

You must enlarge EPSILON for bad grain statistic (e.g. 5µ grains and both primary plus secondary soller slits). Otherwise INDEX will fail: EFLECH finds ghost peak multiplets. Such multiplets are caused by errorneous peak shape due to bad grain statistic.