Ignore:
Timestamp:
May 23, 2013 9:43:12 AM (10 years ago)
Author:
srkline
Message:

Updated the NGB attenuator tables with Katie's new measurements from May 2013.

Made some changes to Auto-Fit to make it more useable. Will turn it "on" when documentation is finished.

Changed how the random deviate function is calculated to improve the estimates of multiple scattering. Now the q-range of integration is shifted to lower q for the USANS calculation (using very large objects). Also, the number of points used in the integration is chosen adaptively until a constant value of scattering cross section is obtained. There are still some functions where the integration fails (power laws). These will always fail. My best solution currently is to flag the times when the XS>100. Then the estimates are invalid. The 1D simulation is fine, the 2D is reasonably fine, but anything to do with multile scattering is incorrect.

File:
1 edited

Legend:

Unmodified
Added
Removed
  • sans/Dev/trunk/NCNR_User_Procedures/Reduction/USANS/U_CALC.ipf

    r807 r911  
    904904                // -there are only a few points in the USANS range, and extending it lower caused issues before. 
    905905                CalculateRandomDeviate(funcUnsmeared,$coefStr,wavelength,"root:Packages:NIST:SAS:ran_dev",sig_sas) 
    906                  
     906 
     907                if(sig_sas > 100) 
     908                        DoAlert 0,"SAS cross section > 100. Estimates of multiple scattering are unreliable. Choosing a model with a well-defined Rg may help" 
     909                endif                    
    907910//              if(sig_sas > 100) 
    908911//                      sprintf abortStr,"sig_sas = %g. Please check that the model coefficients have a zero background, or the low q is well-behaved.",sig_sas 
     
    930933 
    931934                Print "Sig_sas = ",sig_sas 
    932                  
     935                Print "Estimated Transmission = ",estTrans 
    933936                 
    934937                Duplicate/O qvals prob_i 
Note: See TracChangeset for help on using the changeset viewer.