Si XIII He-like complex using hewind 14Oct08 atomdb says: 6.6479 6.6866 6.7403 and I'm going to use (for now) Maurice's phiratio value of 17.8 fitting the continuum 6.3:6.5; 6.9:7.0 XSPEC12>iplot PLT> wdata /home/cohen/xspec/2007/zpup/sixiii_665_dataonly.dat ======================================================================== Model powerlaw<1> Source No.: 1 Active/On Model Model Component Parameter Unit Value par comp 1 1 powerlaw PhoIndex 2.00000 frozen 2 1 powerlaw norm 9.26004E-04 +/- 1.05122E-04 ________________________________________________________________________ C-statistic = 357.95 using 352 PHA bins and 351 degrees of freedom. Warning: Cstat statistic is only valid for Poisson data. XSPEC12>goodness 1000 nosim 85.00% of realizations are < best fit statistic 357.95 fit not that great - try eliminating 6.3-6.4 portion of the spectrum (the fitting range was pretty large to be using a flat function) XSPEC12>ignore 0.-6.4 Model powerlaw<1> Source No.: 1 Active/On Model Model Component Parameter Unit Value par comp 1 1 powerlaw PhoIndex 2.00000 frozen 2 1 powerlaw norm 9.88921E-04 +/- 1.31268E-04 ________________________________________________________________________ C-statistic = 240.94 using 232 PHA bins and 231 degrees of freedom. Warning: Cstat statistic is only valid for Poisson data. XSPEC12>plot XSPEC12>goodness 1000 nosim 84.70% of realizations are < best fit statistic 240.94 XSPEC12>notice 6.3-6.5 XSPEC12>ignore 6.5-** ======================================================================== Model powerlaw<1> Source No.: 1 Active/On Model Model Component Parameter Unit Value par comp 1 1 powerlaw PhoIndex 2.00000 frozen 2 1 powerlaw norm 8.90481E-04 +/- 1.28484E-04 ________________________________________________________________________ C-statistic = 232.17 using 240 PHA bins and 239 degrees of freedom. Warning: Cstat statistic is only valid for Poisson data. XSPEC12>goodness 1000 nosim 67.60% of realizations are < best fit statistic 232.17 XSPEC12>error 1. 2 Parameter Confidence Range (1.0000000) 2 0.000822 0.000962 (-0.000068,0.000072) XSPEC12>error 2.7 2 Parameter Confidence Range (2.700000) 2 0.000780 0.001010 (-0.000110,0.000120) let's fit the complex itself over a relatively wide range (at first) 6.57-6.84 looked, and cut to 6.58-6.82 ======================================================================== Model hewind<1> + powerlaw<2> Source No.: 1 Active/On Model Model Component Parameter Unit Value par comp 1 1 hewind q 0.0 frozen 2 1 hewind taustar 0.415249 +/- 0.144983 3 1 hewind u0 0.665626 +/- 2.21879E-02 4 1 hewind h 0.0 frozen 5 1 hewind tau0star 0.0 frozen 6 1 hewind beta 1.00000 frozen 7 1 hewind betaSob 0.0 frozen 8 1 hewind G 1.02083 +/- 7.98912E-02 9 1 hewind numerica 0 frozen 10 1 hewind anisotro 0 frozen 11 1 hewind rosselan 0 frozen 12 1 hewind expansio 0 frozen 13 1 hewind thick 0 frozen 14 1 hewind Z 14 frozen 15 1 hewind shift "mA" 0.0 frozen 16 1 hewind velocity "km/s" 2250.00 frozen 17 1 hewind phiratio 17.8000 frozen 18 1 hewind verbose 0 frozen 19 1 hewind norm 1.12000E-04 frozen 20 2 powerlaw PhoIndex 2.00000 frozen 21 2 powerlaw norm 9.26000E-04 frozen ________________________________________________________________________ C-statistic = 335.64 using 284 PHA bins and 281 degrees of freedom. PLT> wdata /home/cohen/xspec/2007/zpup/sixiii_665_hewind_smooth.dat XSPEC12>goodness 1000 nosim 92.60% of realizations are < best fit statistic 335.96 *Fit's not that great (should we double check phiratio?) As with the *Mg XI complex, "fit" has trouble settling on the true minimum, but *what it settles into is functionally equivalent (but with a C value *about 0.3 higher) XSPEC12>error 1. 2 Parameter Confidence Range (1.000000) 2 0.290318 0.558866 (-0.124933,0.143615) XSPEC12>error 2.7 2 Parameter Confidence Range (2.700000) 2 0.224349 0.665072 (-0.190902,0.249821) XSPEC12>error 1. 3 Parameter Confidence Range (1.0000000) 3 0.640910 0.686409 (-0.024716,0.020783) XSPEC12>error 2.7 3 Parameter Confidence Range (2.700000) 3 0.626862 0.699747 (-0.038764,0.034121) XSPEC12>error 1. 8 Parameter Confidence Range (1.000000) 8 0.941732 1.101534 (-0.079097,0.080705) XSPEC12>error 2.7 8 Parameter Confidence Range (2.700000) 8 0.899172 1.155364 (-0.121656,0.134535) infinite loop of "better" fits when running 'error' on the norm XSPEC12>steppar 19 1.e-4 1.2e-4 20 C-Statistic Delta norm C-Statistic 19 349.44767 13.48283 0 0.00010 347.20405 11.23921 1 0.00010 345.17071 9.20587 2 0.00010 343.34277 7.37793 3 0.00010 341.71853 5.75369 4 0.00010 340.29396 4.32912 5 0.00011 339.06585 3.10101 6 0.00011 338.03088 2.06604 7 0.00011 337.18586 1.22102 8 0.00011 336.52766 0.56282 9 0.00011 336.05325 0.08841 10 0.00011 335.75965 -0.20519 11 0.00011 335.64396 -0.32087 12 0.00011 335.70337 -0.26147 13 0.00011 335.93510 -0.02974 14 0.00011 336.33646 0.37162 15 0.00012 336.90481 0.93997 16 0.00012 337.63759 1.67275 17 0.00012 338.53228 2.56744 18 0.00012 339.58642 3.62158 19 0.00012 340.79761 4.83277 20 0.00012 XSPEC12>log /home/cohen/xspec/2007/zpup/sixiii_665_hewind_smooth_35by35.log Logging to file:/home/cohen/xspec/2007/zpup/sixiii_665_hewind_smooth_35by35.log XSPEC12>steppar 2 0.01 4.5 35 3 0.48 0.99 35 OK - after doing this, I found that I'd fit this complex back in late July 2008; the basic results are very similar, but the fit quality is better. This appears to be because I used a section of the continuum that was bigger, closer to the complex itself, and better balances between the red and blue side. The upshot was a higher continuum level (though not by much), and that seems to have made the difference. For reference, check out the files with 66479 in their names.