Monday, February 26, 2007

chandra response matrix generation in xselect

Updated the perl script xsl_chandra_acis_makeresp to replace old ftools with new heatools and improve the search for the bad pixel file. Outstanding issues are that mkacisrmf doesn't work with old event files (fails to find a relevant CALDB file) presumably because these event files have to be reprocessed in CIAO 3.4; the current advice is to use mkarf for point sources and mkwarf for extended sources - at present we use mkwarf for all sources; there should be a mechanism for the user to give their own bad pixel file.

bug in kerrbb

There is a memory handling bug in kerrbb (due to my error) which can cause a seg fault if multiple response matrices are in use and the first one does not have the largest number of energy bins. This is fixed in both v11 and v12.

Thursday, February 22, 2007

filter intensity in xselect

Another iteration on the filter intensity perl script in xselect. Replaced most of the calls to old ftools with their heatools analogs. This seems to fix a problem with long filenames.

Wednesday, February 14, 2007

bug in extractor when combining event files

Fixed bug reported by Phil Evans when creating a new event file from multiple input event files with different WCS values. If the sky X and Y columns were integers the new values were calculated as an INT after doing the transformation instead of a NINT causing a possible one pixel shift. Also did some optimization of the code. v4.74.

Update: Also found that the goodfwcs logical variable in EXTWCS was not being initialized which was causing the position recalculation not to occur in some circumstances. v4.75

Keywords: HEAsoft, extractor.

Tuesday, February 13, 2007

xspec error with the migrad method

There was a bug in xspec v12 when using the error command with the migrad method. If there were frozen parameters the wrong parameter number was being passed to the MINOS command.

A useful tip to know when using migrad is that the standard MINUIT output is written to the file mn_output.log in the current directory.

Handy C++ tip: there is a function, distance, associated with maps to return the distance in the first index of the map between two iterators. So if we have a map called mymap and an iterator called iter then the position in the map list of iter is 1+distance(mymap.begin(),iter).

Keywords: HEAsoft, xspec

keV-Angstrom conversion factor in xspec

Both v11 and v12 xspec use a conversion factor from keV to Angstrom of 12.39854. Using the current standard values of constants (CODATA(2002) via NIST) this factor should be 12.3984191. Apparently the difference is noticeable when plotting grating spectra in wavelengths. Made this change in both v11 (bug fix 11.3.2ac) and v12.

Keywords: HEAsoft, xspec

text size in the PGPLOT PS driver

Another tweak of the new PGPLOT PS driver to get the text size closer to that appearing on the screen with other drivers. Also, moved the code to calculate the correction to the string length from grlen to the driver.

Keywords: HEAsoft, pgplot

Saturday, February 10, 2007

ranlux in xspec11

xspec11 crashes when compiled using gfortran apparently due to a problem with ranlux. So replaced xanlib/sysdep/ranlux.f with C code based on the gsl version but with gsl-specific parts stripped out. Note that only the RLUXGO and RANLUX Fortran calls are supported and that RLUXGO ignores the final two arguments (K1 and K2) ie the restart option is not supported.

Keywords: HEAsoft, xspec

Tuesday, February 06, 2007


Portegies Swart et al. test using modern Graphics Processing Units to perform N-body calculations and conclude they show some cost and reliability benefits over specialised hardware.

Brownstein & Moffatt argue that the bullet cluster results are consistent with modified gravity and not with dark matter.

Evrard et al. use an ensemble of heirarchical clustering simulations to argue that sigma_8 = 0.8 in contrast to the best-fit number of 0.69 favored by recent WMAP and SDSS analysis.

Lodato & Natarajan derive the seed function of mass BHs at z~15.

Krongold et al. use an XMM observation of NGC4051 to track changes in the ionization state of the warm absorber and deduce that it comes from an accretion disk wind not the dusty torus.

Schafer provides a statistically rigorous method of estimating luminosity function using truncated data.

Rossetti & Molendi fire back in the controversy over the existence of a non-thermal tail in the X-ray spectrum of the Coma cluster.

Medvedev examines evaporation rates of hot gas from clusters as a possible explanation of recent S-Z results and concludes that for the mechanism to work magnetic fields must be non-existent or highly ordered.

Seto & Cooray discuss a technique for placing limits on primordial BH densities by using pulsar timing measurements to look for impulsive accelerations on the Earth.

Linder & Miquel warn against applying model selection procedures in cosmology without adequate thought about physics.

Hickox & Markevitch look at Chandra deep field data excluding all known HST and Spitzer sources to provide a limit on X-ray emission from the WHIM.

Fusco-Femiano et al. continue the data about the Coma hard X-ray emission.

Chen et al. use ROSAT and ASCA observations of HIFLUGCS clusters to investigate the statistical properties of cooling cores.

Zavlin reviews what is known about thermal emission from neutron stars.

Johnstone et al. analyze the Spitzer 10-37 micron spectra from off-nuclear regions in NGC1275 and 4696.

Monday, February 05, 2007

extractor and string columns

Phil Evans reported a problem that when combining two event files a string column in the output was garbage for all rows copied from the second input file. This happened because the input files had differing WCS so the slow copy method had to be used and this read all columns as doubles. Fixed the problem by checking for string columns and reading/writing them as strings. Note that strings can have a maximum length of 256 characters. Fix is v4.73.

Keywords: HEAsoft, extractor