SDAST Meeting 44, 29th May 2012 at DTU-Space North These are additional notes to supplement the presentations that are archived along with these notes. Participants: Niels Joergen Westergaard (NJW); Soeren Brandt (SB); Erik Kuulkers (EK); Jerome Chenevez (JC); Carl Budtz-Joergensen (CBJ); Carol Anne Oxborrow (CAO); Lucia Pavan (LP); Niels Lund (NL); Silvia Martinez-Nunez (SMN), remotely. Welcome (NJW) - Start of meeting delayed by setting up webmeeting with Silvia in Spain - Use https://connect.forskningsnettet.dk/SDAST44 Summary of NJW's presentation for which there was no PowerPoint presentation: The Nuclear Spectroscopic Telescope ARray (NuSTAR) launched in June 2012 carrying the first focusing hard X-ray (5 - 80 keV) optics to orbit. The on-ground calibration was performed in a long-beam facility at Columbia University. During assembly of the optics, mechanical surface metrology was acquired to map the reflecting surface of all mounted substrates. The surface information is utilized in a ray trace code along with a model for X-ray scattering. The scattering model is based on data acquired off spare flight substrates at BNL. The ray trace is checked against the calibration results and used for predicting the imaging properties of the telescope for X-ray sources at infinite distance. JEM-X Status and INTEGRAL news (SB) - Presumably we would get extension to 2016 but various factors have worsened for INTEGRAL though its fate is still linked to that of XMM - Changes in ESA's internal accounting make extensions appear more expensive, though this would be partially offset by ground station coverage holes - Aging of instruments is problematic for the IT though not for ESA ISDC News (LP) - OSA 10 in preparation both as test software and documentation - Critical package for JEM-X is j_ima_iros which is not yet delivered - Some suggestions for software improvements for OSA11, like different binning ranges without re-running the software - HEAVENS: http://www.isdc.unige.ch/heavens browsing tool for the archive - Can save the fits files that you view for various sources - Shortest recommended time bin for lightcurves is one SCW according to ISDC - But much shorter binning is really necessary - HEAVENS covers all the ISDC instruments, not just INTEGRAL, so settings have to be useable across several missions, including Planck - Cannot search archive by revolution - LP will see if this can be done by the software team - Documentation is only preliminary at the moment AI44.1 - All developers to look through the preliminary documentation and comment on it. Due mid June News from ESAC (EK) - Was Earth observation successful: yes, it did point at the correct target but background was unstable so very hard to analyse. - Now preparing for next Earth observation in August - a standard procedure for EOs being prepared. So next one should be much smoother. Involves alot of work, especially for MOC - A standard procedure was prepared before the Earth pointing, but that didn't yet contain the lessons learned from the actual observation. That is now incorporated. - OMC still had some manual commanding, but ISOC hopes to get that done automatically in August. - Startrackers don't work when they point at the Earth, so essentially no steering for several hours. Pointing has to be reconstructed after the observation, which is what went awry last time. - j_ima_source_locator,j_q_identify have been updated - The Galactic bulge monitoring pages have been updated. This involved lay-out, as well as JEM-X analysis scripts. Now using j_ima_source_locator,j_q_identify - X-ray sources from JEM-X on the left of the movie and gamma-ray sources on the right. - 1E1740: hard source, persistent blackhole - There will also be a 3-D movie that shows flying through the galaxy past all the INTEGRAL sources. Will be unveiled in October in Paris - AO10 TAC meeting 4-6 June, AO10 starts in january 2013. AI44.2 - Turn their presentations into pdf files and send to either CAO or NJW Flux Error Estimation (NJW) - chi-squared estimates were not good enough - P is the poisson distribution over about 30,000 pixels in the JEM-X instrument - Z = log (P), so minimize Z instead of maximising P - Error on background is smaller that on source because source is so weak. Note that found source and background slightly offset from input values. - Method assumes that you know the PIF, but in real life you don't it perfectly. - And then there's the collimator to take into account - Further discussion will be needed Temporal IMOD Corrections (NJW) - ARF should be the same for all observations - Corrections should then be put into the image and the flux estimate - Also correction of electronic efficiency at low energy - All three components determine final results - Corrected results shouldn't have the spurious 7-8keV feature - NJW hopes all this will be implemented for OSA10 - There will be new IMOD tables for OSA10, though software doesn't rely on the new tables AI44.3 - On NJW: to deliver new IMOD tables for OSA10 Cross-correlation (NJW) - NJW recently sent Elisabeth Jourdain some new curves - Seems like we're almost there - It wasn't possible to make a 'first principles' spectrum - i.e. no spectrum yet made based on detector properties alone - In all calibration efforts NJW has disregarded the variation in the Crab itself, which mostly seems to affect higher energies - Variation of Crab in our energies is about 5%, which is almost within our range of uncertainty IMA IROS version 4.0 (NL) - PIF-selected images are improved, with reduced noise around strong sources - PIF option is specified by writing 'PIF' in parameter `userImagesOut' - Fewer spurious sources found in PIF mosaic images - NL mostly working on LOFT for the past year - Still need to understand the Crab SCWs with strongly outlying fluxes - Flux errors are even more underestimated at high energies - Flux error problems are probably even worse for crowded fields and weak sources - the cases presented are for the Crab, which should be good - Further sanity checks on the errors should be done - Bottom line is the j_ima_iros cannot be delivered in a new form for OSA10 right now. If OSA10 is significantly delayed, it may be possible. Deadtime verification (SB) - Nothing new, but we still think that our deadtime is higher than calculated - Not entirely clear how determination of the actual deadtime might be done - Still lots of work to be done on this problem New software for JEM-X ICAL calibration (CAO) - One new piece of software: an IDL procedure for correction patching of NRT gain history tables which used to be done manually by CAO - j_cor_gain v9.0 which now includes a countrate dependent smoothing model and was delivered to ISDC for inclusion in OSA10 in March - There are still difficult revolutions but they can be dealt with much quicker than before because CAO can set the various options on the two pieces of software just by looking at the Xe line position throughout the revolution, produced from the NRT data with the NRT gain history tables Crab flux using various methods (JC) - Mosaic_spec developed by Roland Walter for IBIS, not for JEM-X, so the various options and parameters may not be suitable for JEM-X images, which many users do. - Widthmode option has three settings, for which zero value is to let a free psf value. - PSF should not be specified in pixels because you cannot get low enough with the allowed values. - LP suggests that there should be a documented warning about using this tool with JEM-X data - Can also let free the position of the source, but the detection significance drops by doing this - j_src_lc is the other tool that can be used (this was Stefan Larsson's software) to get fluxes. - It is possible to make lightcurves with timebins much smaller than a science window (e.g. 1000sec) - This gives completely different flux values in counts/sec. - LP says this is the problem with short time bins. - QLA also has this problem that different software tools give different flux values, even for something as strong as the Crab. - Significance also changes as flux values change, and people who monitor QLA output never know which numbers are correct. - NL: image generating software does not output a DETSIG value if you don't output an image too. - The next method is j_collect_spectra - Do the counts per sec include the presence of the mask and correct for it? - NL's original software tried to match up results from both JEM-X units, at a time before we worried what the images might be used for. - This last method does not include the effect of the ARF - Seems like the whole area is a quagmire of conflicting definitions and interconnected instrumental effects that may or may not be corrected for by the different flux determination methods. - Seems to show that there's a consistent error at the very low energy end. - NL hasn't worked on the electronic efficiency since revolution 800, so if this has changed since then we shouldn't be too surprised by the variations we're seeing - Carl says the electronic efficiency shouldn't vary with temperature or time. But evidently it does, especially below 5keV AI44.4 JC, SB, NL and CBJ to determine why the different methods give different results, and what can be done about it. AI44.5 NL,JC To ensure that their software gives fluxes in counts/cm^2/sec - Positional accuracy error really blows up at low source significance. After-meeting note: It turns out that the significance obtained from the images depends on the version of j_ima_iros which is used... Jérôme has used OSA 9. See his updated figure. Miscellaneous known issues etc. (LP) - Should we flag as BTIs the first few SCWs of every revolution? - Some few (5 or so) revolutions that are short and affected by re-switch on and/or solar flares should also be flagged as BTIs - LC suggests a new flavour of BTIs to flag beginning of each revolution - Could flagging at beginning of each revolution be automated? We could set limits on the gain (chans/keV) so that very low gain - It is agreed that BAD_FLUXES will be name of new BTI flavour, for periods where images can be used for source finding, but spectra are poor and gain calibration is uncertain, e.g. at beginning of each revolution AI44.6 LP to send an email describing how she would like the new BTI to be implemented. Focus of software development (ALL) - First j_ima_iros, including flux determination. Imaging part is quite mature, but fluxes are still a problem - Next, j_ima_src_lc needs work - j_mosaic is already delivered for OSA10 - j_cor_gain is in very good shape for the time being - IMOD tables will take a couple of months to update Known Issues and Documentation (LP) - http://www.isdc.unige.ch/integral/osa_doc-10.0 - Check the documents (see AI above) - Changes are written in blue so easy to find AI44.7 All SDAST to read the Known Issues list and check that it is current and complete Person Power (ALL) - CAO expects to use less than 50% of her time on JEM-X - LP will use about 50% of her time - CBJ will continue to monitor incoming IPF files - NL very busy with LOFT, but aiming for using 1/3 of his working time on JEM-X - SB as PI but also very busy with LOFT just like NL and its not clear how this will develop, but will be very busy with LOFT at least for the rest of the year. Whether this continues depends on ESA decision at the end of the year. - EK will still be JEM-X liaison at ESAC, but he is also committed to LOFT - JC is paid by PRODEX for this year, so technically he'll be 100% on JEM-X, but for teaching - NJW will be spending about 30% of his time on JEM-X and will be very busy with NuStar for the next few months. - Mission extensions assume that Instrument Teams will continue their usual level of participation until the end of the extension. - Institute's official policy is to support the mission until it ends Next Meeting - Should we hold it in connection with the October meeting in Paris - Meet for a few hours, over supper maybe, and see if that's enough and if not we'll arrange another meeting