SDAST meeting #76: Thursday 22nd May 2003 Participating: Peter Kretschmar, Silvia Martínez-Nuñez, Jérôme Chenevez, Sami Maisala, Stefan Larsson, Carol Anne Oxborrow, Niels Jørgen Westergaard Subjects discussed: Status of PRP, COR and DEAD levels Status of BKG level ARF table Status of BIN_I and IMA levels Status of BIN_S and BIN_T levels Weighted event extraction ISDC software versions Status LCR and SPE levels Results of Crab observations/spectra Status of scripts: planned new feataures Status of IMOD group STATUS values and masking Dead anode status: tools for dead anode detection Personnel rotation at ISDC AI list Next SDAST meeting in Helsinki Action Items: AI020522_1 ALL June 2003 Everyone to tell Jerome about a) flux units OPEN and b) holiday plans **** Logging Started : Thu May 22 13:02:14 MEST 2003 Hello everyone END Hi Carl Anne, Niels Joergen and Sami ! END > Hello everyone! END Thank to you, Silvia, to take these days. CArl = CAROL ! Pardon ! END What about you Sami? You have not answered my email... END No problem, Jerome ! ENd not yet. I have to first look about my schedule. END > This looks like a really good turnout - let's hope Peter joins us soon. END What about Stefan ? END > I haven't heard the slightest thing from him in months, so there's a good > chance he doesn't know there's a meeting. END *** peter (~pkretsch@crab.unige.ch) has joined channel #jemxadr Hi There sorry for being late - my watch is off by a fee w... by a few minutes. END Hi Peter! > Hi Peter - do you know if Stefan's joining us today? END Probably, Carol Anne ! END Sorry, no idea. I guess we should start. END > Okay, I'll start the ball rolling. *** larsson (~larsson@tenma.dsri.dk) has joined channel #jemxadr Hi. Sorry to be late. END > For PRP level there's a SCREW on j_prp_verify to output event histograms > Hi Stefan, I'm just starting on PRP status. > but that's not very high priority. All the otehr PRP components look good > though we maybe need an update of ECAL-MOD to deal with these non-linear > fast anode amplifier/ADCs. Again the offline version of j_calib_adc which > can make me a nice ECAL history table is in the works but low priority. > For correction, j_cor_gain will be redelivered today with sorting of the > gain history entries and corrected linear interpolation. The good news is > that the program can run on either global tables, disordered by REVOL value, > or single-REV tables disordered by OBT. Also the new program produces nearly > identical results with linear interpolation (=/- 1-2 PI channels) as with > the decay smoothing model. The difference is presumably much larger close > to switch on and near anode glitches. Also, comparing the running time of > the program using a global gain history and a single-rev gain history > shows we're on the right track here Peter wit using the single-rev tables: > the global one takes AGES to run! Next piece of work is to implement > Jerome's dead anode corrections. > j_cor_position and DETE-MOD still need updating to handle various different > `bad event' conditions and theres a a minor update on j_dead_time_calc to > include buffer losses in the `input' dead time. > That's all from me. Any questions? END N N N N N n * oxborrow Okay, now it's Sami's turn: status for BKG. END We've been lately made xspec fitting and I think we finally can separate the diffuse and instrument bkg. The main problem is tiny variations of the know peak places different parts of the detector. know= known :) > I think there's a good chance this problem will be eased a little by > implementing the dead anode corrections, Sami. END This might be a good solution and I thin we need new SPAG-MOd tables. oops. thin=think > They're on their way too. END That is nice to hear. The next stumbling block has been the fitting of models in low energies < 6keV. > The real question Sami, is have you any indication that using the background > subtraction improves the imaging and source finding? END That question is for me I think, and the indications from Carl and Niels is yes. The background subtractions needs to be implemented.END I try to deliver the first models for you next week and then we can actually test the influence. * njw Sounds good END Is it possible to have some updating for response files? It will help quite lot of my work. END Great, Sami !I will test you models with the EXO 2030 observations! END > That's good, Sami. I think the gain corrections are working pretty well you = your ! END > now, though clearly more details, like dead anodes can be include in the > SPAG tables, and maybe with some deeper work on glitches and anode 4 > problems I can gain an understanding of any remaining sources of line > broadening. Otherwise Sami, what you've got should be 90% correct. END But what happen under 6 kev ? The rmf that Sami is using is not correct, it isn´t ? END Just one more thing. We have found out that diffuse bkg decreases in a function of radius. How big is that effect ? I have estimated it to be 10% END It seems like there is not much diffuse bkg left the edges of the detector and it is quite impossible to fit anything there. Niels, Joergen, the effect are you talking about is the rmf under 7 kev ? END The one reason is vignetting and I think we have to take count of that in our model. Any comments? N n I think in the case of sources, the effect is bigger than 10 per cent. The wrong rmf introducces a cut-off in the Cyg X-1 spectra around 6 kev, that it is instrumental for sure . END To Silvia: no, this decrease is due to change of solid angle to 'open sky' END Ok, sorry The nforgot my lat comment ! END lat = last ! Pardon ! > Could you say a little more about the problems below 6KeV, because we're Perhaps it is time to throw in a remark on response below 8 keV: > talking about our most sensitive range here. END We have a single pointing (004500860010 if I remember correct) with the Crab on axis and updated rejection criteria. On the basis of that one I have determined a ARF that gives a good fit to the spectrum with a powerlaw down to 3 keV. The photon index comes out a little low (1.95 compared to the expected 2.1) and there is flux missing, the normalization is a factor of 2 down with respect to the expectation. So far I have no reason to change the RMF itself, but this new ARF should be used for observation after 2003-02-27 For the period before that time we have a lot of Crab observations and it is only a matter of time before I have the ARF that applies to that period END Notice: We have also 005800230010 which maybe is even better. END I didn't realize that, are you checking it + END + = ? END Yes too. END How long it will take ? Sorry for stressing about this point ! Willl I be able to use the correct rmf for the special A & A issue ? ENd I think you will END Thanks Niels Joergen ! ENd > Is that all on BKG? Any more questions? END N N N > N N N > In that case, we have NJW with BIN_I and IMA news. END j_ima_shadowgram has been updated with better deadtime and greyfilter corrections. I'm working quite intensely on j_ima_refined_recon right now. Have you looked at the question I sent you this week? END j_ima_src_find is also on my todo list to make dealing with catalog positions better. To Peter: yes, briefly, I'll return to it END Thanks. END I'll also make the background subtraction in j_ima_basic_recon All for now END How is going with j_ima_refined ? Do you need help ? END Actually I think it is easier to finish the first version by myself. Carol Annd (diffcult name you have) and Jerome have taken upon them some of the other stuff I was doing END Ok > Well, some new things have appeared on the todo list, but they're not done > yet! END > Any questions on BIN_I and IMA? END N N n N N > Let's be moving right along to BIN_* with Peter. END Some redeliveries with bug fixes, mostly minor, but I had one really stupid bug in calculating the good time fraction if there was more than one GTI in a ScW. This could really give wrong results for time bins partailly overlapping with "bad times". If some of you have 'profited' from my code in the past for this, we should take a look and see if the error is there also. Note: the upcoming OSA release 1.1 has most but all of my redeliveries included. Another bug that is now fixed is that j_bin_bkg_spectra no longer tries to sort the input IC table ... END Today is not my best typing day ... Unrelated note, mostly for NJW and SL: I have started looking into the weighted event extraction and already bothered Stefan with some questions/ END' END Any questions? N > N I noticed... :-) N Is the OSA on the ISDC standard environment the same as the release you mention Peter? >END Which one do you mean exactly? The recommended way is to set up with: source /isdc/scripts/login --ACTIVITY=OSA Yes. Is this the same as the OSA official release? This gives you the last official release. Good. thank you. END > If there are no more questions for Peter, let's move on to Stefan with To get the very latest mix, add --ISDC_ENV=/isdc/integration/osa_int/current > LCR and SPE END But that environment is not fully tested. END Sorry .. on to Stefan! I plan to make a new delivery next week with a correction of the GTI bug in j_src_lc. Background subtraction in _lc and _spe will not happen until August. END I think that is in line with the general big picture. Do you have any questions regarding LC/SPE? Just to remind you: we plan to have the next formal OSA release in July, with updates up to mid June (we'd really like to do real testing). And then one in October, so deliveries up to mid September. END No questions at the moment from me. END Any one else? N N Sorry - I do have one: when will you support REST data? END Could there be a reason in SPE why I dont find all the Crab counts ? N In any case I'll make a comparison with HK countrate data END To Peter: I dont know.... Depends on priority. Possibly for the september update. END > Re updates: I certainly hope to have everything on my ISSW todo list done > by mid-June. I don't think that's unreasonable or un-do-able END NJW: The fact that the Crab is extended will reduce the net counts > Otherwise, no questions from me on LCR/SPE. Any one else with questions? END No problem - we do not mind having some pieces stable :-) N which are based on the assumption of a point source. I am sure it is significant but not how strong the effect really is. END OK END > Okay, onto scripts with SMN and PK. END Nothing new to say. Peter delivered a new version a couple of weeks ago. Just to mention, that the scripts are very stable. I ran an analysis of 135 ScWs last week, and the ran until the end without problems. May be, Peter, has runned it with more Scws in some of the GPS data. May be, Peter can added something else. sorry runned =run , the ran = tey ran ! Obviously, it is alos no my best day typing ! END Well, I've made recent updates with the main change being to use gain history tables per revolution as default > Its wonderful to hear that the pipeline functions so smoothly now. END and otherwise some default parameter changes. Stability is good, I've run hundreds of Science Windows by now. > Any questions on scripts anyone? END For the July release I'd like to add (reintroduce) the following points: 1. Image mosaic 2. Also in IMA2: src_collect, which gives a summary table of all found sources in the OG 3. Despite the stability, an option "ignore all ScW errors" for jemx_science_analysis (hidden, default=no). END > No shortage of things on your todo list then! Any more on scripts anyone? END To Peter :During my next stay at ISDC ( from 9th of June to 13th of June), I (we) can work on the version for next OSA release ! Since then, I will like to focus on Science. END That's a very good idea. END Perfect. ENd Silvia, how long did it take to run these 135 ScWs? > Next up to bat: NJW on IMOD group. END I ran in our machine, it took 6 hours. END OK. Thanks. END > NJW, anything to say on IMOD? END No real news from my side. Carol Anne and Jérôme are working on ECAL, DETE, and SPAG, I guess we'll hear from them soon END > Depends on your definition of `soon'......Any questions on IMOD? END When will the new ARF be in the IMOD? END It will not be in IMOD, it will be part of RMF-GRP. When I've got a good match to the Crab spectrum before 2003-02-27 I'll release two RMF-GRPs: before and after. END Sorry NJ. Thanks. END Sorry, I had a question about scripts... Is it still necessary to state "STATUS<256" for LCR and SPE? No, anymore ! That option was needed withe the all gain corrections. END Nice! thank you. END > I don't think that option can be very useful because you really need to > make a mask to find the non-zero status values that are acceptable for > your application. I've considered a ranking system where the lower a > number the more you can ignore it, but I don't think this really applies: > what's important for one application is trivial for another. E.g. imaging > vs. spectral analysis. In the future you really should look at the list > of STATUS values in jemx.h (they each correspond to setting one bit in > 2-byte status value) and decide which ones you can live with and which > ones you can't and define your mask accordingly. END The option is there as a fall-back solution for users to override the choices done auotmatically. It was very useful for just that purpose but is no real replacement for the selection specific to the task. END Note that the rowSelect option can be used on everything, not only on status. So you could, for example, also write "DETX > -30.0 && DETX < -33.0" to exclude a 3 mm slice of detector area. END Sorry I flipped the < > ! END Shall we move on? END > Next is JC with dead anode news END It seems there have not been more anodes deaths for a while, but from the very last checks, I think there is a some suspect anodes at 3 different positions which may slowly have become weak... TBC I have made an analysis to look for an eventual dead anodes effect in the Backplane direction. There is nothing to see there. So I guess it is enough to build a new SPAG from the gain factors I got for each anode. I hope there is nothing more to do with that... END I should mention that Reiner has included displays for anodes in IOSM, which at some point could be used in the barn. Nice. Good to know. END For those he could use a good reference set with normalized flux as function of RAWX. He has created some from an empty field done in December or so, but it's not fully up to date. END > Some time in the future you will be getting the new j_prp-verify with > anode population table output, and a little tool to collect these into > an all-mission history of anode activity....but that's not quite yet. END > Any more from you JC? END No. If there is no question... END > No END No n N N N Any questions regarding the rotation scheme? N N Tomorrow I will book my flight tikects. END n Stefan, from when will you be available for this rotation? > Sorry, I still haven't looked at it! I'll do it while I'm waiting for the > compilation check. END From second half of August or so. I'll sennd you a mail of > There doesn't seem to be anything urgent on the AI list, so can I suggest > that we run through it next week. Also, Jerome has tried to initiate some limitations when I have checked my schedule. END > activity on the documentation of the various flux units used. Please be > kind enough to respond to him all of you that use flux units. END OK Stefan. The other: tell me if you have any summer holidays plans. END THANK YOU CAO! JC: July! END > Let's make that an official AI: Everyone to tell Jerome a) all they know > about flux units and b) all about their holiday plans. END > Any other business? END not from me N N N N > Do we need to consider out next SDAST meeting yet? END Where ? and when ??? END Maybe, we said August in Heslinki or something like that, yes? September is my holiday season ... END > July is our holiday season - sunny Sweden! (Sorry Stefan!) END Yes ... ! September will be also my holiday season (first ten days) ! > Thanks for your AI status Peter - I'll tidy up the list END > So it looks like August or October for the next SDAST meeting - personally > I think October is a little too far away. END Mid-august I will be in Lahti (Finland) for Nordic championships I would not mind if it was in connection with that... END I prefer August ! i will like to escape form the hot for a couple of days ! END August 9-14 is blocked with shift duties for me, otherwise it should be fine. END End August then ?? Why not ? End > August it is then - start looking at your calenders people! Any other > issues, ideas, worries, niggles, gossip? END Not for this session END No ... i just wnat to run to take lunch ... please ! ENd "See " you next week ! Good weekedn ! END > Why do we never have any gossip? END Just take it at 12:00 like some others (gruel grin ...) END Bye bye, have a nice day END *** njw has left channel #jemxadr : (njw) hej , hej .... *** silvia has left channel #jemxadr : (silvia) > Okay, there will probably be a chat next week. See you all then. END Tchao!