SDAST CHAT MEETING # 104: Wednesday 18th May 2005 Present: Carol Anne Oxborrow, Peter Kretschmar, Jerome Chenevez, Niels-Joergen Westergaard, Silvia Martinez-Nunez, Stefan Larsson Subjects discussed: Xe line determination of differential gain aging: See figures and email details on the Forum under Technical Notes Does this explain the 7+keV placement of what might be Fe lines? Offering assistence to Ekaterina with her spectral feature Stephane's email about bad lightcurves and source positions from j_ima_iros: SPR to be written on JMX_COSX problem (always 0) Helping naive users get the best out of the software: known issues, improved default parameters, better manual descriptions and more informative log messages would all be a step int he right direction NJW and NL to look into the rest of the problems Known Issues must be updated before OSA 5.0 release Documentation: priority for completion of documents for OSA 5.0 Users Manual (Masha, NJW, all) Must be ready for OSA release Known Issues (Isabelle, all) Must be ready for OSA release SVR (NJW, all) Should be ready soon after OSA release ADD (CAO, all) Must be completed before summer hols NJW to circulate list of required inputs from SDAST for SVR and UM June SDAST meeting: Starting at 13.00 Monday 6th June On the agenda: science results from Silvia, Peter and Jerome `Fiasco' for supper? Supper at CAO's for anyone staying overnight on Wednesday **** Logging Started : Wed May 18 10:51:53 MEST 2005 *** njw (~njw@130.226.216.2) has joined channel #jemxadr > Hi NJW!!! You owe me several Monday meeting summaries........END Yes, as soon as this j_ima_iros stuff stops firing back at me END > Let's hope the shooting stops soon! END > Peter should be joining us soon. END *** larsson (~larsson@130.226.216.181) has joined channel #jemxadr *** peter (~pkretsch@isdcsf2.unige.ch) has joined channel #jemxadr Hello everybody! *** silvia (~silvia@130.226.216.2) has joined channel #jemxadr > Hi, Peter and Stefan - I hope we won't take you away from your teaching for Good morning everyone ! Sorry for the delay ! END > too long, Stefan - I know how very time-consuming it is. END > HI SILVIA! The gang's almost all here. End *** jerome (~jerome@130.226.216.3) has joined channel #jemxadr Hello, all END Hi! Sorry to be late. END You can pay around of beer in Copenhagen to make up for the delay ;-) END > If we're all sitting comfortably, then I'll begin..... > I hope you all got my email, so you know what I've been spending the last > couple of weeks on. The best way to download the figures is from the > Forum, where they're the first entry under `Technical Notes'. Please take > a look. Basically, we can conclude that there's a pronounced, though very > gradual differential gain aging in the detector plate, and that needs > correcting by periodially updating the values of the reference channels for > Xe and the four calibration sources in JMXi-SPAG-MOD. These are the values > I've been struggling to find. Jemx-1 is completed, just JEM-X2 to do now. > Any questions? END No N While it appears quite clear visually, have you made a solid statistical check to see how significant the gain drft is? END > Actually, my work doesn't really demonstrate this statistically, just > that we can see the drift that Jerome's analysis originally showed was > there, which involved lots more data points. Jerome, is the drift significant > statistically? I think we really do need this correction, because I've been > having an interesting correspondance with a Russian lady, who seems to have > the Iron line at 7keV or there abouts. While it's well known that analysing > iron with jemx data will be difficult because of the low energy cutoff, I > think that in her case, we really do have a gain drift that needs to be > corrected. END I am sorry I cannot see your figures Carol Anne... Anyway, I think I can remember that there was about 3% drift from my analysis... I have to check. END OK, but does it hang together? I.e. do you have evidence that during her observations the gain drift was there and not fully corrected? > Yes, the drift is not huge, but it is significant after several years. END Is there data from around her observations showing similar behaviour? Fe lines are rather common so one should standa chance ... END > Not really, Peter. I'll have to see how our corrections improve her results. END The question to be checked is if everyone would have that energy shift or if she for some reason has especially strange data. END > yes, I agree. My first check on the results will simply be to find the > corrected KeV position of the Xe line for as much data as I can get my > hands on - this will be part of the SVR. Then take a look at case of strange > spectral bumps. END I just can say that I have NEVER seen the Xe line so low END Now I'm puzzled, our Russian friend has an Fe line which is too _high_ - right? END Sorry! > yes, the russian lady has what may be a too high Fe line.END I raed Xe line at 7kev... Forget it. END > Don't say things like that Jerome - the very idea gives me goosebumps! END > Any more questions? END N From which revolutions are the data of the russian lady ? END N n > I'm not too sure END > If there's any one out there wants to help with the russian lady's problem > your very welcome to join in! END > Shall we move on to Stephane's email about lightcurve problems? END I'll ask Ekaterina which data she used exactly. NED I'll like to take a look to these data. END > I'd be very glad if someone could help her. Thank you both of you! END Let'e move on Stephane mail ! END > Now on to Stephane's problems: has anyone got any brief or initial > explanations for any of his complaints? END Except for the use of data down to 2 keV - not really END > Thanks for the email to Ekaterina, Peter - I haven't had time to get back > to her since my long email about gain drift and low energy cutoffs. END > NJW and Stefan - any good expert explanations from you. Ideally, he's > just used a load of bad options or a bad combination of parameter options. > That would be a nice solution to the problems! END I have also made a LC from Crab, which looks quite good to me, though the source is about 4 degrees off-axis! But I also get this weird off-axis value of 90 degrees in the output file. END I assume the off-axis is from the output of src_collect, right? END Y es it is. END If so, it is calculated (sloppily) from an angle in the individual SRCL-RES I hope it is not JMX_COSY and JMX_COSZ that are used because their implementation is still lacking END I'd have to check ... Wait 20 secons (or talk on) ... END It is based on COSX. Are you sure you are writing COSX and not the 90 deg shifted one? END Please just check for a few ScWs. END You mean JMX_COSX - I think it has _always_ the value of 0. deg hence the 90 deg I believe - in any case it is a bad value to use since the precision is very bad END You can write an SPR or SCREW on that and we will change that END Well, in the past it worked well enough. So now we have a problem - will this become a Known Issue? END Depends if we have time to update j_ima_iros which in any case is in trouble since Jake can make it fail in strange ways. END OK, if you must update it anyway and it would be easy to write COSX, just do so please so one element that roughly worked before works again - OK? END You do mean JMX_COSX, right ? END yes - I'm just lazy typing END So - who will look in detail at Stephane's results? END I guess I'll urge Niels to take this up - after all it is his flux determination * njw that looks strange in some cases END > Oaky, more comments on this problem? END N n n I would just say that I am afraid our credit at ISDC is going further downn after that bad results... It was as if this Crab LC was the test they absolutely wanted to do before I went -sorry- LEFT ISDC! END > Yes, this doesn't look good - IROS was supposed to be the solution to our > problems, but problems keep having a habit of cropping up. END As CAO said, these bad results can just be due to a wrong parameter settings. I will contact Stephane to know his parameter settings. END > I do think that naive users do tend to get a lot worse results than we do. But Jerome, you said you got a reasonable LC? What parameters did you use? END > I know for sure that if people are having a difficult time getting the If naive users get worse results than we do than we should > gain smoothing to work with a dodgy revolution (lots of stops and starts) - have better default parameters The default ones, I guess! END > that I can always help themd to get better results. END - explain better in the user manual what they should do - improve (for OSA 5+) the problem reporting so thee software warns them in clearer ways. END I agree. I have already the same feeling regarding j_ima_mosaic... END > Yes, I agree about log comments: j_cor_gain produces to much meant for > debugging, and not enough suggestions to users regarding gain history > tables and their editing/pruning/finding. END > Any more comments about this basic problem with naive users? END N n Just one, do we will be able to have all these improvements in the next release ? END > Not for OAS 5.0 END I do not think so... END But at least we should have an update "known issues" that will contain all this kind of problems. END Yes, and explain them in the analysis manula. END > Yes, Known Issues is a very important document. END > Which brings us seemlessly to OSA documentation. For OSA 5.0 we need to > update four basic documents: ADD (me); SVR (NJW); User Manual (ISDC and NJW); > and KnownIssues (All, Isabelle). Is this the agreed division of labour? > Of course we're all expected to contribute to each document. Comments? Do they all have the same deadline? When is it? END Well, User Manual and Known Issues _must_ be ready for the OSA release. The SVR _should_ be ready and the ADD update can come afterwards - but should be doen before you all go on Summer holiday :-) END -- before any of you go on summer holiday ;-) END > I agree with Peter's prioritisation. My contributions to SVR will be forth > coming as soon as I've done the Xe analysis for JEM-X 2: verify the new > misalignment matrix and verify the Xe line positions using the new > corrections. END Very good; in a short time, perhaps hours, I'll go through the old SVR and find where updates are needed and hand out tasks for you. Same with Analysis * njw User Manual END Brrr, sounds like work might come doen the line ;-) END OK, anything else? END n N N N > N > Let's move on to agenda items for June: > I suggest outstanding OSA 5.0 matters: late documentation and making a > list of outstanding known problems to be solved for 5.1 > I already have a thorough discussion of person power on the agenda. > Anything else that's really important? END Doe we still have time to discuss OSA5 at the meeting? Won't it be out around then? END > $100 says we'll still be grappling with documentation! This is not a perfect > world. END > also plans for 5.1 can't be started soon enough! END How do you know there will be a OSA 5.1 and not directly a OSA 6.0? END Do not take it seriuosly :) END > I assume that everyone will have bits and pieces to tidyy up. END > Other agenda items? Background handling? Science results (do we even have > any???) ? END Weel, I am trying. but with all this OSA work, it ai almost impossible! END I can show some science results ! END Good! I'll also show some work-in-progress END > Hurray for Silvia!!! I welcome anything that indicates we're not thrashing > around in vain. END 10' should be enough in my case. END Me too. END > Great stuff - you'll all get your day in the sun. Can we move on to > AI list - anyone done anything lately? Other than all that trivial `ordinary > work'!! END No news from here END Allan will be coming home in July - perhaps we should drag him into this END Nothing comes to mind mam... END > If you remember Allan dropped out of JEM-X when it became apparent that > diffuse sources would not be detectable. But perhaps he's found an > interest in compact sources while he's been in California. END * njw Its worth a try END > I think we'll have a deeper trawl through the mysteries of the AI list > next chat - so be warned!!! END > Okay, AOB -- anyone got some AOB?? END N N What is on the menu? END Jsut one. To NJW: j_ima_iros.txt has not be updated. END > Sorry I don't have digestives to share today, Jerome. END Do not you mean apperitives? We have not eaten yet! Yes, I know, Silvia, I hope to get to it soon END -Sorry Silvia to distrub the class. END When do you plan to start the meeting in Copenhagen? END > I assume we'll begin after lunch, about 13.00, and work til 18.00, with > supper somewhere nice/cheap (delete where applicable) at 19.00. What say > you all? END Good! END Let's go to "Fiasco! END I wil lprobalby be in Denamrk from Sunday to Thursday.END > When will you other out-of-towners arrive and depart? Silvia, you must come > and have supper with us one evening, how would Wednesday suit? END Do you really hope the weather be so nice in June here...? END > Of course, anyone else staying on Wednesday night is welcome to come to > `Chez Oxborrow' for supper too. END Fine, I do not have yet any other appoiment !END Sounds tempting, but I fear I won't make that. END I plan to arrive by train Monday lunch and go back late afternoon/evening on Tuesday. END I must leave the chat now ... I've an appoiment with the tax authorities. See you soon. END *** silvia has left channel #jemxadr : (silvia) > Yes, I think we've all got an appointment with our stomachs, so let's call > it a day. Chat with you all again in 2 week's time. END > Please send emails with your arrival and departure times. eND