| There are two main accounts for
the pooled "flexible" observing programs: flexible and observer.
For general purpose linux and data access you need to log into mrt-lx1.iram.es
or mrt-lx2.iram.es. Former is a new dual-processor linux PC with
lots of memory and a big raid hard disk system on which the bolometer data
is stored. On both machines you log in as flexible.
Ask the other pool observers or the operator or astronomer on duty (AoD)
for the password. Start netscape only from the mrt-lx2, we had problems
starting it from lx1. And start netscape from the flexible
account so you have access to the html pages such as this one, and to the
observing log, all of which you find in subdirectory ~/htm, which is linked
to /usr/local/httpd/htdocs/flexible/htm, on mrt-lx1, which is accessible
from the outside world for some privileged sites as http://mrt-lx1.iram.es/flexible/htm.
When you work from one of the terminals, other linux machines, or the HP workstations, try to connect to lx1 or lx2 using ssh flexible@mrt-lx2 if possible. There is also a flexible account on the HPs, but it has a different home, so don't use it. After logging into lx1/lx2, you are in the bash shell. If you like to use nic on lx1, type "gag test" right after logging in -before moving into any other shell- this will enable a nic version that Albrecht Sievers and Helmut Wiesemeyer (Grenoble) have been adjusting to the new array and the new fits data format (you can also access nic on mrt-ux2). If you want to look at the data or store your personal files then create your own directory on the flexible account under your name. If you like to reduce your data with nic, make a subdirectory "nic", e.g., and copy the macros ini117.nic, ini37.nic, fch.nic, nic.fit, and nic.sky from ~/bertoldi/nic. When you open nic (in your nic directory), first call ini117.nic if you are working with the 117 array: nicIf your receive an error message that it cannot locate any data files, it is likely that you are using a wrong nic version, you probly forgot to type "gag test". While observing on the obs console, you should keep a nic session open in order to reduce the pointings and skydips with it. For this one, start nic from ~/bertoldi/nic. With nic you can take advantage of a simple skynoise subtraction, which will improve the quality of your pointing considerably especially for faint pointing sources. See the section regarding pointing below. Here you should also append the nic.sky file with the skydip results using the macro @bt <scannumber> after every skydip. On the lx1/2 machines' flexible account you find links to the data directories. If you like to see the fits data directory, cd fits, if you like to check out the abba raw data, it is in the linked directory abba, and if you want to see the whole data directory for all accounts, go to the link directory data, from where you can cd flexbol to arrive at the flexible bolometer data directory. In there you see the raw data directories named after the observing date, and also the fits and abba directories. You should frequently check these two directories to make sure that the data you are taking is arriving here. Sometimes the processes that copy the data from the VAX and the ABBA PC to mrt-lx1 may break down, there is no more data arriving, and you should bring this to the attention to the operator and AoD. See below for a detailed discussion of the data flow and possible problems. After you logged into the flexible account, you may prefer the tcsh instead of the bash. Type tcsh, and the files .cshrc and .frankrc are automatically executed, which set a number of aliases. You can add your own favourites to .frankrc if you like. A few good ones are "l" which is "ls -FA" and "lct" which shows you a recent files directory listing sorted by time, and "b", which sets you back to the directory you were in before. Information on each project is located in a separate project subdirectory, named by the number of the project followed by the first three letters of the PI. The web pages for flexible are in htm, the observing logs in htm/log. In netscape, check out htm/index.html. Start observingTo start observing: at the observing console (after confirming with the operator) push the OBS icon, log in as observer, and answer for project name either flexbol or flexhet, depending on whether you want to do bolometer of heterodyne observing. Then type OBS, answer the questions (observer: your name, operator: Manolo, Mariano, Teresa, Juan-Luis, Fernando, or David and then initialize the bolometer session by running the macros@wstartDo not forget these initializations, else nothing will work! Make sure that the obs display shows something like Frontends: MBOL2 250. GHz BolGain 10 Wobbler 43 0.52where the wobbler throw may be between 32 and ~70 (arcsec). The gain should always be 10 for the 117-array and 50 for the 37-array, except when you observe a bright planet, for which you reduce it to 2 or 10, respectively. At the end of a session, type SAVE LAST and ask the operator before you EXIT. Unless you want to switch to heterodyne observing or for Tuesday morning maintenance (maybe even then not), you need not to leave the session. If you move to heterodyne observing however, you need to leave obs and better log off the IRAMEA observer account and log in again, and then answer flexhet as the project, which sets up the proper parameters and directory. Note that we observe from the IRAMEA vax. There is a parallel vax called IRAMEF, on which you should log in to access the flexhet or flexbol directories which contain the catalogs and observing macros. To edit a file there you must use the edt editor, from which you exit by typing Ctr Z and quit (quits without saving changes!) or exit (saves changes and then exits). If you are unfamiliar with the edt editor, ask for assistance. Pointing and focusWhen you start after a longer break, pick a strong source such as a planet or secondary calibrator (there is a list of them posted near the obs console), and point and focus on it before you go to a fainter pointing source. Do not start a session with an onoff or a map, since the VAX then will not know the reference channel, which is set up with a pointing or focus. In case you are using the 117-array, the macro wbolo117.pro that you ran when starting the session has aliased the commands:p40 --> point 54 /length 40Do use these aliases if possible, and note that the commands point and focus by themselves will display channel 1 on the ret screen! If you use p40 or focus54 the VAX is told that the reference channel on which we like to point and focus and later do onoffs is channel 54, which is the central array channel in the numbering scheme of the ABBA backend. Note that in the fits data as you will see it in nic or mopsi, this is channel 53 (!!)-- the numbering is slightly different. ABBA channel 39 is a total power channel, i.e., one of the bolometers (the central one in this case) for which the signal is amplified without a high-pass filter (a capacitor), so that it reflects the actual power on the bolometer and thereby the radiation intensity. The normal channels do not show the total power signal, since with the necessary resolution the strong sky power variations would require a dynamic range in the digitalized signal too large to handle. For this reason a capacitor is placed in line, serving as a high-frequency pass filter that blocks the strong low frequency amplitude variations. The three total power signals (channels 118-120 in the fits files and in nic/mopsi) have lower amplification and resolution to stay within the bandwidth of the backend. We use the total power signal for the skydips (bolotip) to determine the zenith opacity. We have dicovered that the central channel of the 117-array is not very sensitive, but that channel 1, which is offset by -37" and -22" from the center, is ~30% better. Therefore for onoff observations we from now on use channel 1. This complicates observations a little bit. When you map, you should point an focus on the central channel 53, and then map, but if you do onofs, then point and focus on channel 1 (it is one in both the ABBA and fits order). In order to tell the telescope to point on channel 1, you need to give it a receiver offset, for which we have an alias: roff1 == set recoff -37 -22and then use the above aliases to point and focus, p140, p160, focus1 etc. To move back to channel 53 type the alias: roff54 == set recoff 0 0A normal pointing is 4 times 30 sec long, i.e. 2 azimut scans, 2 elevation scans on the source, with a scan length of 60 arcsec. If you are fairly well pointed already then 40 arcsec scanlength may suffice. This gives you better sensitivity since the 30 seconds are spent on 40 arcsec instead on 60 arcsec. We noticed recently that under unstable conditions you may get a better pointing when you run it twice as fast and for 8 instead of 4 subscans. This better averages out the baseline drifts. You are sampling not as well, but sufficiently. The pointing command p408 = point 54 /len 40 /time 15 /sub 8takes only about 10 sec longer than p40, coming from larger overheads between subscans. Try it out sometimes and compare it with p40. To load a pointing source, you need not to type in the entire name, just enough for the source to be found properly in the source catalog. E.g. you can type 0221+ for the source 0221+067. However, you can not type 0336- for 0336-019, since - at the end of a line indicates that the line continues. It will also not suffice to type only 0336, since there is a high chance that it will find this string in another source, eg. in the coordinates. Always make sure obs found the correct source before you go on! If your target is far from a pointing source (e.g. > 5 deg) then watch out for gradients in the pointing model, i.e. the pointing correction! Better pick two pointing sources on opposite sides of the target and then interpolate the pointing corrections. Especially for faint (<4 mJy) targets a reliable pointing is absolutely crucial for a good observation! If you are not sure you are pointed well, do not observe at all! Starting this year two of the telescope pointing model parameters, P4 and P5, are automatically upgraded every time you move the telescope by more than 60 degrees in azimut. These two parameters describe the horizontal levelling of the telescope. They are computed from readings from two inclinometers. If you notice that despite moving the telescope by more than 60 deg in azimut recently there appears no message that P4 and P5 were upgraded (by a few arcsec typically), then notify the operator or AoD. It may happen that one of the processes (calcp4p5) on either IRAMEA or IRAMEF that take care of this are not working properly, and they may have to be restarted. If they do not work, then the pointing of the telescope deteriorates. If they work, then the pointing tends to be pretty good and stable. If you notice large (> 5 arcsec) gradients in the pointing corrections as you move around the sky, something is wrong -- start complaining! A way to check whether P4 and P5 were updated recently is to log into IRAMEF as observer, project pointing, then "sd [.inclin]", and "run check_p4p5", which will list the most recent updates with the corresponding date. A pointing source catalog is linked to these web pages, and a hardcopy of it should be lying on the observer desk. The flux values in there give you an idea of the intensity of the source. A strong source has > 1 Jy. In very good weather you can still point at a 200mJy source. In order to get a better pointing reduction, you can open a small mrt-lx1 window on the observer console, login as flexible, type gag test(note that these ini macros read in a scan vonr the fits directory --make sure this scan exists, else the ini does not work. if the file was muved, replace it with some other file) From now on you can use two macros, pp.nic. pp1.nic, and ppp.nic / ppp1.nic, which reduce pointing scans, with or without skynoise subtraction, respectively, e.g. @pp [scan number] (evaluates pointing on channel 53)or @ppp [scan number]where [scan number] is the scan number of the pointing scan you just did. The procedure pp.nic will reduce this pointing applying skynoise reduction, which will typically improve the S/N by a factor 2 or more. Thereby you get a more accurate pointing correction on faint sources. @pp may take 10-30sec to run. For sufficiently strong sources, you can run the faster @ppp, which does not do a skynoise subtraction and should therefore yield results almost identical to those displayed by obs. Another task of @pp and @ppp is to give you two strings with the pointing corrections and the fit parameters, e.g. 10.4 10.2 420 461 mJy (0.26)You should grep the first string and paste it into the log. This is important information to have in the log. The value in parenthesis is the tau correction the program applied to compute the source flux. It takes a recent value from the nic.sky file. If you have not reduced any skydips recently using nic in ~/bertoldi/nic, then it will not find a recent value, and applies no tau-correction, so that there will be a (0.0) appearing. (therefore after every skydip, in your nic session run the procedure @bt [scan number]which will add the tau to the file sky.nic). Then the computed source flux is lower than the actual one. The second string you can grep with the mouse and paste it into the obs window, applying this pointing correction. For shaky pointings, I advise you to use the corrections from @pp, not the ones computed by obs --unless you see that it did not fit the sources well-- have a look at the plot that appears first. Often the pointing correction suggested by the OBS program is reasonable and you might want to adopt it, typing cor f fthen start a focus, and in the mean time you can run @pp to get a better correction from nic which you apply when the focus is finished. Look at the fit and make sure that the corrections make sense! Sometimes the fits are crazy. Do not focus on faint sources because the result can be wrong. As a rule of thumb, for a heterodyne focus the source needs to have more than a few Jy. For bolometer observations, they can be fainter, but should have more than 500 mJy. Note that if the focus was far off, then it is better to repeat the pointing, since the pointing somewhat depends on focus, and vice versa. Note also that the counts you see on the ret for the focus should agree with the peak count you saw in the pointng -- if they are very different, something is very wrong (e.g. you are displaying the wrong channel!) You adopt the suggested focus correction by typing cor * * fwhich leaves the pointing corrections untouched and updates the focus to the suggested value. Watch out that the focus correction is not too large, since in this case the fit may be inaccurate. Then better lower the correction by hand, else you can overshoot the correct focus, if you simply adopt the suggestion. This may sound confusing, but you will know what it means after a while. Check the focus frequently especially during sunset and sunrise, every 30 min if possible. At other times it may only be necessary to check every 1 to 2 hours. Sometimes it is very stable, other times not -- one never knows. It is very important to keep a good focus, else your observations may be useless!!! Focus only on sources strong enough. If the pointing showed the source at high S/N, then this source is good for a focus. If the pointing looked shaky, dont try to focus on this source and look for a stronger one. If you cannot get a good pointing or focus, do not continue observing - the results will be useless (unless you just want to detect a source and dont care about the flux). To point and focus on a faint point source, you should use a wobbler throw of 32", but if the source is strong, or extended (as many of the seconday calibrators), use 42, or even 50". Minimize the telescope idle timeBetween pointing, focus and getting to your source, plan ahead!!! It is easy to waste a lot of telescope time if you only start to think about what to do next after a previous process is finished. It would be better if you knew what to do next before it is finished. This way you minimize the telescope idle time. Also watch out when a longer onoff or map is finished. The telescope does not alarm you, but instead sits idle patiently until you noticed. I (FB) left a small kitchen alarm clock near the obs console. You can compute how long a map or an onoff loop lasts, then set the time, and take the clock with you as you stroll to the kitchen, your room, or wherever, and the clock will remind you when the observation is finished. (1 second observing time costs about 1 Euro!)The OBS ConsoleHere is how your obs window might look like when you have set up the numerous windows:
On the upper left: MonScan, which you can start with the left mouse by clicking "MRT Monitoring". It will start a number of windows, all of which except this one you can iconify again. On the upper right is the Netscape Composer window with the observing log you keep. Below that on the right is nic running on mrt-lx1, using the procedures @pp and @ppp to look at pointings. On the lower left is the obs command window. It comes up with an unnecessarily large font (the one you see on this image), and you can decrease the size of the window with Control(left?? mouse button), selecting medium size. That gives you more space for the other windows. The obs window is rather crowded. I found the order shown here a most useful arrangement of this limited space. On the other windows that you access (see bottom right of hte image) we run xephem in one, and netscape in another. Open the observing logOn the OBS console open a remote window to mrt-lx1 and log in as flexible. Then "cd htm" and "netscape index.html" A netscape browser will appear with the cover page of the flexible program. Look at the links to the observing logs. There should already be a link to the new date, showing an empty log. If you dont find one, use log0000.htm, which is in the same directory, and copy it to the proper name, e.g. log1224.htm, then open it in netscape or right away in the netscape composer. If you want ot edit a page open in the netscape browser, select "File > Edit Page". It is not difficult to use the composer. Open a new log every day -- we usually start a new one around midnight. Netscape is configured that it automatically saves the page you edit every 5 minutes. You can also save it by typing ALT-s. Look at older logs to see how to fill in the information. You dont need to always fill in LST and UT, one is enough, the othr one, once in a while. Do fill in the pointing correction and the focus every time you focus or point. For onoffs, note the RMS of each scan (or most scans) displayed on the screen to the left of the obs console after the scan. Its value gives you an idea of the sky-noise level. Write as many comments as possible. It is always interesting to know these peripheral info about whatever later on to assess the quality of the observation. Note when Nitrogen was refilled, when the system crashes, clouds move in, etc.The logs for the heterodyne projects are in the directory /vis/flexible/logs-het. There is one log for each project.An example for the log-file is log-nnn-yy.txt. Please look at this file in order to see what to fill in the logs. CatalogsThe source catalog files are on the IRAMEA or IRAMEF (they share the same directory) in the observer flexbol or flexhet directories. Our convention is that for the bolometer projects, the catalogs are called by the program number only, e.g. 150.cat for the project by Omont. The corresponding edb catalog file for xephem is 150-omo.edb. The directory on the lx1 or lx2 flexible account with information on this program is called 150-omo. For heterodyne programs, the catalogs are called model-131-01.cat, e.g. The edb file and the directory name on the linux machines still has the same convention as for the bolometer projects.There are a number of pointing source catalogs. The standard IRAM pointing sources are listed in ibol.cat. The flux of the sources in this catalog is monitored regularly . The latest fluxes can be found in www.iram.es/Telescope/pointing_results.txt. (There should also be a printout of this file in the pointing folder, but it is not always up to date). The secondary calibrators are now added to this catalog -- they used to be in a separate second_calib.cat. A larger number of mostly fainter pointing sources are found in alok.cat. The corresponding edb catalog is divided into alok-500.edb and alok-100.edb. Latter are divided into sources with over 500 mJy at 1.2mm as measured in a MAMBO survey a few years ago, and those between 100 and 500 mJy. Note that these quasars are variable and may be stronger or fainter than advertised. The Alok Patnaik survey went only to +20 deg declination. If you are desperate for a pointing sources between -20 and +20 declination, check out alok-new.edb. Here I selected all the VLA 8 GHz sources with fluxes > 1 Jy. Many of them may be strong enough to point at, but we dont know yet. If you try one of them, make a note in the log that this is a new pointing source, with a flux or an upper limit. It is easy to point on a source with >500 mJy. It is more difficult, but not impossible to point at sources down to 100 mJy. You may need to double the subscan number to 8: point 54 /sub 8 /length 40 which takes 8 times 30sec. To look for pointing sources, use the xephem tool that is running in one window on the OBS console. Look for a source near your target. You may want to first visit a securely strong one and use that for focus as well, and then move on to a fainter source nearer to your target and see if you can get a decent pointing on it. On-off observationsStarting March 2002 we do onoffs with the 117-array using channel 1, which is not the central channel. It is located in the second ring away from the central channel. Set the receiver offset to -37 -22, then point and focus, then start the onof:set recoff -37 -22 ( alias: roff1)We recommend that you do not spend more than 30 minutes at a time on a source. Then check the pointing again and move on to another source. Come back to the first source at a later time, best some other day. This ensures that systematic noise patterns on the array average out better. For point sources use a wobbler throw of 32-34 arcsec, unless they are extended, such a secondary calibrators, for which you better use 50". For onoffs on exended sources (a comet, e.g.), you may use up to 70 arcsec, but note that the efficiency drops the larger the wobbler throw is. We use 10 second long subscans before nodding the telescope. For scans with more than 12 subscans, we recommend the asymmetric mode. The number of subscans per scans depends on the weather. If it is stable an low skynoise, use 20 subscans (5 minutes actual observing time). If the weather is unstable, because of clouds, e.g., then use shorter scans, as you can later discard those with clouds in them. Use 12 or 16 subscans then. Do not start an onof observation with a loop right away. Do one scan by itself, and if all looks well, start a loop. Here is how you start: source mysource /cat mycatnow check if all is running well and fill in the log. Often you will notice that you fogot to set something properly, such as the wobbler throw or the gain. Thats why you should not start a loop right away. You can also use the macro onf.pro: @onf 20which is equivalent to the last 3 lines above. After the first scan, you can set up a loop: for i 1 to 5and go for a walk for 25 minutes -- and maybe set the alarm so you return in time. If you want to have a quick look at the data, on mrt-lx1 you can use the MOPSI script oo.MOPSI in ~/bertoldi/calib. cd ~/bertoldi/calibyou will get a colorful plot and in the end a list with the coadded signals and rms, unless someone has fiddled with the macro to make it do something else. Make usre you look for the signal at the right channel, 1 or 53, depending on where you pointed. MappingMaps should not take more than 1 hour, since then you need to check the pointing again. That puts a limit to the map size. For deep field mapping we usually make maps of size 240 x 220" with a 4" spacing between subscans and a speed of 4"/sec. To start such a map, first point and focus on the central channel, then:source LH-0-0 /cat 172This map has 220/4+1=56 subscans of 60 sec each, and thus takes about 60min including overhead. This map is scanned from low to high elevation. If you want to start at high elevation, then use map 240 4 220 -4 /time 60The second argument to the map command it irrelevant -- it used to be the speed, but that is set not through the time of a subscan. Since we map in azimuth-elevation, in equatorial
coordinates the map diverges in one direction, depending on whether the
azimuth step is +4 or -4. If you mosaic with many coverages, try to have
the different overlapping maps as orthogonal as possible, by observing
at different times, or at +4 and -4 steps.
Maps with subscan spacing of 4" are fully sampled for each channel. If you want full sampling only for the whole array you may set the spacing to about 24", which is the spacing of the rows in the array plus 4". Thereby you can scan a much larger area evenly in an hour. A map with map 600 4 460 24 /time 150will make 21 subscans of 150sec and take about an hour. You could also scan at 6"/sec, but i dont recommend that for deep imaging - though we have not really investigated the difference in quality. Note that you will not see the subscan data displayed on the RED screen any longer during a mapping observation. We avoid the problems we had with long subscans by not transferring any data to the VAX, so the VAX does not need to buffer the telescope data during a subscan, which did cause the problem (the memory was too small). If you want to (and you should) look at the data during a map to make sure all is ok and the weather is appropriate for mapping, then reload the scan in nic, mopsi, or FRED once in a while. It should suffice to display one channel, e.g. the reference channel 53: nicFor how to use MOPSI or FRED, see below. For mosaicing a large field, create separate target positions in you catalog, such as in the following example. Do not use offsets if possible. INDIVIDUAL: SNAM SLAM SBET Calibrator ObservationsTry to observe a primary (planets like Mars and Uranus) or secondary (HII region or YSOs such as W3(OH), CRL618, HL Tau ... [see list near the obs console and Ute's memo on secondary calibrators]) calibrator regularly, at least one per day. For this, point at it, focus, followed by a short onoff with 12 subscans and a wobbler throw of 53", or if the source is very extended, a somewhat larger throw of 60".Once in a while (ideally
once a week) we need a map of a strong point source in order to
measure the array parameter, i.e. the relative position of the channels,
and their relative gain.
The minimum size of such a map is given by the array size, 240" for the 117-arry, plus 2 beams, plus the wobbler throw in scanning direction. Thus for a 42" wobbler throw, the min. size is 240+22+42 x 240+22 = 304" x 262"and to be on the save side, around 320"x280". If you want to do this quickly, then run the map with 6"/sec and 6" spacing, but better results would be obtained with 5" spacing and 5"/sec, however the problem with this is that the map takes over 1 hour and pointing drifts may affect the array parameters. For a calibration map with the 117-array first point and focus, do a short onoff and then a map: set wob 42or: map 320 4 280 5 /time 64 (takes 64 minutes)Such calibration maps should be done only under very stable weather conditions (high tau is ok though) at the highest elevation possible for the planets, or if you use a strong QSO, somewhere between 50 and 70 deg. The source should be point-like (e.g. Uranus and Mars, but not Saturn or Jupiter!) and stronger than 10 Jy. There is one pointing source QSO, 1253-055, which is currently ~15 Jy bright, and thereby good for a calibration map if the planets are not visible. Absorber ObservationsIn order to monitor the minimum noise level of the bolometer system, especially of the pre-amplifieres, we need to make observations of some source with minimum noise, for which we place an absorber foam with a metal plate at the bolometer entrance and take normal short observations. This should be done regularly, usually during bad weather when you cannot observe, but can still move the wobbler, and if possible the telescope. Note that the system will not record data if the wobbler does not run -- a simulated wobbler signal does not yet work properly.First ask the AoD or operator to place the absorber at the bolometer entrance. For the 117-array, there is a square plate with two 2 round black absorber plates, one of them glued on the plate, which fit below the cryostat at the entry hole. It must be moved in there CAREFULLY, and fixed with some normal scotch tape to hold it tight. For the 37-array, place the absorber in front of the bolometer window and tape it tight so it cannot vibrate. The solid plate must always be outside, and the absorber material between the plate and the bolometer entry hole To take an absorber "map", choose some source in the sky, doesn't matter which, any pointing source is fine, and replace its name with "absorber". Then, after a quick "pointing", take at least 2 short 3 subscan map with 100 sec each subscan, which takes 5 minutes each. If you cannot move the telescope since it is parked, that's fine. Ask the operator to simulate the observation. source 1226+023 /cat * (any source in the sky, elevation >5 deg, will do)It is also interesting to see what happens when the bias voltage of the bolometers is turned off. There is a switch on the pre-amplifier box (on top of the bolometer) which says "bias". Turn it off, and repeat the absorber measurments. For a few minutes before and after the switching, the preamp signals will drift to some new equilibirum, which you will see when you look at the single pahse signals --not the phase differences. Try to do absorber maps whenever possible. The more the better. They will warn us about possible electronic problems, and give us a bolometer sensitivity measurement. Data display with FREDTo take a quick look at the data, you may use an idl routine package called fred, which you find in ~/bertoldi/idl/fred1. On the mrt-lx1, go to this directory, (open fred.pro in emacs) start csh and set the Display with: setenv DISPLAY mrt-xxx:0.0 . Start idl and then:@makeand three windows should appear showing you the RMS of the data in window 3 (three points for each channel: cross and green diamond for the single phases, red triangles for the phase differences), the signals of phase 0 or 1 in window 1, and the correlation between channel n and channel 1 in window 2. How many channels are displayed depends on what is set in fred.pro as chan_min and chan_max. The plot range of window 1 is given by the variable ymax. A baseline of order base_order is subtracted before display. In successive calls you need not to use the flag /dis, which tells fred to open three windows if they are not yet open. So to show another scan, fred,<scannumber>If the plot range is not appropriate, then change it in fred.pro and recompile this routine with .r fredIf you like to zoom in, you can call the display routine directly: signal,0,1,4,color=1,xl=[0,100],yl=[-5000,5000]which will display phase 0 for channel 1 to 4 from time 0 - 100 sec and signal range -5000 to 5000. If you like to plot the pase differences, then add the flag /pd. If you like to display the total power channels, then type signaldc,0or signaldc,1 for the other phase. Look in the fred.pro for the usage of the other subroutines - if you are familira with idl, then most of it should make sense. FRED is a package under construction, started in december to simply display the 117-array fits data while the other programs nic and mopsi still did not run properly. If you are interested in contributing to the development of FRED, contact Frank Bertoldi. Data reduction with MOPSIFor a quick reduction of onof or map data, we provide two scripts which you find in ~/bertoldi/calib. For onof, use oo.MOPSI or oost.MOPSI, for maps use map.MOPSI. Start mopsi from mrt-lx1 only. oo.MOPSI is intended to co-add scans of faint sources, whereas oost.MOPSI (on-off strong) if for single scans of strong sources such as calibrators. Former applies higher order baseline fits, spike blanking, and correlated noise reduction. Latter only fits a 0-order baseline. Both routines correct for opacity. The taus are written in a file ~/bertoldi/master/taus_02.TAUS if someone kept up reducing them (using the mopsi script ~/bertoldi/master/sky.MOPSI -- this is tricky, so if you are not familiar with mopsi, dont try). If the taus are not available, you can set them explicitly in the oo macros. To reduce onoffs, start mopsiEtcTo print a text file from the VAX use lw_text.The flux calibration is 35 ct/mJy for phase differences for the 117-array, and 4.2 ct/mJy for the 37-array. |
