From monnier@umich.edu Sun Dec 15 18:44:25 2002 Date: Sat, 14 Dec 2002 18:13:52 -0500 (EST) From: John Monnier To: rafael@ipac.caltech.edu Subject: IOTA Summary: 2002 Oct 01 - 15 (fwd) IOTA Summary Report : 2002Oct01-15 (visible fiber run) ------------------------------------------------------- JD Monnier and E Pedretti Outstanding Problems/Issues 1. 60Hz (+odd harmonics) on piezo tiptilt signals seen in telescopes before going into piezo-driver-boxes. 2. Oscillations of Tiptilt South Mirror both with tiptilt loop on and off [the problem is currently mitigated through extra damping imposed by foam+shims behind mirror!!]. Beware of this when taking data! 3. Color problems on left-hand monitor of iota17 (can't display startracker and IDL on the same monitor). 4. Telescope A is not well focused. Not sure about Tel C. Unfortunately, its clear that the secondary modules need reworked -- not only are all the axes confusingly coupled, but some of the picomotors can not even be used remotely due to too much force (e.g., tilt pico B on tel A). [old news] 5. netscape on iota17 hangs all the time. Sometimes the xinterface is locked up and one has to log-in remotely to kill all the netscape processes. 6. A number of computer configuration problems. iota16 doesn't seem to see any printers. xv,xfig,ghostview, etc. are either not installed or don't work on some (most) of the computers. 7. IDL GUI's don't work right on iota16. Probably because of a needed motif patch for 2.8. 8. The current computer naming scheme (iota*) is not very useful (and is confusing). Perhaps we can rename some computers -- e.g., controller, analysis, etc. visitor,.. 9. Observing efficiency limited by: a) lack of framegrabber for star acquisition b) no way to do sense or control yaw from computer after "long delay" moves. 10. OT crashes a few times a night, seems to be associated with running scripts. 11. Better approach and methods needed for solving and archiving "Delay System" calculations. 12. A sudden thunderstorm revealed the lack of a proper shutdown procedure. We need posted on the wall a checklist of what equipment to unplug (for lightening protection of sensitive electronics) and also the method for shutting down computers (the NFS server caused most of the computers to hang during shutdown). We fortunately recovered from the harsh power-downs, but it required some unix tricks on iota17. 13. There is one mysterious failure mode which has happened to us twice. In this mode, the fringes just "disappear" and they can not be found again. Even after rehoming long and short delay lines the fringes are gone. We found that we had to reboot the vmes to find fringes again. We don't know what get corrupted to enter this failure mode, but would be interested in hearing other reports. 14. The XDisplay help programs (Fast startracker utility Startracker2) works, but I tried re-compiling it and the recompiled version doesn't work. Something if probably amiss with the X11 libraries? 15. Ettore and I both have noticed flashes of light (spark?) out by the airconditioner units. We don't know what they are from, probably the airconditioners. Not sure if we should be worried about big sparks or not. 16. The disk on iota16 makes a funny noise every once in a while -- like when it is accessed after not being used a while. Doesn't sound normal to me -- like aggressive humming for a second. 17. Ettore thinks the Sun graphics cards are pretty crappy, because there is fair-amount of flicker and other artificts on the Sun flat screen displays. Maybe they are not optimally configured? -------------------------------------------------------------------------- Summary of "solved" problems and other possibly interesting information 1. IDL DAQ works on new hardware and OS upgrade. 2. Eliminated (we hope) IDL hangs during data acquisition (due to usleep() timer command in C-program, not IDL) 3. Sped up fiber explorer by passing shmem pointers in and out of call_external C routines (also should eliminate the crash-problem) 4. Documented that the IR dichroics in use were actually 50/50 beamsplitters, and replaced with correct optics. 5. Documented Altera temperature sensitivity. Control room must be below 73 degs in order to avoid bad pixel during PICNIC reads. ** USE AIR CONDITIONING ** [this confirmed previous experiment in Spring] 6. Timing problem on vme solved (to be confirmed) by Kamal. The clock reads exactly one-second late on the OT, but I hope this is the expected delay (is it kamal?) 7. For the visible fiber experiment, we found calibration more stable if the entire lab is airconditioned during the day. However, be sure to turn off during observing -- air currents dramatically affect fiber coupling stability. 8. Strong improvement in dome seeing by opening louvres and lidflaps well before observing. Opening doors improves ventilation significantly when wind is low [maybe we can install hooks to hold doors open, since they blow/bang around during windy times] 9. The new alignment targets put in during the shutdown are very useful. 10. We got lots of nice visible data, observing sources down to V 5.5. Some very excellent seeing this run, although clouds were a significant bother periodically. 11. The new computer situation is a big improvement (and flat screens allow more space to work), but they are not set up completely yet. [helper utilities like gv, xfig, xv, etc. are either not functioning or installed]. The additional IDL license was very useful. 12. Small user-friendly enhancements to IDL data acquisition program. Fringe display now includes the average power spectrum, in order to help users detect tiptilt oscillations, or/and other problems. Changed default camera parameters to 1MHz clocking and 4mu-sec delay. This was shown by Pedretti (2002Jul07) report to yield significantly better readnoise values (old values: 390Khz,15.330 mu-sec)