Date: Sun, 11 Feb 2001 06:58:28 -0600 From: "Eric H. Neilsen, Jr." Subject: 2.5m obslog 20010210 (MJD 51051) APACHE POINT OBSERVATORY SDSS 2.5M OBSERVING LOG Saturday February 11, 2001 (MJD 51951) OBSERVING TEAM: =============== Jurek Krzesinski (APO) Eric Neilsen (FNAL) Scot Kleinman (APO) Alan Uomoto (JHU) PT VISITORS: ========= Mamuro Doi (JPG) PT Masataka Fukugita (JPG) PT THE PLAN: ========= Shakeup! Check that the exposure times in the headers is correct for altered exposures (PR 1793) and that the calibration lamps are properly turned off when exposures are aborted (PR 1812). Build a new fiducial table, and obtain a pointing model. If the weather seems if it can support it, make another attempt at an astrometric calibration scan. SYNOPSIS: ========= Checked out corrections to SOP: exposure times in headers for aborted and modified exposures are still off, and lamps and flat field screens are turned off / closed appropriately if a calib is aborted because not all the lamps come on. We attempted to build a new fiducial table. Taking the data went well, but the data for the Az and Alt looked very strange, showing a strong, nearly linear shift with time. As the Rot fiducial table was the only one we felt confident in using, and was very close to that already in use, we did not install new fiducial tables. We found that the logs which should be copied to commish automatically are not (PR 1636 is still not finished off). goDrift now acts appropriately (and informatively) when the imager amps are on (PR 1765 appears to have been addressed). GENERAL COMMENTS (see Problems below for details): ================================================== New code to get the exposure time in goStare from the spechMech's causes goStare to fail. I modify goStare on the branch, and it seems to work. I tag and install v3_29_2 with this fix. (More details are in the "Problems" section, below.) If you get logged out of the laptop by the telescope, logging back is made interesting by the color map of the login screen. (Text and background seem to be the same color.) We open the doors and 5:15, and remove the building at approximately 6:30. We begin the night by installing cartridge 2, plate 413. As we were preparing to observe (at 19:35), the watcher interlocks display reported a rotator problem: rot_motor_VI_R_check. The mcp menu shows everything to be fine, however. Shortly thereafter (19:39), the interlocks display shows an alt error as well, but the button turns green before I have a chance to track down what caused it. Forcing an update of the interlocks display causes the alt error to return. The alt problem is tracked down to: alt_motor1_VI_R_check. The mcp menu still shows no problems (no breaks or aborts, |voltages|<500). We try sending the rotator from 0 to 1 degree using the track command in a telrun session, and it works without a problem. The mcp menu still shows everything as being okay, and the watcher is still complaining (despite a forced update). As everything seems fine, we march bravely on despite the interlocks complaint. At 20:07:07, we notice the interlocks display is all green once again. Note that a grabInstruments seems to be required to run goStare, even if we are only taking a bias. This is a serious problem, if we are required to run the goStare in every SOP session! I speculate that the new mechanism for getting the exposure time requires asking the mechanicals. As 0 second exposures should not be interrupted, and changed, simply assigning an exposure time of 0 when a bias was requested may be a reasonable solution. We go to to the field for plate 2, but cannot even find the fk5 star. Given that we don't need to have aquired the field to do the tests we need, we go strait to the tests. (Our inability to find stars may just be due to the clouds, which have arrived in force.) We start exposure 8646 with doNextScience 60, and stop it with and endStare. The header of exposure 8646 shows an exposure time of 0, still incorrect. We start exposure 8647 with doNextScience 60, pause it with pauseStare, and resume it with resumeStare 90. Using exposure remaining, I'm not completely convinced that it really did extend the exposure time. The header of exposure 8647 shows an exposure time of 60. I do not know if this is correct or not, as I am not convinced that resumeStare really changed the exposure time. To test if resumeStare really is altering exposure times, we start exposure 8648 with doNextScience 60, pause it with pauseStare, and resume it with resumeStare 240. The header of exposure 8647 shows an exposure time of 60. It based on exposureRemaining command issued before and after each of the above, it is clear that resumeStare 240 did extend the exposure time, but not to 240 seconds. Perhaps it extended the time to the end of the exposure to 240 seconds from the time of the start of the exposure, including the time that the exposure was paused in that 240 seconds. To test PR 1812, Jurek went down to the telescope and unplugged one of the flat field lamps, and a preCalib was attempted. It prompted appropriately when it could not turn them all on. When "y" was typed at the prompt, it waited longer, as it should have. When "no" was typed, it turned off the flat field lamps, opened the flat field screens, and aborted appropriately. I'd still like it if it recognized a spelled out "yes" as yes, but otherwise I think this is working fine. We then attempted to build new fiducial tables. Taking the data for the fiducial tables went smoothly. The begin and end times of our fiducial sweeps were: Axis Begin End Alt 981865683 981866668 Az 981867079 981870821 Rot 981871150 981874543 The plots of the rotator fiducial data look like they are in good agreement with the existing model. The scales for the Az and Rot look very close to the existing values as well. When plotted against time, the fiducial errors for Alt and Az vary strongly with time. In the case of Alt, the pos1 error shifts by about 120 encoder ticks each focus sweep. We do not understand why there should be such a large, abrupt shift corresponding approximately to the time we shift from negative to positive motion near the lower limit. The pos2 error appears to be much better behaved. In the case of Az, the pos1 and pos2 errors both shift linearly with time, by over 1000 encoder counts over the course of the sweeps. Because it provides the only readings which do not seem suspect, we build a fiducial table with the Alt pos2 readings. The encoder positions appear to be offset by approximately 60 degrees from the existing (pos1) table. We do not apply the new table. As it appears to be too cloudy to observe anything (no stars are visible) we close. Checking out the new plate database, we do indeed find that a priority field has been added. Most priorities are currently set to 3, with a few set lower. When looking at further information on a plate, the "back" button does not seem to work. The "qa" field now shows up as expected, and even seems to have reasonable values. Next, we check out PR 1636. These logs have not been copied over to commish (at least not since 51843, and apparently not automatically), but they have been generated on 'host. Examining the code, I can see 2 problems. First, it will only ever copy over files from the day before. If for some reason idSysLog is not run the day after the generation of an idSysLog file, that file will never be copied. Secondly, and more importantly, it assigns the origin directory to the variable "old_dir", and then looks in "$olddir/$old_logfile". I am a little surprised we haven't noticed iop complaining about undefined variables as a result of this. I notice that idCCDLog does not appear to have the second problem, but there is no evidence of cdCCDLogs being copied over to commish either. On to testing PR 1765. First, we use: setImagerAmpsOff to turn off the imager amps (at Sun Feb 11 11:37:08 UTC 2001). We then do a goDrift -bias, and the expected menu shows up. We have it abort. We then setImagerAmpsOn, iack, and do another bias. It works as expected. We note that all the id nodes in watcher complained about the aborted image (no surprise). On closing, we notice that the airflow meter at the 2.5m basement (close to the air pumps) shows the upper most value (over 20). Other presure gauges, such as those for the T bar latches, look okay. We leave a note for Larry, who is on site. DATA ACQUISITION (see IOP/SOP output below): =========================================== IMAGING: -------- Run Stripe Strip Flavor Lambda Begin End ----------------------------------------------------- SPECTROSCOPY: ------------- UT Exp Time flavor comment =============================================== OFFSETS: UT Cart Plate getclock Az Alt Rot Scale TAPES: Goes to FNAL: JL3618 Stays at APO: JL3619 MIRROR NUMBERS ============== PRIMARY: -------- Scale: 1.000000 MIGS TONIGHT NOMINAL Axial A: 0.0489 0.0150 Axial B: -0.0630 -0.0510 Axial C: 0.0209 0.0300 Trans D: 3.7500 3.7560 Lateral E: 1.9040 1.9100 Lateral F: 0.4045 0.4095 GALILS Commanded: 4900. -2750. 450. -6650. 8750. 8300. Actual: 4900. -2750. 450. -6650. 8750. 8300. SETMIR VALUES primDesOrient: 0.00 -4.40 10.70 586.50 240.70 primOrient: 0.00 -4.56 10.53 586.58 240.03 SECONDARY: ---------- Focus: 0.0 Air temp: 2.4 C Altitude: 70.00 MIGS TONIGHT Axial A: -1.1460 Axial B: -1.1480 Axial C: -1.2100 Trans D: 0.2790 GALILS Commanded: 1425950. 1425950. 1425950. 9350. 10550. Actual: 1426320. 1424916. 1425432. 9350. 10550. SETMIR VALUES secDesOrient: 1131.25 0.00 0.00 0.00 -328.89 secOrient: 1130.92 -0.02 -0.43 -0.28 -328.84 FOCUS LOG: ========== UT Cart Plate Primary Secondary Scale Focus Az Alt Temp Wind (C) spd. dir WEATHER SYNOPSIS ================ Mostly clear in the beginning of the night, but clouds are approaching from the north west. Very cloudy for most of the night. WEATHER DATA: ============= Wind Time Temp F Dewp F MPH Direction Dust 16:01 39 -1 17 238 (WSW) 584 17:01 35 -2 16 227 (SW) 618 18:01 35 -9 15 188 (S) 1392 19:01 36 -10 19 217 (SW) 2234 20:01 36 -7 12 249 (WSW) 1804 21:01 35 -4 13 183 (S) 1471 22:01 35 -2 15 201 (SSW) 1240 23:01 35 -1 18 197 (SSW) 1115 00:01 36 2 19 199 (SSW) 1268 01:01 34 13 17 193 (SSW) 2246 02:01 34 12 19 221 (SW) 2082 03:01 34 13 14 227 (SW) 2304 04:01 34 8 14 224 (SW) 1699 05:01 34 0 14 244 (WSW) 860 LN2 SCHEDULE: ============= Spectrographs filled at ~12:00am. Spectro dewar weight 164 lb at 05:25. Imager dewar weight 148 lb at 05:25. TELESCOPE STATUS: ================= Stowed at 30 deg alt, cartridge 2 on telescope. Spectro auto-fill connected. INTERLOCKS: =========== none bypassed SOFTWARE USED: ============== IOP/SOP: v3_29_2 Watcher: v2_8_14 MCP: v4_5_0 TPM: v2_2_2 AstroDa: v14_34 TCC: v2.5.0b2 PROBLEMS IN DETAIL: =================== The initial sop "goStare 0 -init -bias" of the night complains: Error: can't read "mechData(sp2_last_exp_time)": no such variable Doing a parray on mechData, I notice that the index is sp2_last_exp.time, rather than sp2_last_exp_time. I change highCommand on the branch so that it uses sp2_last_exp.time, and has mechData as a global variable in goStare. This seems to allow goStare to work. I tag and install v3_29_2 with this fix. DATA LOGS FROM IOP/SOP: =======================