Subject: March 15,2000 Date: Thu, 16 Mar 2000 00:20:30 -0700 (MST) From: ank@apo.nmsu.edu (Atsuko Nitta Kleinman) Wednesday Mar 15 2000 Apache Point Observatory SDSS 2.5m Observing Log (MJD 51619) OBSERVING TEAM: =============== Dan Long (APO) Steph Snedden (APO) Atsuko Nitta Kleinman (APO) Bryan Laubscher(LANL) PT VISITORS: ========= Don Holmgren (FNAL) Robert Lupton esq. (PU) Chris Stoughton (FNAL) THE PLAN: ========= Exercise the DAs and break things. SYNOPSIS FOR NIGHT: =================== We exercised the DAs and tested the MCP/TCC communication. TELESCOPE STATUS: ================= Nothing mounted. DIMM: ----- INTERLOCKS: =========== no interlocks bypassed. SOFTWARE USED: ============== IOP/SOP: v3_11_1 Watcher: v2_7_1 MCP: v4_0_1 WEATHER SYNOPSIS (see below for hourly data): ============================================= CLOUDY Wind Time Temp F Dewp F MPH Direction Dust Sky Precip. 16:01 51 7 25 265 (W) 787 17:01 50 6 18 234 (SW) 888 18:01 45 3 19 264 (W) 950 19:01 43 4 19 262 (W) 848 20:01 44 6 14 266 (W) 818 21:01 43 10 21 268 (W) 1024 22:01 42 11 17 283 (WNW) 3603 23:01 42 12 20 290 (WNW) 5027 00:01 41 12 17 310 (NW) 4094 GENERAL COMMENTS (see Problems below for details): ================================================== We followed the procedure in the document (a new and improved cheat-sheet which is located at http://sdsshost.apo.nmsu.edu/iop/iop.procedures.html) Chris is writing. We cut new version of IOP using the above procedure. DA-fest ------- We put fresh tapes in the drives at about 18:00. Tape drive 6-1's cleaning light came on. We put in a fresh tape without cleaning the drive and it came up correctly. We set up the archiver and went through the preparation for observing checklist. This is part of Chris' document "iop.procedures.html". Changes to any procedure will require updating and checking in this document via CVS. We then started taking bias drifts (with camera in the doghouse) and Don deliberately caused some drives to fail, so we could practice recovery procedures and check that auto-recovery worked properly. Don is developing a written procedure for the cases where statusDump archEnable comes up 1 (archiver enabled) or 0 (disabled). We ran endNight 51619 and it completed successfully. TCC/MCP Communication Test ran by Dan and Robert ------------------------------------------------- Robert and I spent some time this evening working on TCC/MCP communications. We found a bug in how the TCC command AXIS INIT is handled. Apparently the MEI buffer was not being flushed. The consequence being that if a slew was in progress when an AXIS INIT was issued, the move continued even after the INIT. This has been fixed and will show up in the new version of the MCP. See PR 902. Some testing was also done on how superceded slews were handled. The following were found to work properly. Slew superceding a slew in progress Rotate command superceding a rotate command in progress Rotate command superceding a slew in progress. The one combination which did not work was Slew command superceding a rotate command. When this was done, the rotator would slow to a stop and not respond to further commands to move. An axis init was required to clear the problem, but curiously the state of the rotator showed it was still in closed loop. See PR 903. Some more exploration of the known bug of azimuth and rotator oscillations was done. The problem is that when at or near zenith there are sometimes oscillations in the azimuth and rotator axes. The behavior was further defined tonight. The oscillations only occur when both axes are in motion. If either axis reaches its destination and stops, the oscillations stop. This is true when the TCC is controlling the slew. It is also the case when the MCP alone requests the move. It happens over a wide range of acceleration and velocity limit parameters. It seems that when azimuth and rotator are moving about nearly the same axis, they begin to interact with each other, going into oscillation. This interaction apparently occurs at a fairly low level since when the commands are issued from the TCC or the MCP, the behavior is the same. If the update rate of the servo loops for azimuth and rotator is the same or are multiples of each other, I can see how a resonance gets started as an error correction in one causes an error term to be created in the other. Is it possible to make the servo update rates for azimuth and rotator enough different from each other that it becomes difficult to excite the resonance between them? See PR 904. Finally I spent some time tonight exploring how different velocity, acceleration, and jerk limits effected telescope performance. I found that azimuth and rotation velocity limits could be increased significantly without causing amplifier trips, ( changes in ) oscillation behavior, or running into the windscreen. Acceleration was increased for the rotator, and jerk was reduced for all axes. Here are the old and new values for axelim.dat and the relevant values of default.dat Was Axelim.dat Default.dat Az El Rot Rot Vel 1.4 1.25 2.0 1.0 Accel 0.1 0.1 0.1 0.05 Jerk 0.4 0.4 0.4 0.1 Is Axelim.dat Default.dat Az El Rot Rot Vel 2.0 1.5 2.0 2.0 Accel 0.2 0.1 0.25 0.25 Jerk 0.2 0.2 0.2 0.2 DATA ACQUISITION (see IOP/SOP output below): =========================================== IMAGING: -------- 1268: bias 1269: bias 1270: bias 1271: bias SPECTROSCOPY: ------------- 3139: bias MIRRORS: ======== PRIMARY: -------- MIGS SHOULD BE Axial A= 1.9315 1.9315 Axial B= 0.5565 0.5570 Axial C= 0.5565 0.5570 Transverse D= 0.4085 0.4075 Lateral E= 1.8985 1.8980 Lateral F= 0.3985 0.3990 GALILS Commanded: 000004850, -000002350, 000000100, -000009500, 000000050, -000000100 Actual: 000004850, -000002350, 000000100, -000009500, 000000050, -000000100 SETMIR VALUES primDesOrient: primOrient: SECONDARY: ---------- Focus:0.00 Temp:5.7C Altitude:30 MIGS Axial A= 0.1110 Axial B= 0.1500 Axial C= 0.1590 Transverse D= -0.0800 GALILS Commanded: -000176900, -000189800, -000200050, -000006250, -000011000 Actual: -000209977, -000216025, -000247363, -000006250, -000011000 SETMIR VALUES secDesOrient: secOrient: LN2 SCHEDULE: ============= Imager dewar swapped in the afternoon Spectro dewars filled at 16:30, 0:30 PROBLEMS: ========= WEATHER DATA: ============= see WEATHER SYNOPSIS (see below for hourly data): DATA LOGS FROM IOP/SOP: =======================