GBT Commissioning & Operations Meeting 10 October 2003 AGENDA 0. Special time for next week's meeting (16 Oct, 8:30a) 1. Az Track Status & other telescope news -- Bob A. 2. Observing news -- Ron 3. Spectrometer status -- Rich 4. Spectral Baseline, Front-end, and IF work -- Roger 5. Software status -- Nicole, Mark 6. Schedule -- Carl 7. Project scheduling -- John 8. Any other business PRELIMINARY REPORTS 0. Special time for next week's meeting (16 Oct, 8:30a) Because of the Penn Array Critical Design Review next Friday, the Commissioning and Operations meeting will be held at a special time, 8:30 a.m. on Thursday, 16 October. A video connection to Charlottesville will be available. Phil 1. Az Track Status & other telescope news Az Track: Work continues on the finite element analysis and materials research. No milestones achieved. Structural Defects: A contract has been awarded to Modjeski and Masters to collect and interpret strain gauge data. -- RA 2. Observing news Over the last week we've seen a very wide range of projects for the telescope. VLBI, pulsar, spectral-line, K-band commissioning, spectrometer checkouts, PTCS, proposal checkouts, .... Items for possible discussion are: o Antenna problems are still occurring at about the same rate as I reported in the previous week. o The Spectral Processor may be failing a bit less than before the cable changes but the difference appears to be slight. o A setup error after a Spectral Processor reboot wasn't caught and the observer lost about 4 hours of time. o The holography run had to be cancelled when the correlator showed insufficient power. This may be due to the failure of the "LO Receiver" which produces the 500 MHz and 5 MHz signals in the receiver room. We have scheduled further tests of the system. o A series of problems during proposal checkout on Tuesday and Wednesday night. The items I think we need to note are: 100 MHz rails in the Spectrometer, further failures of the 5 MHz/1 PPS signals from the "LO Receiver" (fixed on Wed.), failures in GO and IARDS, and restrictions due to the recent installation of Red Hat 9. Except for the reported human error, astronomers lost very little time this week. Most of the remaining lost time was during tests of one kind or another. -- RJM 3. Spectrometer status A few instances of 1024-lag-long steps in the autocorrelation function have been noticed both by Karen and Rich. This problem is very intermittent. They happen in only one of two phases in a CAL ON/OFF type of observation, and always between certain lags. This implies that the problem is in or after a particular LTA. On Friday the potentially offending LTA will be moved to see if the problem follows. Some progress has been made in figuring the problems with the High Speed Sampler "improvements". The problems seem to be related to op-amp and resistor noise coupling into the 1600 MHz VCO. (The op amp and associated resistors are changed in this upgrade.) We are presently in the middle of an experiment to see if simply cleaning up the power supplies to the VCO module is sufficient to clean up the 1600 MHz sampling clock to the required level. This would save a bit of work in board modifications. One of the reasons that progress has been so slow is that the only piece of test equipment sensitive enough to thoroughly test the modifications is the GBT Spectrometer. LTA redesign is continuing. These two weeks we have been working on a design for an FPGA adder with Xilinx. So far we have finished 70% of it. Spigot scripts have been improved by Ford. It is now possible to download required xilinx personalities as part of the scripts. This eliminates a tedious step using the Device Explorer. Also, scripts that set the spigot card mode have been rewritten to take a "spigot mode" parameter. This drastically reduces the number of scripts, and should make the system less human-error prone. Plans for next week: - Spigot testing with Escoffier and Kaplan to figure out why the system sometime hangs after certain commands. - Continued work on the LTA redesign and High Speed Samplers. after certain commands. -- Rich and Holly 4. Spectral Baseline, Front-end, and IF work Repaired Q-band amps are now back into the system, and the receiver was cooled on 10/9. Cold testing will resume. It has been a bad week related to the Ku-band receiver. Tests with the spectrometer showed essentially the same noise spikes with the new thermal gaps as were seen previously. This is disappointing. After some thought, we planned to put the receiver back on the GBT so that careful tests on cold sky could be run with the spectrometer. However, during final checks prior to installation, one of the cryogenic LNA's failed. There appears to have been a mechanical bond failure within the amp, but that is speculative until the amp can be inspected. The LNA has been removed and sent to CDL for repair. I have started to write a specification for a S-band notch filter for the Satellite Digital Audio Radio Service (SDARS) downlink band. -- RDN 10/9/2003 5. Software status Single Dish Development IPT #54 - Friday, October 10, 2003 Today ends week 1 of 6 in the current development cycle, which is the 8th cycle in 2003. The Plan of Record for the current development cycle is now available from the wiki at Software.PlanOfRecordC82003. This cycle our detailed software status will be kept online on the Plan of Record page, making it easier to keep up to date. This replaces the status board which was in the 2nd floor hallway. Please visit this wiki page if you would like to see additional details or background information concerning work in progress. The major issue for this week has been attempting to make progress on the spectrometer balancing problems. This was originally started in response to Karen's observing reports, which reported that it was taking far too long to balance the spectrometer. Mark discovered the bottleneck and patched in a fix on Monday 10/6 AM, which was tested and approved by Karen. Ron also reported in his observing Monday night that the bottleneck appeared to be gone. However, Roger experienced problems with his balancing on Monday afternoon, and Toney reported extensive problems balancing on Tuesday 10/7. In response to Toney's issues, version 3.16 of the spectrometer code only was restored and he was able to balance properly. So the bottom line is, v3.17 with the balancing patch provides the solution for K band commissioning, but negatively impacts other observers (particularly those using fast samplers). In order to progress on this issue any further, Ron has agreed to set up a meeting next week to discuss and recommend a solution to the problem, as well as to capture a list of subtleties that must be considered when attempting to balance. In the meantime, v3.17 with the speed optimization will be restored immediately prior to K band commissioning on Sunday, and left in during the week while troubleshooting is done. All those who are negatively impacted during the week are encouraged to balance manually as a workaround. Antenna manager upgrades work initiated last cycle is progressing well and will be made available for PTCS tests coming up in the next couple of weeks. Now the antenna manager directly controls the subreflector in "actuator" coordinates. (Previously the servo system did the XYZ to actuator conversions.) This change is required to allow calibration corrections which only make sense in actuator coordinates. An antenna pointing/focusing database schema is nearing its final form, and the revision of pointing algorthims, pointing residual compensation, and focus control are complete, and being tested. Configuration testing has continued on schedule and additional work, adding prime focus receivers which was scheduled for this week, has been completed and tested for PF1. Work on configuring for the K-band receiver has started as well. Regarding K Band calibration, both Jim and Bob feel that an error in the measurements of Tcal in the lab may be producing errors as large as 50%. They are working with Ron to try and find a reasonable resolution as we prepare for K Band observers and program checkouts starting within two weeks. Operational support this week included changing the LO1 Router so that the 1PPS/5MHz control bit would not be overwritten, and adding a check to verify antenna control. -NMR 6. Schedule Last Week ======== Observations for: BU025, GBT03B-014, GBT03B-037, GBT03B-026, GBT03A-014, GBT03A-016, GBT02A-012, GBT02A-069 , GBT03B-015 Completed: BU025, GBT02A-012, GBT03B-014, GBT03B-037 October ====== Scheduled hours [backup]: Astronomy ~ 290 [147] Maintenance ~ 132 Test & Comm ~ 311 [63] Un assigned ~ November ======= Scheduled hours [backup]: Astronomy ~ 270 [64] Maintenance ~ 102 Test & Comm ~ 230 [9] Shutdown ~ 36 Un assigned ~ 81 -- RCB 7. Project scheduling October 6th Planning Meeting Minutes Rm 241 10:00 A.M. 0) Observer comments O'Neil's report The first two hours were spent at L-band running a checkout for projects 3C-23 and 3B-19. The rest of the night was spent on K-band and spectrometer checkouts. The project checkouts went fine (although further data analysis is necessary). The k-band checkout was rained out around midnight & I only got reliable data for the 18-22 GHz feed. this data also needs to be analyzed before more can be said. Problem Summary: (1) Cannot put spectral processor into standby (2) Spectrometer balance still takes too long (3) Spectrometer typically started first scan 2-6 seconds late (4) Power dip seen in channel R1 at K-band (5) The J1 path for the spectrometer, when run through the high speed samplers produces garbage. Minter's report Summary: Ingrid Stairs was observing pulsar(s) in M80. After a little confusion during setup concerning LO2 settings the observations were started at about 12:30 pm. This was 1:30 hours after the setup time was scheduled with most of the time going to the fact that the source was not above the horizon. Problem Summary: (1) PSRGUI did not set the LO2 values properly for the BCPM (2) BCPM doesn't stop when scan is aborted by bad antenna trajectory (3) Optical Driver 8 has low power (3) Unannounced Schedule Changes [1] This week's schedule Spectrometer time fully booked. We will juggle things to allow testing of all important projects. We will steal some time from Melinda for checking out stuff as needed. It was agreed that all the time can't go to checking out the configuration tools. [2] October Observing Schedule discussions There was a great deal of discussion about PTCS time this week, and we decided to use the time that Richard gave up to test the configuration tool and the spigot card scripts. The evening hours will be allocated to the astronomy projects that were already on the schedule. [3] November Observing Schedule discussions Not much discussion here. [4] GBT development planning Impressions from meeting The project presentation day was good, if long. It is useful for all to be able to know what is going on, even if they are not responsible for a particular project. Day/Time for 6 month look meeting? We decided to put this off until the week of the 13th. It has been set for Thursday, October 16th. Cycle 8 reports from Division Heads (as needed) Software Electronics Scientific Staff Operations Mechanical Dennis and Bob were missing for this topic, but all others are satisfied with the work assigned for cycle 8. [5] AOB The planning meeting agenda and minutes will be put on the web. The verbose observers' reports will be left there for the record, to abbreviate the commissioning meeting minutes somewhat. -- JF 8. Any other business