Minutes of the GBT Commissioning Meeting Friday, 3 January 2003 AGENDA 1. Spectral Baseline work -- Rick 2. PTCS project work -- Richard 3. Az Track work -- Bob A. 4. K-band and other commissioning -- Ron 5. Project and observing program scheduling -- John, Carl 6. Software status -- Amy 7. Spectrometer status -- Rich 8. AIPS++/Dish status -- Joe M. 9. Any other business Minutes 1. Spectral Baseline work The X-band (8-10 GHz) receiver was remounted on the antenna on 12/30/2002 and scheduled to support observing programs from Jan 5 to 23. The receiver contains interim InP amplifiers; the final units were received but not in time for installation. Receiver total power sensitivity to cryo temperature changes is greatly improved in lab with new amps, but this is not yet verified on the antenna. Baseline tests will resume on this receiver shortly. 12/17/02 On the L-band (1.4 GHz) receiver a 1.6-MHz ripple was detected in cold sky (on-off)/off spectra, where the 'on' has the subreflector displaced by 1/4-wavelength in the +Y direction (toward the main reflector). An FFT of this spectrum showed that this ripple has a range of frequencies from about 1.35 to 1.65 MHz. This corresponds to a multipath difference of 180 to 220 meters. This which is twice the distance to the circumferential panel gaps. A stronger 9.1-MHz ripple with a peak-to-peak amplitude of about 0.2 K was also seen. This period corresponds to the distance between the feed and the subreflector. 12/26/02 The 1.6-MHz ripple on cold sky was investigated in more detail at 1.4 GHz using an 1/8-wavelength subreflector displacement. As expected, the ripple amplitude is stronger than with a 1/4-wavelength displacement since the multipath interference between the feed and the circumferential the panel gaps involves two subreflector reflections. The 1.6-MHz ripple is considerably stronger in receiver channel Y whose linear polarization is perpendicular to the panel gaps. Scattering from the gaps should be strongest in the perpendicular polarization case. Reflections over nearly the full range of panel gap distances were seen in the spectrum FFT. No significant dependence on sky noise amplitude was seen in pointing the telescope to 10 and 15 degrees elevation. The 9.1-MHz ripple in the cold sky, was stronger with 1/4-wavelength offset than with 1/8-wavelength offset. This, too, is what was expected since the extra path length involves only one subreflector reflection. 12/26/02 Good stability of a continuum source's spectra at 1.4 GHz was seen over about 50 minutes with slow deterioration thereafter. The 2.33-MHz ripple, due to reflections between the subreflector and main reflector, shows up in the quotient spectra after 50 minutes as one might expect from motion of the subreflector with respect to the main reflector as the elevation changes. Using one continuum source to calibrate the spectrum of a weaker source worked reasonably well with the caveat that the 2.33-MHz ripple is not canceled if the sources are at different elevations. It looks like using simple difference spectra (on-off) instead of (on-off)/off works better for carrying the calibration from one source to another, possibly because the changing system noise with elevation is not involved in the arithmetic. 12/27/02 Observations at 1.4 GHz to look for changes in continuum source spectra at different positions in the beam were taken on this day. We are interested to know whether the telescope and receiver respond differently to extended and point sources. Data analysis is in progress. 12/18/02 At C-band (5 GHz) the 2.33-MHz ripple in continuum source spectra is barely detected in a 5.4 Jy source. Its amplitude shows a wavelength-squared dependence when compared with L- and S-band (1.7-2.6 GHz), as is expected from multipath reflections between the subreflector and main reflector. 12/30/02 observations similar to the L-band cold sky focus offset and continuum source spectra were done at C-band, and the data analysis is still under way. A preliminary look at the data shows the reflections from the circumferential panel gaps (1.6-MHz ripple with cold sky focus offset). The wider spectrometer bandwidth produces better distance resolution on the reflections. Norrod and Fisher met with Pospieszalski, Bradley, Srikanth, and Webber on 12/20/02 to discuss power transfer and LNA noise characteristics of the GBT receivers. The CV engineers suggested that the variation of continuum source Ta to Tsys ratio with frequency could be due to small impedance mismatches in the feed-OMT-LNA combinations. They will do some feasibilty calculations to verify this, and the Green Bank crew will continue to measure the GBT receiver characteristics with particular attention to the differences between LNA's with and without input isolators. -- JRF 2. PTCS project work Work on various activities (PTSC system design, Active Surface Project Improvements, etc) proceeded at a low level over the Christmas/New Year period, but there is nothing specific to report. The PTCS project had two commissioning periods since the last meeting: 23/24th December and 31 Dec/01/02 Jan. Both periods were plagued by bad weather. A significant fraction of 23/24th was lost due to high winds until ~7pm on 23rd, a four hour gap for 2C35, and then snow starting ~9:00am on 24th. In total about 12 hours of observations were made; these were used to check efficiencies and beam offsets as a function of elevation at 20GHz for the K-band receiver. The data have not yet been reduced. The telescope was used from 5pm 31st to 5am 2nd to perform an all-sky pointing run, with sources chosen to probe relative pointing offsets as a function of offset distance and time. These data were taken at C-band due to extremely poor weather (high humidity/rain) and in retrospect may be of marginal usefulness, although this remains to be seen. The data have not yet been reduced. -- RMP 3. Az Track work Bob reported that work was proceeding on the bolt-loading test. The instrumented bolt with strain guages is being constructed. The zinc paint has been obtained and Harry is drawing up plans to paint the top of the base plate and underside of the wear strips on a track segment. 4. K-band and other commissioning Ron reported that only one commissioning session planned for the coming week will be on Saturday but the reports indicate that the weather will not be good for high frequency tests and we'll have to revert to a low-frequency backup project. Frank has conducted most of the commissioning work over the last week or two, and submitted a couple of reports: ==== Activities over Dec 27-29 included K-band commissioning, Spectrometer mode checkouts, and some observing make-up time. 1. Beam Balancing (project code: TKband_Dec27) at K-band. Found gain settings in the K-Band reciever that equalized power in beams 1 and 2. Did this by hand over the range 18-20 GHz, then decided it should be automated. Results seemed rational for the LCP channels, but not RCP. 2. Make-up time for Fernando Camilo at 800 MHz Observed projects AGBT02A_062 (J2229+6114) and AGBT02C_034 (3C58) This all worked sucessfully. 3. Focus and pointing checks (TKPOINT_DEC28) at K-band. Used the "focusSubreflector" procedure to do a series of peak scans at a range of Y-focus, and alternated these for beam 1 and beam 2, following a calibrator from low elevation to high elevation. These data can be used to address questions of the beam offsets vs elevation, and corrections to the focus tracking model. No reduction done on these yet. 4. Tests and make-up observing for E.Churchwell et al (AGBT02A_18) at C-band -- all sucessful. 5. Spectrometer checkouts of multi-bank 200 MHz mode (TSPECTEST_13) using K-band receiver and observations of W3(OH). Found a problem with M&C that prevented dish from making sense of banks other than A. Also found a problem with scaling in dish. After consulting with J.McMullin and M.Clark, both problems were fixed. 6. Did several continuum maps at K-band of 30'x30' size, scanning both in RA and Dec (TKMAP_DEC29). Objects mapped were: NGC7027, CassA, Tycho's SN, 3C147. ==== Status of testing spectrometer modes. Of the high priority spectrometer modes, hardware testing has been done on all of them. Most problems, except for baseline problems, have been fixed or are being worked on. Astronomical checkouts remain to be done on: -- 8-sampler and multi-bank modes at 12 and 50 MHz bandwidths. -- multi-bank modes at 200 and 800 MHz bandwidths -- 1W2 and 1W1 modes at 800 MHz. In general, if we ignore baseline problems, the results look pretty good. Some problems with measured rms values may have been due to a scaling error in the manager, which has been fixed. We need to repeat some observations to check this. [In discussion, Ron suggested that we need a system with more follow-through when commissioners find a problem with the system. These problems are sometimes not addressed for a long time. It was agreed that the best approach is to address this through a fault-tracking system (which needs to be implemented). Priorities will be set at the Project Planning meetings. It is likely that the person who noted the problem originally would be assigned to do repeat tests after the fix. Joe McMullin noted that this scheme has worked well for the AIPS++ group. The problems that Frank noted above with the Spectrometer checkouts were successfully resolved in this manner. Both Roger and Amy stated again that their groups would like to be called at the time a problem is occurring, as it is often difficult or very time-consuming to reproduce the problem later.] 5. Project and observing program scheduling From John: Baseline work continues to use all the time scheduled. X Band receiver was put back up with the interim amplifiers, in order to support the Goldstone radar experiments. Will need to come back off after this month's observing. C Band needs to come down for a similar amplifier retrofit. In addition, its refrigerator is acting up. Should schedule it off as soon as we can. Spectrometer Spigot card testing has been scheduled for 4 hours per week for the next 6 weeks. Exact times and days will float to allow our higher priority work to continue. Software group will write some shell scripts to support the loading of personalities in a crude but simple way. January targets for the high-priority projects have been given to me. I'm in the process of generating the one month and 6 month schedules from them for public consumption. Carl continues to improve upon the resource calendar. From Carl: Observing summary: - past/present-future. Last Week ======== 2C35 (Titan radar) completed (16 days in last part of Dec) 2C34, 2A62 Pulsar makeup time 2A18 makeup - completed Next week ======== 2C49 X band Radar of Mercury starts 2A52 Pulsar observations January ===== Astronomy - 35 hours + 86 hours of backup time Tests & Commissioning - 571 hours Maintenance - 113 hours Un-assigned - 26 hours February ======= Astronomy - 60 hours Maintenance - 108 hours Tests & commissioning - 132 hours Un assigned - 372 hours 6. Software status Software Development Division Update #18 - Friday, January 3, 2003 We are in Week 3 of a 5 Week Development Cycle, preparing for the release of M&C v3.10 on January 15, 2003 Start thinking about future M&C releases such as Wednesday, February 12, 2003 (v3.11) - Get your requests in prior to the release planning deadline on January 13, 2003 - Use the Software Project Office online to submit your request (To be unveiled in January) Data Export Improvements (Filler): Eric has a fairly full-featured (and fast) query implementation, FITS file save capability, and an implementation of F. Schwab's Van Vleck correction. Eric is currently working on making usability enhancements and will need some scientist involvement when he returns next week. Baseline Investigation Support: Work on the continuous real-time total power display is near completion. The MR for enhancing the DCR so that engineers/observers have the ability to record continuum data when observing with another backend has been approved and implementation has begun. Antenna Control Unit Improvements: Joe and Melinda have implemented the contents of the original MR and are currently testing them. Further improvements are deferred until the next release cycle and require additional clarification. Multibank Capability: Some Issues raised in the code review and through testing by R. Lacasse are now to be released in 3.10 rather than as a patch to 3.9. Work arounds for these issues are in place until the release. The debug file generation will appear as a patch to 3.9. Regression Test Suite: Waiting until Jan 13-14 to initiate regression testing on MCv3.10 post-AAS. Software Project Office Online: Project Office web site unveiling was moved to sometime in January in order to incorporate suggested changes, which will make it more responsive to user needs. Configurator Development: Melinda has added the K and Ku band receivers to the Configurator. The changes are awaiting acceptance testing by R. Maddalena. Q Band Receiver: Deferred by Ron; Release will occur approx. March 2003. Engineering Measurement System: Project Charter underway. The charter will be completed and presented by Ray Creager in January. Active Surface Improvements: Project Charter has been reviewed and comments are currently being incorporated into a final product. Amy and Melinda have begun work on redesigning the 100ms control loop based on the requirements specified in the MR. Amy expects to lead a code review of the entire AS code base during the week of Jan. 13. -- AS 7. Spectrometer status The low speed sampler upgrade is continuing. 24 units have been modified and tested in the system. 11 remain. Preparation for the upgrade of the high speed samplers continues. Many components have been ordered and are in house. A few still need to be ordered. One part has been discontinued and we are looking at alternatives. Low level ripple in the auto-correlation function of some modes seems to be related to correlator chip blanking. Longer integrations than 40 seconds are required to confirm this and verify the fix. However, while performing multiple integrations within a scan with the intention of averaging them, errors in the autocorrelation function were seen in three of ten integrations. A few plots will be shown to illustrate this. The cause of these glitches will have to be sought before we make further progress on the low level ripple! No progress was made in looking for the large glitches flagged by Dana and Frank a few weeks back. We await the re-installation of the software feature which allows us to look at "raw lags", and which was pruned out of the software in the last release. This will be included in the next release, due out Jan. 15. -- RL 8. AIPS++/Dish status o New stable functionality - available in /home/aips++/stable1 o Status page: http://aips2.nrao.edu/daily/docs/programmer/buildstatus.html This is under construction o Report card for snapshot available there soon (reported last meeting). o Looking at K-band data o Change time of AIPS++ report to earlier in the meeting to allow more details. o January targets - require investigation - will need to convene a special meeting on the multi-x implementation. o Defects resolved: Use http://aips2.nrao.edu/ddts/ddts_main to examine AOCso04229 renaming does not work with catalog gui AOCso04149 cannot calibrate scans with 1 integration per scan AOCso04248 scaling is off for new weighting scheme AOCso04171 add exposure time to header to see averaged data AOCso03920 filled file does not have specified name (solution as next) AOCso04188 incorrect instructions for opening previously created ms AOCso04209 reports errors if there is a message AOCso03859 Unzoom gives error message AOCso02553 Some basic enhancements in the baseline program AOCso03870 Flagging Channel 1 AOCso03434 Problems using the zoom feature with the pgplotter AOCso04166 can't handle flagged data AOCso04117 Signal and Reference Quotients AOCso04112 confusing behavior - multiple RM entries AOCso04115 view;flag;plotr;flag sequence hangs AIPS++ -- JM 9. Any other business Phil has asked Carl to prepare a forward look for discharging the observing proposals that we have in the queue, which amount to over 4000 hours of time. We must balance observing time against commissioning and maintenance needs, and will aim for best scientific return in all respects. This plan will be discussed at upcoming Commissioning Meetings and at GBT project planning meetings. The Commissioning Meeting will be held again next Friday, 10 January, at 10 A.M. PRJ 4 Jan 2003