Run Coordinator report: Spring 2017 r1
Dear collaborators,
Please find below my report as Hall D run coordinator for the 01/26-02/08 2017 period. If you want to skip the details, there is a few lines summary at the end of this mail.
The Spring 2017 run has begun very well. It is expected to be the actual start of physics production for the GlueX experiment, E12-06-102.
Beam started to be restored to Hall D on Jan. 31. This was preceded (Jan 30th) by a successful re-commissioning of the five pass RF separation, although only one beam (Hall D's) was used for this test. Establishing that 5-pass separation operates smoothly with several beams may happen at the end of this run with Halls C and D at 5 and 5.5 passes, respectively. Hall D was operated with RF separation until Feb. 5th. Then after a trip of the 5-pass separator, it was decided that the separator was tested long enough and that it was faster to restore the beam with magnetic separation. The success of this commissioning is important for Hall D since Halls A, B and C are expected to run at 5th pass in Fall 2017.
We spend Feb. 1st. assessing the beam quality. It was initially not nominal:
- Harp scans showed that the beam was vertically focused on the diamond rather than the collimator. This degraded the beam polarization and the photon flux in the hall. This problem was fixed later (see below).
- The radiation levels are different from the previous runs. The photon detector on the floor under the end of the tagger magnet (rad102_p1) registers a larger background than usual: This background (i.e. the radiation level with radiator retracted) was about 0.2 mrad/hr/nA but it is now 0.5 mrad/hr/nA. This is consistent with a radiological survey that found a hot area at the entrance of the tagger beam dump. The beam may be scraping at the constriction of the beam line located there. The radiation levels near the goniometer are, on the other hand, lower than in Fall 2016 by more than an order of magnitude.
- We see "satellite peaks" on the profiler. Those are independent of the radiator thickness and, contrary to the main photon beam, do not activate the fibers of the profiler. This is a sign of beam scraping upstream of the radiator that produces soft photons that make these peaks. Satellite peaks were seen during the previous runs, but not as marked.
These two instances of scraping (upstream and downstream of the radiator) do not affect the quality of the data but produce extra radioactive background in the tagger and an annoying background for the profiler signal.
On Feb. 2nd, we started checking out the DAQ and level-1 trigger. The tagger microscope was re-commissioned because new fibers were installed to replace old ones with poor light yield. These new fibers are showing excellent performances. Then, the detectors were checked out and data were gathered for calibration. There was a few problems that delayed the beginning of our program, mainly due to three CHL 2K cold box trips/maintenance and problems/investigation of ARC power supply breakers. Those problems were solved and the program restarted in the evening of Feb. 3rd. In particular, the successful level-1 trigger and DAQ optimizations allowed us to run at 70kHz with 96% livetime in standard production configuration. This is a milestone for GlueX.
On Feb. 4th, the diamond JD70-100 (58 micron) was aligned with 4 angle configurations (0 and 90 degree, and 45 and 135 degrees). It initially yielded a photon beam polarization of about 35% according to both the TPol and the \rho preliminary analyses (this is to be compared to 40% in Spring 2016 with three different diamonds, and with 30% in Fall 2016 with the same diamond but a beam focused on the diamond rather than the collimator). After the beam vertical convergence was improved (see below), the coherent peak shape analysis indicates an increase of polarization of possibly 5% relative. We are waiting for the TPol and \rho analyses to better quantify this improvement.
Once the diamond was aligned (Feb 4th), we started gathering production data on the four diamond orientations, as well as on the Al. radiator (about 10% of the production data). Data are being taken at a DAQ rate of 30KHz, to compare them with the Spring 2016 data that were taken with this rate. The data taking was interspersed with harp scans to monitor the beam profile, and with several other tasks. The main ones were: 1) Continue to fine tune the level-1 trigger and of the DAQ. 2) Data taking on empty target (this will be done weekly). Vertex reconstruction for the empty target run is excellent (comparable to Spring 16). 3) TAC run on Feb. 7th. Although it had been scheduled for latter in the week, the TAC run was done (opportunistically) ahead of time: the solenoid had tripped (see below) and, since TAC runs do not need magnetic field, it was decided to do it during the ~8h needed to ramp-up the solenoid to nominal field (1350A). Preliminary analysis shows that the TAC run was successful.
The beam y-convergence was fixed on Feb. 6th. It took most of the day but the beam is now converging on the collimator rather than on the diamond. As mentioned, the beam polarization seems to have increased, but the most noticeable effect was that the photon transmission in the hall was improved by at least 50%. This allows us to run at about 100 nA to reach 30kHz rather than about 150nA. Low energy background in the hall was also decreased significantly (however, the radiation levels in the tagger vault did not change).
The only noticeable incident we had is a solenoid current dump on Tuesday 02/07/17 near 14:00. This was due to a compressor shut-off button being accidentally pushed during routine maintenance work in the Hall D refrigerator building. This triggered the current dump. The solenoid current was ramped-up back to its nominal value of 1350A by 23:30. To mitigate the impact of the incident, it was decided to do the TAC run while the solenoid ramped up. In all, this incident cost us no more than 2h of lost time. Lesser events on the Hall D side are crashes of the LH2 target IOC (almost daily), and DAQ crashes (almost hourly). We had also two trips of the Pair Spectrometer power supply, which had so far been very reliable (only one trip in the past years).
At the end of my run coordination tenure, we were gathering production data with the Collaboration effort focusing on checking the data quality, and insuring that online and offline monitoring are efficiently done. Reconstruction of physics quantities is good, given the available preliminary calibration.
Mark Dalton took over the run coordination duties on 02/08.
In all, we gathered 5.3 billion triggers (1.3b para, 1.1b perp, 0.8b "135 deg", 1.2b "45 deg", 0.9b unpolarized). This is already the same order than the analysis-worthy data gathered in Spring 2016. 18% of the time we are allocated for the Spring 2017 run has been spent, and we actually operated 6.5% of the total allocated time (51 days). This represents an efficiency of 36% compared to the expected overall ~50% efficiency. This is in fact very good for a start of run: run efficiency typically increase as the accelerator becomes stable and preparation and systematic checks are over. (To compare, in the previous runs the efficiency during the first week was 0%, since physics-quality beam was never available in time.) This good start is due in part to the very good accelerator operations. In particular, the beam was available almost in time (only 2 days late, part of it spent in the necessary commissioning of the 5-pass separator) and although Hall B ran concurrently with Hall D for several days, the multi-hall operation went smoothly without hall interference.
A brief summary of this report is: The start of the Spring 2017 run went extremely well compared to our other runs. In the first 10 days, we did all tasks necessary to start GlueX production, and all the polarized data expected to be taken were gathered. The amount taken (5.3 billion triggers, mostly polarized) is comparable with the analysis-worthy data taken in Spring 16, but of somewhat lower polarization (maybe ~35% vs ~40%). The accelerator and hall are working well and there was no major incident. We are now focusing on establishing the data quality and efficient data monitoring. The rest of the run will be focused on taking data, probably at larger rate (50kHz), interspersed with a few tasks for systematic studies (empty target runs, TAC run(s), a day or so of data taken with the new 17 micron diamond, straight-track runs) as well as a few tests of Levels 1 and 3 triggers to prepare for future high luminosity runs.
Best regards,
Alexandre Deur
