Summary of the Fall 2016 GlueX Run

From Hall D Ops Wiki
Revision as of 19:09, 14 February 2017 by Deurpam (Talk | contribs) (→‎Brief summary)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Fall 2016 GlueX Run

Brief summary

Due to problems, the run was reduced from 2.5 months to less than a week of data taking. It occured between Dec. 15th 2016 and Dec. 21st 2016. Little physics data were taken, but the DAQ, Trigger optimization as well as studies of their performances at high luminosity. A new diamond (JD70-100, 58 μm) was aligned and used successfully. In addition, the photon beam residual polarization at low photon energy was studied with reasonable statistics for both the rho and TPol polarimetries.

  • Beam conditions:
    • Energy: 11.640 GeV
    • Frequency: 500 MHz
    • Intensity: up to 1.5μA. Typically, between 0.2 and 0.5 μA.
    • Photon polarization at top of coherent peak: about 30%
    • Quality: Radiation levels were good. The tuning was not optimal for production: Due to lack of time, we had to be content with the electron beam focused on the diamond rather than the main collimator (it may be the cause of the lower photon polarization).
  • Run conditions:
    • Solenoid at 1200 A
    • Actual run calendar time: about 5.6 days (from 12/15/16 10:00am to 12/21/16 04:02am)
    • Electron beam integrated charge: 0.054 C
    • Run numbers: 21965 to 22170
  • Revised goals:
  1. Trigger and DAQ development check.
  2. High luminosity trigger and DAQ performance study and optimization.
  3. Checkout of a new diamond and of the goniometer, in particular investigating whether we still have the parallel-perpendicular rate difference.
  4. ToF/CDC voltage studies.
  5. Enough data taking for global checkout/calibration.
  6. Investigate using the Triplet Polarimeter the significant (5%) polarization seen at low photon energy by the rho analysis. (No polarization is expected there according to leading-order QED predictions).
  • Results:
  1. Trigger and DAQ development check was done.
  2. High luminosity trigger and DAQ performance study and optimization were done, resulting in the hability to run at 50 to 70 kHz for the Spring17 run.
  3. Diamond JD70-100 was aligned, yielding 30% of polarization at the top of the coherent peak. A small (5 to 10%) parallel-perpendicular rate difference was seen, to compared with 30% seen previously with the smaller JD1s50 diamond.
  4. ToF/CDC voltage scans were done.
  5. Production data taken. Overall, we took 2.4×1010 tot triggers (1.5×1010 para triggers. 1.0×1010 perp triggers. No Al. radiator triggers.)
  6. Polarization at low photon energy: Triplet Polarimeter yields a 3.5% +/-1.1%. The rho analysis yields 3.5% +/-0.5%. These numbers are preliminary, as of 01/20/17. This -seen with a 30% polarization at the top of the coherent peak- agrees with the 5% seen in Spring16 with a 40% polarization at the top of the coherent peak.
  7. Details and other results are given in the longer summary below.
  • Problems:
  1. Several trips of the solenoid power supplied occurred while ramping it up. To be safe, the run was done at 1200A rather than at the planned 1350A.
  2. A bad FCal crate (crate 10) severely limited the DAQ performance for production.
  3. Beam tuning was not adequate for production and not optimal for our polarimetry study at low photon beam energy.

More detailed summary

Beam schedule and operation:

This Fall run was expected to be the first part of the physics run of GlueX, with two and half months of running. However, due to two technical problems, the Fall 2016 eventually became a one week preparation run for the Spring 17 run. The first major problem that caused this was a failure of a power supply in CEBAF's arc 7 that precluded 5- and 5.5-pass running. This cost Hall D to delay its foreseen start of operation until around Thanksgiving. The second problem delayed our start further. It was due to a vacuum failure in the RF separator that allows to provide 5-pass beam to Halls A, B and C, while delivering beam to Hall D. Since Hall A required 5-pass for its program, Hall A operation was incompatible with Hall D's one and it was decided to finish the Hall A 5-pass program before possibly running Hall D. The beam delivery to Hall A was very efficient and their 5-pass kinematic point was completed well before the end of the Fall run. This allowed the start of the Hall D operation on Dec. 12th, concurrently with 4-pass beam delivery to Hall A. After setting-up the beam, we eventually got usable beam on Dec. 15th, near 11am. Beam was terminated on the 21st, 4am. One difference with the previous runs was the slightly lower beam energy: we ran at 11.640 GeV. The beam quality was good (radiation levels seemed slightly better than the previous runs) but the tuning was not optimal (more on that below). The beam was operated at 500 MHz (the initial plan was 250 MHz).

Goals of the run:

The initial goal was to take the GlueX low luminosity data. In addition we wanted to check several technical points, the most important being the developments made on trigger and DAQ and testing them at high luminosity. After the reduction of our beam time, the goals were revised to:

1) Trigger and DAQ development check. 2) High luminosity trigger and DAQ performance study and optimization. 3) Checkout of a new diamond and of the goniometer, in particular investigating whether we still have the parallel-perpendicular rate difference. 4) ToF/CDC voltage studies. 5) Enough data taking for global checkout/calibration. 6) Investigate using the Triplet Polarimeter the significant (5%) polarization seen at low photon energy by the rho analysis. (No polarization is expected there according to leading-order QED predictions).

Results of the run:

The 5.5 days of running allowed us to achieve most of the goals above, which places us in an excellent position for the Spring 17 run. Overall, the accelerator efficiency (counting it over the allocated 7-days operation) was 53% and the overall accelerator + Hall D operation was 51% efficient. This is good, for multiple Hall operation. Overall, we took 2.4×1010 tot triggers (1.5×1010 para triggers. 1.0×1010 perp triggers. No Al. radiator triggers.) Regarding our goals:

1) Trigger and DAQ development check: This was checked early on in the run and no special problem were encountered.

2) Trigger and DAQ at high luminosity: This constituted the bulk of our spent time: Several problems and bottlenecks were discovered and addressed. Overall, the DAQ information transmission can safely reach 1 Gb/s, leading it to be able to operate at 50 kHz with acceptable livetime. Further optimizations of the thresholds and readout windows should allow to increase the DAQ rate up to 70 kHz. This was achieved with a luminosity of about half of the maximum luminosity expected for the final GlueX run. In addition, the trigger+DAQ was tested at various beam currents in order to map its behavior and to be able to extrapolate to the high luminosity run condition. L3 trigger was studied as well, although we had time to complete only half of the planned L3 study.

3) Checkout of a new diamond and of the goniometer: The goniometer did not have any operation problem during the run. Our goal was to align the old 50 micron diamond (JD1a50_ and the new 60 micron diamond (JD70-100). However, it was difficult to align the old diamond so this was gave up due to our tight schedule. Although not a critical loss, it deprives us from a known reference for this run. On the other hand, JD70-100 was quickly and successfully aligned, allowing the delivery of the polarized photon beam early on in the run. The parallel-perpendicular rate difference is much less for the bigger new diamond (between 5% to 10%) than for the small old 50 micron diamond JD1s50 (around 30%). This gives credence to the assumption that during the Spring 16 run, we were missing part of the old diamond in the perp configuration because of its smallness and because of some unwanted difference in the diamond location between the para and perp diamond setups. (This particular position problem was fixed during the break between the Spring and Fall runs).

4) ToF/CDC voltage studies: The high voltage scans were done successfully.

5) Enough data taking for global checkout/calibration. Enough data were taken, although only a small fraction of what was expected. This is due to the fact that we needed extend the studies on the trigger/DAQ at high luminosity, and that we needed more time for the Triplet Polarimeter study. We were able to checkout the detectors and have enough data to calibrate them. A run in long-mode acquisition was also taken.

6) Investigate using the Triplet Polarimeter the polarization at low photon energy: About 6 to 7 solid hours of special-trigger -high efficiency- TPol data were taken, mostly at 500 nA. This was just enough for the investigation. Triplet Polarimeter yields a 3.5% +/-1.1%. The rho analysis yields 3.5% +/-0.5%. These numbers are preliminary, as of 01/20/17. This -seen with a 30% polarization at the top of the coherent peak- agrees with the 5% seen in Fall16 with a 30% polarization at the top of the coherent peak. This unexpected polarization may be interesting enough to warrant a physics article.

In addition, a number of other goals were achieved:

a) The new vertical motion of the main collimator was tested successfully and used to optimize more conveniently the photon flux transmission. b) The pair spectrometer (PS) signals were gathered at different values of the PS magnetic field. (PS magnetic field scan). c) Data were gathered with and without a new lucite shield for the inner part of the TOF. d) The Compton calorimeter was parasitically and satisfactorily tested. (This detector is for the future PrimEx program in Hall D. It is not related to the glueX program. The test was not a full commissioning but rather a first checkout of its behavior.) e) We took some data with the tagger quadrupole on (-4.2A). f) Accelerator conducted parasitic commissioning of our position Fast-Feedback. g) We tested beam delivery up to 1.5 uA. However, much of the run was done at 500 nA or less.

Problems encountered:

Not all the goals were achieved. As mentioned above, no data were taken with the old 50 micron diamond, only a fraction of the data in standard running conditions were taken, and only half of the L3 studies were done. Furthermore, we had planned an empty target run and a Total Absorption Counter flux measurement. None were done. Although it would have been desirable to get these various data, these were of lower priority and not critical, which is why we gave them up.

We/Hall D beam-operations also had some problems during this run:

1) The goniometer motors stopped working soon before the initial date for the start of the run. The goniometer was sent back for repair to the company that built it. While the motors were repaired, the controls were altered and it took time and efforts to be able to drive the motors again. Furthermore, the limit switches are not operational, which means that we are at risk of losing reference positions (implying among other things that the diamonds would need to be realigned: a time-expensive task). The goniometer operated without problem during the run. But its overall reliability is questionable. Furthermore, we still don't know what caused the initial motor problem. It is unlikely we can improve this situation before the start of the spring 17 run, although we may fix the limit switch problem.

2) The solenoid tripped 3 times when we tried to ramp it up to 1350A 10 days ago (those trips were not quenches but ramp-downed erroneously triggered by the quench detector). This was very unusual because it had rarely been seen in ramp-ups (although commonly seen in ramp-downs). Eventually, the solenoid was ramped-up more slowly to avoid the problem, and only to 1200A to minimize the chance of trips. The Fall 16 run was thus operated at this lower current. Shortly after the beam termination, the solenoid was successfully ramped-up to 1350A and stayed there for 30 min before being (on purpose) ramped-down. This problem will be investigated and solutions (beside a slower ramp rate) are being discussed. This should not be a problem for the Spring 17 run.

3) DAQ rate during production runs was severely limited by a problem with a FCal crate (crate 10). This problem did not impair the DAQ/trigger tests nor the TPol data taking. It was closely studied during the run and solutions to the problem are being discussed. This may imply replacing the crate. Whatever solution is chosen, this problem will be fixed for the Spring 17 run.

4) The beam tuning was adequate for all the studies we conducted during this run. However, it would not have been adequate for production runs. The problem is that the initial beam was creating too much background at the level of the radiators. This was solved by focusing the beam on the diamond position. However, the consequence of the solution was to defocus the (projected) electron beam at the level of the collimator, where it should instead be optimally focused. This dilutes the polarization of the photon beam. Preliminary studies indicate that the polarization at the top of the photon coherent peak was 30% for this run. It was about 40% in Spring 16 with the old JD1s50 diamonds or with the two new 20 micron diamonds that were tested (JD-118 and JD-119). While it is possible that the lower polarization is due to a somewhat defective new diamond, it is more likely that it is due to the beam defocusing. Again, this was not a problem for this test run. It will also not be a problem for the Spring 17 run since it would "simply" required more time to tune the beam. However, we had no interest in investing in such tune time during our short Fall 16 run.

5) We measured a large bleed-through from the Hall A beam. This was expected. While it may have contributed to higher backgrounds and difficulties to run at our higher currents (requiring our slit to be more open, and thus more bleed-through), it did not seem to impair our run significantly.

6) RF separation. This prevented accelerator to run Hall A and Hall D concurrently. There is no plan to run 5-pass beam to Hall A, B or C in Spring 17. Furthermore, the 5-pass separator was repaired and tested. The 4-pass separator is working well (this is what we just used). Finally, the 3-pass separator was tested yesterday and appears to work fine. (The program in Spring 17 in other halls will be at 3-pass). Consequently we do not expect the separation to be a problem for Spring 17.

In spite of the points above, our week of was very successful and fruitful. To achieve this meant a very intense and efficient work from the Staff, Hall D Users and accelerator Staff. Their full dedication to the run and their proficiency made this week a success (and made my RC work easy).

Retrieved from "https://halldweb.jlab.org/hdops/wiki/index.php?title=Summary_of_the_Fall_2016_GlueX_Run&oldid=2629"