General Meeting Summary 8/22/13

NPS DESIGN PREPARATIONS

• For design optimization angles and distances for PbWO4 detector needed in experiments were collected

• Angles < 30 deg require front of detector between 3m and 12 m (DVCS/SIDIS: 3m and 6m)
• Angles < 25 deg (other side of SHMS) require front between 2.5m and 3.5m

• Next, options for rails to cover the range between 2.5m and 12m on both sides will be explored with designers.

• at forward angles there may not be sufficient clearance in horizontal direction for detector to go to 3m due to interference with beam line. Plan is to optimize magnet to fit into as small as possible distance from target to maximize acceptance.
• Stray fields will be studied once the optimized bore size is known

• Design of temperature controlled frame, e.g., in situ curing system will be examined in more detail as well

MRI PREPARATIONS: BUDGETARY QUOTES

• PbWO4 vendors and potential vendors have been identified and/or contacted (Crytur/Czech Rep., Russian companies, JFCC/Japan)

• Crytur initial quotation: anticipated price is in the range over 40USD/cm3, or 3300 USD/pc for a 20x20x200mm crystal. They received a similar request to ours from GSI.
• Russian companies: have not yet responded
• JFCC: need to know desired percentage of Pb and WO4. Apparently many companies have reduced production Pb-containing glass because of environmental concerns. They also offered to help identify addtional sources
• SICCAS: not yet contacted. Based on the paper by R. Novotny, The electromagnetic Calorimetry of the PANDA detector at FAIR. J. Phys.: Conf. Ser. 404 012063 (2012) GSI is also considering procuring additional crystals from this company (or perhaps Crytur?)

• To contact additional vendors it would be good to have a set of uniform specifications for the PbWO4. These could be based on the specs CMS used

• It would also be interesting to know what GSI's technical specifications are, e.g., radiation hardness, polishing quality factors, etc.

• PMT vendors have been contacted. Current plan is to use the 19 mm diameter Hamamatsu R4125 model. It's gain is about 8.7E+5 at 1.5kV max. anode voltage and rise time 2.5ns. This PMT was also used in Primex and we have designed a new active voltage divider that improves the performance of the PMT by a factor of 25.

• Discussion about size of crystal, and in particular if there could be savings by changing to larger crystal size (savings in electronics/PMTs). However, larger crystal size would result in loss of resolution. It is also not clear if there would be a significant gain in savings. Based on earlier discussions with SICCAS larger crystals seem to be more difficult to make and it may also be more difficult to control their quality. Furthermore, increasing the PMT photocathode may scale with cost, and so may actually not result in savings. Quantitative reasoning for dimensions 2.05cmx2.05cm is that these are optimal for background/resolution eff. These dimensions are also used by many other experiments/facilities. We decide to keep the ~2cmx~2cm crystal size.

• Discussion about possible strategies to stay within the "small" MRI proposal bugdet limits.

FEEDBACK FROM HALL C AUGUST MEETING

• Discussion of three comments Vardan got on his talk:

• Magnet stray fields and beam steering problem: this issue was already addressed in earlier discussions with Jay Benesch. See also our reply to the PAC40 TAC comment from WACS below.

TAC Comment 5. The beamline will need to have magnetic shielding and a downstream dipole corrector. Stray fields up to 4m radius need to be accounted for when designing the beam line. Post-target correctors need careful design to avoid additional backgrounds from beam scraping.

Reply: Indeed it is important to avoid beam scraping and to minimise the effects of stray particles. We are working on this issue with Jay Benesch of the Accelerator Operations group and believe that, with magnetic shielding, the stray field integral can be reduced from ~100 Gauss.m to ~10 Gauss.m. With an active corrector the stray field integral can be reduced to a negligible level.

• Limitation of DAQ capabilities: 20 Mb/s too much based on Qweak data flow. However, it seems that in the mean time rates of 50-100 Mb/s have been achieved with fADCs in, e.g., tests in Hall A. This comment was also subject of our PAC40 TAC comments. Our replies are listed below.

TAC comment: Given the (potential) amount of data generated by the > 1000 blocks of the NPS, additional hall resources for online data processing and storage may be required. As a benchmark, the Q-weak experiment took data at a rate on the order of 10 Mb/s this put significant strain on hall computing resources.

Reply DVCS: Reading 26 samples of the FADC modules (104 ns sampling window) implies an event size of 64 bytes per channel. We expect cluster size to be around 25 blocks and we do not expect an average of more than 3 clusters per event, that is 75 blocks. With these assumptions and the integrated DIS cross section in the HMS acceptance, in most of the kinematics we can read every single HMS trigger with a data throughput of less than 20MB/s, which we have successfully done in the past in Hall A. Two particular kinematics with high DIS cross section will require a coincidence trigger with the PbWO4 calorimeter in order to reduce the data acquisition rate below this limit.

Reply WACS: The PbWO4 crystals of the NPS have a pulse length of ~30 ns, so that 16 samples (16x4 = 64 ns) from the 250 MHz flash ADC will be suffcient to characterize the pulse form and base line. Monte Carlo simulations of the EM shower induced in the calorimeter suggest that the bulk of a shower will be contained in a 3x3 cluster of crystals. Online, a cluster will be signaled by a hit in a single crystal which exceeds a threshold of 25% of the exepcted Compton sacttered photon energy. If this and the surrounding 8 crystals are read out, a cluster will generate 9 x 16 = 144 data words or 288 bytes of data. If one extends the sweep of neighboring crystal to a 5x5 array, then the cluster read out will extend to 800 bytes of data. Read out the the FADCs is controlled by FPGA hardware, programmed to recognize where a hit has occurred and read out only the relevant group of FADC modules, so that the generation of large amounts of non-useful data is avoided. It is expected that the total cluster read-out will be a factor of 2 larger than this value as a result of the inclusion of auxillary words in the FADC readout, leading to a conservative estimate for the NPS event size of 2 kB. Expected physics singles rates in the NPS for Compton scattering, ep scattering and pi0 photo production are given in Table 4 of the proposal, for a threshold corresponding to 25% of the elastic scattering peak energy. Atomic processes will produce low energy photons and electrons which will be suppressed very effectively by the 25% threshold on the trigger. The total singles trigger rates for the NPS, which is dominated by these background events, are given in Table 1 of the proposal. Since the trigger will be formed by the HMS, the maximum data throughput required will be at kinematics 5A. From experience gained during the previous Hall C WACS experiment, we expect the total HMS rate to be around a factor of 5 larger than the physics rate of 350 Hz for this kinematic point and the HMS event size to be less than 2 kB. Assuming a 4 MHz rate for the entire NPS at a 25 % threshold we therefore expect a maximum data rate of ~6.0 MB/s, which should be well within the capabilities of the online DAQ and data storage facilities.

• Source of fADCs: physics division. Note that this is a little different from our previous MRI, where the total number from physics division had not been sufficient

NPS COLLABORATION MEETING

• Discussion about date and location of our next NPS collaboration meeting

• Based on our last discussion the month of December seems to crowded for another meeting, so standalone meeting in November seems best
• Dates in range 11 - 19 November would have minimal conflict with other conferences/meetings. Doodle poll will be sent to select meeting date(s).
• Location at JLab has some benefits as majority of people are already there. This may also give us additional opportunity for feedback from Hall staff. CUA/DC is possible too (travel is relatively easy, but lodging is more expensive than NN). Ohio would also be possible.

ITEMS FOR NEXT MEETING:

• NPS design - info on detector front minimum distance from design studies

• Budgetary quotes for PbWO4 and PMTs

• Hamlet will send set of uniform specs based on CMS that can be used when inquiring
• Hamlet will remain in contact with Crytur and also contact several Russian companies
• TH will remain in contact with Japanese vendors
• Carlos is checking with Orsay about GSI's specs on PbWO4. Rolf may be able to get additional information too.
• Charles is checking with Hamamatsu or other vendors on R4125

• Continue discussion of NPS project funding strategies, e.g., budget optimizations

NEXT MEETING: THUR, 12 SEPTEMBER at 9AM (EDT)