EC Tests at UVa, 2015

Summary and Next step:

  1. Achieved sub-100 ps timing resolution with LASPD bar coupled to 55 degree light guide using optical cement.
  2. Reproduced IHEP preshower tile results of about 70 photoelectrons, and achieved a record 78 photoelectrons.
  3. Test tile response with DDK fiber connectors.
  4. Determine and apply position-dependent time-walk corrections.

09/25/15 04/10/15 04/8/15 #2 04/8/15 #1
03/27/15 03/20/15 02/10/15 01/26/15 01/15/15 01/09/15 01/06/15

September 25th - October 1st, 2015

  1. Time resolution measurements using 3-bar test with Fine-mesh PMT:
    Using the EJ-200 plastic scintillators (5 cm x 5 cm x 30 cm) with each end coupled with optical grease to a R9779 PMT for trigger. The middle scintillator bar was viewed by a R9779 PMT on the left-side of the bar and a 2-inch fine-mesh (H6614-70) PMT on the right-side of the bar.
  2. The HV settings for the PMTs are as follows:
    - Top right HV = -1590 V
    - Top left HV = -1025 V
    - Middle right HV = +1250 V
    - Middle left HV = -1425 V
    - Bottom right HV = -1000 V
    - Bottom left HV = -1235 V

  3. For run 1591, the HV of the FM-PMT was increased to 1450 V.
  4. For the set of measurements, the bars were stacked vertically with respect to each other. The three trigger bars were oriented parallel to each other. The vertical spacing is about 24.2 cm between the top of the top bar and the top of the bottom bar. The trigger was generated by forming an "AND" of the left and right PMT signals of the top and bottom scintillator bars. The top right bar's signal was delayed by 16 ns with respect to the other signals so that this PMT would always carry the trigger time. The middle bar was not part of the trigger.
  5. For all measurements, the PMT signals were split from the dark box with one signal passed to a discriminator channel set at -50 mV. The other signal was sent through 80 feet (123 ns) of delay cable before being sent to the QDC module.
  6. Conclusions and Observations:
    1. The timing resolution of the three bar measurments was determined using the procedure discussed in the CLAS12 FTOF report.
    2. For each run, the QDC pedestals were deterimined in order to subtract the offset from the QDC response. The pedestal was isolated by placing an anti-cut on the TDC for the corresponding QDC channel. Table 15 summarizes the pedestal and MIP response for each PMT.

      3. Table 15: Timing resolution data using three bar method with five Hamamatsu R9779 PMTs and one H6614-70 fine-mesh PMT
      run#
      		 Discriminator
      threshold [mV]       		Pedestals
      		 measured MIP positions
      		 comments
      1590
      		 -50
      		 file1
      		 file2
      		 Trigger: "AND" of four pmts
      top and bottom bars
      1591
      		 -50
      		 file3
      		 file4
      		 Trigger: "AND" of four pmts
      top and bottom bars

    3. Then the time from the TDC was plotted versus the pedestal subtracted QDC response (1/sqrt(QDC - pedestal)): runs 1590 and 1591. The top right PMT is self-timed, and hence there is no time dependence wrt the QDC signal.
    4. A profile histogram was created so that the distribution could be fit: runs 1590 and 1591. The slope from the fit was used to correct for the time-walk effect.
    5. A cut was also placed on the pedestal subtracted QDC distributions to remove the low energy events.
      For run 1590, the following QDC cuts were chosen:
      - Top right cut = 290 channels
      - Top left cut = 330 channels
      - Middle right cut = 300 channels
      - Middle left cut = 840 channels
      - Bottom right cut = 290 channels
      - Bottom left cut = 280 channels

      For run 1591, the following QDC cuts were chosen:
      - Top right cut = 260 channels
      - Top left cut = 330 channels
      - Middle right cut = 1020 channels
      - Middle left cut = 840 channels
      - Bottom right cut = 290 channels
      - Bottom left cut = 280 channels

    6. The initial time-walk slopes were optimized by varying one slope and holding the other 5 fixed until the timing resolution was minimized. This procedure was repeated for each PMT.
    7. Using the expression T = 0.5*(tt + tb) - tm, where tt, tm, and tb represent the time the particle passes through the top, middle, and bottom bars, respectively, the time resolution of the bars was determined.
    8. The time can be expressed in terms of the times of PMT signals as T = 0.25*(ttr + ttl + tbr + tbl) - 0.5*(tmr + tml), where the subscript 'r' and 'l' denote the right-side and left-side pmts, respectively, for that particular bar.
    9. Table 16 contains a summary of the time resolution results. The σ from the fit represents the convolution of all six PMTs in the system. The reference counter (middle scintillator bar) resolution is obtained by σm = sqrt(σT2 - 0.5*σr2), which is the value shown in Table 16 with σr = 58 ps from previous measurements.
    10. The timing resolution of the fine-mesh PMT can be deconvoluted by using σm2 = 0.25*(σmr2 + σml2), and σml = sqrt(2)* σr = 82 ps, since the left-side PMT is also a R9779. Solving the expresion for σmr (the fine-mesh PMT), yields σmr = σFM-PMT = 104 ps (run 1590) and σmr = σFM-PMT = 83 ps (run 1591).
    11. The higher gain on the FM-PMT for run 1591 helps to reduce the sensitivity to the time-walk effect, and we obtain the same timing resolution as achieved when using all R9779 PMTs.
    12. Table 17 shows the time response for the Hamamastsu R9779 and H6614-70 PMTs. The TTS is about a factor of 1.75 worse for the fine-mesh PMT.
    13. Using the achieved time resolution for each PMT and the light decay constant of the scintillator (τsct = 2.1 ns), the number of photoelectrons can be estimated from σPMT2 = (σTTS2 + τsct2)/NPE, which is about 640-660 photoelectrons for each PMT.

    Table 16: Timing resolution results using three bar method with Hamamatsu R9779 PMTs and 2-inch FM-PMT
    run#
    		 Raw Resolution [ps] Time Walk Corrected
    Resolution [ps]     		Time Walk Corrected with QDC cuts
    Resolution [ps]     		Time Walk Corrected with High-amp QDC cuts
    Resolution [ps]
    1590
    		 T Spectrum: 250
    		 79
    		66
    		 65
    1591
    		 T Spectrum: 213
    		 71
    		58
    		 57


    Table 17: Hamamatsu R9779 and H6614-70 (2-inch FM-PMT) Time Response
    PMT Type
    		 Anode Pulse Rise Time [ns] Electron Transit Time [ns] Transit Time Spread [ns]
    R9779
    		 1.8
    		 20
    		0.25
    H6614-70
    		 2.5
    		 9.5
    		0.44

April 10th - 21st, 2015

  1. MIP Response and Time Resolution of FASPD bars:
    Using the 5-mm thick scintillators from Eljen Technology in the shape of a trapezoid. The smaller bar is 232 mm long with sides 100 mm and 120 mm wide, and the larger bar is 444 mm long.
  2. The bars were first wrapped in aluminized mylar, and the covered in black tape.
  3. Optical grease was applied inside the scintillators' grooves.
  4. For the smaller bar, one 1.0-mm diameter, Y11 (200 ppm) fiber was used. The fiber was cut from the 500-m spool with a length of 212 cm. After the fiber was embedded into the scintillator groove, the fiber ends were cut off so that only 20 cm of fiber was not embedded in the tile. The fiber ends were then polished with polishing paper from JLab.
  5. The fiber inside the groove has 3 turns for an approximate embedded length of 168 cm and a total length of 208 cm, which includes the length of fiber outside the scintillator.
  6. For the 444-mm long bar, two 1.0-mm diameter, Y11 (200 ppm) fiber were used. The fibers were cut from the 500-m spool with a length of 240 cm. After the fiber was embedded into the scintillator groove, the fiber ends were cut off so that only 10 cm of fiber was not embedded in the tile. The fiber ends were then polished with polishing paper from JLab.
  7. The each fiber inside the groove has 2 turns for an approximate embedded length of 220 cm and a total length of 240 cm, which includes the length of fiber outside the scintillator. The combined embedded length for the two fibers is 440 cm.
  8. After the fibers were embedded, the bars were first wrapped in aluminized mylar, and then covered in black tape.
  9. Optical grease was coated onto the pmt cathode surface, and the four fiber ends were placed up against the coated cathode. The fibers were held rigid by placing them inside of a DDK connector ferrule.
  10. Conclusions and Observations:
    1. The new data can be compared to the SDU tile tests from January begining at bullet #7, in particular, the results in Table 8 for runs 1525 and 1526. In that configuration, a 1-mm diameter fiber of length 111 cm was used with 2.5 turns. The number of photoelectrons, averaging the two results together is 11.2.
    2. Using an attenuation length of 3.5 m for the Kuraray fibers, we expect an reduction in amplitude by about 25%. However, the number of fiber turns increased from 2.5 turns to 3.0 turns, which will partially compensate the loss due to attenuation. The measured result is a loss of about 18% between the two measurements.
    3. For run 1554, the signal from the FASPD bar was sent through the 50-50 splitter to measure the timing resolution. The procedure outlined in mid-January was followed to determine the timewalk corrections and the timing resolution using the three bar method. However in this case, the FASPD was only viewed by one PMT. The equations were adjusted to account for this fact. Also the FASPD's light was collected by a XP2262, whereas the trigger bar PMTs are R9779's. The XP2262 has slightly worse timing resolution. The achieved timing resolution after time-walk corrections is about 2.5 ns.
    4. The N.P.E. for run 1554 appears to be consistent with runs 1525 and 1526. However, the S.P.E. is sitting on the tails from the pedestal and MIP in 1554. More than likely the true location is a bit higher.
    5. Applying cuts on the QDC spectra, did not have much affect on the timing resolution. The spectra can be viewed from the link in Table 14.
    6. Run 1555 is with the 444 mm long FASPD bar. The PMT signal was sent through the 50-50 splitter, and the PMT HV was increased to -2200 V in an attempt to help isolate the single P.E, though the SPE was not isolated.
    7. Run 1556 is with the 444 mm long FASPD bar, but the 50-50 splitter was bypassed and the PMT HV was decreased back to -2100 V. The SPE is seen, but it is sitting on top of the tail of the multi-photoelectron peak. Normally, the SPE at this gain is between channel 29 to 31, which would give 10.4 to 11.2 photoelectrons for the MIP.

    Table 14: FASPD MIP Response and Timing Resolution Results
    run#
    		 Bar Length (cm)
    		# fiber turns
    		 fiber length embeded (cm)
    		 measured single p.e. position
    		 measured MIP position
    		 measured # p.e.
    using single p.e. position     		Time resolution (ns) comments
    1552
    		 23.2
    		 1 fiber
    3 turns     		168
    		 30.9
    		 QDC Spectra: 295 9.5 ---- XP2262 PMT, S/N 24561
    HV = -2100 V
    1554
    		 23.2
    		 1 fiber
    3 turns     		168
    		 13.1
    		 QDC Spectra: 145 11.1 T Spectrum: 2.5 XP2262 PMT, S/N 24561
    HV = -2100 V
    50-50 Splitter to TDC
    1555
    		 44.4
    		 2 fibers
    2 turns each fiber     		440
    		 ----
    		 QDC Spectra: 261 ---- N/D XP2262 PMT, S/N 24561
    HV = -2200 V
    50-50 Splitter to TDC
    1556
    		 44.4
    		 2 fibers
    2 turns each fiber     		440
    		 35.4
    		 QDC Spectra: 324 9.2 N/A XP2262 PMT, S/N 24561
    HV = -2100 V
    Direct to QDC, no splitter

April 8th - 10th, 2015

  1. IHEP Tile Reproducibility Checks:
    Using the IHEP Preshower hexagon tile with two 1.0-mm diameter, 1.5-m long Y11 (200 ppm) fiber. The fiber ends were cut off so that only 16.5 cm of fiber is not embedded in the tile. The fiber ends were polished with polishing paper from JLab.
  2. Optical grease was coated onto the pmt cathode surface, and the four fiber ends were placed up against the coated cathode. The fibers were held rigid by placing them inside of a DDK connector ferrule.
  3. Optical grease was also applied in the grooves of the tile.
  4. Two different PMTs were used with serial numbers: 24561 and 26799 to check reproducibility and the possibility that the noise seen in S/N 24561 might affect the results that were obtained with this PMT.
  5. Even though the noise level appears to have increased in PMT 24561, the number of photoelectrons using the two PMTs are consistent at the 3% level, which is within the measurement uncertainties. So the results with PMT 24561 should be reliable.

    6. Table 13: Reproduction of Multi-fiber test with IHEP Tile
    run#
    		 # fiber turns
    		 fiber length embeded (cm)
    		 measured single p.e. position
    		 measured MIP position
    		 measured # p.e.
    using single p.e. position     		comments
    1549
    		 2.5 each fiber
    		 155.4
    		 26.0
    		 QDC Spectra: 2419 93.2 XP2262 PMT, S/N 24561
    HV = -2050 V
    1550
    		 2.5 each fiber
    		 155.4
    		 28.4
    		 QDC Spectra: 2726 96.0 XP2262 PMT, S/N 26799
    HV = -1950 V

April 8th, 2015

  1. Yield Analysis of Preshower Tile Measurements:
    Using the measurement results from the IHEP, CNCS and Kedi tiles, an analysis of the SPE and MIP yields was performed to study the relative yield between the two peaks.
  2. For this analysis, only the measurements using the XP2262 PMT with serial number: 24561 were considered.
  3. Since run 1526 with the IHEP tile, the yield of the single photoelectron peak has grown relative to the MIP peak. This analysis will provide a quantitative value for the size of the change in yield across the preshower measurements.
  4. The yields for the SPE and MIP peaks were determined by integrating the area under the histogram distribution. The same bin size was chosen for all runs with the following sizes:
    - SPE: 16 bins with QDC values 0 to 120
    - MIP: 100 bins with QDC values 0 to 4000.
  5. After run 1539, the HV of the PMT was from -2100 V to -2050 V, since the MIP amplitude was close to the saturation point of the QDC. Before lowering the HV, the relative yield of the MIP had decreased by about a factor of 10 from the earlier IHEP tile measurements. After lowering the HV, the relative yield improved so that the decrease was between a factor of 2 to 6 compared to the earlier IHEP measurements.
  6. The fact that the relative yield of the MIP improved after decreasing the HV appears to indicate that the decrease in yield might be related to the PMT being noisy. There also seems to be a trend that the PMT's performance is worsening with time, since the relative yield appears to be slowly dropping with time.

    7. Table 12: SPE and MIP Yields from IHEP, Keid, and CNCS Tiles
    run#
    		 Tile
    		 SPE Yield
    		 MIP Yield
    		 Relative Yield
    = MIP/SPE     		QDC Integration Ranges
    1526
    		 IHEP
    		 353
    		 13051
    		 37.0
    		 SPE: 10-60
    MIP: 1480-4000
    1539
    		 Kedi # 1
    		 1596
    		 6136
    		 3.8
    		 SPE: 10-60
    MIP: 1480-4000
    1537
    		 Kedi # 2
    		 5176
    		 20637
    		 4.0
    		 SPE: 10-60
    MIP: 1480-4000
    1542
    		 Kedi # 3
    		 463
    		 8387
    		 18.1
    		 SPE: 10-53
    MIP: 1000-4000
    1540
    		 Kedi # 4
    		 1383
    		 14508
    		 10.5
    		 SPE: 10-53
    MIP: 1000-4000
    1543
    		 CNCS # 1
    		 3038
    		 24602
    		 8.1
    		 SPE: 10-53
    MIP: 1000-4000
    1544
    		 CNCS # 2
    		 720
    		 9019
    		 12.5
    		 SPE: 10-53
    MIP: 1000-4000
    1546
    		 CNCS # 3
    		 3535
    		 37082
    		 10.5
    		 SPE: 10-53
    MIP: 1000-4000
    1547
    		 CNCS # 4
    		 1426
    		 8624
    		 6.0
    		 SPE: 10-53
    MIP: 1000-4000
    1548
    		 CNCS # 4
    		 1297
    		 9935
    		 7.7
    		 SPE: 10-53
    MIP: 1000-4000

March 27th - April 8th, 2015

  1. 2-fiber Preshower Tile from CNCS Tests:
    Using the CNCS Preshower hexagon tiles with two 1.0-mm diameter, 1.2-m long Y11 (200 ppm) fiber. The fiber ends were cut off so that only about 19-20 cm of fiber is not embedded on each fiber end.
  2. Optical grease was coated onto the surface of pmt cathodes, and the four fiber ends were placed up against the coated cathode. The fibers were held rigid using either a DDK ferrule.
  3. Optical grease was applied in the grooves of the tile.
  4. The 5 cm thick scintillator counters used in the 3-bar tests formed the trigger. The HVs remained the same for these counters. For the preshower tiles, two XP2262 PMTs were used with HVs = -2050 V and -2000 V.
  5. Conclusions and Observations:
    1. The surfaces of the CNCS tiles have some scratches and blemishes.
    2. The width of the groove is 1 mm.
    3. For CNCS tiles, the groove depth is 6 mm. The grooves are 1 mm wide for both tiles. The diameter of the groove is 9 cm.
    4. The length of the groove outside the circular region for the fibers to pass outside the tile is about 2.7 cm.
    5. For CNCS tile #2, two of the fibers had pulled away from the cathode's surface. However, it has been consistently seen that the system using this PMT yields a lower number of photoelectrons.
    6. For run 1544, the fiber ends were trimmed and repolished for CNCS tile #2, since one of the fiber ends was bent. Also the PMTs used for the two tiles were swapped to see if the results for CNCS tile #1 can be reproduced with the poor performance PMT.
    7. The performance of CNCS tile #1 greatly improved after swapping the PMT and trimming the fiber ends. On the other hand, tile #2's n.p.e. decreased by about 38%.
    8. For run 1546 with CNCS tiles #3 and #4, the PMT with S/N 23963 was swapped to the PMT with S/N 26799, since the previous test consistently shows a lower photoelectron yield using the S/N 23963 PMT. Also tile #4 had a piece of plastic that needed to be cut away inside the groove in the region where the circular part and the straight end come out on the left-hand side of the tile.
    9. For run 1547, the PMT and tile configuration was swapped. CNCS tile #3 was paired with PMT S/N 26799, and CNCS tile #4 was paired with S/N 24561 to test reproducibility with different PMTs. There was about a 9% improved for the tile #4, which is more consistent with tiles 1 and 2. However, tile #3 dropped by about 20 photoelectrons. It was found that for tile #3, the fibers had slide towards the top of the cathode's surface. The fibers were repositioned to the center of the cathode, and a new run was started.
    10. For run 1548, after repositioning the fibers, CNCS tile #4 gained about 4-5 photoelectrons, but it seems to be mostly related to the SPE changing from 25 to 24 QDC channels. For CNCS tile #3, 80 photoelectrons was achieved, but this is still about 14% low compared to a high of almost 93 photoelectrons achieved previously.

    Table 11: CNCS Tile, Multi-fiber test results (two Y11, 1mm dia, 1.177m length)
    run#
    		 Tile
    		 # fiber turns
    		 fiber length embeded (cm)
    		 measured single p.e. position
    		 measured MIP position
    		 measured # p.e.
    using single p.e. position     		comments
    1543
    		 CNCS #1
    		 2.5 each fiber
    		 151
    		 24.1
    		 QDC Spectra: 2011 83.4 XP2262 PMT, S/N 24561
    HV = -2050 V, Tyvek wrapping
    1543
    		 CNCS #2
    		 2.5 each fiber
    		 151
    		 26.6
    		 QDC Spectra: 1024 38.4 XP2262 PMT, S/N 23963
    HV = -2000 V, Tyvek wrapping
    1544
    		 CNCS #1
    		 2.5 each fiber
    		 151
    		 26.2
    		 QDC Spectra: 1588 60.6 XP2262 PMT, S/N 23963
    HV = -2000 V, PMT swapped compared with run 1543
    1544
    		 CNCS #2
    		 2.5 each fiber
    		 151
    		 25.4
    		 QDC Spectra: 2151 84.7 XP2262 PMT, S/N 24561
    HV = -2050 V, PMT swapped compared with run 1543
    1546
    		 CNCS #3
    		 2.5 each fiber
    		 151
    		 25.0
    		 QDC Spectra: 2320 92.8 XP2262 PMT, S/N 24561
    HV = -2050 V
    1546
    		 CNCS #4
    		 2.5 each fiber
    		 151
    		 28.9
    		 QDC Spectra: 2151 74.4 XP2262 PMT, S/N 26799
    HV = -1950 V
    1547
    		 CNCS #3
    		 2.5 each fiber
    		 151
    		 29.1
    		 QDC Spectra: 2109 72.5 XP2262 PMT, S/N 26799
    HV = -1950 V
    1547
    		 CNCS #4
    		 2.5 each fiber
    		 151
    		 25.1
    		 QDC Spectra: 2032 81.0 XP2262 PMT, S/N 24561
    HV = -2050 V
    1548
    		 CNCS #3
    		 2.5 each fiber
    		 151
    		 28.4
    		 QDC Spectra: 2271 80.1 XP2262 PMT, S/N 26799
    HV = -1950 V
    1548
    		 CNCS #4
    		 2.5 each fiber
    		 151
    		 24.2
    		 QDC Spectra: 2077 85.8 XP2262 PMT, S/N 24561
    HV = -2050 V

March 20th - 27th, 2015

  1. 2-fiber Preshower Tile from Kedi Tests:
    Using the Kedi Preshower hexagon tiles with two 1.0-mm diameter, 1.2-m long Y11 (200 ppm) fiber. The fiber ends were cut off so that only about 20.0 cm of fiber is not embedded on each fiber end.
  2. Optical grease was coated onto the surface of pmt cathodes, and the four fiber ends were placed up against the coated cathode. The fibers were held rigid using either a plastic block from the machine shop or a DDK ferrule.
  3. Optical grease was applied in the grooves of the tile.
  4. The 5 cm thick scintillator counters used in the 3-bar tests formed the trigger. The HVs remained the same for these counters. For the preshower tiles, two XP2262 PMTs were used with HVs = -2100 V and -2025 V.
  5. Conclusions and Observations:
    1. The surfaces of the Kedi tiles have some scratches and blemishes.
    2. For Kedi tile #1, the groove depth varies from 6 mm to 7 mm. The groove depth of Kedi tile #2 appears to be roughly uniform at 6 mm deep. The grooves are 2 mm wide for both tiles. The diameter of the grooves is 9 cm. Tile #3 and #4 have the same dimensions and groove sizes compared to tile #2. Kedi tile #3 has a small chipped at one of the hexagon corners, and it has a larger chip at the middle of one of the hexagon's sides: image1 and image2.
    3. The length of the groove outside the circular region for the fibers to pass outside the tile is about 2.4 cm compared to 3.5 cm for the IHEP tiles.
    4. When embedding the fibers and wrapping the tiles in Tyvek paper, the fibers were particularly hard to keep in the grooves while wrapping. This could be due to using fibers cut directly from 500-m spool and/or due to the grooves not being deep enough.
    5. As was seen previously, the ferrule provides the better connectivity to the PMT surface.
    6. After run 1537, it was found that some of the fiber ends had an air gap between the fiber and the cathod surface. The PMT viewing Kedi #1 was recoated with optical grease, and the fibers were better positioned to touch the surface of the PMT's cathode. However, the SPE was difficult to find for run 1538 with Kedi tile #1, and it appeared to shift considerably in channel number (30 ----> 40).
    7. For run 1539, the DDK connector was used to hold the fibers from Kedi tile #1 and the PMT used for the measurements on Kedi tile #2 was used instead. Using the same connector for the fibers and the same PMT results in a consistent n.p.e. (84-87) at the 3% level.
    8. Even though a DDK connector was used for the fibers from tile #3, the n.p.e. is low compared to the others. It was noted that the fiber lengths were slightly different, which caused at least one fiber to pull away from the cathode's surface. For run 1542, the fiber ends were repositioned to obtain a better coupling to the PMT's cathod surface and the fiber ends were trimmed to be the same length.
    9. For all four Kedi tiles, the maximum difference in the n.p.e. is (84-91), with an average of 87.2 and a spread of about ± 4%.

    Table 10: Kedi Tile, Multi-fiber test results (two Y11, 1mm dia, 1.177m length)
    run#
    		 Tile
    		 # fiber turns
    		 fiber length embeded (cm)
    		 measured single p.e. position
    		 measured MIP position
    		 measured # p.e.
    using single p.e. position     		comments
    1537
    		 Kedi #1
    		 2.5 each fiber
    		 151
    		 29.4
    		 QDC Spectra: 1469 50.0 XP2262 PMT, HV = -2025 V
    Tyvek wrapping
    1537
    		 Kedi #2
    		 2.5 each fiber
    		 151
    		 33.2
    		 QDC Spectra: 2801 84.4 XP2262 PMT, HV = -2100 V
    Tyvek wrapping
    1538
    		 Kedi #2
    		 2.5 each fiber
    		 151
    		 33.0
    		 QDC Spectra: 2847 86.4 XP2262 PMT, HV = -2100 V
    Reproducibility of Run 1537
    1539
    		 Kedi #1
    		 2.5 each fiber
    		 151
    		 32.0
    		 QDC Spectra: 2786 87.1 XP2262 PMT, HV = -2100 V
    Swapped Connectors and PMTs
    1540
    		 Kedi #3
    		 2.5 each fiber
    		 151
    		 32.0
    		 QDC Spectra: 1749 54.7 XP2262 PMT, HV = -2025 V
    Tyvek wrapping
    1540
    		 Kedi #4
    		 2.5 each fiber
    		 151
    		 24.2
    		 QDC Spectra: 2205 91.0 XP2262 PMT, HV = -2050 V
    Tyvek wrapping
    1542
    		 Kedi #3
    		 2.5 each fiber
    		 151
    		 25.0
    		 QDC Spectra: 2176 87.0 XP2262 PMT, HV = -2050 V
    Trimmed Fiber ends

February 10th - 18th, 2015

  1. DDK Fiber Connector Tests with IHEP Tile:
    Using the IHEP Preshower hexagon tile with two 1.0-mm diameter, 1.5-m long Y11 (200 ppm) fiber. The fiber ends were cut off so that only 14.5 cm of fiber is not embedded in the tile. The fiber ends were polished with polishing paper from JLab.
  2. Optical grease was applied in the grooves of the tile.
  3. The 5 cm thick scintillator counters used in the 3-bar tests formed the trigger. The HVs remained the same for these counters. For the preshower tile, a XP2262 PMT was used with a HV = -2100 V.
  4. For these tests, the WLS fibers were placed in a DDK connector in one ferrule, and 4.5 cm clear fiber pieces were placed in anothre ferrule. The two ferrules were then meshed together inside the box component of the DDK connector. The two ferrules are held in place with clips.
  5. For the first measurement, only optical grease was used to couple the two sets for fibers inside the ferrules.
  6. For the second measurement, optical cement was applied to the two sets for fibers inside the ferrules to keep the fibers rigid.
  7. For the third measurement, a third ferrule was used to optically cement one end of four 2-m clear fibers, which was then meshed together with the ferrule holding the WLS fibers.
  8. Conclusions and Observations:
    1. The MINERvA collaboration measured average connector transmissions of about 87.0% with a standard deviation of 4.8%: arXiv:1305.5119v1.
    2. Without optically cementing the fibers into the connectors, the fibers can move, which causes less than ideal coupling between the WLS and clear fibers.
    3. The measurement with with the third ferrule measures the convolution of both the DDK connector transmission and the attenuation of the 2-m fiber length. It is assumed that the DDK connector transmission was 84% as with the previous test with 4.5 cm clear fibers.
    4. A range of values is provided for the transmission, since only once has about 78 photoelectrons been acheived.

    Table 9: DDK Connector Multi-fiber test results (two Y11, 1mm dia, 1.177m length)
    run#
    		 # fiber turns
    		 fiber length embeded (cm)
    		 measured single p.e. position
    		 measured MIP position
    		 measured # p.e.
    using single p.e. position     		Transmission NPEmax = 71-78 comments
    1529
    		 2.5 each fiber
    		 155.4
    		 31.5 ± 0.8
    		 QDC Spectra: 1406 44.7 57-63% XP2262 PMT, HV = -2100
    DDK Connector using grease
    1530
    		 2.5 each fiber
    		 155.4
    		 31.4 ± 0.6
    		 QDC Spectra: 1876 59.8 77-84% XP2262 PMT, HV = -2100
    DDK Connector using cement
    1532
    		 2.5 each fiber
    		 155.4
    		 30.3 ± 0.6
    		 QDC Spectra: 1229 40.6 67.8%1 XP2262 PMT, HV = -2100
    DDK Connector using cement
    coupled to 4x 2-m clear fibers
    1Assuming 59.8 p.e. after transmission through the DDK connector. This number should represent
    the percentage of photoelectrons after attenuation through 2-m of clear fiber.

January 26th - February 10th, 2015

  1. Reproduction of Multi-fiber test from September 2014:
    Using the IHEP Preshower hexagon tile with two 1.0-mm diameter, 1.5-m long Y11 (200 ppm) fiber. The fiber ends were cut off so that only 20 cm of fiber is not embedded in the tile. The fiber ends were polished with polishing paper from JLab.
  2. Optical grease was coated onto the pmt cathode surface, and the four fiber ends were placed up against the coated cathode. The fibers were held rigid using a plastic block from the machine shop.
  3. Optical grease was also applied in the grooves of the tile.
  4. The 5 cm thick scintillator counters used in the 3-bar tests formed the trigger. The HVs remained the same for these counters. For the preshower tile, a R9779 PMT was used, though it was found the gain was too low even at -1700 V (97% of maximum) to isolate the single photoelectron peak from the pedestal noise.
  5. Additional tests were performed either with the previously used XP2262 PMT or a R11102 PMT.
  6. Conclusions and Observations:
    1. Even with the XP2262, the light collection was poor. The tyvek paper was originally wrapped on the preshower tile in September 2014. The paper was worn and light could easily leak from around the edges of the paper. With a fresh piece of tyvek paper, the light collection dramatically improved.
    2. For run 1522, the XP2262 was swapped with a R11102 set to -1400 V (93% of maximum). The relative gain between the two PMTs based on HV setting was about a factor of 10 in favor of the XP2262. The measured MIP position with the two PMTs was actually only a factor of 5 higher for the XP2262. However, the light collection efficiency from the fibers to PMTs is very sensitive to connection to the PMT surface. Hence, it cannot be guaranteed that the light collection was the same between the two runs.
    3. Runs 1525 and 1526 were taken back-to-back without any changes in the system. The single p.e. peak was not promienent in run 1525, though run 1526 does have about 2.7 times more statistics. 78 photoelectrons is an achieved record for the preshower tile.

    Table 7: Multi-fiber test results (two Y11, 1mm dia, 1.177m length)
    run#
    		 # fiber turns
    		 fiber length embeded (cm)
    		 measured single p.e. position
    		 measured MIP position
    		 measured # p.e.
    using single p.e. position     		comments
    1516
    		 2.5 each fiber
    		 155.4
    		 Not Found
    		 QDC Spectra: 519 ----- R9779 PMT, HV = -1700 V
    Tyvek wrapping from 09/2014
    1518
    		 2.5 each fiber
    		 155.4
    		 31.4 ± 2.2
    		 QDC Spectra: 1617
    		 51.5
    		 XP2262 PMT, HV = -2100 V
    Tyvek wrapping from 09/2014
    1521
    		 2.5 each fiber
    		 155.4
    		 27.5 ± 1.2
    		 QDC Spectra: 1767
    		 64.2 XP2262 PMT, HV = -2100 V
    New tyvek paper
    Recentered fibers on PMT
    1522
    		 2.5 each fiber
    		 155.4
    		 Not Found
    		 QDC Spectra: 355
    		 -----
    		 R11102 PMT, HV = -1400 V
    1525
    		 2.5 each fiber
    		 155.4
    		 36.6 ± 2.1
    		 QDC Spectra: 2503
    		 68.4
    		 XP2262 PMT, HV = -2100 V
    Reproducibility of Run 1521
    1526
    		 2.5 each fiber
    		 155.4
    		 32.1 ± 2.1
    		 QDC Spectra: 2499
    		 77.8
    		 XP2262 PMT, HV = -2100 V
    Reproducibility of Run 1525

  7. 3-mm and 5-mm SDU Tile Fiber tests:
    Using the SDU square tiles with one 1.0-mm diameter, 1.5-m long Y11 (200 ppm) fiber. The fiber ends were cut off so that only 20 cm of fiber is not embedded in the tile. The cut fiber end was polished with polishing paper from JLab.
  8. Optical grease was coated onto the pmt cathode surface, and the two fiber ends were placed up against the coated cathode. The fibers were held rigid using a fiber connector from DDK.
  9. Optical grease was also applied in the grooves of the tile.
  10. The 5 cm thick scintillator counters used in the 3-bar tests formed the trigger. The HVs remained the same for these counters. For the SDU tiles, a R11102 PMT was used, though it was found the gain was too low even at -1400 V (93% of maximum) to isolate the single photoelectron peak from the pedestal noise.
  11. Additional tests were performed with a previously used XP2262 PMT.
  12. SDU tile dimensions: 10 cm X 10 cm X 0.3 cm or 10 cm X 10 cm X 0.5 cm (length X width X thickness), the groove diameter is 8 cm with 4.1 cm long straight grooves for the fiber to pass outside the tile.
  13. The 3-mm tile has a 2-mm deep groove, and the 5-mm tile has a 3-mm deep groove.
  14. These tests were conducted in parallel with the preshower tile tests with the SDU tile lying on top of the preshower tile.
  15. Conclusions and Observations:
    1. For run 1521, the "MIP" was isolated by cutting on the preshower tile MIP, which helped to greatly suppress the pedestal of the SDU tile. However, it is likey that this peak is a combination of MIP and the single p.e. peak.
    2. For run 1522, the R11102 was swapped with the preshower XP2262 set to -2100 V. The relative gain between the two PMTs based on HV setting was about a factor of 10 in favor of the XP2262. The peak at channel 30 is probably the single p.e. peak, and the small peak near channel 60 is probably the MIP location, though there is very little signal above the single p.e. peak.
    3. Data for the 5-mm tile with 2.5 fiber turns were taken at two HV settings: -1900 V and -2025 V. Both runs show a consistent result of about 11-11.5 photoelectrons.
    4. Data for the 5-mm tile with 1.5 fiber turns appears to indicate about 7 photoelectrons.

    Table 8: SDU Tile, Fiber test results (one Y11, 1mm dia, 1.177m length)
    run#
    		 # fiber turns
    		 fiber length embeded (cm)
    		 measured single p.e. position
    		 measured MIP position
    		 measured # p.e.
    using single p.e. position     		comments
    1521
    		 1.5
    		 45.9
    		 Not Found
    		 QDC Spectra: 8.26
    		 ----- R11102 PMT, HV = -1400 V
    3 mm tile
    1522
    		 1.5
    		 45.9
    		 30.0 ± 0.5
    		 QDC Spectra: 59.7
    		 2.0
    		 XP2262 PMT, HV = -2100 V
    3 mm tile
    1525
    		 2.5
    		 71.0
    		 14.7 ± 0.6
    		 QDC Spectra: 167.3
    		 11.4
    		 XP2262 PMT, HV = -1900 V
    5 mm tile
    1526
    		 2.5
    		 71.0
    		 33.5 ± 0.6
    		 QDC Spectra: 365.2
    		 10.9
    		 XP2262 PMT, HV = -2025 V
    5 mm tile
    1528
    		 1.5
    		 45.9
    		 31.0 ± 0.4
    		 QDC Spectra: 227.2
    		 7.3
    		 XP2262 PMT, HV = -2025 V
    5 mm tile

January 15th - 26th, 2015

  1. Time resolution measurements using 3-bar tests Part 3 with LASPD bar and 0 degree and 55 degree light guides:
    Using the EJ-200 plastic scintillators (5 cm x 5 cm x 30 cm) with each end coupled with optical grease to a R9779 PMT for trigger. The middle bar is a 2 cm (thick) x 8.3 cm to 14.0 cm (wide) x 57.0 cm long bar. The narrow bar's end was wrapped so that only the 5 cm center section was exposed to couple to the the PMT. The wide end of the bar was attached to a light guide in the following configurations:
    * 0 degree light guide with optical grease (2 cm x 14 cm to 3.9 cm diameter, 5.05 cm long with 3.9 cm diameter x 1.27 cm disc)
    * 0 degree light guide with optical cement (2 cm x 14 cm to 3.9 cm diameter, 5.05 cm long with 3.9 cm diameter x 1.27 cm disc)
    * 55 degree light guide with optical cement (2 cm x 14 cm to 3.9 cm diameter, with 3.9 cm diameter x 1.27 cm disc at 55 deg angle)
    The same HV settings from the previous tests were used.
  2. The configuration of the bars was the same as the previous tests; however, the LASPD bar was approximately centered wrt the trigger bars.
  3. For all measurements, the PMT signals were split from the dark box with one signal passed to a discriminator channel set at -50 mV. The other signal was sent through 80 feet (123 ns) of delay cable before being sent to the QDC module.
  4. Conclusions and Observations:
    1. For each run, the QDC pedestals were deterimined in order to subtract the offset from the QDC response. The pedestal was isolated by placing an anti-cut on the TDC for the corresponding QDC channel. Table 5 summarizes the pedestal and MIP response for each PMT. For the trigger bars, dedicated pedestal runs were taken with the aid of a pulse generator to monitor the pedestals' locations.

      2. Table 5: LASPD bar with Light Guides and Hamamatsu R9779 PMTs, Timing resolution data using three bar method
      run#
      		 Discriminator
      threshold [mV]       		Pedestals
      		 measured MIP positions
      		 comments
      1501
      		 -50
      		 ped spectra
      		 MIP spectra
      		 Trigger: "AND" of four pmts
      top and bottom bars
      0 degree light guide optical grease coupling
      1502
      		 -50
      		 ped spectra
      		 MIP spectra
      		 Trigger: "AND" of four pmts
      top and bottom bars
      0 degree light guide optical grease coupling
      1507
      		 -50
      		 ped spectra
      		 MIP spectra
      		 Trigger: "AND" of four pmts
      top and bottom bars
      55 degree light guide optical cement coupling
      1511
      		 -50
      		 ped spectra
      		 MIP spectra
      		 Trigger: "AND" of four pmts
      top and bottom bars
      0 degree light guide optical cement coupling

    2. Then the time from the TDC was plotted versus the pedestal subtracted QDC response (1/sqrt(QDC - pedestal)): runs 1501, 1502, 1507 and 1511. The top right PMT is self-timed, and hence there is no time dependence wrt the QDC signal.
    3. A cut was also placed on the pedestal subtracted QDC distributions to remove the low energy events.
      For run 1501, the following QDC cuts were chosen:
      - Top right cut = 330 channels
      - Top left cut = 440 channels
      - Middle right cut = 240 channels
      - Middle left cut = 300 channels
      - Bottom right cut = 370 channels
      - Bottom left cut = 350 channels

      For run 1502, the following QDC cuts were chosen:
      - Top right cut = 340 channels
      - Top left cut = 470 channels
      - Middle right cut = 240 channels
      - Middle left cut = 260 channels
      - Bottom right cut = 390 channels
      - Bottom left cut = 370 channels

      For run 1507, the following QDC cuts were chosen:
      - Top right cut = 310 channels
      - Top left cut = 440 channels
      - Middle right cut = 300 channels
      - Middle left cut = 300 channels
      - Bottom right cut = 340 channels
      - Bottom left cut = 330 channels

      For run 1511, the following QDC cuts were chosen:
      - Top right cut = 300 channels
      - Top left cut = 420 channels
      - Middle right cut = 240 channels
      - Middle left cut = 320 channels
      - Bottom right cut = 340 channels
      - Bottom left cut = 330 channels

    4. A profile histogram was created so that the distribution could be fit: runs 1501, 1502, 1507 and 1511. The slope from the fit was used to correct for the time-walk effect.
    5. The initial time-walk slopes were then optimized by varying one slope and holding the other 5 fixed until the timing resolution was minimized. This procedure was repeated for each PMT.
    6. Using the expression T = 0.5*(tt + tb) - tm, where tt, tm, and tb represent the time the particle passes through the top, middle, and bottom bars, respectively, the time resolution of the bars was determined.
    7. The time can be expressed in terms of the times of PMT signals as T = 0.25*(ttr + ttl + tbr + tbl) - 0.5*(tmr + tml), where the subscript 'r' and 'l' denote the right-side and left-side pmts, respectively, for that particular bar.
    8. The timing distributions for data are shown in four panels in Table 6:
      - the top-left panel is without any corrections
      - the top-right panel includes the global timewalk correction for each PMT applied
      - the bottom-left panel include the global timewalk corrections with low-amplitude QDC cuts.
      - the bottom-right panel include the global timewalk corrections with low- and high-amplitude QDC cuts.
    9. The timewalk corrections improve the resolution by a factor of 2-3, and an additional 15-20% is achieved with the QDC cuts. At high QDC amplitudes the statistics are poor, and the timewalk corrections extrapolate into this region, which can cause a tail in the timing distribution. Before optimizing the time-walk coefficients, placing high-amplitude QDC cuts could improve the timing resolution by 30-40 ps. After optimizing the coefficients, the high-amplitude QDC cuts provide only a slight improvement of 2-5 ps.
    10. Table 6 contains a summary of the time resolution results for each run. The σ from the fit represents the convolution of all six PMTs in the system. The counter (scintillator bar) resolution is obtained from the expression: σm = sqrt(σT2 - 0.5*σr2), which is the value shown in
      Table 6, where σr = 58 ps.

    Table 6: LASPD bar with Light Guides and Hamamatsu R9779 PMTs, timing resolution results using three bar method
    run#
    		 Raw Resolution [ps] Time Walk Corrected
    Resolution [ps]     		Time Walk Corrected with QDC cuts
    Resolution [ps]     		Time Walk Corrected with High-amp QDC cuts
    Resolution [ps]
    1501
    		 T Spectrum: 304
    		 139
    		118
    		 115
    1502
    		 T Spectrum: 323
    		 141
    		 121
    		 119
    1507
    		 T Spectrum: 288
    		 115
    		98
    		 96
    1511
    		 T Spectrum: 283
    		 128
    		 106
    		 104

January 9th - 15th, 2015

  1. Time resolution measurements using 3-bar tests Part 3 with LASPD bar:
    Using the EJ-200 plastic scintillators (5 cm x 5 cm x 30 cm) with each end coupled with optical grease to a R9779 PMT for trigger. The middle scintillator bar was replaced with a 2 cm (thick) x 8.3 cm to 14.0 cm (wide) x 57.0 cm long bar. The bar's ends were wrapped so that only the 5 cm center section was exposed to couple to the the PMTs. The HVs for the six PMTs were approximately gain matched.
  2. The HV settings for the PMTs are as follows:
    - Top right HV = -1590 V
    - Top left HV = -1025 V
    - Middle right HV = -1360 V
    - Middle left HV = -1425 V
    - Bottom right HV = -1000 V
    - Bottom left HV = -1235 V

  3. The area of the reference bars (5 cm x 5 cm x 30 cm) is 150 cm2, and the area of LASPD bar is 635.55 cm2. The area of the LASPD bar is about 4.25 times larger than the trigger bars.
  4. For the set of measurements, the bars were stacked vertically with respect to each other. The three bars were oriented parallel to each other. The vertical spacing is about 24.2 cm between the top of the top bar and the top of the bottom bar. The LASPD bar is resting on a small tape spool (~ 2 cm high) to raise the bar so that it is centered vertically with respect to the R9779 tubes. The trigger was generated by forming an "AND" of the left and right PMT signals of the top and bottom scintillator bars. The top right bar's signal was delayed by 16 ns with respect to the other signals so that this PMT would always carry the trigger time. The middle bar was not part of the trigger.
  5. Since the LASPD bar is approximately twice the length of the trigger bars, two measurements will be done where the LASPD bar is moved relative to the trigger bars. During the first measurement, the LASPD bar's right end is located at the same position as the trigger bars' right ends. For the second measurement, the left ends of the bars will be matched.
  6. For all measurements, the PMT signals were split from the dark box with one signal passed to a discriminator channel set at -50 mV. The other signal was sent through 80 feet (123 ns) of delay cable before being sent to the QDC module.
  7. Conclusions and Observations:
    1. For each run, the QDC pedestals were deterimined in order to subtract the offset from the QDC response. The pedestal was isolated by placing an anti-cut on the TDC for the corresponding QDC channel. Table 3 summarizes the pedestal and MIP response for each PMT.
    2. For run 1494, the first run, it was found that the QDC and TDC information was mismatched due to a DAQ error. After run 1494, the VME crate was rebooted, and after run 1495, the CODA crl file was modified to look for and clear any extra words in either module after an event has been readout.
      3. Table 3: LASPD bar Timing resolution data using three bar method with Hamamatsu R9779 PMTs
      run#
      		 Discriminator
      threshold [mV]       		Pedestals
      		 measured MIP positions
      		 comments
      1494
      		 -50
      		 file1
      		 file2
      		 Trigger: "AND" of four pmts
      top and bottom bars
      QDC and TDC information out of synchronization
      1496
      		 -50
      		 file3
      		 file4
      		 Trigger: "AND" of four pmts
      top and bottom bars
      Right-side of bars aligned
      1498
      		 -50
      		 file5
      		 file6
      		 Trigger: "AND" of four pmts
      top and bottom bars
      Left-side of bars aligned

    3. Then the time from the TDC was plotted versus the pedestal subtracted QDC response (1/sqrt(QDC - pedestal)): runs 1496 and 1498. The top right PMT is self-timed, and hence there is no time dependence wrt the QDC signal.
    4. A cut was also placed on the pedestal subtracted QDC distributions to remove the low energy events.
      For run 1496, the following QDC cuts were chosen:
      - Top right cut = 330 channels
      - Top left cut = 350 channels
      - Middle right cut = 320 channels
      - Middle left cut = 240 channels
      - Bottom right cut = 370 channels
      - Bottom left cut = 340 channels

      For run 1498, the MIP amplitudes were higher for three of the PMTS, and the following QDC cuts were chosen:
      - Top right cut = 320 channels
      - Top left cut = 430 channels
      - Middle right cut = 450 channels
      - Middle left cut = 400 channels
      - Bottom right cut = 350 channels
      - Bottom left cut = 350 channels

    5. A profile histogram was created so that the distribution could be fit: runs 1496 and 1498. The slope from the fit was used to correct for the time-walk effect. For run 1496 and the middle left PMT, a third order polynomial was used instead, since there was a transition around 0.048. Using this correction, vastly improved the timing resolution.
    6. Using the expression T = 0.5*(tt + tb) - tm, where tt, tm, and tb represent the time the particle passes through the top, middle, and bottom bars, respectively, the time resolution of the bars was determined.
    7. The time can be expressed in terms of the times of PMT signals as T = 0.25*(ttr + ttl + tbr + tbl) - 0.5*(tmr + tml), where the subscript 'r' and 'l' denote the right-side and left-side pmts, respectively, for that particular bar.
    8. Added on January 27: The initial time-walk slopes were then optimized by varying one slope and holding the other 5 fixed until the timing resolution was minimized. This procedure was repeated for each PMT.
    9. Updated on January 27: The timing resolution for the two runs (1496 and 1498) are quite different, and this result is reflected in the significant increase of the MIP values for top-left, middle-left and middle-right PMTs. A significant increase in light collection efficiency is directly related to improvement in the timing resolution.
    10. The timing distributions for runs 1496 and 1498 are shown in four panels in Table 4:
      - the top-left panel is without any corrections
      - the top-right panel includes the global timewalk correction for each PMT applied
      - the bottom-left panel include the global timewalk corrections with low-amplitude QDC cuts.
      - the bottom-right panel include the global timewalk corrections with low- and high-amplitude QDC cuts.
    11. The timewalk corrections improve the resolution by a factor of 2-3, and an additional 15-20% is achieved with the QDC cuts. At high QDC amplitudes the statistics are poor, and the timewalk corrections extrapolate into this region, which can cause a tail in the timing distribution. Before optimizing the time-walk coefficients, placing high-amplitude QDC cuts could improve the timing resolution by 30-40 ps. After optimizing the coefficients, the high-amplitude QDC cuts provide only a slight improvement of 2-5 ps.
    12. Table 4 contains a summary of the time resolution results for each run. The σ from the fit represents the convolution of all six PMTs in the system. The reference counter (scintillator bar) resolution is obtained by σm = sqrt(σT2 - 0.5*σr2), which is the value shown in
      Table 4 with σr = 58 ps.

    Table 4: LASPD bar timing resolution results using three bar method with Hamamatsu R9779 PMTs
    run#
    		 Raw Resolution [ps] Time Walk Corrected
    Resolution [ps]     		Time Walk Corrected with QDC cuts
    Resolution [ps]     		Time Walk Corrected with High-amp QDC cuts
    Resolution [ps]
    1496
    		 T Spectrum: 314
    		 149
    		130
    		 128
    1498
    		 T Spectrum: 284
    		 110
    		 88
    		 83

January 6th - 9th, 2015

  1. Time resolution measurements using 3-bar tests with R9779 PMTs Part 2:
    Using the EJ-200 plastic scintillators (5 cm x 5 cm x 30 cm) with each end coupled with optical grease to a R9779 PMT for readout. The HVs for the six PMTs were approximately gain matched.
  2. For the set of measurements, the bars were stacked vertically with respect to each other. The three trigger bars were oriented parallel to each other. The vertical spacing is about 24.2 cm between the top of the top bar and the top of the bottom bar. The trigger was generated by forming an "AND" of the left and right PMT signals of the top and bottom scintillator bars. The top right bar's signal was delayed by 16 ns with respect to the other signals so that this PMT would always carry the trigger time. The middle bar was not part of the trigger.
  3. For all measurements, the PMT signals were split from the dark box with one signal passed to a discriminator channel set at -50 mV. The other signal was sent through 80 feet (123 ns) of delay cable before being sent to the QDC module.
  4. For run 1480, the PMT HVs were set modestly low. For run 1484 the HVs were increased by about 100 volts, and for run 1487, the HVs were increased by approximately another 100 volts. However, when the HVs were adjusted to better match the the MIP response, the HV channels were incorrectly adjusted, which caused large differences in most of the MIP values (from 525 up to 900 QDC channels) for all six PMTs.
  5. Conclusions and Observations:
    1. The timing resolution of the three bar measurments was determined using the procedure discussed in the CLAS12 FTOF report.
    2. For each run, the QDC pedestals were deterimined in order to subtract the offset from the QDC response. The pedestal was isolated by placing an anti-cut on the TDC for the corresponding QDC channel. Table 1 summarizes the pedestal and MIP response for each PMT.
      3. Table 1: Timing resolution data using three bar method with Hamamatsu R9779 PMTs
      run#
      		 Discriminator
      threshold [mV]       		Pedestals
      		 measured MIP positions
      		 comments
      1480
      		 -50
      		 file1
      		 file2
      		 Trigger: "AND" of four pmts
      top and bottom bars
      1484
      		 -50
      		 file3
      		 file4
      		 Trigger: "AND" of four pmts
      top and bottom bars
      HV increased to move MIP to ~ 450 channels
      1487
      		 -50
      		 file5
      		 file6
      		 Trigger: "AND" of four pmts
      top and bottom bars
      HV increased to move MIP to ~ 600-700 channels

    3. Then the time from the TDC was plotted versus the pedestal subtracted QDC response (1/sqrt(QDC - pedestal)): runs 1480, 1484 and 1487. The top right PMT is self-timed, and hence there is no time dependence wrt the QDC signal.
    4. A profile histogram was created so that the distribution could be fit: runs 1480, 1484 and 1487. The slope from the fit was used to correct for the time-walk effect.
    5. Interestingly, the data from run 1487, do not show any significant timewalk effect. Updated on January 13: It was found that the TDC and QDC information from the two modules were not correlated with each other. This can be easily seen by cutting on the TDC. The pedestal should be completely removed, but it is not for at least runs 1487 and 1494. More than likely this is a DAQ problem. One or both modules during the run have extra words for one or more events. Since this data is not cleared before the next event, the TDC and QDC information fall out of synchronization with each other. The CODA crl file was updated to clear any extra words that are found before moving on to process the next event.
    6. A cut was also placed on the pedestal subtracted QDC distributions to remove the low energy events. For run 1480, the cut was placed at channel 153 for each PMT signal.
      For run 1484, the following QDC cuts were chosen:
      - Top right cut = 300 channels
      - Top left cut = 350 channels
      - Middle right cut = 350 channels
      - Middle left cut = 350 channels
      - Bottom right cut = 360 channels
      - Bottom left cut = 360 channels

      For run 1487, the following QDC cuts were chosen:
      - Top right cut = 430 channels
      - Top left cut = 750 channels
      - Middle right cut = 500 channels
      - Middle left cut = 440 channels
      - Bottom right cut = 670 channels
      - Bottom left cut = 560 channels

    7. Using the expression T = 0.5*(tt + tb) - tm, where tt, tm, and tb represent the time the particle passes through the top, middle, and bottom bars, respectively, the time resolution of the bars was determined.
    8. The time can be expressed in terms of the times of PMT signals as T = 0.25*(ttr + ttl + tbr + tbl) - 0.5*(tmr + tml), where the subscript 'r' and 'l' denote the right-side and left-side pmts, respectively, for that particular bar.
    9. The time resolution is then expressed as σref = sqrt(2/3)* σT, where σT is determined from a fit to the timing distribution.
    10. Added on January 27: The initial time-walk slopes were then optimized by varying one slope and holding the other 5 fixed until the timing resolution was minimized. This procedure was repeated for each PMT.
    11. The timing resolution for the two runs (1480 and 1484) are quite different from each other, which indicates there is a dependence on PMT gain. Without applying QDC cuts to remove the low energy plateau, a tail is still seen even after the time-walk corrections are applied: 1480 and 1484, though the timing resolution does improve by about 50% and 100%, respectively, with the time-walk correction applied.
    12. By applying QDC cuts, the long tail is greatly reduced in both the distribution after time-wallk corrections: 1480 and 1484. The QDC cut was not applied to the left-hand side plots, in which the time walk correction is also not applied.
    13. For run 1487, there is no difference when the timewalk correction is applied. Finally there is only a slight improvement when the QDC cuts are applied. Clearly the timing distributions are not gaussian, and the fit quality is poor.
    14. Table 2 contains a summary of the time resolution results for each run. The σ from the fit represents the convolution of all six PMTs in the system. The reference counter (scintillator bar) resolution is obtained by multiplying σ by sqrt(2/3), which is the value shown in Table 2.


Table 2: Timing resolution results using three bar method with Hamamatsu R9779 PMTs
run#
 Raw Resolution [ps] Time Walk Corrected
Resolution [ps] 		Time Walk Corrected with QDC cuts
Resolution [ps]
1480
 287
 T Spectrum: 184
 T Spectrum: 169
1484
 T Spectrum: 202
 66
 54-58
1487
 T Spectrum: 178
 N/A
 N/A