Tests of a
A Time-of-Flight System for
Trigger Considerations
Tests of a
A Time-of-Flight System for
Trigger Considerations
To optimize the performance of TOF counters of the geometry we plan to use in
the CDT location we have begun testing prototype counters at Fermilab. The
first three prototype counters are 40 
40 mm in cross section and
approximately 300 cm long. The scintillator for two counters was Bicron BC408,
for the other Bicron BC404, all were cut and polished at the factory. The
counters were wrapped in foil and black paper. We first tested both the BC404
and the BC408 with R2490-05 Hamamatsu PMT glued on each end. This tube is a 2"
diameter version of a fine mesh dynode tube similar to that which we plan to use
in the final system. Its performance both in and out of a magnetic field has
been well studied and documented 9. The photocathode diameter of the
R2490 is 38 mm which means that it accepted 70% of the light produced in the
bar. We also tested the BC408 bar with two Hamamatsu R5946 PMT's. This tube is
1.5 inches in diameter with a 27 mm diameter photocathode and is small enough to
be used in the CDT location. Finally, we tested one BC408 bar with CPC light
concentrators (more on this later) and two Hamamatsu R5946 PMT's. The test setup
is shown schematically in Figure 19. Cosmic ray trigger
counters above and below the TOF counter were mounted on a movable track so
they could be positioned along the length of the TOF counter. For most of the
results discussed here the top trigger counter used a 40 
20 
50 mm piece of BC408 scintillator and was equipped with an RCA8575. The
scintillator was oriented with the 20 mm dimension along the length of the 3 m
counter and the PMT was horizontal. The lower trigger counter used a 40 
40 
50 mm piece of BC408 scintillator read out with a Hamamatsu R2490-05
PMT. The bottom counter was oriented such that the 50 mm dimension and the PMT
were vertical. The anode signals from all four PMTs were split with a resistive
divider. One split signal was connected to an Ortec model 934 constant
fraction discriminator, the other to a LeCroy 2248 CAMAC ADC. The signals from
the constant fraction discriminator were passed through 64 ns of delay and then
were connected to the stop inputs of a LeCroy 2228A CAMAC TDC (50 ps least
count). The coincidence of the trigger counters provided the common start to
the 2228A. The relative timing of the trigger counters was adjusted so that the
counter with the R2490-05 always determined the timing of the coincidence.
Figure 20 shows a typical pulse height distribution from one of
the PMT's from a cosmic ray run with 4000 triggers. A run with this many
triggers takes about 20 hours. The top two plots corresponds to the two trigger
counters, ADC C2 the bottom and ADC 1 the top. Clear minimum ionizing peaks are
observable. The bottom two plots show pulse height distributions for PMT's on
the 3 m counter, the left tube (ADCL) and the right tube (TDCR). Again, minimum
ionizing peaks are apparent. To measure the resolution we use only those events
in which the ADC values are within the minimum ionizing peaks. For this run, we
required ``MIP'' cuts of 400 < ADC C1 < 800, 400 < ADC C2 < 800, 100
< ADCR < 400 and 90< ADCL < 390 counts.
To determine the best scheme for measuring the arrival time of particles at the TOF array we have tried several methods of combining the measured times from the PMT's on each end. A typical set of time distributions for the left and right PMTs of the 3 m counter are shown in Figure 21 after ``MIP'' cuts. For this run the top trigger counter was placed at the middle of the 3 m BC404/R2490 counter. We observe similar time values from the two PMT's as expected. The RMS widths of the distributions are 3.2 and 2.8 counts, respectively. A TDC count corresponds to 50 ps so the time resolution of an individual PMT is about 150 ps. The bottom left plot of Figure 21 shows the distribution if we use the average times of the left and right PMT's. The RMS width is now 2.3 counts or 115 ps. The improvement over single tube measurements is presumably due to both the improved photoelectron statistics from both tubes and because averaging the left and right times reduces time smearing due to variations in the position of the cosmic ray as it passed through the the trigger counters.
Figure 19: Prototype TOF counter and cosmic ray test setup
Figure 20: Typical PMT pulse height distribution from
a cosmic ray run with 4000 triggers. The top two plots corresponds
to the trigger counters, ADC C2 the bottom one and ADC C1 the top one.
The bottom two plots show pulse
height distributions for PMT's on the 3 m counter, the left tube (ADCL) and the
right tube (ADCR). Minimum ionizing peaks are apparent.
Figure 21: Time distributions for the a) Left PMT b) Right PMT
c) sum of the times from both PMTs and d) Weighted average time using
both PMTs. Data is from the 3 m long BC404 prototype counter. Tubes are
R2490-05, readout via
a constant fraction discriminator, and LRS 2228A 50 ps least count TDC
When the time measurement is limited by the photoelectron statistics, we expect
that the time resolution (
) is related to the pulse height (ADC) in
the PMT by 
. By weighting each time measurement
by the pulse height in that PMT we should be able to minimize the width of the
resulting combined time distribution. The bottom right plot of
Figure 21 shows the distribution of the average of the times
measured by the PMT's on each end of the 3 m counter weighted by their pulse
heights. The RMS width is 2.4 counts, which is nearly identical to that of the
simple average. This is not unexpected for a run with the trigger counters at
the middle of the 3 m counter because two tubes have almost the same pulse
heights at this location. We note however that this technique does improve the
resolution over the simple average when the cosmics pass through locations away
from the middle of the counter. Superimposed in the plot is the result of a
Gaussian fit, which gives an RMS width of 2.4 counts (120 ps).
Figure 22, and Figure 23 show these time
resolutions measured at several locations along 3 m counters fabricated from
BC408 and BC404 respectively and instrumented with R2490 PMT's.
Figure 24 is a similar plot for the same BC408 bar
but instrumented with the smaller R5946 tube. For these figures we used data
taken with a small top trigger counter 20 
40 
50 mm to minimize
trigger counter width effects. We observe that the difference between the
trigger counter time and the pulse height weighted average time of the two PMTs
exhibits a time resolution of less than 110 ps nearly independent of the
position of the trigger counters for both the BC408 and BC404 bars readout with
R2490 PMTS. We find no significant difference in performance between BC404 and
BC408. The resolution of the BC408 bar instrumented with the R5946 is about 135
ps. In this case we expect the R5946 results to be worse than the results with
the R2490 since this tube has a smaller photocathode and collects only about
half as much light as the R2490. The fact that the resolution with the R5946 is
not 
worse than the R2490 indicates that other significant factors in
our test setup other than photostatistics contribute to measured resolution.
Figure 22: Time resolutions measured at several locations along the 3 m counter.
Scintillator is BC408. Phototubes are Hamamatsu R2490.
Figure 23: Time resolutions measured at several locations along the 3 m counter.
Scintillator is BC404. Phototubes are Hamamatsu R2490.
Figure 24: Time resolutions measured at several locations along the 3 m counter.
Scintillator is BC408. Phototubes are Hamamatsu R5946.
Figure 25 is a plot of the difference in arrival time (t
) between PMT signals measured on each end of the
BC408/R2490 counter versus the location of the trigger counter. From the slope
of this distribution we can measure the effective propagation velocity of
scintillation light from cosmic rays as it propagates down the length of the
counter. The effective velocity is very close to 15 cm/ns.
Figure 26 plots the pulse height (minimum ionizing peak position)
of each PMT of the 3 m counter as a function of the location of the trigger
counter along the bar. The measured points are nearly straight on the
semi-logarithmic scale indicating an exponential decrease in pulse height with
increasing distance from the PMT. Using three points in the middle, the
effective attenuation length of light in the 3 m bar is about 225 cm.
Figure 25: Time difference between PMT's measured at several locations
with cosmic rays along the 3 m counter. The slope represents the effective
propagation velocity. We find 
= 15 cm/ns.
Figure 26: Pulse height (minimum ionizing peak position) of each PMT of the 3 m
counter at several locations along the 3 m counter.
From this we determine the effective attenuation length of the counter to
be 225 cm.
Il'ja Mikhailovich Frank was born in Leningrad on October 23, 1908, the younger son of Mikhail Lyudvigovic Frank, a Professor of Mathematics, and his wife, Dr. Yelizaveta Mikhailovna Gratsianova He attended the Moscow State University as a pupil of Vavilov, and graduated in 1930. In 1931 he became a senior scientific officer in Professor A.N. Terenin's laboratory in the State Optical Institute in Leningrad, and in 1934 he joined the P.N. Lebedev Institute of Physics of the U.S.S.R. Academy of Sciences as a scientific officer. He was promoted firstly to senior scientific officer and, in 1941, to his present position as officer in charge of the Atomic Nucleus Laboratory. Since 1957 he has simultaneously occupied the post of Director of the Neutron Laboratory of the Joint Institute of Nuclear Investigations.
|
|
| free web hits counter |
This is my BrainyGoose:
United States, IL, Chicago, English, Italian, Genry, Male, 21-25, bodybulding, swiming.