The ADC full-scale is 8192 counts above pedestal. The cosmic
signal was NOT put off-scale, so the slowest neutrons we want to detect are
certainly less than 8192/40=200 counts above the pedestal right now. When
the bars were gain-matched in Dec '02, a note in the CC/NC logbook says the
cosmics were put at 2000 counts above pedestal. Therefore the slowest
neutrons are 50 counts above pedestal.
The neutron CFDs are set to 20.9mV threshold, which I believe is
as low as they go. For a 20mV square signal that's 30ns wide (a guess of
what a real signal might be equivalent to), the integrated charge is
12pC. The ADC is 50fC/count, so this slowest neutron would be 240 counts
above the ADC pedestal. That means we need to put the cosmic peak around
9600 counts above the pedestal, or off-scale on the ADC.
So John is right, we're not seeing the slow neutrons now, even
though the thresholds are as low as they can go, because the tube voltages
are set too low, to put cosmics at 25% full scale. Of the 32 tubes, 2 are
over 2900V already, 2 are over 2700V, 6 are over 2500V. The HV power
supply won't go over 3kV, so we'll have to replace a few tubes if we want
to increase the gain by a factor of 4. Or change the split fraction so
that more signal goes to the discriminator than the ADC -- the cosmic peak
would then be on-scale in the ADC, you just wouldn't have an ADC
measurement for the slow neutrons. That's ok, we get the energy/momentum
from the timing anyhow. We could also try attenuating the tube signal by a
factor of 2 to gain-match with cosmics at 4800 counts above the pedestal,
then removing the attenuator? Or gain match with a source?
The phototube signals may be larger in the D Tunnel now than last
December when the gain-matching was done. At some point there were 50ohm
feedthrough terminators on the tube bases, and those are gone now. Worth
running cosmics to check the gain match voltages and also where the cosmic
peak appears.
Karen
At 04:59 PM 9/24/2003 -0400, John Calarco wrote:
>Here is a short note on setting thresholds in the neutron bars. Similar
>considerations will hold for the LADS bars, but the numbers need to be
>adjusted for the thicknesses.
>
>For the Ohio n-bars, they are 10 cm thick to neutrons and 22.5 cm thick
>to cosmics. It has always been planned to use cosmics as a calibration
>standard, including using a "cosmics" trigger between the top and
>bottom bars. Note that this trigger still allows for cosmics to pass
>through a much larger thickness than the 22.5 cm vertical dimension
>as long as they are still in the plane of the n-wall. A cut on the
>time difference between the n-bar ends in the top and bottom bars will
>pick out (almost) vertical cosmics.
>
>Assuming we pick out almost vertical cosmics, they lose 45 MeV in each
>n-bar. So the question to ask is: what is the useful range of neutron
>energy deposition in the n-bar?
>
>Assuming that the most interesting region for BLAST to contribute to
>a measurement of Gen is in the low Q^2 region, like 0.1 to 0.3 GeV^2,
>we need to optimize for that region. Let's assume we want to go as low
>as Q^2 = 0.05 GeV^2. Neglecting the Fermi momentum in the deuteron for
>the moment, the corresponding neutron energy is 25 MeV. The corresponding
>momentum is a hair less than 220 MeV/c. Taking into account the Fermi
>momentum spreads that momentum by +/- 50 MeV/c. So we can have neutrons
>with momentum as low as 170 MeV/c, corresponding to neutron kinetic
>energies of 15 MeV. The energy loss in the n-bar comes from mainly
>s-wave scattering off protons which produces a proton spectrum that is
>theoretically uniform from 0 to 15 MeV. However, whereas the conversion
>from energy deposition to light is essentially 100% efficient for
>relativistic particles (electrons, cosmics), it is much less efficient
>for lumbering behemoths such as slow protons. There exists an old rule
>of thumb that gives the light yield in this energy range as 0.6*E - 1.3
>meaning a 15 MeV proton produces light equivalent to a 7.7 MeV electron
>energy loss. Given that the light spectrum from 15 MeV neutrons will be
>a roughly uniform distribution, therefore, from 0 to 7.7 MeV electron
>equivalent, if we want 6/7 of the theoretical efficiency offered by the
>10 cm thickness of the n-bar, we need to set the threshold of the
>discriminator to the equivalent of a 1.1 MeV electron energy loss, or
>about 1/40 of the cosmic signal.
>
>This is relaxed by about a factor of 2 if we only care about Q^2 of
>0.1 or higher. However it is my suspicion that, at present, the thresholds
>on the n-bar discriminators are MUCH higher than this and that this is
>probably the explanation of Vitaliy's weak neutron yield in comparison
>to Igor P.'s thesis from NIKHEF.
>
>
>--
>John R. Calarco
>Dept. of Physics
>Univ. of New Hampshire
>Durham, NH 03824
>phone: (603)862-2088
>FAX: (603)862-2998
>email: calarco@unh.edu
This archive was generated by hypermail 2.1.2 : Mon Feb 24 2014 - 14:07:30 EST