This document discusses the impact of electronics deadtime on physics asymmetries during Gen98.
The electronics deadtime during Gen98 was caused by the high rate of the neutron
bars. Each OR of the output of eight L/R coincidences (Coinc 4516) was sent
to a discriminator (Disc PS706) resulting in ten lines, each counting with typically
400 kHz. The width of the signals was typically 50-60ns (depending on overlap).
This width was required by the logic delay units (LDU) to work properly. The
output of the ten discriminators had a width of 
19 ns and was
sent to a Logic Fan. The output of the Fan with a rate of 4 MHz was sent to
the 8LM where the coincidence with the HMS signal (width 11 ns) was formed.
The discriminators were running in a non-updating mode which induced a certain
probability that a ``real'' bar pulse was ``eliminated'' by a preceding
random pulse and thus did not make the coincidence with the electron. The size
of the effect can be roughly estimated: To first order it occurs whenever a
random pulse appears within 60ns in front of the ``real'' pulse. The probability
that this occurs at 400kHz is simply 60
10-9
4001030.025.
The effect is partly compensated if the ``real'' pulse follows very close
to the preceding random pulse or if on one of the other 9 lines either another
``real'' or another random pulse jumps in to make the coincidence anyway.
In a post experiment setup we simulated the experimental conditions by means
of artificial random signals and measured the magnitude of the effect to be
DTel0.005Bar Rate [MHz], which is in
agreement with the expectations outlined above. An earlier document 1 describes the technique that was used in this test and its results.
As long as the Bar rate does not show a helicity dependence the impact on the physics asymmetry is zero, because both helicity states are affected equally. An asymmetry in the bar rate could be caused by a beam current asymmetry or the intrinsic asymmetry in the production of the bar singles.