Log entry time 22:58:21 on November 16, 2008

Entry number 249280

This entry is a followup to: 249161

Following last halog, there is a better hypothesis to explain the 
vibration of count difference between two gated scalars. And it's 
confirmed that the accumulating count difference on scalar set (spin=-
1;Hel=-1) is likely to apply to all its high rate channel (fast clock, 
d10, u10). For low rate counting, it's OK (ex. L1 acceptance rate)

----------------------------------------------
   hypothesis explaining the vibration
----------------------------------------------
If bigbite scalar is read 200us later than the one on LHRS, then it 
will cause the 20 count vibration discussed in Halog # 249161. 

Taking scalar set (Spin=+1, Hel =+1) as example: 
If the scalar read is during Helicity=1, then both scalar on bigbite 
and LHRS is counting. And if LHRS is read out 200us ahead, then bigbite 
scalar will read 20 more count on 103kHz clock. However, when read out 
during Helicity= -1 state, both scalar is not counting. So it dropped 
back to the previous base line.

Besides, the slow random walking of scalar count diff (on order of 
5count/min) could due to accidentally missing or gaining a clock count 
during one Helicity window. It should be normal. And the ungated scalar 
is free of this effect.

--------------------------------------------------------------
Counting problem on (spin=-1;Hel=-1) is global to this scalar
--------------------------------------------------------------
Last halog discussed that fast clock counting difference is climbing @ 
2count/scalar read on scalar set of (spin=-1;Hel=-1). This climbing is 
also seen at d10 and u10. But for slower frequency count, like the L1 
Acceptance count, is free of this problem. Please see plots below.

d10 counting difference is also climbing for (spin=-1;Hel=-1)

L1 Acceptance count is consistant on same sclar set