Log entry time 07:17:23 on November 16, 2008

Entry number 249161

This entry is a followup to: 249159

Followups:

• 249280 11/16/08 22:58 Jin Huang Update on scalar synchronization check
• 249438 11/17/08 21:45 Jin Huang The Scalar Counting Difference Climbing is due to missing on LHRS scalar

A 103kHz fast clock is sent to both LHRS and bigbtie scalars. The 
difference between clock counts of two scalars are no more than 10^-5 
level. However there is some interesting structure below 10^-5.

The count difference between the two scalars are checked to be 
** count difference  <1ms for a 1M/50min run for scalar sets of 
ungated, (++),(+-) and (-+) .
** For scalar sets of (--), The count difference of bigbite - LHRS is 
climbing @ ~10 count/s = ~2 count / scalar read = ~2 count /100 event. 
This climbing is also correlated to data taking rate. 

Besides, there is a ~200us = ~20 counts vibration in counts diff of 
gated scalars. It's correlated to helicity state. For, ex: for scalars 
(spin=-1,hel=+1), the bigbite scalar will read extra ~20 more count 
when the scalar event is read in helicity=+1 state. It appears as if 
gate for left HRS scalar is applied 200us later than the one on Bigbite 
(although the cable delay is 3 order of magnitude smaller).

See plots below & in Last entry. run # 3308 is taken as an example.

Scalar count difference for (spin=-1, helicity=-1) sets. It's climbing when scalar is actived

20 count vibration when scalar is actived. Same for 4 gated scalar. Disappear on ungated scalar