Email from Jim:

From birchall@jlab.org Thu Aug 26 07:31:36 2004
Date: Mon, 16 Aug 2004 21:20:22 -0500
From: Jim Birchall 
To: Shelley Page , Roger Carlini ,
     Wim van Oers , Mark Pitt 
Subject: Times

Roger,

There are three factors affecting the running times compared with the 
TDR, all three increasing running time:

1) Event rate from 760 to 620 MHz, time going as 1/rate

2) Q^2 from 0.03 to 0.028, time going as 1/(Q^2)**2

3) The value of Qw used in the TDR was 0.0756, but the value on the 
recent "pull plot" I have is 0.0716. I used 0.0716 whereas earlier I'd 
used 0.0756. I guess this is a disadvantage of not quoting absolute 
values of errors. We should use a consistent value for Qw. Time goes as 
1/Qw**2.

Reverting to 0.0756, my running times should be multiplied by 
(0.0716/0.0756)**2. The counting statistics on Qw for 3300 hours at 620 
MHz and Q^2 = 0.028 becomes 2.7% of 0.0756, compared with the TDR 
target value of 2.8%. We'd get 2.8% of 0.0756 in 3070 hours. We get the 
rest of the way down to 2200 hours by increasing the rate back to 760 
MHz and Q^2 to 0.03.

Jim