Main INDEX
Monthly INDEX
PREV
NEXT
User name R. Michaels
Log entry time 09:45:08 on July25,2001
Entry number 67509
This entry is a followup to: 67492
keyword=re: charge asy algorithm
I agree about the improvements, and they will be ready in a few days, we hope.
The new algorithm, which is in the works (thanks, Antonin Vacheret)
will do the following better things.
1) Runs online (ET system) constantly and produces feedback every
10 minutes. This is the optimum timescale given the noise structure.
2) Each feedback value is based on 2 passes through the data. One pass
to get the rough average and sigma, and the other to cut out the
beam trips by rejecting tails. This is essentially what Seonho suggests.
This could also be done on the existing algorithm, but we're close to
releasing the new one anyway. It looks like the ramp up time is bigger
than what the code assumes.