From blecherm@alphamb2.phys.vt.edu
Tue Sep 10 10:31:14 2002
Date: Tue, 10 Sep 2002 10:22:23 -0400 (EDT)
From: Marvin Blecher
To: laurie littenberg
Subject: simulation for meeting

Some time ago, Mike Sivertz asked if I would examine the possibility of vetoing triggers from decays in the downstream (DS) highvacuum pipe (HVP). Specifically, he suggested that I calculate,

<X> = Sum_i(X_i*E_i)/Sum_i(E_i)
<X**2>= Sum_i(X_i**2*E_i)/Sum_i(E_i)
Sig_X = sqrt(<X**2> - <X>**2)

and similar expressions for <Y>, <Y**2>, and Sig_Y.Instead of X_i or Y_i, I used nX_i and nY_i, where these are the slat numbersof the scintillators that make up a scintillator layer. There are 20 suchslats in a layer. Thus X,Y are proportional to nX,nY. I don't use the Uslats in this analysis.

Mike supposed that good events entering the PR at theupstream face (z=1384.06) might look different from decay events that enterthe PR at z>1384.06. Note the decay vacuum center is at z=1150. In my analysis I started 1000 gammas from (0,0,Z). The values of Z were 1150,1200,1250,1300, 1350, and 1400. Each gamma had an energy of 200 MeV and was going in thedirection given by theta = 28.54 deg and phi = 45 deg. The table belowshows where the gamma rays enter the PR for a given Z:
 

In the attachedfiles I show histograms of Sig_nX for all layers of the PR. The scatterplot of Sig_nX vrs layer number for each layer involved in the shower isless informative. Those for sig_nY are similar. The Z values on the figure are too low by 50 in each case (oops!) These plots show a difference between good and background gammas, but unless we give up on good events that enter close to the DS HVP, I don't think an on-line veto of decay events canbe worked out from these variables.

Hi, Here are the rest of the angles (theta = 45, 60, 75deg) that I did for the Mike Sivertz effect. It is true that sig_x (seelast week's contribution) is greater for signal events at the same angle, but there is no way to safely make a veto from sig_x that won't kill theacceptance. Vetoing K_L-> 2 or 3 pi0 in the DS HVP at the trigger levellooks like a really difficult problem. I'll look at what fraction are vetoedby the outside veto around the PR that I presently have in the geometry.

Date: Thu, 12 Sep 2002 10:35:21 -0400 (EDT)
From: Marvin Blecher
To: douglas bryman, michael zeller, mike marx, laurielittenberg
Subject: simulations

Hi, After seeing that Mike Sivertz's idea wouldn't work at the trigger level, I ran 10K K_L->pi0 pi0 events. The decays were required to take place in the DS high vacuum pipe HVP. All gammas were loaded onthe stack. I wanted to see what fraction were not vetoed and where thevetoing was coming from. The results are 8% are not vetoed. The main vetosare the DS CPV after the calorimeter 57% and the CPV lining the DS HVP20.5%. This stresses the importance of the CPV after the CAL and that itshould be sensitive to gammas. Also if a crystal liner is put between thebeam pipe and PR, it will be very busy. Trying to add its energy back intoshowers, that is by over-riding veto info, is dicey and wouldn't allowuse of this liner at the trigger level for veto.

Tonight I'll run K_L -> 3pi0.

 Date: Fri, 13 Sep 2002 11:03:37 -0400 (EDT)
From: Marvin Blecher
To: laurie littenberg
Subject: your requested figure

Hi, Here is the Mike Sivertz testcase that you and David requested. I started 1000 events at X=50, Y=0,different Z in the DS HVP except for Z = 1350 (end of decay region). Theyall had theta = 75 deg and phi = 90 deg. I plotted sigma_y.

Date: Thu, 19 Sep 2002 13:33:20 -0400 (EDT)
From: Marvin Blecher
To: dave jaffe
Subject: corrected figures

Hi,

I corrected the error that David found in my subroutine that calculated second moments. The new results on sigma_y for  theta= 30^o, 45^o, 60^o,and 75^o(all with phi = 90^o) are attached. They look different fromthose with the mistake, however, the result remains the same. They arenot useful to distinguish decays in the DS HVP from possible signal gammasat the trigger level. I'm going to dig into this problem as my next project.

hi Marvin,

I assume that X=50, Y=0 is true for all figures despite the title on the theta=75^o. I disagree with your conclusion.A cut at sigma_y = ~1 gets rid of about half the Z=1400cm or 1425cm eventsfor theta=75^o while retaining a large fraction of the Z<1350cmevents regardless of the angle. It would now be interesting to look atsuch distributions for KL decays in the DS HVP that shoot two or more photonsinto the PR. By the way, Ermanno sent a copy of his trigger simulationcode. It's available at http://www.phy.bnl.gov/~djaffe/KOPIO/Trigger/ He does everything with ntuples. I don't know how that jibes with youranalysis.

Best regards,
David

Date: Mon, 23 Sep 2002 10:23:38 -0400 (EDT)
From: Marvin Blecher
To: michael zeller, laurie littenberg
Subject: Kpi2 non vetos

Hi, I cannot send you a file with 10 events that don't veto and with every step printed out. My mail service doesn't have enough memory. However, I can summarize for the first 5 events where the firstcharged particle was generated. This is for K->pi0 pi0 in the DS HVP. Qstands for the quad.

event 1: PR Q2, PR Q1, PR Q2, Cal Q3
         2: PR Q2, PR Q4, PR Q2, PR Q2
         3: PR Q3, Barrel only 17 MeV, Cal Q3, PR Q1
         4: PR Q2, Cal Q4, PR Q2, PR Q4
         5: PR Q4, PR Q2, Cal Q3, PR Q4

Hi, I looked at 5 events of K->pi0 pi0 pi0 in the DS HVPthat did not veto. The first charged particles from each gamma were as follows: (note DSPI) is the DS low vacuum pipe I put in D4. It has 1" thick walls of Al.

event 1: PR Q4, PR Q1, PR Q3, DSPI, PR Q2, PR Q1
         2: PR Q3, cal Q1, PR Q2, cal Q1, DSPI, PR Q2
         3: PR Q2, cal Q2, Cal Q4, PR Q3, PR Q3, PR Q4
         4: PR Q4, cal Q2, PR Q3, PR Q2, Cal Q3, cal Q1
         5: Cal Q2, Cal Q3, Cal Q4, PR Q1, Cal Q4, Q2

It looks like we could veto these events if > 2 quads had energy. Working out some sort of fast veto will be the next project.

Mike - Is is possible to send me a sketch of your proposed DS gamma veto?

Next installment