[g13] linear pol

[Ken Livingston]

Hi Nickolay et. al.,
Some responses below.
 
 
nivanov at jlab.org wrote:
> Hi Ken and other g13 Users,
>
> I have 2 questions about the available calculations of the photon linear
> polarization for g13.
>
> 1. Presently there are 2 sets of quick and rough estimates of the mean 
> polarization. The first set is presented in the Daria's message by 15 July
> 2009. The second one is given in Neil's polarization tables at
> http://nuclear.gla.ac.uk/~kl/g13/poltables.tgz
> (see the file meanpol.txt in cbremROOT directory).
> Unfortunately, there is a difference (up to 15%) in the results presented
> in these sources. So, my first question is: What predictions for the
> polarization are more reliable? What numbers we should use in our
> analyses?
>   
For the moment you can take your pick. Each is as reliable as the other, 
and neither is good enough to be the basis of a CLAS analysis or paper. 
I'm working on an improved method of analysing the coherent brem data 
and producing more reliable tables. I'm also writing a CLAS note on that.

We will also need to look at Darias results for systematics and 
comparison. In the next pass of cooking of g8b some code will be run to 
make this easier.

I think there's no point in worrying too much about the absolute value 
of the polarization untile that is done.

First, the current coherent edge position is in the EPIC bank which is 
read out every 2s. This bank is in the cooked BOS files and the root 
DSTs, but probably not in the hbook ntuples.
 
> 2. In Neil's polarization tables (pol_tables_*.dat), the ANB predictions
> for the polarization for different coherent edges are presented. However,
> to have reliable predictions for a real experiment we need also the
> coherent edges distributions for this experiment. The second question is:
> How and where we can get the coherent edges distributions for g13b?
>   
The current coherent edge position is in the EPIC bank which is read out 
every 2s. The variable is called "coh_edge". The relevant variables in 
that bank are (pasted from bosdump):
...
value: 1692.354 char1: coh_edge                      
value:    1.000 char1: coh_plane     //need to convert to int 
                   
value:    2.000 char1: coh_radiator          //need to convert to int  
...

And here are meanings of the integer values for coh_plane and coh_edge

#define PARA 0                  //defs for polarisation plane
#define PERP 1
#define AMO  2
#define CHANGING 3
#define ALUMINUM 0               //defs for radiators
#define AMORPHOUS 1
#define DIAMOND 2
#define UNKNOWN 3
#define AUTO 4

 This bank is in the cooked BOS files and the root DSTs, but probably 
not in the hbook ntuples.

In the next pass we'll write these out to files for each run, and make 
some edge histories and distributions.
> Unfortunately, Ken, your explanations of this topic given at
> http://www.jlab.org/ccc/mail_archives/EXPERIMENTS/g13/CURRENT/msg01026.html
> are not quite clear for me:
>
>   

>> For the files you analyse for, say 1300MeV, make histograms of the
>> distribution of coherent edge positions (which you get every 2s)
>> for para and perp.
>>     
>
>   
>> Also, make a photon energy distribution of para and perp photon energy
>> for your good tagger hits for the events which make it through your
>> cuts.
>>     
>
> Could you explain in more detail what the g13 users should do to get the
> coherent edges distributions?
>   
You need to make the distribution yourself from in data analysis. For 
every event that makes your cuts fill up a para of perp histogram.
eg. for 1.7GeV coh edge data something like this:

...
...
const Float_t EDGE = 1.7;
polhists polDists[2];    //make array of 2 hists for edge position
                         //name     title      bins min       max    
polDist[PARA] = new hist("paraDist",paraDist", 100, 0.8*EDGE, 1.2*EDGE);
polDist[PERP] = new hist("perpDist",perpDist", 100, 0.8*EDGE, 1.2*EDGE);
...
...
int currentPlane;
int currentRadiator;
Float_t currentEdge;



while(getNextEvent()){

    if(isEPICS()){     //if  EPIC event, read values
       readEpicsInt("coh_plane",&currentPlane);
         readEpicsInt("coh_radiator",&currentRadiator);
         readEpicsFloat("coh_edge",&currentEdge);
       continue;       //because there's nothing other than EPIC in an 
EPIC event
    }
   
    if(myChannelCuts()){                         //if passed all my cuts
       if(currentRadiator!=DIAMOND) continue;    //and radiator == diamond
       polDist[currentPlane]->Fill(currentEdge); //increment  the 
relevant histogram
 
    ...
    ...
    }
}

...
...



I hope to make something more user friendly soon.

Cheers,
Ken

> Thank you very much,
> Nickolay
>
>
>
>
>
>
>   
>> Hi Yordanka,
>> Had a talk about the linear pol with Neil. The current tables are based
>>     
> on a representative sample from each run period, so we don't have a
> distribution for each run. I'll produce one this week, but need to run a
> local farm job to do so.
>   
>> Cheers,
>> Ken
>>
>>
>> --
>> =======================================================
>> Ken Livingston
>>
>> Dept. of Physics & Astronomy,        Tel: +44 141 330 6428
>> University of Glasgow,               Fax: +44 141 330 5889
>> Glasgow G12 8QQ.
>> Scotland. UK.
>> =======================================================
>>
>> _______________________________________________
>> g13 mailing list
>> g13 at jlab.org
>> https://mailman.jlab.org/mailman/listinfo/g13
>>
>>     
>
>
>
>
>
>
>
>
>
>
>
>   


-- 
=======================================================
Ken Livingston          

Dept. of Physics & Astronomy,        Tel: +44 141 330 6428
University of Glasgow,               Fax: +44 141 330 5889
Glasgow G12 8QQ.
Scotland. UK.
=======================================================