[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: single end hits




Hi Blake,

I imagine the reason it was not done before is because it is not  
possible to calculate z and t.  This causes problems also since (I  
believe) the reconstruction corrects for attenuation on the cell level  
not on the full cluster level.  Furthermore, I suspect with the KLOE  
geometry there was no strong motivation to try to improve it.

You may be able to do this leaving the main hunk of the KLOE  
reconstruction code (which has some confusing data structures) fairly  
unchanged.  There is an initial loop in the beginning the prunes out  
all the single ended hits.  You just need to stuff them somewhere and  
then make a loop at the end that tries to put them back onto the lists  
of clusters.  You can use the cluster z position to correct the energy  
of the single ended hit.

The first step would be just to try to enhance the performance by  
adding single hits to existing showers.  In this case  
DBCALShower_factory produces the same shower list effectively however  
all the energy values a tweaked a little for the single ended hits.   
(You could in principle also adjust the positions, but we know at  
leading order energy is the most important, and the corrections from  
single ended hits will be small.)

Of course, with any change to the algorithm you'll need to adjust your  
calibration.  (You will be explicitly adding back reconstructed energy  
instead of estimating the energy lost using some calibration constant.)

I expect incorporation of single ended hits to only make a difference  
at low energy near the end of the BCAL.  It will help to rescue the  
resolution for those 200 MeV photons in the end of the BCAL.  We know  
this is the "magic region," but you might want to evaluate maximum  
potential gain before embarking on this adventure.  Come up with some  
figure of merit and run some MC with the threshold in the BCAL set  
very low and the fiber attenuation length set high.  This might give  
you a minimum number of single ended hits and a "best case scenario."   
If the gains are tiny, you might rather spend your time working on  
some toy amplitude analysis studies!

Cheers,

-Matt


On Dec 30, 2008, at 5:06 PM, Blake Leverington wrote:

> Hey all,
>
> So a question before I start trying to implement single ended hits  
> into the bcal code, is there any reason that there has to be 2-ended  
> hits such as some catastrophic error or is it just that a few things  
> aren't able to be calculated for those hits, such as z and t or just  
> need to be calculated differently, like e? I'm referring to the  
> variables in DBCALShower_factory. I have a general idea how to go  
> about this, and it seems fairly simple, but I was just checking to  
> make sure there isn't anything down the line that will be affected  
> that people are aware of.
>
> -Blake