Re: FCAL test

[Matthew Shepherd <mashephe@indiana.edu>]

On Sep 16, 2007, at 11:57 AM, Elke-Caroline Aschenauer wrote:

> I'm not completely sure I can exactly follow your argument.
> If I have bigger blocks like the ones from Hermes 9*9*45cm^3 I contain
> more of the shower in one block, I need to take this into account  
> in my shower
> algorithm, but that should be possible. Do I miss something.

The problem with the blocks is not that they are bigger, but that  
they are not a precise integral multiple of the current block  
dimension.  By construction this means it will introduce gaps either  
in the boundary of the surrounding blocks or elsewhere in the array.   
By removing blocks from the array one only creates holes that are  
very precise multiples of 4 cm.  What goes in those holes (other  
blocks, supports, etc.) must match the dimensions exactly or else it  
will create gaps in the stacking of the rest of the array.

> Actually Beni, what did you calculate from the radphi data how many  
> days
> does it need till the blocks are completely blind. Of course all of  
> this
> has to be folded with the phase space of photons from pi0 we are
> interested in.

I think this step is key.  I know Beni has done this calculation that  
basically shows days-weeks before damage is noticeable.  I believe on  
the page I sent Richard estimates a block half life near the beam  
hole of about 800 hours.  We need to check all of these.  Maybe one  
way to do this is run side by side geant simulations with GlueX and  
RadPhi geometries.  We could then use these simulations to estimate  
how to scale RadPhi experience to GlueX.

> May be all of this is not a serious problem as the dose
> should drop with 1/r.

I believe the drop is much faster than 1/r.

> But I think we really want to look into more detail
> in this, because it means we have an energy dependent gain, this  
> can in
> principle be calibrated out if you have a gain monitoring system  
> along the
> length of the blocks and with different wave length. But that is a
> complicated one and gain monitoring systems which measure absolute are
> very very complicated and expensive.

We will have gain monitoring along the length of the blocks.   
Absolute transmission is not as important as relative transmission.   
Past experience I believe shows the proposed system does a great job  
at uniformly illuminating the front of the blocks to get good  
relative transmission.  We need to think about the wavelength  
dependence though.

> Matt could we put the whole thematics on one of the next calorimeter
> meetings.

This is a good idea.

Matt