[Solid_baffle] [Solid] revised draft agenda for SoLID collaboration meeting 5/14-5/15

[Zhiwen Zhao]

hi, Rich

Here I try to recall some of the histories.
The Babar baffle made by Eugene was based on the idea of preserving a single phi slice of acceptance
by following electron bending in field.
It basically worked only because Babar and CLEO coil and iron flux design is similar and we only 
need to
1 more degree in phi and more layers and photon blocker before EC to make it better, so it became
"babarmore1" with photon blocker as used in pCDR.
It has passed FOM test with 11GeV beam, but Paul didn't like its FOM with 6GeV beam.
The current background and trigger rate is based on it.
Overall, I think we can do a better design to tuning to current CLEO field.
But the design approach is easy to follow and proved to work.
So I tried the "5555" baffle, did some FOM test, but not background and trigger test.
I think someone can easily pick up and continue the work.

The "following electron bending in field" approach certainly will produce baffle slit more and more 
curvy
from first to last plane because the field effect.
My guess is this requires more detector coverages. but our current budget for GEM and EC assuming 
cover all areas.
The other approach Seamus use can have different shape, maybe more straight. And it may be better 
for detector
coverage?
but I think no matter what approach, a baffle needs to pass FOM, backgroud and trigger test to be 
considered ok to use.

I think only after a basic design of baffle shape is done, we can improve details on edge angles, 
materials etc.

About baffle parametrization, we certainly need to improved the current design using small blocks.
but the guideline can't be just reducing parameters.
I think we should really consider how it will be made or machined.
It may be not too early to discuss this with engineer how we can do direct transfer from design to 
production
to ensure the precision of the final product.

Other news about magnet, it will need to be changed!
hallA engineering gives us feedback that they need more iron for supporting
and SIDIS setup certainly now is too tight to fit everything in.
And we haven't cut holes for any cable out yet.
I will try to start the discussion during the meeting

I am CCing this to solid_baffle at jlab.org
Currently Rakitha, Zhihong, Zhiwen, Seamus, Rich and Paul, Lorenzo are on the list
We should use it for real discussion to keep everyone informed.

Thanks

Zhiwen

On 5/12/2015 12:04 PM, Richard S. Holmes wrote:
>
> On Mon, May 11, 2015 at 7:05 PM, Seamus Riordan <riordan at jlab.org <mailto:riordan at jlab.org>> wrote:
>
>     Just the standard different optimization things I want to try, it's a short list.  Putting
>     things into an FOM fitter and letting it run, tungsten tipped, angled edges.  That also includes
>     exploring shorter target options.
>
>
> I'm a little hazy on what was done a couple years ago but as I understand it (educate me if I'm
> wrong) there was a focus on minimizing apertures for straight through photons from target to
> detectors. The resulting CLEO baffles produced worse backgrounds than the BaBar baffles that were
> not even intended for the CLEO field. We now know a lot of what we're seeing in our backgrounds is
> from 1-bounce photons and from hadrons producing showers in the baffles. Is that why previous
> optimization didn't work?
>
> Of course the backgrounds we get depend on the generators we use. That implies optimizing against
> one generator will give different results than optimizing against another. So baffle optimization is
> to some extent tied in with the questions we have about the generators we should use.
>
> Also our picture of the backgrounds in the GEMs changes if we take advantage of the gaps between
> sectors and let the photon peaks fall there; we also can block the corresponding areas of the EC.
>
> So we need to be careful with our choice of optimization methods, criteria, and constraints. And we
> need to think about how we parameterize the baffles. The Perl scripts use 40+ numbers to describe
> each baffle. I suspect we need, in a sense, both fewer and more numbers — fewer to describe the
> baffles we have currently, more to incorporate features like slit liners and angled edges.
> Intelligent choice of parameterization is probably at least half the optimization battle.
>
> Of course this can go far beyond what you can cover in half of a 10-minute talk, but I think these
> are things to mention and to start discussing.
>
> --
> - Richard S. Holmes
>    Physics Department
>    Syracuse University
>    Syracuse, NY 13244
>    315-443-5977