Dear Alessandro, Pat and Stephen,


Based on the Excel tables of v4 by Alessandro, I have tried 
to create drawings and a table of simulation volumes for on-
barrel services. I kept the total masses exactly same as 
those given by Alessandro. In the process of making and 
placing the simulation volumes, I have encountered many 
crashing cases. Therefore I had to make wild approximations 
in shapes, sizes and places of volumes. Below is my brief 
explanations. Please take a look of drawings (new) in my web 
page of
   http://atlas.kek.jp/si-soft/geometry.html
in which you can pick up new drawings (such as rphiLoer-B3 
new) and a table (OnBarrelServicesVolume_v1.xls) in box of
 gOn Barrel Services : For simulation geometry". 
In these drawings and table, you see some new volumes names 
because I re-named them just for simplicity. We must agree 
on namings sometime soon. 

(1) Cylinder : The length (z) is now set to 153.0 cm.

(2) EndFlange: Alessandro splits to two parts, but I just 
made it into one box for simplicity. 

(3) EndClamp: Its position (rrr) is moved 4 mm away from the 
cylinder outer radius, in order to avoid crashing with power-
tapes and straight harness. 

(4) CoolingEnd: Alessandro tables have two separate objects: 
Bend Loop Sector and Inlet-Outlet Assembly, which 
distributed along the edge of the barrel cylinder. I try to 
find the spaces for them, not conflicting with module edges 
and EndClamps. But not easy. Also if we keep these two 
objects separately, we need to specify the divisions and 
locations in phi direction. As a conclusion of my try and 
errors in drawings, I like to propose to combine two objects 
into one volume of a cylinder each end, as drawn in 
rphiCoolingEnd-B3.pdf. This scheme flattens out the phi 
irregularity of cooling inlet-outlet, but it makes the 
program very simple. In addition, I had to make the length
(z) of the CoolingEnd smaller and the position moved out in 
z because of crashing possibility with modules, which you 
can see from drawings of rzLeft-B3.pdf and rzRight-B3.pdf. 

(5) BottomBracket : no change from the current one.

(6) PowerTapes: I made the thickness (r) to 0.3 mm each 
layer because it include optical fibers. There positions are 
moved up radially in order to place the StarightHarness 
underneath (see next).  

(7) StraightHarness: Again after try and errors, I like to 
place it underneath the power tapes as shown in the 
drawings. The thickess (r) is arbitrary chosen to be 1 mm. 

(8) TopBracket-Dogleg (Dogleg, simply): I like to accept the 
simpler version of Alessandro, namely combining TopBracket 
and Dogleg since there are very close in space, and for 
simplicity. I like to keep the current dimension of 30mm
(phi) x 36mm(z) with no strong reasons to change. 

I hope these explanations as well as the drawings and the 
table are clear enough for your understanding. With these 
shapes, sizes and locations given here, I hope Pat can start 
implementing the new software version. If necessary, I can 
give PAT more numbers for programming necessity. I would be 
extremely happy to accept any of your questions, comments 
and suggestions.

Best regards,

Taka from CERN