Rack Layout: Criteria

Murrough Landon - 2 July 1999

0pt 0pt 0pt

• Minimise the overall latency
• Ergonomics, Maintainability
• Cable routing paths
• Holes in shielding wall
• Allow for future flexibility

Rack Layout:
Latency

Current Critical Path

0pt 0pt 0pt

• Furthest (which?) FE electronics [60m$\rightarrow$]
• Receiver stations [10m?$\rightarrow$]
• PreProcessor Modules [7.5m$\rightarrow$]
• Jet/Energy Modules [7.5m?$\rightarrow$]
• Energy Sum Merger Modules [7.5m?$\rightarrow$]
• Central Trigger Processor [?]
• TTC Distribution [?]
• Critical (eg LAr) TTC Crates [?]
• Longest TTC Fibres [80m$\rightarrow$FE]

Less Critical Paths

0pt 0pt 0pt

• Muon and Cluster Processors (about 2-4 BC to spare, ie 10-20m of cable)

Rack Layout:
Cable Routing

External Rules

0pt 0pt 0pt

• Route to/from detector is via four (small, slanted) holes in shielding wall The proposed layout is far from ideal, going to both sides of both floors) Try to renegotiate with safety group
• Meanwhile TC would like to increase the maximum length!
• The path from rack to rack is normally down and under the false floor...
• ...but front panel to front panel is OK
• (Almost) no cables between floors (except via long route)

Rack Layout:
Cable Routing

Internal Requirements

0pt 0pt 0pt

• Receiver crates will have all +ve or -ve eta; and either barrel or endcap
• Each PPM crate spans all of eta and needs input from four receiver crates
• Each PPM crate feeds three CPM and three JEM crates
• Each CPM/JEM crate takes input from six PPM crates

Rack Layout:
Front/Back Panels

Input and output connectors of the various trigger modules (along the trigger data path) are as follows:

Module	Input	Output
Receivers	Back?	Front?
PPMs	Front	Back
CPMs	TBD	N/A
JEMs	Front	N/A
CMMs	Front	Front
SMMs	Front	Front
CTP	Front	Front
TTC	Front	Front

Front panel inputs are more easily disturbed, but might offer shorter cable runs.

Rack Layout:
Ergonomics

0pt 0pt 0pt

• Would like no more than two 9U crates per rack
• Centrally located for easy access and maintenance
• Other services (6U crates, bulk cabling, DCS, scopes, computers, monitors, network) above, between or below
• Quadrant architecture suggests a 2*2 layout is ideal...
• ...and doesnt take much extra cable path (because of the fanout to multiple crates)
• The (probably) eight crates of receiver modules are also likely to be naturally arranged in pairs rather than threes
• Separate racks for each system during installation and commissioning phase

Rack Layout:
Flexibility

0pt 0pt 0pt

• Critical latency path might change
• Maybe more subdetectors will want TTC crates close to CTP
• We might need more nearby crates than we expect
• ...
• Aim for a layout which can accommodate some growth