5 November 2002 J.Pater - Systemtest Report 1

Endcap Systemtest

• Current Grounding/Shielding status• Recent Upgrades to sector• Measurements with 6 Modules

– Basic performance– First tries at noise injection

• Plans

5 November 2002 J.Pater - Systemtest Report 2

Three Grounding Schemes• All 3 schemes have:

– continuous shielding VME crate to Cu sector enclosure,– (~) all power lines AC connected to shield at PPF1,– system of interconnected foil rings on disk face.

• Differences:– “Shunt Shields In”

• Built-in shunt shields on module + isolation layer on cooling block

• all modules’ DGNDs connected to shield at PPF0– “Shunt Shields (~) Shorted”

• Shunt shields still there but electrically bypassed• DGNDs not connected at PP0

– “Doubleshort”• Shunt shields ‘shorted’• DGNDs connected together at PPF0

• All 3 schemes perform equally well with no noise injected

5 November 2002 J.Pater - Systemtest Report 3

Measurements with 3 Modules

5 November 2002 J.Pater - Systemtest Report 4

Injection onto Tapes

“Shunt shields in” (black) vs.

“Shunt shields shorted” (red) “DoubleShort”

scheme(Note different vertical scale)

5 November 2002 J.Pater - Systemtest Report 5

Injection onto Pipes

“Shunt shields in” (black) vs.

“Shunt shields shorted” (red) “DoubleShort” scheme

(Note different vertical scale)

5 November 2002 J.Pater - Systemtest Report 6

Ground Scheme Comparison with3 outer modules: Summary

• A hybrid G&S scheme seems to work to best against injected noise:– Short around shunt shields i.e. DC-connect

all GNDs together at cooling circuit– Also connect all DGNDs together at PPF0 – NB this is based on one measurement;

needs verification.

5 November 2002 J.Pater - Systemtest Report 7

July/Aug: Upgrades to System

• Now can simultaneously read out 6 modules of different types– Many problems with connectorised power

tapes and Molex patch panels

• Shielding improvements:– Added missing filtering caps at PPF1 (Vdd-

shield, Vcc-shield)– Improved screening of 50 tapes– Added Cu sheet PPF1-thermal enclosure; used

to improve interconnection of cable shielding

5 November 2002 J.Pater - Systemtest Report 8

6 Modules on Sector

Position O11 not working

Back of disk: one module in position M10

5 November 2002 J.Pater - Systemtest Report 9

6-Module Run, no noise injection

Blue = in test box at instituteYellow = alone on sectorPurple = 6-module run

5 November 2002 J.Pater - Systemtest Report 10

6 Modules, noise injection onto cooling pipes (shunt shields in)

• Little or no pickup on inner and middle modules

• More pickup on outer modules than before

• Not understood, but…

• … electrical connection of cooling pipes to outside world is not well simulated in systemtest – needs improvement

5 November 2002 J.Pater - Systemtest Report 11

6 Modules, noise injection onto power tapes (shunt

shields in)• Injected similar signal as in

3-module case – no pickup seen (!)

• Injected MORE – very little pickup seen

(see plot to right )

• Recall grounding/shielding improvements:– more filtering at PPF1

– better shielding and connections between PPF1 and sector box

Recent work suggests that the solidity of the electrical connection between PPF1

and the thermal enclosure is crucial.

5 November 2002 J.Pater - Systemtest Report 12

Near-future Plans• Upgrades: as soon as is practical:

– Make cooling pipes’ exit from thermal enclosure more realistic

– Install new on-disk tapes, PPF0s• (provided initial tests show satisfactory performance)

• Repeat injection with all 3 G&S schemes– To verify previous observations with outer modules– To see what happens with middles and inners

• If “doubleshort” scheme still best:– Remove shunt shields from modules to verify that it is not

needed.– Address what to do electrically at cooling block.

• Check: is pickup ~simply function of proximity to tape?– Test other outer positions, upper and lower– Vary distance silicon-to-tape. Need add-on for cooling block.