from the Meeting at the
These minutes can be found at:
o We had a meeting at UW to discuss specifically the grounding issues related to the inner silicon layers. These notes briefly summarize the main conclusions and action points of the meeting. There were three formal presentations: one by Kazu on his layer 0 studies, one by Marvin on general grounding issues, and one by Colin on the FEA mechanical model of the L0 structure.
Layer 0 Studies (Kazu)
o Kazu summarized the results of his recent grounding studies of the layer 0 prototype module. Details can be viewed at http://d0server1.fnal.gov/projects/run2b/Silicon/www/smt2b/Documentation/Minutes/May_05_meetingUW.ppt or http://d0server1.fnal.gov/projects/run2b/Silicon/www/smt2b/Documentation/Minutes/May_05_meetingUW.pdf
o Under the best circumstances a S/N ratio of 11 is obtained on the bench with a layer 0 module. The design goal aims for a S/N greater than 12 before irradiation. Noise levels are demonstrated to be strongly dependent upon the details of the grounding. The best performance is observed when only the hybrid is grounded. When only the sensors are grounded the noise is extremely high. This points to the importance of a low inductance connection. An interesting study which corroborates the importance of the low inductance connection was clearly demonstrated when Kazu added additional ground wires to the hybrid ground connection, varying the number of wires and their length. .
General Grounding Issues and Implications (Marvin)
o Marvin reported on his interpretation of numerous observations made on CFT and SMT readout of the detectors at DAB. His scanned presentation can be found at http://d0server1.fnal.gov/projects/run2b/Silicon/www/smt2b/Documentation/Minutes/Marvins_smt_cft_noise.pdf . Some of Marvin’s observations are:
o SVX oscillations are observed after the preamp reset in the SMT, CFT and CPS resulting in pedestal variations of a few counts for the SMT
o All channels have a similar response; the SVX2 chip is not ground referenced. As such is the case, the most likely culprit for this effect is a bounce in the ground.
o The CFT sees a similar effect in the SIFT readout (discriminator readout).
o Preamp reset is probably not the only signal that the SVX is sensitive to in this fashion, it just happens to be phase locked so that we directly notice the impact of the ground shift in this case.
o CFT readout also suffers from split pedestals in alternating ticks. This looks like a harmonic of RF/4, and Marvin suspect that this is also attributable to ground bounce.
o The concern (with respect to the Run IIb SMT project) is that the current design ties the analog and digital grounds together, so the current design may also be sensitive to these problems (although there are certainly significant differences between the SVX 2 and the SVX 4 which may reduce the sensitivity to this issue)
o Marvin emphasized that it is good practice to separate analog and digital grounds, but this would require more layers in the hybrid, making the hybrid more challenging to fabricate to specs and increasing costs. The design of the digital cable would also have to be modified to accommodate this separation of grounds.
o An Analog Front End Board for the CFT has been modified to change the inductance between the analog and digital ground, to attempt to reduce this effect (and verify the interpretation), and that AFE board has been installed on the platform. Studies performed in the near future should provide additional insights
Conclusions and Action Items
o The L1 shell will have an extension from the cooling manifold at approximately z=600 to z=815 and an end cap to serve as a Faraday shield for Layer 0
o Breakdown voltages to be verified to insure adequate gaps
o Sensors will tentatively be pulled back at z=0 cm from the end flange by 1.2 mm
o Copper in the ground mesh will be pulled back in the kapton by at least 2 mm, and plating trace will likely be etched back from edge
o An additional pad will be provided on the ground mesh for the castellated structure at the high z end. This will be used to connect the grounds of the sensor support structure to the hybrid support structure.
o The grounds of Layer 0A and 0B of the hybrid support structure will be connected using small strips that will snake through the cooling lines. UW will determine what the best possible way is to provide this connection
o All contacts will be gold plated
o 1 mil kapton will be used for the ground mesh and lengths will be evaluated to eliminate trimming
o six strips of ground mesh will be provided on the carbon fiber hybrid support structure. UW will provide the dimensions of the strips.
o Ground mesh in the castellated structure on layers A and B will not overlap. The gap between the two will be protected by Kapton in the vicinity of the HV filter card, and this Kapton will be installed at Fermilab.
o Gold plated pins will tie the L0 to the L1 structure
o Thermal cycling tests will be performed on the proposed conductive epoxy joints (at UW)
o ELMA sensors will be diced to match the HPK specifications to perform grounding studies
o Fermilab will decide on the shape of the end flange for the connection between Layer 1 and 2. A notched flange was proposed.
o Within two weeks, UW will evaluate whether the location for the radiation monitors at z=815 mm inside the layer2-5 support cylinder is acceptable.
o Concerns regarding possible interferences at manifolds need additional investigation, and cooling tube diameter will be evaluated
o electrical coupling between radial membranes and the support cylinders needs to be properly evaluated and addressed
o various installation concepts were discussed
o Henry and Bill will provide a set of specifications for Layer 0 and Layer 1 to which the support structures have to conform
Link to the Run IIb silicon page
Links to minutes archive