Minutes of the Tracker Technical Meeting April 19, 2000 Present: UCSC: Robert, Masa, Jose, Wilko, Gwelen, Alec SLAC: Martin, Elliott, Eduardo, John, BJ, Hartmut, Gary Trieste: Guido, Fernando, Christian, Julio, Michela Hytec: Erik, Eric, Scott Guido announced that Trieste has detectors from ST. They are still in the wafer. Trieste want to test them both before and after dicing. Michela said that they planned to cut them by scribing, but concern was expressed that this would not be a good test, since in production the dicing will be done with a wafer saw. She agreed that at least some of them should be sent back out for sawing. Regarding the SLAC detectors, this week ST is testing capacitors. Next week they will test currents. Michela will find out a definite delivery schedule. Data sheets will be sent with detectors (by email). Eric Ponslet summarized the document that he wrote on making low-CTE converters (attached to this mail). In the meeting he discussed 4 options: 1. a lead-carbon-matrix composite, in which carbon fibers are imbedded directly in the lead. 2. A sandwich of prepreg carbon-fiber panels and pb foils cured together. 3. Already cured composite panels bonded to lead. 4. Lead and invar bonded together. Here the invar contributes substantially to the radiation lengths. He tried to select the material thickness to achieve a CTE of 2.6 per million with 25% radiation length total. The figures of merit then were the percent increase in thickness and mass with respect to pure lead. The 1st option would offer the best performance, but the development cost would probably be too much for us. Next best is option 2, but there is an issue of thermal stress in cooling down from the curing temperature. Many thin layers are needed in order to mitigate this problem. Gwelen asked if one might use a room-temperature prepreg. Eric acknowledged that that could be a possibility. Option 3 could be more work to produce, and one pays a price of extra adhesive in the assembly. Option 4 has a severe mass penalty, and concern was expressed about the use of such a low-Z material for a substantial part of the converter. Another concern in some of these options is that the increased thickness will impact the PSF by moving the conversion points away from the detectors. Eric considered both mid-range and high-end fibers. The option 2 requires about 14 layers Pb and 16 layers prepreg. He guessed at a cost of about $100/sq ft for the high-end carbon fiber material. It was agreed that if possible, CERN should try to prototype option 2. They should first review Eric's note, then select materials and fabric thicknesses that they prefer or have available. Then Eric should re-run calculation for those materials. Eric will send instructions (by way of Guido?). Eduardo talked to Ed Garwin about CTE measurement at SLAC. Derick looked through a catalog for strain gauges and found some for about $600 for a single-channel system. Eric stressed the need high quality amplifiers. They paid about $1500 for a single-channel and about another $1500 for a 10-channel multiplexer. He suggested the vendor "Vishay", which is an industry standard. Erik will send information to Eduardo. Gwelen suggested that a computer interface would be best, but Eric said that the computer interfaced amps are lower quality. Gary and Derick will set up the system at SLAC, using existing environmental chambers. Hartmut talked about the new committee set up to look into a timeline for fixing tracker parameters. The issues are the overall footprint (mass), strip pitch (power and S/N), and implant width (micro-discharge and capacitance). There was a kickoff phone meeting Tuesday, and a report is due May 11. Trieste needs information for field calculations: oxide charge, oxide depth, and implant depth. This information should be sent to piemonte@trieste.infn.it and copied to Guido. Regarding mechanical samples, Hamamatsu has lots of BTEM type detectors (about $100 each), which are 6.4cm by 10.7cm. Only a few 9.5cm mechanical samples are available now, but Hamamatsu is now running 9.5cm square detectors in 100 Si. Hartmut recommended that we get as many of those large ones as possible. Gwelen would like lots of BTEM ones, so that extensive production testing can be done using some of the fixturing already available (such as work holders). He suggested building the first prototype of the new ladder fixture for that size. It was concluded that Eduardo and Gwelen should get together and decide on a strategy for this. Hartmut talked about some analysis he did of signal-to-noise. He fit Wilko's data for single-plane occupancy and also found that he can explain the full plane occupancy (OR of 1596 channels) from the singles rates within a factor of 2. This indicates good uniformity across channels. To get a noise figure, he plotted the log of the occupancy as a function of threshold squared (since the dependence is approximately the tail of a gaussian). This fit very well to a line, allowing him to calculate a noise sigma. He gets a noise of about 0.2 fC or 1220 electrons. He now would like to look at the slope with respect to capacitance by using the few short strips (2 detectors instead of 3) on the layer. There was some surprise at this result, but Robert explained that the old published noise slope was obtained by a threshold scan involving an external pulser, which was eventually realized to disturb the system ground so much that it adds a significant contribution to the noise. The occupancy fit is a more reliable measure, and closer to what we are really concerned about, anyway. Also, spice simulations, using as input our measured transistor noise spectrum, always gave lower noise predictions than the old measurements. Robert recalled that the simulations agreed fairly well with Hartmut's and Wilko's new results. Robert mentioned that Dave Nelson at SLAC is concerned about crosstalk and would like measurements and/or simulations. Eduardo presented the latest design for a gluing jig. It looks a lot like the old one, but with some improvements: 1. The posts have vacuum, with 4 separate circuits. 2. A cam pushes against the edge of each detector to press them against the alignment pins. 3. A lot of space is left underneath for a UV lamp. They are working on methods to apply glue by hand in more controlled way. One concern is to use a jig to bring the detector down into the glue film vertical and lift it straight back out. Gwelen would like a dedicated face-to-face meeting at SLAC to discuss this fixture. They will fix a schedule for that offline. Robert stressed the important for INFN institutes to review and agree on the fixture as well, since our plan is that all assembly points should use same procedures and fixtures. Eduardo ordered some samples of NASA approved UV glues. He is calling about a dozen companies for more information. A SLAC material engineer is going to help with understanding the requirements. Gwelen would like to be included in the glue discussion. Gwelen mentioned that shorts between the back planes of adjacent detector in ladders is an issue, because it makes debugging more complicated. Such shorts were commonly seen in BTEM ladders. There was some discussion of the issue of using 2 glues for detector mounting. Robert said that the only concern is not to put too much silver (high Z) into the tracker volume, but we need to evaluate quantitatively the amount of silver that would be involved, the impact on performance, and the impact on assembly of having to use 2 adhesives, one loaded and one unloaded. Maybe we can use carbon loading instead. Robert verified with Martin that the famous 10 mils in the hybrid backing plate have NOT been removed from the drawings (and should not be). Jose described progress on the psf analysis. He has checked matching of the tracker to calorimeter. There was a problem involving a 90 degree rotation between the two. He solved this problem, which was an issue of definition of coordinates. He needs to talk to Dan and Berrie about making a drawing of the coordinates to clarify things. Jose started to write some documentation on how to use the program, to help our French collaborators. He has problems with putting the latest version in CVS. Now the tracker and CAL reconstruction codes are working together. Hopefully the French can take over the CAL code now. The CAL energies seem to make sense, but need more checking. Jose needs to make a few more checks before doing the PSF analysis. There is no root tree output yet, but Robert thought that the immediate priority should be to get some results on the PSF. Wilko thinks that the root tree for calorimeter information needs to be reordered to make a consistent coordinate definition and avoid more confusion in the future. This need to be discussed with Richard. Eduardo reported that they found a machine on campus to measure shear forces and shear strength. It can measure down to about 400 psi. It takes only about .AN= hour for an experiment. They do need to make a jig to use with it. It was agreed that this should be used to test the adhesive for detector mounting. Eric mentioned that standards exist for doing such measurements, which should be consulted. Eduardo was aware of at least some of them. Gwelen would also like to measure peel strength of the hybrid from the tray closeout. *B Hartmut reported that 111 Si and 100 Si really have the same bulk strength. This was determined by further research. Hence this is not an issue we should worry about. Erik mentioned that there are baseline drawings now on the Hytec ftp site. He also still needs facesheet material properties, which is a open action item from the last meeting. Issues 1. How many, and what type of, mechanical samples should be purchased now? 2. What is the cross-talk between adjacent channels in the system? How far is it from threshold? 3. Does the edge gluing procedure result in too many shorts between detectors? Is this really a problem for us? 4. How much silver loaded epoxy can we tolerate in the tracker volume? Should two adhesives be used for detector mounting, on loaded and one not? 5. There is a lot of confusion still in coordinate definitions in the ROOT data files. Action Items: 1. Eduardo: discuss possible changes in the CAL root structure with Richard. 2. Gwelen and Eduardo: schedule a meeting at SLAC to go over the ladder jig design. 3. Gwelen and Eduardo: work out a strategy for the use of mechanical samples in the near term. 4. Hartmut: send information on the oxide thickness, implant depth, and oxide charge to Trieste. 5. Eric: send instructions to Trieste/CERN on how to proceed with the SuperGLAST converter development. 6. Michela: find out the detector delivery schedule from ST and send to SLAC.