CALICE MAPS Preliminary Design Review, RAL, 05/05/06 ==================================================== Present: Jamie Crooks, Paul Dauncey, Tim Jones (external reviewer), Anne-Marie Magnan, Yoshi Mikami, Mark Prydderch (external reviewer), Renato Turchetta, Mike Tyndel, John Wilson By phone: Nigel Watson Notes: Paul There were presentations from Renato on a general overview of MAPS sensors and from Paul on the requirements on MAPS set by the ILC and CALICE baseline design. Project Specification: Renato led the discussion on this document and the associated financial summary and two GANNT charts. The deliverables list does not include the laser system as this already exists in RAL/PPD. The recommendation sent before the PDR to add test reports for each round to the deliverables list was accepted. The GANNT chart does not specify the people involved as it is effectively all Jamie's work except for the last few items. There will be some aspects done by Renato but Jamie will partition these out as appropriate at the time; the detail is not yet known. It is clear the design schedule is very tight and items (such as the non-wirebonding pads and DAC) are being delayed until the second round to keep to time. There was discussion on whether it would be better overall to put more into the first round, even if it took longer and delayed the fabrication, because this would make the second round less risky. It should allow the second design to proceed faster also, which might compensate for the delays in the first round. However, the detailed schedule is uncertain because of the issue of a shared fabrication run, which would restrict fabrication to particular (as yet undefined) dates. This is probably the limiting factor on the sensor design period if we go with a shared run. When these dates are known, Jamie will put as many features onto the first round sensor as time allows. [Note added after the meeting; the financial spreadsheet numbers are incorrect as both RAL Technology and RAL PPD underspent on effort last FY.] Technical Specification: Jamie led the discussion on this document. The reviews all relate to one project so there can only be one Preliminary and one Final Design Review. Hence, the reviews with these names for beginning the second round design and finalising the first round design, respectively, will be called Intermediate. Some costs for Foundry B had been obtained by Renato. These were for a shared-wafer run giving 60 dies in the first round and a specific run of twelve 8-inch wafers in the second round. These prices in principle allow a choice of either Foundry A or B. There are significant design benefits in using Foundry B because of the 0.18mu feature size. However, there would be some risk as it would be the first RAL Technology design to this specification. However, this was not thought to be a major drawback; the structures are better modelled now than previously so the design software is more reliable. It was decided the benefits of 0.18mu outweighed the risks as long as the cost for both rounds of fabrication could be fixed from the start, ensuring they could both be done within the budget. A DAC was thought very desirable in the first round. In case of problems, the output would not be connected directly to the rest of the sensor but connected to an external pin which could then be connected to another pin when the DAC performance was proven. The issue of what happens for signals very much bigger than a MIP (which could occur due to slow, highly ionising particles) was raised. The pixel will presumably saturate but the time to become sensitive again was not known. Also, the possible rate of such large signals had not been studied. While it is likely the rate is low, it should be investigated. There will be some degree of exploitation of the experience from the MI3/OPIC design. This has a DRAM and comparator and so is similar to some degree. However, the specifications are somewhat different and working in 0.18mu means the actual design will have to be redone. Testing Specification: Paul led the discussion on this document. The first sensor will have four types of pixel design and the question was raised of powering them on separately. This would allow the power for the different designs to be measured. However, this would introduce extra complexity and the power for the different pixel designs could be measured from the test structures. It was not clear if a failing design would prevent readout of the working pixels; this should obviously be avoided if possible in the design. The physical constraints on the sensor holder in sec 8.1 arise from the focal depth of the laser from the base of the optics unit. The control board could be an FPGA development board as mentioned in the plan but another possibility which has recently arisen has been to use the IDAQ board developed at Imperial for a different MAPS sensor project. Project Management, Risk Management and Stakeholder Plans: Renato led the discussion on these documents. Due to a misinterpretation of the meanings of the labels, all the risks were labelled as Amber although none are currently seen as being a problem. These should be changed to the correct label, Green. Many thanks to the two external reviewers for helping with the PDR.