CALICE MAPS Meeting, RAL, 21/04/06 ================================== Present: Jamie Crooks, Paul Dauncey, Anne-Marie Magnan, Yoshi Mikami, Renato Turchetta, Mike Tyndel, Giulio Villani, Nigel Watson Minutes: Paul Minutes of last meeting: Following the meeting, Yoshi pointed out a mistake and corrected the statement concerning the energy ratio value of the epitaxial layer to bulk, which is in fact consistent with the expected value. Paul has removed all sensitive cost information relating to the foundries from the minutes and slides. Giulio should do the Siena talk and should contact Marc Weber directly. Preliminary Design Review: This will take place on Fri 5 May, starting at 2pm. We will hold it in the usual meeting room in R76 and will call Nigel by phone as he will be at SLAC. On Mike's recommendation, we have got Tim Jones from Liverpool to come along as one external reviewer. Renato has asked Mark Prydderch to be the second but has not had a reply yet. If Mark in not available, other possibilities are Nicola Guerrini and Marcus French. The proposed agenda for the PDR would be a physics introduction by Paul and a technical introduction to MAPS by Renato, before then going though each of the documents. Jamie will lead the discussion on the main document, the Technical Specification, and it would be worth putting some slides together to steer this, even if just a collection of the relevant plots. This document could take up to an hour to go through. We will aim this review only at the first round of sensor production (ASIC1 in the documentation) with a further PDR next year for the second round. There was a discussion on whether the documents should be for the whole of the project (i.e. WP3) or just the RAL part of this. Given all the parts are very inter-related, it was thought the review should cover the whole project. This means the work done outside of RAL (or even RAL Technology) should be covered as subcontracted out work, even though the same people are the customers. The status of each of the other four documents was discussed: o Project Specification: Jamie said most of the content for this document was known, except for the testing parts. These will now be put into a separate document (see below). o Risk Management Plan: For the OsC, WP3 has one specific risk, concering failure of a sensor round. This was purposely left ambiguous as it was intended to cover anything that would mean we would have to redo the fabrication (errors in design, manufacturing, etc). The OsC risk register had other risks, not specific to any particular WP, concerning retention of critical personnel. These are all relevant to the PDR risk register also. In addition, other risks identified were; schedule, inability to test the sensor and inability to use the same technology in the second round as the first. This last risk could arise because the cost of the second round is too high. Hence, the fabrication contract should be for both rounds, so as to fix the total price before the first round. o Project Management Plan: This will apply to RAL Technology only. Again, Jamie saw no sigificant issues outstanding for this. o Technical Specification: Jamie had distributed a first draft before the meeting (see usual web page). There were various comments which will be incorporated in the next version. In particular, the possible variations for the four different types of pixel should be mentioned. (Mike commented that most work should be targetted toward one design and the others should be a lower priority; otherwise it is less likely any will work well.) The document also mentioned Foundries A and B, although B was thought too expensive. Renato will get an update on the costs by the end of next week, so we will know by the time of the PDR if B is still excluded on cost. All documents will need to be released on Fri 28 Apr. It was decided that the testing should be specified in more detail than the Project Management Plan document would allow and hence a fifth document, a Testing Specification, would be needed. Paul will edit this together with input from Giulio and Jamie on their test requirements. It should include a list of things which need to be tested as well as a description of how they will be tested and by who. The tests divide into four sections; basic (RAL Technology), laser (RAL PPD), source (Imperial) and cosmics (Birmingham). The basic tests are to verify the functionality of the sensor. These include measuring the response time of the electronics, the dark noise rate, the configuration and data I/O and the comparator. The threshold will need to be set externally as there will not be a DAC on the sensor in the first round. For performing automatic threshold scans, then the test system will need a DAC on the sensor PCB. The threshold value will be likely to be around 10mV. One critical issue for the other tests is having the pixel mask, so only particular regions of the sensor can be read out. Jamie had not planned for this in the first round, but will now include it. The laser system is intended to allow comparisons with the sensor simulation. Therefore, these tests will focus on properties such as charge diffusion, collection time, crosstalk and MIP sensitivity. The sensor will probably have a lot of metal layers on the surface so the number of places where the laser light can reach the silicon may be limited. However, the laser can be operated at a wavelength of 1064nm, for which silicon is very transparent, so the laser could be shone on the underside. The laser system exists and has been used although RAL PPD are hoping to get their RA earlier than the original date of Apr07, so that the RA can learn the system and make any changes which may be needed. The system has a movable stage so things like a uniformity scan should be straightforward. An absolute calibration should be possible using other silicon detectors with a known response. There is no analogue output from the sensor but the signal size can be deduced from a threshold scan. Similarly, the charge collection time can be determined from a timing scan of the comparator. The source measurements will be with a high-Q beta source which gives a reasonably fast electron, resulting in a signal not far from a MIP. Some scintillator triggers will be needed and possibly a thin sheet of lead absorber to remove low energy electrons knocked out of the sensor. The source setup could also be used (without the source itself) as part of a magnetic field test to check performance, particularly dark noise rate, in a B field; 1T would be the minimum of interest, although 4-5T would be preferable. A usable magnet has not yet been found. The cosmics test would involve a stack of sensors and a simple cosmic scintillator trigger. The rate would be low (1/steradian/min) so this would be a dedicated measurement which would run for several weeks. Sensor design: Besides the preparations for the PDR, there has not been much other work on the design. Jamie and Renato are considering centralising the memory for several pixels in one dead area, which would lump the inefficiency in one place. If these regions are small and the spacing between them is small compared with a typical shower spread of ~1cm, this should not be a problem. However, a more detailed physics simulation study will be needed. Sensor simulation: Giulio reported that the disks for the simulation farm has still not been installed, although they were supposed to have been ready several weeks ago. Giulio had been in contact with the IT people but Renato asked for future requests for such work to go through him as he would be able to apply pressure. He thought the disks were probably at RAL and would be installed within a few days. This is clearly a high priority now and Renato will make sure it gets done soon. Renato will send Giulio the technical information from Foundry B next week if it does not look too expensive. In addition, the deep N-wells can be simulated; Renato will send the relevant information also. Physics simulation: Yoshi showed some slides on progress on the MAPS implementation in GEANT4 (see usual web page). He sees that the total number of pixels hit with just longitudinal segmentation drops when looking only at the epitaxial hits compared to the original, which needs further investigation to understand. He has also now started to divide the silicon pads transversely and so now is able to measure the number of hits in pixels which are the correct size for MAPS. Assuming the number of pixels which would fire is with a factor of two of the number of hits, then this allows an estimate of occupancy in MAPS due to physics. Yoshi and Nigel will now try to simulation the main physical sources of pixel hits; Bhabhas, mini-jets and machine background. These numbers will be very useful for the PDR, for Paul's talk. Anne-Marie showed a slide outlining the plans for the post-GEANT4 digitisation simulation (again, see the usual web page). This would add noise and apply a threshold, with the code implemented within the Marlin framework. Next meeting: The PDR, at 2pm on Fri 5 May, in R76 as usual.