ATLAS Level-1 Calorimeter Trigger Hardware Progress Meeting - 15 February 2000 Minutes Present: Bruce Barnett, Eric Eisenhandler, John Garvey, Norman Gee, Tony Gillman, Bob Hatley, Murrough Landon, Viraj Perera 1. Birmingham hardware Unfortunately Richard was ill and therefore unable to be present, but he had e-mailed some relevant information which Tony reported on his behalf. The ATLAS note on the backplane studies has been slightly revised, and it will hopefully make it through the submission process this week (w/b 14th February). The CPM prototype specifications are currently being updated, and a new draft will be ready next week (w/b 21st February). There will be at least one more draft version available before the Preliminary Design Review. Software control of the TTC system is being worked on at present, and if Scott and Bill need further help they should call on Bruce or Murrough. Richard also points out that he needs one of the CERN TTCrx Test Cards - arrangements will be made to get one sent to him in Birmingham. On the general topic of transmission studies, Bob reported that the 100m drum of halogen-free coaxial cable from CERN is now at RAL, and he will assemble it with appropriate connectors into four 25m sections for the final burst-mode G-link tests. These tests will require appropriate firmware and software to be prepared, so Viraj and Norman will look into the requirements and report at the next meeting. 2. CPM prototype review Tony noted that the date for this Preliminary Design Review had unfortunately been originally proposed for Monday 20th March, but as this clashed with the next UK Group meeting the PDR would be moved to Friday 17th March. The proposed reviewers are: Bruce Barnett Paul Hanke Viraj Perera Uli Schaefer Tony will contact Paul Hanke and Uli Schaefer to invite them to participate in this review. (N.B. In discussions following this meeting it was agreed to postpone the review date until after the Mainz meeting (to avoid overload!) - Monday 10th April was proposed.) 3. ROD prototype status Viraj reported that the System Design group in ID had now obtained the PCI core, and James is making good progress with its evaluation. This process must be completed by 5th April, by which time it must be purchased. Richard had heard from Pedja Jovanovic that CERN were also purchasing a PCI core, thought to be from a third-party. In case this could be beneficial to us, Norman will follow it up with Pierre Van de Vyvre of the ALICE DAQ group. The overall design work for the ROD module itself is proceeding very well, and is still on target for completion and submission to the Drawing Office by the end of March. 4. CP and Serialiser FPGA status Ian has completed design of the Serialiser chip, and is now working on the corresponding de-serialiser function in the front-end of the CP chip. There is now a preference for using a Virtex-E device for the Serialiser, and a corresponding new proposal for the corresponding Test Card to test the Serialiser and CP chips together. Viraj described the essential features of this proposal, which would be for a full VME module supporting the existing LVDS and/or G-link daughter-cards. This new module would be only of similar complexity to the double CMC test card proposed originally, which would have resided on the existing DSS module. One major advantage of this proposal would be the provision of direct VME access to the Serialiser and CP chips, which would be impossible for the CMC daughter-card option. As its flexibility could allow for wider use in other projects, additional design effort would be provided, so there would be no extra cost or time penalty to the level-1 project. Viraj was asked to prepare some more details of this proposal (cost, design effort, timescale, etc.) for our next meeting. Viraj also reported that the CP chip specification document needs updating, following the review. The question was raised about whether we have sufficient technician effort available for the forthcoming phase of module design and testing - at present it is felt that we do, but the situation will be monitored closely. 5. CMM status Norman described the most recent version of the draft specification document for the Common Merger module. It covers a good deal of the module functionality but still with several gaps and questions of both a functional and technical nature. Also, it does not cover the electronic engineering details that will be needed to be added by Viraj after he reviews it. The goal is to have an essentially complete draft available by the time of the Heidelberg "brainstorming" meeting on 3rd March. 6. TCM prototype Bob described the latest version of the TCM specification document, which has been updated following comments from several people. It is now unclear whether the module should be of 9U or 6U format, in the light of recent concerns about the cost of 9U modules containing fine-pitch BGA packages. Norman agreed to get the most recent information from CERN about the status of the proposal for a standard ROD crate, which we may decide to adopt. A discussion of crate systems will be put on to the agenda for the Mainz meeting. Bob asked for comments on the proposed architecture for the CAN-bus control function of the TCM, and Norman agreed to get information about the current plans of the DCS people at CERN, hopefully from their web pages. In the absence of an actual Detector Control System for our prototype tests, when a standalone PC would be used instead, we must not forget that a certain minimum level of software will be needed in order to operate the control system. Bob also outlined the most important features of the TTCdec cards, which will reside on all future trigger modules, receiving TTC signals from the TCM via backplane or cable links. The aim will be to have "UK-approved" specification documents for both the TCM and the TTCdec available for discussion at the Heidelberg "brainstorming" meeting on 3rd March. Tony mentioned that he had received an e-mail from Thomas Toifl at CERN, who has been testing the new D-MILL version of the TTCrx chips. Five chips were measured, two of which showed the following jitter figures (see first table below): At 5 V supply voltage: Optical input power (dBm) rms jitter (ps) p-p jitter (ps -27.6 93.3 704 -25.6 43.5 286 -23.5 34.7 236 -21.4 38.1 304 -19.3 35.5 296 -17.2 38.8 411 -15.1 32.2 240 -13.0 28.2 220 The remaining three chips showed about a factor two worse jitter figures, although these were correctable by reprogramming an internal parameter. However, when these three chips were operated with a 3.3V supply voltage their jitter behaviour improved (see next table below): At 3.3 V supply voltage: Optical input power (dBm) rms jitter (ps) p-p jitter (ps) -31.6 182.1 1354 -29.6 81.3 773 -27.6 48.7 636 -25.6 38.8 363 -23.5 36.0 345 -21.4 30.6 273 -19.3 28.2 273 -17.2 30.2 309 -15.1 27.6 291 -13.0 25.6 264 -11.1 36.2 291 -10.7 28.8 273 It was stressed that it was still too early to draw any general conclusions from these measurements, as chips in BGA packages would be needed and a much larger statistics sample measured. Tony has requested that the ten prototype TTCrx chips which were promised to Level-1 should be reserved for us, and a Purchase Order for them will be raised as soon as possible. 7. Summary of Software Management meeting - 15th February Norman gave a brief outline of the conclusions from this meeting, which covered the following areas: software effort at the three UK institutes, diagnostics software development, data acquisition software and computing hardware relating to VME systems. 8. Alternative link architecture There was some further brief discussion of the advantages and difficulties of the Virtex- E/Paroli link architecture, but no conclusions were reached. 9. Agenda for Heidelberg "brainstorming" meeting - 3rd March Some ideas for inclusion in the informal agenda for this meeting were discussed.The topics covered CMM issues, backplane configurations, TTC distribution, VME standards, crate standards, 9U vs 6U boards (costs, etc.), Virtex-E/Paroli-type architectures . Tony will circulate a draft agenda for general comments before the meeting. 10. Any other business. There was none. Next meeting - Tuesday 29th February at 10:00 in Room S8, Poynting Building, at Birmingham University. Tony Gillman