Report of the TDC Review Committee Committee: Rick Van Berg Bill Foster Henry Frisch Tony Liss (chair) Aseet Mukherjee Harold Sanders Michael Schmidt Terry Shaw The TDC review committee met on January 24, 2001 in a follow-up review to evaluate the status and operational experience of the Michigan TDCs. The agenda and charge are attached below. The committee was very pleased with the enormous progress that Myron and his group have made in addressing a large array of problems that have plagued the TDCs. Thanks and congratulations are due to a large number of people who have helped to identify problems and find fixes for them. The list of people who have contributed to this effort is long, however we wish to give special thanks to the efforts of Ron Moore, Eric James, Arndt Meyer, and of course Myron. Over all, the 8 ECOs seemed to have solved all the problems that were present at the last review. There are a few remaining, relatively minor (by comparison), issues. It would be surprising if there were not still a few bugs here and there, and the committee is confident that they will be solved in due course. The current plan to order 125 new LVDS boards and 75 new ECL boards seems reasonable. Myron stated that his plan is to manufacture 20 boards in order to test the revised design. We assume that all 200 PC boards will be ordered, but only 20 stuffed initially. The committee was split on the wisdom of ordering all at once due to the psychological pressure to stuff the remaining boards before the initial 20 have really passed muster. Given the absolute necessity that the 200 new boards actually do fix all the known problems, we have several recommendations regarding their production. First, we recommend that one or two independent engineers take a close look at the schematics, artwork and net lists prior to placing the order to ensure that all the changes are implemented as intended. Second, we recommend that after the 20 boards have passed muster in a test stand, that they be installed on the detector at B0 to gain operational experience prior to assembly of the remaining boards (or possibly, prior to placing the order for the remainder). Since several of the problems seem to be associated with cross talk, we are particularly concerned that the new design be thoroughly tested. We recognize that since installation at B0 would likely occur during the summer, it will represent an interruption into what will hopefully be smooth running conditions. However, given the importance of making sure that these boards will function adequately, and in particular to insure the peaceful coexistence of the new boards with the old, we think the operational testing should be a prerequisite to proceeding with the production of the remaining boards. We are somewhat concerned about the conflict between building and testing the new run of boards and the continued repair of the existing boards. We recommend that Myron provide the project managers with a detailed plan for delivery and maintenance of TDCs for the period after March 1. We recommend (and Myron did as well) that the TDC task force, which has been so successful in cataloging and understanding TDC problems, continue to operate. The task force should work on continued understanding of TDC failures and should be the repository of all information about these failures. Cathy reported that a study of single channel failures indicates failure rate of approximately 4% per month. It is currently unknown if this is a transient effect or not, and if it is due to chip failures, via failures, or what. The TDC task force should set as one of its priorities the understanding of this problem. There was much discussion about the limits in readout speed for the TDCs, and Myron presented some measurements of TDC processing time vs. occupancy. The COT occupancy during the commissioning run was higher than expected, but at least part of this was due to noisy channels that will presumably be fixed. On the other hand, the SVX and all its cables are now installed. If the COT occupancy is as expected (based on extrapolations from the CTC), then the TDCs are fast enough to allow the DAQ system to meet the 300 Hz Run 2a readout spec. If, however, the occupancy is significantly higher than expected, then the TDC readout may force more DAQ dead time than we had planned. For Run 2b, based on studies by the Yale group, the TDC processing must be done in less than 500 microseconds in order for the DAQ system to operate with less than 15% deadtime for a 1000 Hz readout rate into L3, assuming 45 kHz L1 Accepts, 20 microseconds for L2 decisions, etc. (because of the complexity of the DAQ system, you cannot simply invert the 500us to get the readout rate). There are many variables involved, and Arndt has been working on optimizing the readout algorithm. The situation is somewhat unclear, and one of the priorities with the first Run 2 data should be to optimize the TDC readout and understand its limitations. We know that this is already in the plans for the DAQ group. The decision about whether we can run at very high luminosity with these TDCs must await the conclusions of these studies. It is clear that the demands on the Michigan group will continue to be heavy in the coming months. In addition, there is the possibility that senior post docs may move on before the project is completed. This presents a serious schedule risk that has no easy solution. We recommend that all efforts be made to train a wider spectrum of people in CDF on the intricacies of the TDCs. Finally, Myron requested and we endorse the identification of at least two people resident at Fermilab who are in charge of TDC operations. These people would be responsible for the smooth operation of all the TDCs in the system, simple in situ debugging of the boards, and the decisions about when a board needs to be swapped out. The TDC operations people could also be members of the TDC task force, or could be separate. In either case they will need to couple closely to the task force. §§ Agenda: Exec Session - 15 min Intro and Charge - Bob K (10 min) TDC Status - Myron Design Problems and Solutions (ECOs) Operations Experience and Plan for Installation Readout Performance and Future DAQ Tests Plans for New Board Production Transition from Task force to Operations Exec Session Closeout Charge: Charge to TDC Review Committee CDF will conduct a follow-up status review of the Run II VME TDC sub-project. A review was conducted in Sept 2000 which focussed primarily on the manufacturing defect which resulted in broken vias on the boards. Following that review the committee recommended that CDF purchase new TDC boards to replace boards with bad vias. It also recommended that a task force be set up to address the technical problems with operation of the boards. Following the recommendations of that review, a TDC task force was formed in October. This group has been very effective in uncovering and fixing the known problems with the TDC's. We currently believe that nearly all these problems have been satisfactorily resolved. In addition, we have begun the process to purchase a total of 200 new TDC modules: 125 LVDS and 75 ECL. The purpose of this review is to evaluate the current state of the TDC project from a technical standpoint. The committee should evaluate whether all the design problems have been addressed and complete tests of solutions have been carried out. The committee should recommend an overall course of action to the head of the CDF FE/Trig subsystem, the CDF upgrade project managers, and the CDF operations department heads that they believe will lead to reliable operation of TDC's at CDF for the duration of Run II.