July 30, 1998

A review was held on May 7 and 8 of the PHENIX muon systems, the muon tracking system and the muon identifier system. The review covered the mechanical and electronic construction of these chamber systems. The committee consisted of Chen Yi Chi (Columbia University), Thomas Hemmick (State University of New York at Stony Brook), John Haggerty and Venetios Polychronakos (Brookhaven National Laboratory), and William Edwards (Lawrence Berkeley Laboratory). (Edwards, however, was present only for the first part of the review. The other members of the committee are PHENIX collaborators, except for Polychronakos.)

The committee was impressed with the progress on both the muon tracker and identifier systems, which is obviously the result of much thought and effort. Both the muon tracker and identifier groups have produced mature prototype chambers, and the muon identifier group has begun manufacturing them. The conceptual designs of the electronics for both is very far along, and some prototype work has been done, including a preamplifier ASIC for the muon tracker.

The major areas of concern were the following:

1. The scaffolding inside the magnet for the muon tracker could be complicated by safety considerations, so design and review by the appropriate members of the RHIC ESC should move forward as quickly as possible.
2. Muon tracking simulation should be used to assess the effect of not instrumenting anodes in the muon tracker.
3. Consideration should be given to ways to instrument at least station 1 of the muon tracker anodes at low cost.
4. The conceptual design of the muon tracker electronics should be examined to ascertain a) whether the number and size of PC boards is appropriate, and b) whether the relatively long signal path for unamplified signals is workable.
5. The engineering team for the muon tracker FEE should be augmented by the addition of an individual experienced with low noise wire chamber electronics in at least a consultant capacity, and the size of the engineering team should be kept as small as possible to minimize the total design cost.
6. The muon identifier chamber factory must be closely managed to ensure completion of the chamber installation before the shield wall construction begins without compromising the chamber quality or the safety of the technicians and students in the factory.

The muon tracking mechanics have reached a very mature stage in terms of design details and prototype performance measurements. The muon tracking team is to be congratulated on excellent progress in finalizing designs, demonstrating required performance in prototype chambers, and significantly advancing the realism and performance of the simulations. These accomplishments convinced the committee that the muon tracking mechanics team is fully capable of delivering on all aspects of what is arguably the most ambitious (channel count), large (station 3), and high precision (s=100 mm, stability = 25 mm) detector system in PHENIX. Curiously, the best position resolution (100 mm) was reported for the largest prototype chamber and the worst result (130 mm) came from the smallest one. Although the station 1 result is out of spec, the committee is confident that the muon-tracker team will improve its performance to match station 3.

One possible concern raised by the committee was gaining safety approval of scaffolding to be used inside the muon magnets. The tracking group would be well advised to start this review process at the earliest date. Also, quality assurance procedures must be established for the chamber factory. In general, the construction schedule was judged as aggressive with little or no contingency for solving problems that may arise during construction, such as acquiring the necessary safety approvals.

Tremendous effort has been placed on the tracking software in recent months. The results of this effort were quite impressive. In particular, the committee was pleased to see that recent, considerably more precise simulations show nearly the same performance of the muon systems as was predicted in the CDR. In addition to a reevaluation of the physics performance, the committee was treated to the 3D event displays and track reconstruction analysis under a large, albeit incomplete, set of simulation conditions.

The committee realizes that early simulation efforts are often misleading. On the one hand, incomplete or not yet finely tuned algorithms give an overly pessimistic view of reality. On the other hand, real backgrounds are always worse than simulations. The latter was shown specifically when comparing two GEANT thresholds. For "high" thresholds, 32% of station 1 clusters share more than one track while for "low" thresholds 50% share more than one track. The most detailed simulations shown were all performed using high thresholds. The committee noticed one curiosity of low statistics. For the full detector system 90% efficiency was found. Removing one gap in station 3 increased the efficiency to 92% and removing the anodes further increased the efficiency to 93%. Although this performance is quite good, "problems" were listed for the low threshold + no anode simulations. Since the strip width of all chambers is the same, occupancy falls as the detector size increases. For that reason the committee points out that anode readout is most critical for station 1 and additionally could be critical ONLY at station 1.

Since the track density in the chambers falls rapidly away from the piston, one easy way to decrease occupancy (and acceptance) is to lower the voltage on the innermost anode wires effectively turning them off remotely. Although time will bring anode instrumentation, and running experience will teach us where best to place additional shielding, the earliest running will clearly be the toughest. It would be quite useful to determine the impact on overall acceptance in the "inner anodes off" configuration since this may very well be the conditions for our first data set.

Clearly the muon-tracking group is making significant progress on tracking algorithms. A suggestion that might be useful to pursue is that since the cathode resolution is so high, radial locations of tracks could, in principle, be determined better by the cathodes than anodes. In practice, since avalanches only occur at the anode wires, the radial position spectrum of tracks images the wires themselves. Perhaps this information could be used to test the validity of cluster candidates as an additional aid for the highest occupancies.

The cost estimate of the electronics has more than doubled since the last estimate. It is recognized that the new bottoms-up estimate is more realistic and more reliable, but concern remains that the assigned contingency may still be too low given the fact that most of the components are in a rather early stage of development. The team lost several key engineers and there was much reorganization of task assignments in the last year which resulted in a considerable amount of expenditures which resulted in only a conceptual design for the electronics; this will clearly reduce the amount of money available to carry out the actual design. This suggests that very close attention must be paid to control spending, particularly in minimizing the amount of money spent on electrical engineering, and maintaining a small, efficient design team.

Recent analysis of cost totals and profiles have led to decisions to defer certain parts (readout of one gap at station 3, all anode readout electronics) of the muon tracking system. In addition to overall progress and performance assessment, the committee was charged to evaluate these specific deferral decisions. These evaluations are made on the basis of presently available simulations as well as prototype performance tests. The committee was firmly convinced by the detailed simulations that the deferral of readout of one gap from station 3 will not adversely affect the performance of the muon tracking system. The committee was uncertain whether the same could be said of anode deferral, specifically for the station 1 chambers. In addition to further simulations, the committee recommends that the muon tracking systems place a high priority on instrumenting the anodes for at least one plane of the station 1 detector at the earliest possible moment in the PHENIX physics program and that possible ways to implement this should be investigated. The idea of repackaging muon id FEE for anode readout is very good, and should be pursued by a PHENIX task force to evaluate the technical feasibility.

The proposed cathode readout consists of 16 slots of 3U X 160mm EURO crate standard. The crate consists of seven different types of PC boards, each of which will read out 256 channels of cathode strip. This kind of design has the benefit of modularity, but integrating the design of many different types of boards can easily increase the total cost of the project and add to the schedule delay. The muon tracker team should clearly spell out the considerations which led to the present design, so that they can be compared to alternatives in which there a fewer individual boards with more functions. The choice of EURO card modules may itself place an unnecessary burden on the design, since it constrains the components to one side of the PC board.

The front end integration appears rather risky, with unamplified signals traveling rather significant distances and through several connectors before reaching the active circuits. The signal processing requires a demanding 1% charge measurement and any risk of excess noise should be eliminated to the extent possible. At the very least, this signal path must be prototyped as soon as possible under realistic noise conditions in order to assess the extent of the problem.

It is necessary to rapidly and assuredly overcome these two major concerns before it becomes necessary to begin designs. First, the number and size of boards should be evaluated, and second, the signal path of unamplified signals should be examined.

In the meantime, the work on understanding of AMU/ADC, AMU address manager and data formatter should continue, given that most of design blocks have originated at the ORNL I&C division. The PHENIX management team should have close contact with LANL FEE team. This will help to integrate the FEE team into the PHENIX online system and short circuit any communication barriers.

The muon tracking FEE system engineering group is quite inexperienced with the construction of large, low noise wire chamber electronics. They will need to collaborate closely with the mechanical design group to make certain that issues of grounding and shielding are properly and uniformly addressed. It would be very desirable to augment the engineering group with a physicist or engineer who has experience with low noise wire chamber electronics.

Difficulties with the funding profile were mentioned several times as reasons to defer actions to FY1999. Although this makes sense for major procurements, every effort should be made by the PHENIX management to identify resources so that design, development, testing tasks are not delayed at this critical phase of the project because of the lack of (rather modest) needed resources. Completion of the development stage, for most of the muon tracker FEE is expected towards the end of the calendar year. The committee strongly recommends that, at that time, another independent review focused on the readout technical progress, cost, and schedule should take place in order to more accurately assess the project.

Both electronic and mechanical groups should be congratulated for their progress. Production is taking place in the factory, and the design and prototyping of the electronics is far along. Both have large teams which are well managed.

The mechanical design of the muon identifier is obviously credible and buildable, since 3 of the 60 panels have been installed in the detector at the time of the review. The collaboration is clearly focused on the task, and has made great strides in the last few months in going from design to fabrication.

The schedule is extremely tight, however. The present rate of chamber production would result in all chambers installed in the magnet steel only a few days before the shield wall construction begins, so any unanticipated problems will result in either an incomplete muon identifier system or delay in the construction of the shield wall. Every effort must be made to anticipate problems in advance, first, so that holidays, vacation schedules, and availability of materials is taken into account well in advance. There seems to be a good system for that in place in the factory, but it is very important that John Tradeski and the other senior BNL technicians are part of the planning and decision making.

Even with all the planning in the world, however, the chambers are simply not ready soon enough. Dr. Ozaki rightly insists that some schedule contingency be built into the construction and installation schedule, and the only way to do this would appear to be to reduce the amount of time it takes to construct one chamber. The only possibility that makes sense to expedite the construction of the chambers is to make sure there are no bottlenecks to production, which means crane training for weekend workers, and overtime for technicians and union workers, particularly a welder and sheet metal worker, who may be needed for critical tasks.

A serious concern about trying to speed up such a complicated process is that it will compromise either the quality of the chambers, which are undergoing much less testing before installation than may be wise or customary, or the safety of the construction crew, many of whom are relatively inexperienced physicists and students. Obviously, PHENIX is better served by safely building good chambers than by slipshod work on cheesy chambers.

In general, the design of the electronics looks credible, and the team assembled to build it capable. There are some open issues about the location of the electronics racks and the gas system, and the identity of the designer of the gas system, which should be resolved.

It was pointed out that sixteen Readout Cards (ROC's) are sufficient to read out two full gaps on one end; this fits in less than one full crate. It would seem to be very desirable to produce enough modules for this for the PHENIX Engineering Run (January-July, 1999) despite any pressure to spend money on muon tracker electronics at the same time, since instrumenting even two gaps of muid electronics gives PHENIX some capability for measuring beam related halo, and gives the muid electronics a very realistic shakedown run.