STAR NEWSLETTER NUMBER 45

1. From the deputy spokesman:

Reported by Tim Hallman

One of the main STAR events this past month was the STAR Collaboration Meeting, and it was truly exciting to see so many STAR Collaborators --old and new-- here at BNL. Several people commented that they felt this meeting had a somewhat different atmosphere and intensity, probably related to the fact that the start of data taking is now barely 2 years away. We still have a great deal of work to do, particularly in the area of software, and will need to rely heavily on effort within the Collaboration if we are to be successful.

There are two related developments which occurred at the STAR meeting. The first is that Torre Wenaus, the new Software and Computing Project Leader for STAR, requested that each collaborating institution designate a software representative who can comment both on the needs and the available manpower for their institution. Torre is preparing a list of critical software tasks that must be done, and in the very new future both he and STAR management will be attempting to match manpower and tasks within the Collaboration.

The second point regarding software is that, as part of beginning a final push towards readiness for day 1 physics, there will be a comprehensive review of the STAR software effort in the early fall. The review will cover aspects of online and offline software, as well as software infrastructure, and will be aimed at determining what we need to accomplish in the next two years to be ready for the start of data taking, and to realize the STAR scientific program in the long term.

Two other events of interest occurred in the STAR Council meeting. The first is that the two-photon and peripheral physics program proposed by Spencer Klein was officially endorsed as part of the STAR scientific program. Spencer is now reviewing with the DAQ and trigger groups what, if any, modifications may be necessary to optimize the performance of STAR to pursue the peripheral physics program. Simulations to further document what needs to be done to effectively extract these scientific results from STAR data will continue.

The Council also was very pleased to see the tremendous progress made by the French group from Nantes in documenting the impact and potential scientific benefit of adding a fourth layer of silicon strip detectors to the STAR inner tracking system. The Council is hopeful the Nantes group will continue to pursue this proposal vigorously to respond to the remaining questions and concerns that have been identified, leading to a successful conclusion in the near future.

Finally, this past week (July 29-31) there was a review of the proposed RHIC computing plan conducted by the Department of Energy. The review was very constructive, and the Committee made a number of important recommendations. As a general comment, the Committee felt that the overall RHIC Computing Strategy is fundamentally sound. The detailed recommendations will be included in a forthcoming email once they are official. However, there were two points made during the review that directly impact STAR and which deserve mention.

The first is that the Committee was very skeptical that the correct level of effort and expertise that is expected to be contributed from the RHIC experiments could in fact be realized. In the case of STAR this is 3-4 FTE's within the current RHIC Computing Facility Model. A comment made by the DOE in response to this concern, is that if this model is not workable, then given the priority that must be given to computing at RHIC it may be necessary to tax the existing relativistic heavy ion program in Nuclear Physics to find the resources to support additional hires by the RHIC Computing Facility. One conclusion is that it is clearly within STAR's interest to work aggressively with Bruce Gibbard to explore whether or not this model is workable, and whether the issue of contributed RCF manpower can be solved.

The second point was that the experimental requirements for RHIC computing resources need more study, and that a joint task force including membership from RCF and the experiments should be formed immediately to study and document the computing requirements for RHIC. As part of this effort, RCF personnel should indicate when a given requirement drives the system architecture forcing irreversible decisions to be made, and the experimental personnel should indicate the level of uncertainty in the requirement estimates, as well as the cost to the physics program where requirements connot be met. As a consequence, it will be a matter of some urgency within STAR to review our estimated requirements in consultation with RCF personnel as well as representatives of the other experiments to try to optimize the functionality and architecture of the RHIC Computing plan.

Excerpted from the STAR Monthly Report for June 1997.

Project Management Summary and Highlights

Jim Thomas will be moving to BNL in July. This represents the first of many moves to take place in the near future. Torre Wenaus has come on board STAR along with an on-line person. Other on-line software candidates are being interviewed.

The Project Plan was updated this month. All the latest project information has been included and the contingency picture looks healthy.

TPC Summary and Highlights

Magnet Summary and Highlights

During the installation of the main coils into the magnet core, it was noticed that uncured epoxy was leaking down the ID of coil sn 04 from at least three locations. The coil vendor was contacted and they stated that cosmetic repairs had been made to the surfaces of coils, in particular ID, where there were voids under the outer ground wrap. They gave assurance that the coil was sound and that the problem stemmed from human error in forgetting to add both parts of the epoxy. It has been decided that the coil is of sound structure since it has passed all electrical tests as well as the stress of installation rigging, and that it should be kept installed in the magnet. This is a cosmetic issue only. The resin will be allowed to continue to weep out, and at a later date the uncured resin will be cleaned up and the weep holes sealed with a room temperature curing epoxy.

Electronics Summary and Highlights

Systems Test

With the cooling installed and running, the FEE pedestals appear stable over periods of days. However, when the FEE is left untriggered for many minutes, the pedestals in the first event are shifted by several counts. The readout board reset has been changed to generate a dummy event for the SCA to reset the pedestals and eliminate the problem.

For slanting tracks, the observed crosstalk is worse than expected. This crosstalk does not appear to affect the resolution; it may affect the data volume.

The observed resolution for long tracks (those covering all 32 outer pad rows) is worse then predicted. When these tracks are divided into shorter segments, the resolution improves. This problem is likely due to the poor electrostatic character of our makeshift field cage, and should disappear with the full TPC.

A few more data mapping problems were fixed in June. Tests will continue even though the bugs appear to be all gone.

Data runs occasionally crash. Because these crashes are rare, less than once a day, this does not appear to present a problem as long as the final DAQ software is robust enough to recognize and reject these bad events.

We plan to move the system test electronics up to Building 77 in early July. There it will be installed in the TPC for the cosmic ray test.

Front End Electronics (FEE)

Two of the fiber-optic transmitters have failed. The manufacturer has repaired these units, and improved their testing procedure. In addition, they plan to switch to VCSEL (vertical cavity surface emitting lasers), which they claim are ten time more reliable than the current diodes.

The cabling has been slightly rearranged on the inner sector to reduce the length of cable between the master readout boards and their extensions. This was done to improve the reliability of the data transfers from the extensions.

Production of power supply chassis continues at LBNL. The power cables have been ordered.

FEE will have a final design review on July 2nd.

Data Acquisition (DAQ)

Debugging of the Prototype Mezzanine Board (SUNNY) is complete. Design of the next version, including ASICs is underway. A full speed, full-scale version of the receiver card data distribution system has been built and debugged. This version utilizes a new HP GLink chip which requires only +5V, removing a major contributor to power dissipated on the receiver card.

An extensive evaluation of SCI as a candidate for the DAQ inter-crate network has been started using the Dolphin PCI-SCI bridge cards. While some features do not seem to work as advertised, the card is still a viable candidate Software design has started at the implementation level. Some prototype code will be exhibited at the DAQ working group meeting.

A first test of the MVME2604 as a SCIS master was carried out. Taping to a slow tape drive (Exabyte) has been demonstrated.

Trigger

Slow Controls

The LeCroy and VME hardware and EPICS has been installed at LBNL for use in the TPC Cosmic Ray test. Responses to the initialization/configuration states of this subsystem has been added to the State Manager.

Tests over the HDLC link have continued in the system test room for the debugging of the TPC readout boards. The HDLC EPICS software has been made available to the Brahms experiment at RHIC.

Computing Summary and Highlights

Simulations

The simulations effort to understand the effects that the beam-pipe design has on the SVT, TPC, and FTPC performance is nearly completed. The results are presently being collected and written up.

The GSTAR and g2t binaries, libraries, and includes were released under SL96b.

Analysis Software

The TPC padmonitor was upgraded. Several problems were corrected, and multiple cursor and histogram features were added. The offline FEE location map was updated to reflect the most recent location of the TPC system test electronics for the inner sector.

PKG tfc was upgraded to include the extraction of the gain corrections used online for use in offline gain corrections; clarification between straight pedestal subtraction and threshold selection for zero suppression; and 8-to-10 bit data "uncompression" option.

The TPC tracking evaluation package, tte, was modified to include dE/dx evaluation and installed under irix and solaris. An SVT calibration package, sal, was created and a module was installed for doing SVT wafer alignment corrections.

The global tracking evaluation module was converted to the GSTAR/g2t/STAF environment. The reconstruction package and module, exi, was converted to the GSTAR/g2t/STAF environment.

Work was started on the offline production prototype code under the offline analysis branch /ofl in package /ds1. A special directory branch was created under AFS to organize calibration and configuration files for use by this prototype production code.

Software Infrastructure

On-line software

Greetings from Long Island. The weather here on Long Island was fairly typical for July. Many warm days with highs in the mid eighty degree F range and lows in the upper sixty to lower seventy degree F range. Scattered over the month we also had a few periods with hot days and high relative humidity. The weather didn't effect me greatly in July however as I was in meetings for essentially the entire month! In July we had RHIC 97, which consisted of a weeklong summer school and a half week symposium on recent Heavy Ion Physics results, the STAR collaboration meeting, a STAR review of the Endcap Electromagnetic calorimeter design, and a DOE review on the RHIC Computing Facility and RHIC Offline Computing Plans.

The STAR collaboration meeting was very well attended. For some of the Pleanary sessions essentially every seat in the Large Seminar room in the BNL Physics building was filled. Something new that I noticed about this meeting as compared to past collaboration meetings was the increase in the number of graduate students starting to become involved in STAR.

Most of the work out at the Assembly building this month consisted of aligning the positions of the magnet coils (along the axis of the solenoid) and preparing the fifteen upper backleg return steel bars for installation. In August, once the magnet coil leads are clamped into place, the remaining magnet backlegs will be installed.

Konstantin Olchansky joined the STAR group at BNL to work on online software.

7. New STAR NOTES since the last Newsletter