Jefferson Lab > Physics
Privacy and Security Notice

All Staff Memos

Date: May 10, 2001
To: Distribution
From: Larry Cardman and Andrew Hutton for the Nuclear Physics
Experiment Scheduling Committee
Subject: Accelerator Schedule: May 2001 – June 2002


Attached is the accelerator operations schedule for the period through June 2002. The overall operations continue to be constrained to a level of approximately 30 weeks/year as a result of the laboratory’s operating budget level for FY2002.

The Jefferson Lab Nuclear Physics Experiment Scheduling Committee developed the schedule. Committee members are: Larry Cardman and Andrew Hutton (Co-Chairmen), Hari Areti, Roger Carlini, Bernhard Mecking, Kees de Jager, Mike Seeley, Charlie Sinclair, Dennis Skopik and Will Oren. Franz Gross provided advice. As has been the norm, a number of meetings of this committee were necessary to resolve conflicting requirements and to ensure that sufficient resources would be available at the laboratory to properly stage and carry out each of the experiments. The schedule was derived by looking at the requests for major installation work in the experimental halls, evaluating the number and kinds of people needed, and then scheduling to minimize overlap. The schedule request forms were useful in identifying the detailed requirements of each experiment. Information on other laboratory engineering priorities was included to ensure that the required preparatory work could be completed in time. This provided a rough overview of when each hall would be down.

Each hall leader took the requests for running time submitted by the experiment spokespersons and prioritized them based on the PAC recommendations and other considerations as outlined in the scheduling committee charter. Scheduled time for Hall C was again calculated using an estimated overall efficiency of simultaneous hall and accelerator operation of 50%; this value is consistent with last year's experience for the hall. The same 50% was used for Hall B scheduling, even though the actual availability was higher; since most running in Hall B involves fractions of very long run groups, cutting the scheduling as tight as possible has no long term benefit, and we take advantage of higher running efficiency (when achieved) to complete a larger fraction of a run group's program. Scheduled time for Hall A was calculated using an estimated overall efficiency of 60%, consistent with recent experience. In a number of cases the scheduled beamtime has been adjusted to reflect significant changes in facility capabilities since the time of PAC approval of the experiment; the most obvious of these is the availability of high polarization beams with significantly higher current than was the case a few years ago. The final schedule was then reached by a series of compromises in running order within each experiment and between halls to work around incompatibilities.

To optimize polarized beam running, we have scheduled many weeks of operation at "unusual" energies that are consistent with good polarization in multiple halls. The details vary from run period to run period and hall by hall. In the worst case(5 days in July), the effective polarization delivered to a hall will be reduced to about 82% of the nominal maximum available from the cathode. This reduction is due to the angle at which the polarization vector will be set relative to the beam direction in the hall in a compromise that will optimize delivery to all halls. Details on the constraints associated with polarized beam operation are discussed in a note at the end of this memo.

The standard section at the end of this memo on "the meaning of priority on the accelerator schedule" is included for reference; there were no changes in the policy this cycle. All users with running experiments should read it carefully.

The schedule attached represents our best effort to optimize the physics output of the laboratory consistent with our resource constraints and the technical evolution of the accelerator and the experimental equipment. In the material that follows, we outline the technical considerations that influenced the scheduling, and outline the planned program.


Tests of 6 GeV operation were carried out in August 2000. While the accelerator briefly reached 6 GeV, 107 μA, the RF trip rate was extremely high and would not have been acceptable for physics running. In addition, there were many klystron failures (one per day of the test) due to running them at a higher voltage to obtain more power. A series of action items have been established, and improvements are being carried out in the May downtime. Another 6 GeV test has been scheduled in September 2001. Until then, the accelerator will be operated for physics at up to 5.7 GeV, a compromise between reaching new physics and achieving acceptable availability.

The polarized injector now has two fully operational, horizontally mounted polarized guns. All beam operation, polarized or unpolarized, is now conducted with high polarization cathodes. When polarized beam is not required, shorter wavelength lasers are used to take advantage of the higher quantum efficiency at these wavelengths. This has been very successfully demonstrated during the current running period: "unpolarized" beam up to 130 mA has been delivered to Hall A while high polarization beam was delivered to Hall B at 5 nA. Typical values for the "feed-through" from Hall A to Hall B are about 50 pA. The photocathode lifetime in the new horizontal guns is excellent. The current value is over 35,000 mA-hours. Though this long lifetime makes absolute statements difficult, our current experience is that the cathode deterioration can be completely removed by a simple heat treatment and reactivation. This cathode recovery can be accomplished during a normal maintenance period. This implies that a single cathode could be used essentially without limit. During the past year, over 280,000 mA-hours were delivered from the polarized guns. For the first time, measured polarizations of 80% were measured and it is hoped to maintain this level.

During January, a test run was conducted with a new Ti:sapphire laser, which delivered ~400 mA of high polarization beam. The final version of this laser will allow this current to be more than doubled. This laser was installed during the August shutdown giving us the ability to deliver high polarization beam at full current to one hall. A second laser of this type will be prepared, to allow high polarization operation to both Halls A and C. Based on the very successful operation of the new guns, and the installation of this new laser, we have removed the capability of delivering beam from the thermionic gun.

We expect to be operating for at least part of the upcoming period with a single laser rather than three. This will minimize polarization dilution of Halls with low current (all Halls will have their polarization vector aligned so bleedthrough will not affect the polarization). We also expect to be able to maintain the beams to the different Halls on strictly identical trajectories, reducing beam trips due to particle loss (a common problem with three lasers caused by drifting apart of the three laser spots on the cathode). The down side is that there will no longer be independent tuning of the three beams and if one hall requires the beam to be ramped following an RF trip, then all halls may be similarly delayed. Additionally, access to a hall effects beam delivery to the other running halls. Users with runs during single laser operation should review the relevant section of the memo on "the meaning of priority" later in this memo. These operating modes will be closely studied to optimize the physics results for all Halls.

The three-week maintenance cycle that was started last year was a considerable improvement and we tried extending the time between full (12 shift) maintenance periods even further – to roughly once per month. This was sufficiently successful that we will continue this cycle. During the 4-week cycle periods the (now monthly) maintenance/ development cycle remains the same, but the weekly, non-invasive maintenance periods will be extended from 4 to 12 hours every Tuesday (7am - 7pm). This is required to maintain the complement of active RF cavities at a high level to reduce the unwanted RF trips. The details are provided below in the section on Maintenance/ Development for reference.

Hall A

Since the last schedule was released, Hall A completed successfully four experiments: E97-111, E99-007, E94-104 and E98-108. E97-111 (Templon/Mitchell), was aimed at a systematic probe of short-range correlations via 4He(e,e'p)3He. E99-007 (Brash/Jones/Perdrisat/Punjabi), extended the GpE / GpM ratio measurement via polarization transfer to Q2 = 5.6 (GeV/c)2 using a considerable part of the focal-plane lead-glass calorimeter as an electron calorimeter on the floor. After completing these two experiments, two aerogel Cerenkov detectors were installed in the focal plane and the lead-glass calorimeter reassembled. Experiments E94-104 (Gao/Holt) and E98-108 (Baker/Chang/Frullani/Iodice/Markowitz) then started in Jnauary of 2001. E94-104 will study the fundamental process photo pion production in 2H and 4He. E98-108 will perform kaon electro-production measurements including an L/T separation up to Q2 = 3 (GeV/c)2.

After completion of E98-108, Hall A will use the April/May shut down to remove the cryogenic target system and install the polarized 3He target system required by experiments E99-117 (Chen/Meziani/Souder) and E97-103 (Averett/Korsch). E99-117 will carry out precise measurements of the neutron asymmetry An1 at large Bjørken x while experiment E97-103 will search for higher twist effects in the neutron spin structure function gn2 (x,Q2).

The tentative experimental schedule for the later half of 2001 had to be readjusted to allow for a delay in the delivery of the septum magnets. In September of 2001 the waterfall target will be installed in order to run E00-102 (Saha/Bertozzi/Weinstein/Fissum), testing the limits of the single-particle model in 16O(e,e'p). The septum magnets are then expected to be ready for installation in late November. Starting in February E94-107 (Frullani/Garibaldi/LeRose/Markowitz/Saito) will then perform a high resolution study of 1p-shell hypernuclear spectroscopy. After installation of the polarized 3He target in May, that experiment will be followed by E97-110 (Garibaldi/Chen/Cates), the low-Q2 (forward angle) extension of E94-010, measuring the Q2 evolution of the GDH sum rule. This will permit us to spread the "overhead" of the septa installation over two experiments.

Hall B

Since the last schedule release, Hall B completed the second eg1 run (polarized electrons on polarized hydrogen and deuterium targets, three experiments).

After the accelerator shut down in April/May the g8 run group (linearly polarized photons on hydrogen target) will take data for three experiments to study vector meson production in and above the resonance region. This will be followed by the completion of the g6 run (photon beam on hydrogen, mainly meson production) and a short test run for the Primex experiment. Following the September down time, the e1-6 run (polarized electrons on hydrogen target to study high-Q2 N* excitations and deep virtual vector meson production, 4 experiments) will take data at 5.7 GeV.

The tentative part of the schedule in the first half of 2002 shows a fourth e1 data run (polarized electrons on hydrogen target at 4 and 5 GeV, 13 experiments) followed by e6 (6 GeV polarized electrons on deuterium target, 2 experiments) to study tagged deep-inelastic scattering. The next run to be scheduled is e2 (polarized electrons on a variety of nuclear targets).

Hall C

Since the last schedule was released, Hall C running has included the completion of two major installation experiments E89-009 (Tang/Hungerford), an investigation of the feasibility of performing hypernuclear physics experiments in which a nucleon in the nucleus is replaced by its strange counterpart, the Lambda hyperon and E93-038 (Madey, Kowalski), a measurement of the neutron electric form factor. A third standard configuration experiment E99-118 (Keppel/Bruell/Dunne), a measurement of the nuclear dependence of longitudinal to transverse cross section ratio at low Q2 has also been comducted.

During the spring/summer of 2001 the Hall will be reconfigured for the execution of the remaining portion of E93-026 (Day/Mitchell), the polarized target measurement of the neutron electric form factor. The G0 experiment (E00-006) will have a dedicated one month installation period starting in September after which E93-006 will resume running for the remainder of calendar year 2001. Dedicated G0 installation begins in Feb. 2002 after the removal of E93-006, with an engineering run scheduled for June 2002. This installation date is subject to the conditions outlined in footnote 10 of the schedule. After the first G0 run a significant period of time will be dedicated to the running of shorter, standard configuration experiments in Hall C.

Information about the Schedule

The accompanying revised schedule is fixed through December 2001, and tentative for the following six months. Because of the complex couplings between the hall operations during polarized beam running, all halls must continue to run in "calendar-driven" mode. The firm schedule for the first half of 2002 (and the tentative schedule for the second half of 2002) will be released in October, following the meetings of the next cycle of the scheduling committee.

Footnotes to the Schedule

We summarize here the detailed footnotes to the schedule. They appear in the rightmost column of the schedule listing, and are listed at the earliest date in the schedule when they are applicable; many extend for a considerable time after they first appear. The first five footnotes apply to the entire schedule. All of the footnotes are repeated here for clarity and information.

  1. When two or three halls are scheduled, the relative priority listed in the schedule (in the order listed from left to right) is the relative priority of the halls. For example, A/B/C means that Hall A is the highest priority, Hall B has second priority, and Hall C has the lowest priority. If one of the halls has an asterisk, it means that its priority is conditional, and the conditions are given in appropriate footnotes at the beginning of the running of the affected experiment. If the conditions are not met, then the remaining two halls will have priority in the order listed.
  2. Energies listed in the schedule for the halls receiving polarized beam are the actual, delivered energies; they include the energy of the injector.
  3. In the "Accelerator" column, a "low" under Pol(arization) indicates planned use of either a bulk or a thin (but unstrained) GaAs cathode, implying that medium (~40%) polarization can be expected. A "high" under Pol(arization) indicates planned use of a strained GaAs cathode, implying high (~75%) polarization can be anticipated.
  4. When polarized beam is delivered to all three halls, it is not, in general, possible to provide pure longitudinal polarization to all users. We have optimized the beam energies to provide the highest longitudinal polarization (generally over 90%) to all halls during extended periods of scheduled two- and three-hall operation with polarization. For two-hall operation we have occasionally used less than ideal linac energy settings when one or more of the halls has a scheduled pass change in order to optimize polarization delivery over the entire run. This avoids the loss of beamtime associated with a linac energy change, and it avoids energy shifts in the hall that has no interest in changing energy at the time of the transition in the other hall. See the note in the "polarization" subsection of the text on the meaning of priority in the schedule; the note is attached below.
  5. In all cases, the orientation of the polarization at the injector will be optimized by setting the Wien angle to a value that minimizes the differences between the hall polarizations (by minimizing the dispersion) so long as this scheme does not result in a reduction of the "sum of squares" figure of merit by more than 2% compared to the optimum figure of merit as determined by summing the squares of the polarization provided to all halls scheduled to receive polarized beam. If minimizing the dispersion results in a loss of more than 2% relative to the optimum figure of merit, we will revert to our earlier algorithm of setting the Wien filter to maximize the overall figure of merit. In all cases involving polarized beam delivery the setting of the Wien Filter shall be fixed throughout the running period unless all parties scheduled to receive polarized beam agree to a different setting.
  6. When polarized beam is provided at a new energy, as much time as necessary during the first shift of polarized beam operation will be used to verify polarization in the halls. This can be done by direct polarimetry in the hall(s) and/or by taking data on a reaction that is adequately sensitive to the beam polarization. By the end of the first shift of production running with polarized beam, the run coordinator(s) for any experiment(s) receiving polarized beam must report to the Program Deputy that they have measured the beam polarization and determined it to be acceptable. Otherwise, a measurement of the beam polarization will be scheduled immediately. When the polarized beam energy is being changed in only one hall (e.g. a "pass change") then that hall should measure beam polarization by the end of the first shift of production running. Further, if the change in settings of the Wien filter are substantial, all three halls should measure and report beam polarization by the end of the first shift of production running with the new setup.
  7. A second major set of tests of the capability of the various accelerator system components to support a final energy of 6.0 GeV energy has now been scheduled for September, 2001.
  8. The "switchover" between teardown of E93-038 and the installation of E93-026 is nominally listed as May 9. Of course, the two activities will be coordinated throughout the transition period between the running of the two experiments.
  9. The Hall A septa installation is now scheduled to begin on November 19, pending achievement of the milestones given in the final readiness review.
  10. The tentative schedule is not complete for the e1 and e6 runs. Additional time for these experiments will be considered once the schedules for major installations in Hall A(septum magnets) and Hall C(GO) are firm.
  11. The February G0 installation start will be reevaluated at the September 2001 scheduling meeting taking into account information from the annual DOE/NSF progress review of the project (June 2001), the second Hall C readiness review (late August 2001), ), and the meeting of technical milestones agreed upon by the laboratory and the G0 collaboration

Additional General Information on Operations and Scheduling Constraints

The accompanying schedule is fixed for the nine-month period October 2000 thru June 2001 and tentative for the following six months. Priorities have been assigned as "firm" for the period of the schedule that is fixed; the tentative priorities set for the January-June 2001 period will be reviewed in September, when the schedule for that period becomes fixed. As noted earlier in this memo, the operation of polarized beams in more than one hall puts severe constraints on our ability to change beam energies.

The Meaning of Priority on the Accelerator Schedule

Generally, the assignment of priority to a hall means that the identified hall will have the primary voice in decisions on beam quality and/or changes in operating conditions. We will do our best to deliver the beam conditions identified in the schedule for the priority hall. It will not, however, mean that the priority hall can demand changes in beam energy that would affect planned running in the other halls without the consent of the other halls. Of course, final authority for decisions about unplanned changes in machine operation will rest with the laboratory management.

The operation of more than one hall at Jefferson Lab substantively complicates the interaction between the experimenters and the accelerator operations group. It is in the interests of the entire physics community that the laboratory be as productive as possible. Therefore, we require that the run coordinators for all operating halls do their best to respond flexibly to the needs of experiments running in other halls. The run coordinators for all experiments either receiving beam or scheduled to receive beam that day should meet with the Program Deputy at 7:45 AM in the MCC on weekdays, 8:30 AM on weekends.

To provide some guidance and order to the process of resolving the differing requirements of the running halls, we have assigned a "priority hall" for each day beam delivery has been scheduled. We outline here the meaning of priority and its effect on accelerator operations.

The priority hall has the right to:

When the priority hall has requested a re-tune, if the re-tune degrades a previously acceptable beam for one of the other, lower priority running halls, then the re-tune shall continue until the beam is acceptable to both the priority hall and the other running halls that had acceptable beam at the time the re-tune began.

Non-priority halls can:

The ability of non-priority halls to request retunes and accesses shall be limited by a sum rule - the total time lost to the priority hall due to such requests shall not exceed 2.5 hours in any 24-hour period. (To facilitate more extended tuning associated with complex beam delivery, with the agreement of the run coordinators for all operating halls, the sum rule may be applied over a period as long as three days, so long as the average impact is less than 2.5 hours/day.) In the event that two non-priority halls are running, the 2.5 hours shall be split evenly between them in the absence of mutual agreement on a different split.

During operations in which a single, 1500 MHz laser is being used to drive the electron source for all 3 halls, when a non-priority hall needs changes to the accelerator state (re-tuning, access, etc.), then all halls currently receiving beam need to agree on the timing of the change, and the shift leader for the priority hall should contact the crew chief to make the formal request. The upgrades to the PSS and MPS system, together with the development of the three-laser drive system eliminate the need for this constraint during 3-laser operation of the source. However, it is necessary to reinstate this constraint whenever, for reasons of source performance in service to the running experiments, a single drive laser is used. It is also necessary to reinstate the constraint on a temporary basis in situations such as a laser failure in which we are forced to operate the polarized source in a non-standard manner.)

All Halls:

Initial Tune-up of New Beams:


Finally, any change in the accelerator schedule that has implications for running beyond one week and/or is not agreed to by the run coordinators for all affected experiments and the accelerator program deputy must be discussed and confirmed at meetings to be held (as required) each Tuesday and Friday afternoon at 4:00 in the office of the AD for Physics.


The four-week maintenance cycle that was started last year has proven to be a considerable improvement. In the present running cycle we will continue to extend the time between full (12 shift) maintenance periods even further – to roughly once per month. During the 4-week cycle periods the maintenance/development cycle remains the same, but the weekly, non-invasive maintenance periods will be extended from 4 to 12 hours every Tuesday (7am - 7pm). This cycle offers improved opportunity for accelerator and injector related maintenance, and should result in even higher availability for physics. The details are provided here for reference.

4-week Maintenance/Development Cycle

The new, 4-week cycle of the schedule includes maintenance/development periods of 11 shifts every four weeks. As can be seen below, the 11-shift cycles have the same structure as the maintenance/development periods of the previous 3-week cycle. However, the weekly, non-invasive maintenance periods will be extended from 4 to 12 hours every Tuesday (7am - 7pm). This cycle offers improved opportunity for accelerator and injector related maintenance, and should result in even higher availability for physics. The details are provided here for reference

Machine Maintenance/Development Schedule – 3 Week Cycle

Weeks with long maintenance periods

Owl Physics Injector Hot Checkout by Ops (no beam) Restore Accelerator

(Halls locked)
Start Physics
Day MD
(Halls Locked)

06:00 Rad Survey


Software priority 12:00 - 20:00

Injector check-out 18:00-20:00

Lock-up when work complete
06:00 Open

Maintenance/Final Repairs
Restore Halls
(Halls locked)

MD on Hall lines
12:00 Diagnosis 12:00 Restore Injector


23:00 Lock Halls
Swing Injector Restore Halls
(Halls locked)
Legend: Physics Running; Halls Locked for Accel. Use

Weeks with no long maintenance period

Owl Physics





Day Physics RF Recovery, Injector work, software tests, MD

20:00 Start Physics
Physics Physics Physics
Swing Physics 20:00 Physics Physics Physics Physics


  1. A major block of time for Maintenance and Machine Development will be scheduled for 11 shifts roughly every 4 weeks.
  2. The first four hours will be spent doing MD that may require beam to the halls. At least one hall will be available for such activities, and more halls will be made available by prior agreement as necessitated by the program of activities planned.
  3. The halls will be open for experimental equipment maintenance from 12:00 on Monday through 23:00 on Wednesday.
  4. A designated period of time is provided for pre-maintenance diagnosis with beam (AES Group, Beam Physics & Instrumentation, and others) so that the maintenance time can be utilized effectively.
  5. Monday Swing and Tuesday Owl will be assigned to the Injector for the foreseeable future.
  6. There are two shifts for maintenance on Tuesday. It is anticipated that mechanical installations will start early and be finished around 6 PM. To avoid interference, the Software Group will have priority from 12:00 – 20:00, and the AES Group will have priority before 12:00 and after 20:00. Injector checkout should occur from 18:00 – 20:00.
  7. The Wednesday Owl, "Hot Checkout by Ops," means that:
    1. System Owners need to write Hot Checkout procedures to be used by Ops.
    2. System Owners will be required to be physically present during Hot Checkout until procedures are provided to Ops.
    3. These procedures will be the basis for troubleshooting and repair documentation.
    4. Systems found to be defective will usually be fixed on the Wednesday Day Shift although major problems may have to be addressed immediately.
  8. Wednesday morning is available for completing installations and for final repair of systems found to be defective during Hot Checkout.
  9. The accelerator will be restored starting 12:00 on Wednesday, continuing through Thursday Owl. Some additional Machine Development may occur during these shifts.
  10. Restoration of beam to the halls will be done on Thursday Day and Swing shifts. This will include Machine Development on the hall beamlines, which would benefit from more dedicated time.
  11. On the weeks when no major maintenance is scheduled, 12 hours will be assigned beginning on Tuesday morning. This will be used for recovering RF cavities to maintain a high active cavity inventory, and for tests of Beam Applications that would profit from frequent, short tests.


For holidays shown on the schedule (such as Thanksgiving) the beam will be shut down at ~noon on the last day shown as beam delivery (e.g. Wednesday noon before the Thursday Thanksgiving holiday). Beam operations for physics will be resumed at 8 AM on the first day after the holiday shown as beam operations, with the first shift devoted to beam restoration (i.e. it will be treated as initial tune-up of a new beam from the point of view of operations).

Energy Constraints on Multiple Hall Operations

The standard constraints for the different energies in the three halls during multiple hall operation are reiterated here for your information. The RF separators are able to extract one beam after each pass or, alternatively, to deliver beam to all three halls after five passes.

Therefore, it is always the case that:

  1. All three beams can have the same energy only on the fifth pass.
  2. No two halls can have the same energy, except on the fifth pass.
  3. Unusual beam energies in one hall will sometimes preclude multiple beam operation and impose shutdowns on the other halls, unless one or more of the other halls can also use a commensurate, unusual energy.

Polarization Constraints on Multiple-Hall Operations

There are only two beam energies (2.115 and 4.230 GeV) at which purely longitudinal spin can be delivered simultaneously to all three halls when the halls have the same energy. There are, however, many combinations of passes and linac energies at which it is possible to deliver beams with precisely longitudinal polarization to two halls simultaneously, and many combinations at which it is possible to deliver nearly longitudinal polarization to three halls. A technical note covering all combinations of 2-hall polarized beam running is available (TN 97-021). Tables of ideal energies for two-hall operation and optimal energies for three-hall operation are available at the url:

You can also determine the dependence of the polarization in all three halls on the Wien filter angle for the actual settings of the accelerator. Experimenters scheduled for periods involving multiple-hall polarized beam delivery should consider the possible impact of a transverse polarization component on their measurements, and provide the laboratory with a maximum allowable transverse component if appropriate. Because of the limitations on beam energies associated with the different combinations of linac settings and numbers of passes delivered to the different halls, we have a great deal less flexibility for changing energies in the different halls during polarized beam running. This is because there are many instances where the nominal linac energy and number of recirculations for the running halls provide reasonable polarization, but where changing the number of recirculations for one of the running halls results in nearly transverse polarization.

In an effort to optimize polarized beam running, we schedule many weeks of operation at energies that are consistent with good polarization in multiple halls. The details vary from run period to run period and hall by hall. In the worst case, the effective polarization delivered to a hall is typically reduced to no less than ~90% of the nominal maximum available from the cathode. This reduction is due to the angle at which the polarization vector will be set relative to the beam direction in the hall in a compromise that will optimize delivery to all halls. For two-hall operation we can optimize the figure of merit for both running experiments by simply setting the Wien filter to a value that results in identical longitudinal polarization components for the two halls. For three-hall operation we have previously used an algorithm that set the Wien filter to a value that maximized the overall figure of merit (the sum of the squares of the polarization provided to all halls scheduled to receive polarized beam). It has been noted that this sometimes results in situations where the delivered polarization is significantly different for the three halls. To "equalize the pain" for three-hall operation, we are adopting a refinement to this algorithm. The Wien angle for three-hall operation will now be set to minimize the differences between the hall polarizations (by minimizing the dispersion) so long as this scheme does not result in a reduction of the "sum of squares' figure of merit by more than 2% compared to the optimum figure of merit. In all cases involving polarized beam delivery the setting of the Wien Filter shall be fixed throughout the running period unless all parties scheduled to receive polarized beam agree to a different setting.