• PSB Main Dipole Magnets:
  • Measurement Setup:
    • Two 6 kA power converters: one for the measurements in the inner apertures and one for outer apertures.
    • All measurements were taken in pulsed mode with 2.5 meters long coil plus ADC acquisition.
    • 3 different side plates configurations were tested: All 4 rings equipped with the original side plates, all 4 rings equipped with the new (laminated) side plates, only the 2 outer rings equipped with the original side plates (mixed configuration).
  • The first measurement campaign aimed at determining the currents needed by the two converters to achieve the same integrated field level in both inner and outer gaps for 1.4 and 2.0 GeV and define the requirements for the Q-trims:
    • It was measured that the difference in current between inner and outer aperture to reach the same integrated field at 2.0 GeV is:
      • 188 A (3.44%) for the original side plates.
      • 74 A (1.38%) for the new side plates installed in all 4 rings.
      • 67 A (1.25%) with the mixed configuration: old side plates in the inner rings (more saturation), and new side plates in the outer rings (less saturation).
    • Similar trend was observed for the measurements at 1.4 GeV.
    • The mixed configuration provides the best performance in reducing the current to reach the same integrated field in all 4 apertures.
  • The second measurement campaign aimed at determining the equivalent magnetic length of the four apertures for the different configurations:
    • The balance of the magnetic lengths for the 4 rings is improved at 2.0 GeV with the mixed configuration, while it is slightly worsen at 1.4 GeV.
    • A slight increase of the equivalent length was observed with the mixed side plates if compared to the original ones.
  • Excitation Curves of the Center and Integrated Field (Bdl):
    • The saturation is measured to be higher in the outer apertures.
    • The saturation was measured to be higher in the integral field due to saturation effects at the end of the magnets.
    • Measurement of the dynamic excitation curves with the original plates and the mixed side plates configuration showed a clear improvement in reducing Eddy current effects with the use of the laminated plates, mostly for the outer apertures.
  • Eddy Current Effects:
    • The comparison was done between the mixed configuration and the one with the new plates installed on all 4 apertures.
    • These values are not directly comparable to those reported in 2011 due to different test conditions.
    • The new plates reduce the amplitude of the eddy currents, while the time constant is not significantly affected both at 1.4 and 2.0 GeV.
• PSB Injection Dipole Magnet:
  • Similar measurements were performed on the PSB injection dipole magnet with the default configuration and compared to results on the PSB main dipole magnets with different side plates configurations.
  • The injection dipole magnet has a smaller current difference between apertures than measured in the main dipole with the original side plates (128 A versus 188 A).
  • Integrated field difference between the two outer apertures and the two inner apertures is at the per mill level, and it is improved with the mixed configuration for the side plates. It could possible to trim the difference between the two inners and two outers with extra shims.
  • The difference in integrated field between outer and inner apertures was found to increase at 2.0 GeV with respect to the 1.4 GeV case.
  • The measurement of the excitation curves showed similar expected trend as for the main PSB dipole magnet, but more saturation effects are measured on the PSB main dipole magnet.
  • The measurement of the integrated field quality showed an homogeneity better than +/- 1 per mill on +/- 80 mm.
  • The measured Eddy current effects were comparable to the ones measured in the PSB main dipole magnet.
  • The profile of the integrated field along the longitudinal direction was compared to simulation:
    • Very good agreement was observed between data and simulation.
    • Field attenuation on the magnet side was measured with respect to the simulation because of the injection beam pipe which goes across the yoke causing more saturation in this area.

• B. Mikulec asked if it is understood why more saturation is observed in the PSB main dipole magnets with respect to the injection ones. A. Newbourough replied that the reason has not been investigated yet.

• The presentation is based on the results from the measurement campaign applied to the 2.0 GeV cycle.
• The configuration for the LIU-PSB main magnets is to be split and powered by two MPC:
  • BHZ Outer (30 x BHZ 1+4, 1 x BHZ INJ 1+4, 1 x BHZ EXT 1+4) and 32 QF all in series
  • BHZ Inner (30 x BHZ 2+3, 1 x BHZ INJ 2+3, 1 x BHZ EXT 2+3) and 16 QD all in series.
• The MPC will drive the currents in the main bending to correct field compensating the differences between BHZ Outer (1+4) and BHZ Inner (2+3) due mainly to saturation.
• Parallel trim converters act on the QF and QD magnets to achieve the correct gradients, while isolated trim converters will compensate for differences between the two BHZ INJ & EXT magnets.
• The BDL and QSTRIP trim converters compensate for differences between rings.
• The evolution of the requirement on the trims is:

(+) Courtesy of J.L. Sanchez-Alvarez who extrapolated the trims from the proposed magnetic cycle. The extrapolation does not take into account separate powering of the inner and outer gaps and saturation.
(++) Because the outer gaps have higher saturation the QF which are connected in series need higher current.
(+++) Since the INJ/EXT magnets have less saturation with respect to the main bending magnets the current goes from +18 A to -45 A to compensate for the difference.

• The introduction of the laminated side plates on the outer gaps will minimise the differences between the inner and outer rings as well as minimise the QF trim requirement.
• All these results do NOT assume any margin.
• A. Newborough contacted E. Benedetto who proposed that a 2/3% margin should be fine.
• In order to have an idea of a few percent tunability, the MPC trim has been scaled to the maximum of 6 kA provided by the MPS for the outer bending magnets. Assuming for 5% margin, the QF(QD) would need -802(-580) A. The margin would be on the MPS current and not on the TRIM current. This does not mean that either the MPS or the magnets can achieve or operate at this value. The margin has to be discussed with the EPC Group. A meeting will take place to refine the values soon.*

• More Details about the PSB INJECTION/EXTRACTION Dipole Magnets:
  • Two new magnets are to be built, one for each of the INJ & EXT regions.
  • The new magnets will include new coils with trim windings to compensate for differences already seen at 1.4 GeV which will become worse at 2 GeV.
  • The existing magnets once removed will be fitted with the new coils to act as dedicated spares.
  • The magnets are to be powered with pairs of apertures connected in series: INJ 1+4, INJ 2+3, EXT 1+4, EXT 2+3.
  • Margin:
    • A factor of 12/10 for difference between # turns on main and trim coils.
    • A margin of 250% (requested by MSC) due the difference in performance of the steel and field given by MPS. To be determined and discussed in case.
    • The following assumptions would lead to a requirement of 230 (60) A for the INJECTION/EXTRACTION outer (inner) bending magnets.
    • If the full MPS of 6 kA was applied there may or may not be enough margin. It depends on the performance of the steel, which is not yet available for testing. A sample of the old steel has been used for chemical analysis.
  • The design of the coil is well advanced and the requirement for the converters have been defined.
  • There are few constraints on the coupled voltage due to the particular variation of the ramp at injection and between 600 ms and 800 ms. These are being discussed with the EPC Group.

• S. Pittet commented that it will be not possible to reach a current of 6 kA especially trying to keep the request of a 10% margin on the RMS current requested by MSC. It is more likely that the upper limit is about 5.5 kA.
• K. Hanke commented that another hard constraint on increasing the current could come from the cooling needs to operate at high current.
• A. Newborough replied that A. Blas also presented a cycle long 900 ms which reduces the requested magnet performance almost to the current ones at 1.4 GeV.
• J. Hansen asked which type of magnets would be the one installed in case of an early Linac4 connection in 2016 as the Vacuum Group is working on the chambers and support for the coils of the new magnets. A. Newborough replied that the injection magnet would be one of the existing spare, but the installation of the new coils is foreseen for the beginning of 2018. Generally, the early Linac4 connection in 2016 is the worse case scenario for the MSC as also the magnet in Section 11, BR.BHZ11, will have to replaced using the second spare. But at that point one of the radioactive magnet will have to be used as reference magnet and this is not an optimal working scenario.

• The new PSB B-train system will be based on the PS B-train upgraded system:
  • A prototype PS system has been deployed in 2014, now measuring in parallel with the old system and giving feedback to the RF system via the WhiteRabbit communication system.
  • It was tested with both protons and ions. The initial tests are encouraging.
  • The system is calibrated to be as close as possible to old B-train and for the moment works only on few selected cycles.
  • The ongoing work include finalization of the FPGA firmware and FESA classes, sensor improvement, test and calibration in all relevant cycling conditions.
  • The goal is to switch on-line for long-term testing sometime mid 2017, such that the old system could be dismissed after LS2.
• Concerning the dB/dt of the 2.0 GeV PSB cycle, which is the main parameter for the B-train, the major constraints come from the ramp up at injection (6 T/s) and the ramp down after extraction (-9 T/s). It is important to take into account the ramp down as the long term goal is to have a continuous integration in order to abstract the B-train system from the timing of the machine which is a source of error during the operations.
• In order to have a resolution of 0.1 Gauss, the measurement have to be streamed out at 600 kHz (within the theoretical peak of the WhiteRabbit, 1 MHz).
• Layout:
  • Two independent system installed in ring 3 and 4. Two independent system to act as a spare installed in ring 1 and 2.
  • Electronic components (and spares) for all approved systems are in production (early 2016).
  • Linux Front-Ends and CTRi timing cards are available as stock components.
  • NMR components already at CERN (or available rapidly).
  • Still being defined: integral PCB fluxmeters (4-10 months to procure), sensor supports, cabling, FESA classes and OASIS diagnostic signals.
• Open issues from last LIU-PSB WG meeting (here):
  • Since the top/bottom asymmetry in the rings is measured to be less than 2E-4, hot-spare duplicated sensors can be added in the bottom rings at little additional cost.
  • The NMR frequency programming would be done via FESA classes (controlled in the CCC).
  • Three full-height racks are needed and will be placed as close as possible to the reference magnet (communicated to D. Hay and included in the plan).
  • In order to perform real time diagnostic and check if there are differences between the apertures or the spares the operational sensors, it will be possible to read back the current coming from the DCCTs. The development is ongoing.
  • The continuous integration provides a more robust system, but it is not yet implemented. The next few months will be dedicated to implement it.
  • Synthetic B-train:
    • Originally not included in the project.
    • It is needed for the RF cavity tuning w/o beam, typically on each restart.
    • It may be needed to run the beam in place of the measured B-train, e.g. in case of faults or for testing purpose.
    • The system is implemented in slightly different ways depending on the accelerator complex. The goal would be to standardize the implementation.
    • The setup is done via FESA software.
    • A proposal is being prepared which foresees a system based on an additional FMC card able to get in input current (either in real time or a preset waveform) and apply a magnetic model to derive the field.
    • A. Findlay commented that up to now the beam could not be accelerated with the synthetic B-train, but recently there have been few developments and the acceleration process would be tested again.
  • Planning:
    • Currently the planning is still rough as there are problems with the resource allocated.
    • The current resources are sufficient for the work on the PSB, PS and ELENA, but for instance not for the work on the synthetic B-train.
    • Concerning the PSB, the plan is to keep performing offline tests (as the measurements shown by R. Chritrin today) until the end of 2016. The following step would be to work on the installation and test until LS2, in order to be ready to perform tests with real beam sometime in the end of LS2. B. Mikulec asked if there is any chance to test the beam before LS2. M. Buzio replied that it is very unlikely to be able to test the system with beam before LS2, but certainly all the possible online and offline tests will be be performed before.
  • An additional concern is that some of the component in the existing B-train are going to be phased out by the end of LS2:
    • The maintenance of the front-end of the measurement system of the synthetic B-train and some accessory cards is being dismissed by the BE-CO group after LS2.
    • If the new B-train system is proven to be working by June 2017 the handover will be done to the TE/MSC Group as the old system could be dismissed.
    • There could be implications in case the old MPC has to be used if there are problems with POPS-B. Three scenarios are possible:
      • Use the existing B-train and reference magnet kept “as is” with the risk of unmaintained(-able) components. → baseline scenario
      • Use the new B-train system with the old power supply. Some details about the DCCT beta normalization are still being discussed with L. Soby and P. Odier but they are not a showstopper. The scenarios are two:
        • Connect the new reference magnet to the old power supplies, which implies pulling 6 kA cables. In this case it would be enough to take a subset of the measurements (from ring 3 instead of all 4 rings) and give them to RF and BI equipment. → Recommended by M. Buzio. B. Mikulec mentioned that it could be difficult to pull out power cables.
        • Use the existing reference magnet. This implies pulling sensor cables from the existing to the new reference magnet (may add noise) or bring the electronics racks of the new system to the existing reference magnet and adapt it.

• Planned Activities:
  • Get a complete "snapshot" of the current system in terms of available documentation, spares and details of the synthetic B-train generation software.
  • Whenever possible split up sensor output for measurements in parallel with a separate acquisition system, so that one could also use the new electronics.
  • Clean up the system by removing old or disconnected sensors inside the existing reference magnet.
  • Define the needed FESA classes and OASIS signals.
  • Keep working on offline magnetic tests in B867. Several open questions to address, for instance if using NMR or FMR as field markers, or perform a full characterization of new reference magnet.
  • Test the new B-train system with both the existing and new reference magnet and whenever available with the POPS-B.
  • Partial installation of new B-train in B361, use existing sensors and/or add new ones in the free gaps and attempt to run in parallel with existing B-train providing feedback to RF.