|
Session 6: Present Status: Measurements Versus Targets and Corrective ActionSession Chair: Francesco Ruggiero, Scientific Secretary: Massimo GiovannozziArjan
Verweij: Present Status and Trends of Cable Properties and Impact on Field
Quality Arjan Verweij presented the status and trends of cable properties. Filament magnetisation affects field quality by inducing persistent currents. According to specifications, magnetisation should be less than 30 mT (for cable 01 at B = 0.5 T and T = 1.9 K) and less than 23 mT (for cable 02 at B = 0.5 T and T = 1.9 K). Different averages between manufacturers are allowed, but the spread is fixed (±4.5 %) for all. Although specifications are not met for single strands, the cable, made by many strands, shows much better performance. Using cables from the same manufacturer for the 4 inner and 4 outer coils of each dipole makes it possible to meet specifications given in LHC Project Report 501 for both the systematic and the random part of the dipole field errors. It is stressed that if the same cable has to be used for each octant, the issue of difference in dipole integrated strength is important. Inter-strand resistance (Rc) is responsible for eddy currents between the strands, induced by a field variation. The target value for Rc is > 20 mW (inner cable) and > 40 mW (outer cable). Higher contact resistance could be tolerable, but then the magnet quench performance could become more sensitive to local strand defects or non-uniform joints. Errors induced by contact resistance are higher than target values when using a linear ramp at 10 A/s: they will be reduced by about a factor of five when using the planned exponential ramp. Therefore, the present value of Rc is adequate for reaching the target values of ramp-induced field errors. Snap-back is due to boundary-induced coupling currents as well as redistribution of transport current. The maximum decay depends of magnetization. No specifications exist for the reproducibility of b3 during snap-back (in fact, no clear physical explanation was available at the time orders were prepared). Spread among the magnets is large and is expected to continue for the rest of the magnet production, since no clear correlation with Rc is observed. It is requested that a clear decision be taken concerning the distribution of magnets and cables form the various manufacturers over the octants. Davide
Tommasini: Constraints on Sorting: Types of Magnets and Interchangeability Davide Tommasini presented the key issues that might constrain magnets sorting. The total number of dipole types is 27 (A, B type, diode polarity, interconnections). However, only four types will be present until storage. At this stage a decision on sorting should be taken, as the choice of interconnection will be practically irreversible. Magnet Evaluation Board will take the decision on magnet approval between WP08 and WP09. At this stage a flag will be defined for each magnet to distinguish among various installation constraints (this point will be further discussed by the Magnet Evaluation Board). The maximum storage capacity for green-type magnets, i.e. those that can be installed everywhere in the ring, is of about 120 dipoles. Up to 60-70 can be stored being accessible individually. There will be more freedom for tackling the installation of the first octant, thus allowing more experience to be gained for next ones. An important point is raised, i.e. in case magnets from different manufacturers have to be installed in the same octant (7-8), a decision has to be taken within two, three weeks (maximum) concerning a possible change of diode polarity. Ezio
Todesco: Status and Trends of Field Quality at 300 K
and Possible Corrective Actions Ezio Todesco presented the status and trends of field quality at warm with a proposal for possible corrective actions. Field quality steering is primarily limited by predictivity of the magnet model. Accurate comparisons between measurements and numerical simulations show an agreement for differential effects within 20 %. Another issue is the reproducibility of collared coils. Systematic effects were observed for pre-series magnets that were re-collared, maybe due to the virgin state. Additional experience is expected based on a third de-collaring of dipole 2002. It is generally agreed that field quality steering should be based on collared coils, as the delay between collared coils and cold measurements is too long (about seven months now, but it is expected to reduce to two-three months). Therefore, warm/cold correlations become a critical issue in field quality steering. Present data show that correlations allow to steering the production for all multipoles and integrated transfer function, with the exception of the a4. Presently, the situation of field quality is as follows: · Non-allowed systematics: within specifications (a4 is a bit tight). · Allowed systematics: b3, b5, b7 are out of specification but by less than one unit (this is an improvement with respect to previous cross-section). With this situation the machine could work. However, a drift in the production could be dangerous. · Randoms: b3, b5, are out of specs, but improvements are expected thanks to a more stable production. Two field quality issues are under considerations, i.e. dipole integrated strength (BdL) and odd multipoles (b3, b5, b7). The first point can be tackled using the magnetic length as steering knob. There is a difference of 20 units in BdL between Firm 3 and Firm 1-2 collared coils. Data should be confirmed by measurements at cold – only two cryomagnets have been tested (see talk by Stephane Sanfilippo). Laminations redistribution between the three firms might be the appropriate solution. The only drawback is the reduction of margin for further corrections. Different strategies could be
envisaged to tackle the second point. Polar shims could be used to
have a limited action on b3, the critical point being the
coupling between b3
and b5.
In principle it allows very fast action on magnet production. Mid-plane
insulation is very efficient, and could push b3
b5 and b7
towards the targets. Copper wedges change is still under study, but
the feedback is very slow (see presentation by Paolo Fessia). Stephane
Sanfilippo: Status of Field Quality and First Trends at 1.9 K Stephane Sanfilippo presented the status of field quality and preliminary trends at cold. Cold measurements are performed without beam screen and its effect on odd multipoles (b3, b5, b7) is deduced from simulations. Nearly all the magnets measured at cold had the first cross section. Results coming from only three magnets displaying the corrected b3 and b5 cross-section were presented. The first item concerns the dipole integrated strength. Two magnets from Firm 3 are clearly above the average of other Firms. However the following magnet measured i.e. 3009 did not confirm this trend. This subject needs further investigations. On the other hand, magnetic length seems to be really stable (apart for a couple of cases that are well understood), thus making it a good candidate for field quality steering handle. Field direction is another important quantity. Several measurement systems (including long rotating coils and single stretched wire) are being crosschecked. For some magnets the data obtained with long rotating coil give the field direction out of tolerances, while measurements performed by stretched wire system display results within the specifications. Further analysis is ongoing to reduce the uncertainty related to rotating coil results. Measurements of the multipole components at injection field reveal that b5 and b7 are outside the tolerance bands. The random component of b3 is also not acceptable, but the spread is due to non-nominal shims used for the first pre-series dipoles and to the different cross-sections measured. At collision, magnets with cross-section 1 have b3 and b5 outside tolerances. The sextupole stands also above the critical limit defined by the sextupole correction capability. The corrective action taken on the coil cross section led to an improved situation for b3 and b5, standing at the limits of the window (for the dipoles measured so far at cold). However for magnets with the second cross section, b7 was found to be outside specifications. For geometric components correlation between warm and cold measurements is rather good. However the distribution of the scatter is not a Gaussian but a multi modal one. More statistics are needed to assess the warm/cold correlation. High-field behaviour of transfer function is well described by the iron saturation and in general there is a good agreement between estimates and measurements for the multipoles. Detailed studies of the harmonics behaviour at high-field show a good agreement between model and measurements for even multipoles (b2, b4) but not for odd multipoles (b3, b5). The source seems to be coils motion under electromagnetic forces. Persistent current errors measured at injection are in line with the expected values for series-production apart from the effect observed on the main dipole that is not fully understood. Ramp rate effects on B1 and on the harmonics at 10 A/s were found to be smaller than expected thanks to the effective production control of the inter-strand resistance. The decay of the magnetization is responsible for a significant spread on low order multipoles and in particular on B1 and a1. A possible explanation could be a difference in the decay properties of the inner and outer coils. In this case a1 would stem from left/right asymmetry. Due to the relevance of this subject, further investigations are foreseen. Stephane Fartoukh discussed consequences on beam dynamics of the magnetic measurements results starting from low-order harmonics. The dipoles produced by Firm 3 show a B1 component systematically higher than the others. The impact on the closed-orbit correctors is non negligible. The best solution would be to act directly on the magnet laminations. In this case it is mentioned that measurements of the longitudinal magnetic centre at cold should be foreseen at least for a few dipoles. This represents an extra measurement: a formal request to Magnet Evaluation Board will be issued. On the other hand, partial solutions could be envisaged too, such as a careful mixing of the magnets at the level of single cell (two per each Firm with Firm 3 dipoles near the cell centre). Indeed, if few arc cells will be equipped with only Firm 3 magnets, the safety margin available for the closed-orbit correctors will be drastically reduced. Another important issue is the field direction. Presently, a discrepancy between the two measurements technique is observed: provided the single stretched wire proves to be the correct one, all the dipoles are within tolerance. It is also mentioned that large values of the local field direction could be an issue. Beta-beating and linear coupling are well under control and correctable, as, according to the running average, b2 and a2 are within tolerances. Dynamic aperture computations for field errors at injection show a mild (almost negligible) dependence on b7, b9, while the random part of b3 is the real limiting factor. However, mixing of the two cross-section types as well as other non-standard components (shims) make the present estimate of b3 random rather pessimistic. With the dipole cross-section 2, the value of the dynamic aperture is 11.2 s - instead of (12 ± 0.5) s - due to the b7 component. Finally, a detailed analysis of different installation scenarios for the pre-series magnets shows no sizeable impact on dynamic aperture. As far as dynamical effects are concerned, the large random decay of B1 and a1 might have a strong impact on closed-orbit feedback system. The average effect of b3 is still large, but less critical than expected, due to a reduction of a factor between two to three. All the harmonics specified are within tolerances at high-energy. Only the systematic b3 is too large, leaving only 0.6 units of safety margin in nominal operation conditions. Of course, the operating current could be slightly increased, but this will prevent future upgrades, such as 9 T operation. Finally, it is mentioned that all these conclusions should be confirmed by including feed-down effects induced by misalignment errors.
Minutes by Massimo Giovannozzi |
|
Last update on 07-04-2003 08:30:00
|