Choose timezone
Your profile timezone:
Sextupole Corrector Magnet Review
→
Europe/Zurich
30/6-041 (CERN)
Minutes of the Sextupole Corrector Magnet Review
CERN, 26th May 2015 room 30-6-041
https://indico.cern.ch/event/396055/
Attendees:
Review Committee: Paolo Ferracin, Juan Carlos Perez, Herve Prin, Davide Tommasini, Fernando Toral (video connection)
Organizers: Paolo Fessia, Ezio Todesco, Giovanni Volpini
1. Introduction
GV acknowledges the attendees and reminds the goal of this review meeting.
The electromagnetic design of the sextupole is now finished and the detailed mechanical design is close to its completion; following the good results of a preliminary single coil test, INFN is about to start the manufacture of the sextupole superconducting coils at LASA lab, INFN Milano, Italy.
Before starting the construction, PF and GV have agreed to ask for an independent assessment of the design. Therefore, the committee members are asked to express their comments and criticism, both during this meeting and later, if possible, in written form.
A detailed presentation of the sextupole design and manufacture status is presented by GV. An overview of the lay-out status in the IR 1,5 is given by PF. All their slides may be found at the link above.
A number of comments and suggestions is provided by the committee. Here we summarize them along with our comments, done during the meeting or afterwards.
2. 2D magnetic design
Evaluate the possibility to move the coil closer to the aperture. It helps to enhance the field in the aperture while decreasing the saturation of the iron pole. The coil would have a very limited internal support, but CIEMAT informs that in superferric magnets that they built this feature did not appear to be detrimental for the performance. The iron pole is not used as a support for the coil. The space between adjacent coils is the minimum necessary to insert the set of screws which pushes the wedges.
Comment: This is a very important suggestion. For this sextupole design it is too late but it will be verified and implemented from the next design (octupole) onwards. An improved version of the sextupole design may be considered at the end.
3. Flux return plate
In the simulations that were done two years ago at CIEMAT, it was not possible to appreciate any significant cross-talk between adjacent magnets. Simulations were performed in Roxie. It is true (slide 12) that the fringe field is reduced by the return flux plate, but as far as it was observed , the adjacent fields have a perfect superposition, and there is no cross-talk at all between adjacent magnets. Models are available for discussion if needed. In a further iteration of the design flux return plate can be removed.
Comment: The Flux Return Yoke (see also slides 61-64) has the following purposes: i) to reduce the cross-talk between magnets, ii) to reduce the attractive forces between magnets, and iii) it has shown also to reduce the harmonics. These aspects will be reviewed in the final design, to verify whether FRY can be left out.
4. Mechanical design and assembly
A general remark made by several attendees was that the presented design appears sometime excessively complex, considering that the mechanics should not be a major issue for these type of magnets.
A number of specific remarks based on the experience gained on the development of analogous magnets at CIEMAT are reported.
1) Avoid to exert any pressure longitudinally on the coil ends, and to limit the pressure on the sides of the straight section of the coils. The experience has shown that even if the contact between wedges and coils becomes weaker at cold, they are in contact during the energization, which has shown to be enough.
2) CIEMAT has used use iron threaded rods for iron stacking. In the proposed design the iron yoke length dependence on the torque was checked. This variation is strongly non-linear. When the contraction was diminishing with the pressure, no more force was applied. In XFEL magnets, CIEMAT used M8 rods with 25 N/m torque. Then, the nuts were tack welded to the rods to block them.
3) End ring (slide 34): CIEMAT comments that the pins are not necessary. Take care with the thermal differential contractions.
Comment: Some aspects of the magnet design were designed for redundancy and further safety. It has been proposed to keep all the components and to decide which are the parts that play an important roles and those that can be discarded or simplified based on the experience of the first assembly and test. The end ring has the function to provide connection i) for the eyebolts ii) for the alignment references and iii) for the magnet-to-magnet connection. It is not a fundamental component and its design or even its presence will be reconsidered once the requirements on these functions are frozen.
5. Coil manufacture
The glass fibre used for the insulation has an intrinsic high friction between wires, so winding tension effects can be difficult to control.
Impregnation tool (slide 39): CIEMAT suggests to use a mould where some pressure can be applied on the straight sections of the coil when it is closed. This would help to eliminate the excess of resin and to avoid volumes filled with pure resin. It is suggested to put the resin inlet and outlet on opposite sides of the mould.
Winding process (slide 41): the gaps between turns should be avoided, CIEMAT offered to provide suggestions in case of need. Coil dimensions (slide 46): to avoid volumes with resin in the corners, plastic filler could be used. CIEMAT can provide more info if needed. It has been suggested to provide current, field and temperature margin at 4.2 K and 1.9 K
Comment: The S2 glass was chosen to guarantee a better sticking between the wire, insulation and impregnation, since we were concerned by adhesion between polymidic insulation and the resin. Spacers at the edges to reduce the resin-rich zones will be implemented from octupole coils onwards. Current, fields etc. at 4.2 K and 1.9 K will be included in the final report.
6. Protection
The protection principle needs to be better defined. The target should be to have a self-protecting system, a scenario that has not be analysed in detail. It is reported that a solenoid wound from a wire with copper-to-noncopper of 1.6 and with a stored energy of 16 kJ performed well under self-protection, so it appears likely that the quadrupole also should be compatible with this.
Using the LASA code and compare the results with the CIEMAT code seems also a good idea.
Our comment: Quench protection computations at LASA are in progress assuming a protection scheme based on a solid state switch (crowbar) which allow the magnet discharge on a resistor, limiting the total voltage across the power supply output at 100 V. Different threshold and reaction times will be considered, starting from 10V x 100 ms, to 20 ms. Once the results for the simulations available, the cross check of key results with the CIMAT code looks interesting.
7. Conclusions
The most important conclusion is that no evident show stopper has been identified and therefore the procurement of the parts and the assembly of the components may go ahead.
Once again, the organizers wish to acknowledge the committee for their valuable contribution in terms of comments and suggestions to our project.
v 11 Nov 2015 Prepared by Giovanni Volpini.
There are minutes attached to this event.
Show them.
