RF-Shielded Bellow Assembly

Introduction: In order to take care of thermal expansion during bake-out, fabrication and alignment tolerances of vacuum chambers, Stainless Steel bellows are installed in the straight sections of INDUS-2 storage ring. But ordinary bellows give rise to RF power dissipation as the corrugation may function as RF cavities. This results in over-heating of bellows which gives rise to heavy localized outgassing load and in some cases, puncturing of bellows. Base pressure ~ 2x10-10 mbar is required in the vacuum chambers to achieve very good beam life-time. Therefore, a UHV compatible RF-shielded bellow assembly has been developed. The RF-shield inside is a flexible mechanical structure, which screens corrugations of bellows, causing the wall current flow smoothly and reduce excitation of higher order modes (HOM). End finger contacts are static contacts which bridge the annular gap between mating flanges at the joints. The photograph of a typical bellow assembly developed for Indus-2 is shown below. There are total 44 no. of such bellow assemblies in the machine.

Design: A typical 150 mm long bellow assembly is designed to absorb an axial stroke of 20 mm compression, 10 mm expansion, transverse off-set of 1mm and 15 mrad of angular misalignment. It has two main sub-assemblies: inner sliding RF-shield sub-assembly and outer bellow sub-assembly. This inner RF-shield sub-assembly preserves the race-track profile of the vacuum chamber to create a uniform beam pipe. A welded bellow sub-assembly maintains vacuum and allows for travel with lateral off-set. The main components of the RF-shield sub-assembly are contact finger, cantilever spring finger and inner tube. The contact fingers are pressed on to the inner tube from outside by the cantilever spring fingers. The electrical contact is kept at the edge of the inner tube made of SS 316L. The step at the contact point is limited to less than 1 mm from beam impedance requirement. Size of the step is driven by two features: the mechanical stability of the inner tube wall and the rounded contact surface at the tip, which ensures that the shield finger will not make a secondary contact on the inner tube. This design feature differs from conventional RF-shield design in which contact finger serves the dual purpose of shield and spring. There are two important design aspects of the finger type RF-shield: (a) contact force and (b) finger slit size. Lower contact force leads to heating and arcing at the contact points. Whereas, larger contact force gives rise to intense abrasion  which in turn results into dust trapping problem having harmful effect on beam lifetime. Contact force ~100 g/finger has been considered for the design. Sizing of finger slits requires compromise between various conflicting requirements of sliding strokes, pumping conductance and  RF impedance. The contact finger is 0.3 mm thick and has a width of 4.9 mm and gap of 0.5 mm. The thickness of the contact finger is chosen based on the requirement that it doesn’t kink (buckle) throughout the stroke. The width of contact finger should not be too narrow; otherwise the spring finger will get out of the contact finger by offsetting. The contact finger makes electrical contact on the outside wall of   austenitic stainless steel SS 316L inner tube having 1 mm thickness at the contact region. The spring finger is 0.4 mm thick with the width of 7.8 mm and gap of 1.0 mm at the root. The tip of the spring finger, where it presses the contact finger has curvature of 5 mm. Geometry of the spring finger is optimized based on the deflection required for given contact force.  All other parts of bellow assembly,  other than fingers are made of SS 316L.

Fingers were fabricated from Beryllium-Copper (C17200)–¼ Hard sheet metal of various thickness as mentioned above. Tensile strength in this temper is ~ 550 MPa. Beryllium-Copper (Be-Cu) alloy has been chosen material of construction for these fingers because it keeps elasticity up to about 200°C after the heat treatment and also has relatively good thermal and electrical conductivity (about ¼ th of copper). Various stages of their fabrication include: (a) EDM wire cutting (b) Press forming  (c) Precipitation heat treatment.

RF-shielded Bellow Assembly

Rf-shield without bellow

Structural Analysis: Finite Element Analysis was done using ANSYS code to optimize the thickness & geometry of the fingers. Buckling analysis was done for contact finger. Calculation indicates that the critical buckling force for the RF shield is higher than the applied load. Analysis assumed pinned ends & a co-efficient of friction of 1 in vacuum. To confirm that buckling is not a problem in this design, a test at the worst case extension & highest load was performed. Buckling was not observed. Cantilever beam analogy is used to calculate the pre-deflection for given contact force.

Testing & Inspection: The spring fingers went through numerous inspections during their manufacturing process to ensure that the pre-deflection was within range needed for the spring force. All forces inspected were greater than 100 g/finger and the range was from 120 g/finger to 150 g/finger.

Plating is essential for the fingers for solid lubrication and  for dust reduction Several combinations for plating of contact finger, SS inner tube and  spring finger tip area were tried for this purpose taking into consideration, the various functional requirements. Finally the combination selected was silver coating on Be-Cu contact finger, Rhodium coating on SS inner tube & Be- Cu spring finger uncoated. Static inner flange contact is subjected to Silver coating. The adhesion of the coating is verified before and  after manufacturing by vacuum firing the part at 450°C. Job is rejected if any surface blisters or swelling is observed.

All the components of bellow assembly are chemically cleaned as per standard UHV cleaning procedure before final assembly.

The bellow assembly has a very complicated structure and has a very large surface area. For the typical one with of 150 mm long assembly, the inner surface area is about 2500 cm², which is about 5 times larger than the normal beam tube with the same length. The gas desorption rate, therefore, should be sufficiently lower than that necessary for the usual beam tube. Out-gassing rate measurement and residual gas analysis have been carried out to confirm the aimed goal. Measured value of outgassing rate was 1.3x10^-12 mbar l/s/cm² after baking at 150°C for 24 hour. The residual gas components were also measured after the bake-out & the most dominant gas component was H₂, other main gases being CH₄, CO and CO_2.