From: 1/4/02 3:36 PM Subject: NTX Magnet Status 1/2/02To: SSYU ?SSYU@lbl.gov?, C_Peters ?C_Peters@lbl.gov?, DLVanecek ?DLVanecek@lbl.gov?, DNBeck ?DNBeck@lbl.gov?, Gary Ritchie ?GRitchie@lbl.gov?, EHenestroza ?EHenestroza@lbl.gov?, Tim Houck ?TLHouck@lbl.gov? 1. Coil Winding First coils were wound Dec 20, looking excellent, with a minimum of cold setting. Nominal coil length has been changed from 50.0 cm to 49.9 cm (between the centers of turns 8b and 8d, along the coil midline) to allow room for extra insulation on return leads. This will improve reliability. Coil form and core dimension remain unchanged. It will be important to make sure the coils do not come out too long; we should make a precise gauge to check the coil length carefully, with the proper spacing between the turns. 2. Magnet laminations These are scheduled to go out for competitive bid with three sheet metal vendors on the list. A draft core assembly drawing is complete and will be checked. A change in end plate material from mild to stainless steel will be considered, in order to reduce eddy currents, at the expense of a very small amount of flux return. S.S. is 4x less conductive than 1020 steel. 3. Mandrel and Coil Form A discussion was held with Ralph Hipple, Derek Shuman, Chip Hollister and Al Salazar on the mandrel. Due to concerns from shops about the limited shrinkage range of thermal methods, plus the necessity to set up and use cooling baths for insertion and removal of mandrel, we will revert back to the original split mandrel design. Dave Beck will detail a design. The trepanned cylinders will still be used and will be cold stabilized. The split will be 1/8" wide, with internal bars screwed to each side of the split. No welding will be done. One bar will have tapped holes. The other will have clearance holes at some of the tapped hole locations for retraction screws. The remaining tapped holes will be used for jacking screws. However jacking is to be avoided once the milled grooves are ready to be machined. High stresses and stress concentrations could easily result in cracking which would seriously compromise voltage holdoff. As such, the final OD of the mandrel should be the same as the nominal ID of the coil form, when the mandrel is in the unstressed condition, neither jacked outward or shrunk inward. The thickness of the mandrel in this condition will be more than enough to take out any out of roundness in the coilform; and by matching ID with OD, this will occur without introducing any overall hoop stress in the coil form. The mandrel will mate with two G-10 endplates, one with a driving stub to chuck to and the other having a center drill for a tailstock center. The G-10 is used to avoid dragging metal into the grooves from the cutter. The driving plate will have two keys in them that mate with the coil form keys to both provide positive indexing of the coil form, and to provide a limited amount of rotational torque if needed (though the turned features may be made first with mandrel jacked out against the ID). Beamtube calculations 1. 2D FEA transient analysis of eddy currents in various beamtube thicknesses was performed by D. Shuman. The field response to a sine wave current input is plotted and shows the field phase lag and amplitude decrease expected. This will be checked for accuracy against the analytic method of Kilpatrick (LBNL E. Note M4435). The results are plotted at: http://www-eng.lbl.gov/~shuman/HIF/NTX/ntx_pq_2d_trans_sin_cur_out.pdf and peak values at: http://www-eng.lbl.gov/~shuman/HIF/NTX/ntx_pq_2d_trans_sin_field_peak.pdf here's a movie of an earlier ramp response analysis showing the field penetrating the beamtube: http://www-eng.lbl.gov/~shuman/HIF/NTX/ntx_pq_2d_trans_3.avi The decrease in overall field is less than 6% for a 2 mm beamtube, and phase lag can easily be accounted for when extracting the beam. The field remains a high quality quadrupole field throughout the pulse. However, any variations in resistivity of the tube, such as along the weld could introduce multipole errors. How much is uncertain, and resistance measurements maydifficult to accurately make. Variations in permeability could most likely be minimized by using (properly welded) 310 alloy S.S. which shows almost no change in what is a very low permeability with cold work (welded seams are "plenished" or rolled on both sides). Since we don't yet know the answers to some of these questions efforts will be made to accurately cost out the composite S.S. lined tube as well as the thicker tube. 2. Earlier estimates (using formulas from Roark and Young) of minimum beamtube thickness to resist buckling were, as suspected, non-conservative. A NASA formula that includes an empirically derived de-rating factor shows that a 1.20 mm thick tube 15 cm in radius, circularly stabilized at 0.6m intervals, will have a factor of safety of 2.0 for vacuum service. A 0.9mm thick tube will have a factor of safety of 1.5. This is not recommended as the tube will need to have sufficient factor of safety when hydrostatic pressure tested at 1.5 atm. Stored energy of the vacuum volume is somewhere around the limit for low hazard pressure vessels (100 kJ). It would be nice if it is under, but detailed volume measurements of the entire vacuum system will tell for sure. Please think about painless ways to reduce the volume in the source box, or elsewhere. --------------------------------------------------------------------- Derek Shuman ?DBShuman@lbl.gov? Mechanical Engineer Lawrence Berkeley National Laboratory Derek Shuman Mechanical Engineer ?DBShuman@lbl.gov? Lawrence Berkeley National Laboratory HTML Mail Mailstop 47-112 1 Cyclotron Rd. Fax: 510-486-5392 Berkeley Work: 510-486-4662 CA Conference Software Address 94720 Default Directory Server Additional Information: Last Name Shuman First Name Derek Version 2.1