To: Distribution 28 August 97

From: Martin Nordby

Subject: IR Engineering and Physics Meeting Minutes: 8 August 97

Hard-Copy Distribution:

Bob Bell	41		Nadine Kurita	18
Gordon Bowden	26		Harvey Lynch	41
Pat Burchat	95		Tom Mattison	17
Scott Debarger	17		James Osborn	LBL B71J
Hobey DeStaebler	17		Andy Ringwall	17
Jonathan Dorfan	17		John Seeman	17
Stan Ecklund	17		Mike Sullivan	17
Karen Fant	18		Uli Wienands	17
John Hodgson	12		Mike Zisman	LBL B71J
David Humphries	LBL 46-161
Roy Kerth	LBL 50-340
David Kirkby	95
Jim Krebs	41

Electronic Distribution:

Curt Belser	Tom Elioff	Lew Keller	Natalie Roe	Dieter Walz
Lou Bertolini	Kay Fox	J. Langton	Ross Schlueter	Rick Wilkins
Adam Boyarski	David Fryberger	Georges London	Ben Smith	Fran Younger
Catherine Carr	Fred Goozen	Rainer Pitthan	Steve St Lorant	Ron Yourd
Al Constable	Alex Grillo	Joseph Rasonn	Joe Stieber
David Coupal	Keith Jobe	Jeff Richman	Jack Tanabe

Vertex Vacuum Chamber Paint Tests

Karen Fant reported on progress on testing of paints for protecting the beryllium Vertex Chamber from water. Tests used 6061-T6 aluminum samples, electroless nickel plated. Paints tested were:

BR127: 10% solids, epoxy/phenolic with 15% strontium chromate; 250°C cure

BR154: 20% solids, epoxy/polyimide, 20% (?) strontium chromate; 350°C cure

Data from CERN predicts that phenolic is damaged by 1 Mrad of radiation dose. The paint tests used a 6 MeV electron source at Nutek (an old Varian Clinac), and samples were irradiated to 25, 50, 100, and 150 Mrads.

Test results showed that there were NOT big differences between the two types of paints. The BR127 turned slightly browner, but started off with a smoother, shiny finish, which followed the contour of the surface finish. There was evidence of some cratering and domes on the surface, but it was not clear what the source of the imperfections was. Coating thickness was 0.0004", and the paint showed good adhesion in peel tests and metallography. Some small cracks (0.1 micron wide) showed up on the SEM investigation.

The BR154 samples showed no change in color, but the initial painting produced a visibly rougher surface, with a matte finish. Coating was fairly flat, not following surface contours. Thickness was 0.0005" - 0.001". The paint appeared to adhere well to the nickel plating, but there was some problem in the plating de-laminating from the base metal. Metallography and SEM shows the surface is very rough, with more voids, flakes, and craters, even in the as-painted samples (no radiation).

Further investigation is needed of the cracks in both paint samples: their frequency, size, depth, and if metal is exposed at the bottom of the crack. Also, the nickel-plating quality needs to be investigated. Conclusions to-date are that both appear to stand up to the very large over-dose of radiation. Each had some surface integrity problems, but the BR127 had a much cleaner, smoother surface. Also, BR127 is more the industry-standard paint for corrosion protection. Right now, it is still the first choice for paint.

Q2 Magnet Status

James Osborn reported on status of the Q2 Magnet. The endplates are 50% done. They have been rough-cut, annealled, and machined, with the final step being EDM'ing the inner profile and machining final cut-outs.

The stacking fixture is complete, and is waiting laminations. The vendor has not quite met the 0.001" profile tolerance on the pole geometry. He can laser-cut the lams. to twice this tolerance, and is tweaking his cutting process to try to improve on that.

James is looking at the effect of loosening the profile tolerance on the pole tip geometry. Halbach's perturbation paper predicts greater field errors than James' analyses suggest, possibly because this is not a completely iron-dominated magnet. James perturbed the profile of the pole and coil together in a full POISSON model with otherwise quad symmetry. This showed that changes in harmonics are not very sensitive to pole tip profile. A 0.005" displacement of 1 pole produces a 10^-4 harmonic for n = 6. More work needs to be done on this.

Two coil winding mandrels are complete, and test coils are being wound now, using pre-preg tape. They should start building production coils in mid-August.

Q2 Water Manifolding

James and Scott Debarger discussed current plans to manifold Q2. The manifolds have been moved towards the I.P. to make more room for bending the Syn-Flex hose. The minimum bend radius is 2.5" in the current layout,while the manufacturer recommends 2" min.

The manifolds can not be easily plumbed in parallel, because magnet coil flow direction is set by routing of the coil leads. This means manifold cross-connects are needed to join them together. These are soldered elbows which stick further towards the I.P., and encroach somewhat into the space under Q2 for cable routing. Supply and return hoses come off the HEB side of the magnet, and drop below SK1. These are 0.5" diameter Syn-Flex hoses.

Because Q2 must be slid transversely to remove the manifold from the cut-out in the LEB side of the Raft, room must be left on the HEB side so the magnet does not run into anything.

Cable space under the Magnet must include provision to slide the magnet. The minimal cross-sectional area available under Q2 is 260 cm^2, which occurs at the bend in the Raft, where the Q2 manifolds stick under the magnet. This does NOT include clearance for sliding the magnet. Total area of the cabling going through this region is 160 cm^2, so we are getting tight in the region.

Q1A Dipole Ring Assembly Update

Mechanical Distortion

Andy Ringwall reported on distortion of the support collars for the Q1A slices. CMM measurements of 3 slices shows the rings are out-of-round by 0.06" (+/- 0.03" from theoretical radius). This correlates somewhat with rough measurements taken pin-to-pin across the slice. The deformation is elliptical, with the minor axis in-line with the pins. Calculations predict distortion due to magnetic forces should be 0.06" on the diameter, which is roughly what we are getting, but this does not include stiffening from the magnet blocks. Some of the deflection could be from assembly loads, also. Andy plans to machine new pin holes, and slot out clearance holes for the threaded rods.

Effect of Permeability Variations

Stan Ecklund modified his MATLAB model of the Q1A to iteratively solve for not just effects of mu not equal to 1, but also mu-parallel and mu-perp being different values. mu-parallel = 1.065, and mu-perp = 1.15. This produces an n = 3 harmonic of 1% at r = 6.35 cm. This was alluded-to in Halbach's paper, but not quantified. It is related not just to the ratio of mu's, but also their absolute value. It is not clear if this effect is worsened when magnet slices are stacked (the MATLAB model assumed an infinitely-long magnet).

As-Built Slice Harmonics

Mike Sullivan reported on the actual harmonics of slices built. The average b1 = 0.017 T with the average b3 = 0.0011 T. The measured odd harmonics are higher than even ones, and are 10 times more than Stan's MATLAB code predicts. These appear to be due to fabrication errors, mostly. More analysis needs to be done on the as-built harmonics

These minutes, and agenda for future meetings, are available on the Web at:

http://www.slac.stanford.edu/accel/pepii/near-ir/home.html