To:		Distribution	8 Jan 96
	From:		Martin Nordby
	Subject:	Minutes of Near IR Engineering Meeting of 5 Jan 96


Q1 Permanent Magnet Block Testing at SLAC

Andy Ringwall presented preliminary test results from permanent magnet 
material tests by the SLAC Magnetic Measurement Group (Zach Wolf, et al).  
They tested Samarium Cobalt cubes from three manufacturers:  Electron-
Energy Corp., Shin-Etsu, and Vacuumschmeltze.

The north- and south pole faces of the blocks was magnetically mapped using a 
Hall probe mounted on a slide.  The block was positioned in a fixture to ensure 
that its position was repeatable, but no attempt was made to ensure that the 
plane of the scanning probe was parallel to the face of the block.  The probe 
data suggests that the block faces were not, indeed, parallel to the scan plane, 
since the field tends to drop off to one edge of the block.  Andy plans to re-
measure the blocks to investigate this.

The Halbach formulas theoretically correlates the overall magnetization vector 
(magnitude and orientation) with expected field strength at any point on the 
surface of the block.  The goal of these measurements is to attempt to correlate 
these empirically, then compare the results with the expected values to look for 
local field non-uniformities.  These local errors have been seen in P.M. material 
used for insertion devices at LBL, according to Dave Humphries.

For simple cubic geometry, the Halbach surface integral simplifies to a closed-
form solution (found in the Shin-Etsu literature).  Using the Hall probe 
measurements and the closed-form equation, predicted values for residual 
induction are:

	EEC material	9810 Gauss
	Shin-Etsu		9870 Gauss
	Vacuumschmeltze	10,098 Gauss

The EEC material had the most uniform field contour, but all values fell short of 
the catalog numbers for Br.  Andy thinks that this is due to angularity errors in 
the field orientation which produce a cosine-effect error in the magnitude of the 
normal component.  Another possible source of error is nearby magnetic 
material in the slide assembly or the table.  This will be checked.


Rough Helmholtz coil measurements have also been started.  Running an FFT 
of the time-domain results from a rotating block gives the magnitude and phase 
angle of the residual induction vector, while eliminating much of the imprecision 
of the measurement apparatus.  There is some problem in reproducing the 
phase angle, but the measurement method looks sound.

Curiously, the FFT showed a number of higher harmonics from the 
measurement.  Because the Helmholtz coil is so large (1 meter diameter) this 
could not possibly be due to small non-uniformities of the P.M. block.  Possible 
sources are systematic noise in the measurement system (encoder imprecision, 
etc.), or vibration of the rotating platen in the Helmholtz set-up.  Andy is 
investigating this.

The next step is to calibrate the Helmholtz coil using a known dipole field, then 
take absolute, accurate, repeatable measurements of the blocks.  Also, re-
measure the blocks with the Hall probe set-up, and see how the measurements 
correlate.

This needs to be done to understand what we need from our P.M. blocks 
regarding uniformity, and what we should specify in the Technical Spec and 
purchase order.  Finally, in conjunction with buying the prototype blocks, we 
should plan to purchase a few trapezoidal blocks to prove out the measurement 
method before the prototype blocks arrive.


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

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