Mike and all, I've been pondering how to build the energy chicane and hope to write up the details of a preliminary concept soon, but meanwhile I think I found a good solution to the BPM mover problem: we could use the roller cam type movers designed for FFTB. These are purely mechanical movers with large load capacity, that can be designed with ranges of several millimeters, have excellent repeatability, no backlash, and can be read out with varous encoders to 0.1 um. The system I have in mind would be comprised of: - four identical magnets - BPM's sets with x, y, and phase cavities on separate movers - BPM sets before and after the magnet chicane and one (or two) BPM sets between the middle two magnets. This plan is similar to that outlined in the IPBI presentations with the main differences being that four identical magnets are used instead of three magnets, and the BPM's inside the chicane are between the middle magnets on a leg that is parallele to the main beam, rather on the two legs at an angle to the main beam. The energy is determined from the horizontal offset of the middle BPMS with chicane on. This arrangement avoids two serious issues that arise if the BPM's are put in the legs after the first magnet and before the fourth magnet. - with chicane on the BPMS would have to be moved horizontally and Be rotated to be aligned with the beam. Such compound motion would be difficult to do accurately. - the BPM's are sensitive to the beam angle within the cavities. Beam at an angle to the cavity axis excites the TM210 mode at 90 degrees out of phase from the signal from a beam that is offset but aligned with the cavity axis. This causes problems. With the BPM's between the middle two magnets, the main motion is purely horizontal, and the beam does not have an angle wrt the cavity axis. The system would operate similar to the techniques used at FFTB using beam based alignment. The basic idea is to first set the bends off (or at some well measured value close to zero) and put the beam straight through. Using the movers put all BPM's on the beam line in x and y. Depending on resolutions and other errors, this could take some small number of pulses with 2820 bunches, and could be automated. Then within minutes energize the bend magnets and move the middle BPM's to the displaced beam position (horizontal bend). The energy is determined from the Bdl and locations of the bends, plus the x offset of the middle BPM's compared to the bend off position just measured. The whole measurement cycle would take a few minutes. The BPM calibrations would all be measured using the movers to move the BPMs a known distance, as often as required to certify the gain. The energy measurement would be extracted from combinations of BPM data with zero field and with bends on, using the measurements before and after the chicane to remove the beam jitter. The required precision of 10**-4 would be achieved by averaging over some number of beam pulses determined mainly by the BPM resolution. In this scheme, as in the original IPBI version, most of the longish term drifts (longer than a few minutes) do not enter, because the answer is obtained from a difference of averages over pulses taken within minutes. The main problem is to remove ( or measure and correct) for drifts in parameters (like stray fields, motion of magnets, ...) within the few minutes measuring cycle. Most long term drifts from temperature changes, slow motions of magnets, etc would not affect the measurement and faster changes (vibrations, etc ) would be averaged over. With accurate movers and encoders for x and y on each set of BPM's the gain calibrations are easily made, and the stability and drifts of the electronics can be studied often. The key to this method is to use beam based alignment to establish the BPM positions with field off, and be able to accurately move and measure where the BPM's find the beam with field on. For this its important that the major move be mostly in the horizontal and not involve rotations. Small adjustments in y would be allowed. This could be done with the FFTB syle movers and encoder systems.... which you can read about here: An overall description of the FFTB instrumentation is given in: epaper.kek.jp/p95/ARTICLES/TAG/TAG01.PDF Development of the magnet movers and the stable support stands are described in KEK-Preprint-2002-49 which is at lam27.iae.kyoto-u.ac.jp/PDF/7P-44.pdf The mechanics and principles of operation of the movers are described in: SLAC-PUB-95-6132 A presentation by Mike Brown of SLAC on the propsed electronics systems for operation of the movers at NLC is given in: www- project.slac.stanford.edu/lc/local/Reviews/Lehman99/pdf/Browne_Mover_El ect.pdf A nice description of the signal from beam angle in the BPM's (cylindrical ones in this case but with similar effect in the rectangular ones we are using) can be found in: KEK-Preprint-98-188 that you can get from lcdev.kek.jp/Conf/HEACC98/133.PDF Ray