Minutes of KEK-SLAC collaboration, 12 June, 2003

Agenda is listed below;

"KEK design (1)", Y.Honda

First, Honda showed the requirements for the mover; (1) dynamic range and resolution of position are 100um and 1um, respectively, (2) those of tilt are 100 urad and 1urad, respectively and (3) active position-mover with 1um dynamic range and 1nm resolution at > 100Hz. He assumed that the initial alignments were 100um and 100 urad. At this moment, only Y and Y' (vertical) movers are only considered for simplicity.

A cylindrical BPM is set on a V-block, where all machining accuracies shall be less than 10um. For an example, a mechanical center measured by outer surface and an electrical one of the cavity can be coincident with in a few um. This situation can be realized for a newly fabricated BPM, while the present BINP-BPM might not be in this case.

One question was raised as "How do you control or align a rotation on the V-block?" If the initial alignment is not enough for the rotational error, a X-mover should be required without active feedback system. Tolerance of the rotation error can be obtained from this beam rum.

The V-block sits on the active mover consisting of two elastic hinges and a piezo device. The active mover is set on the Y' mover which also has an elastic hinge and a piezo devise. The 100urad-range can be obtained with 10um dynamic range of the piezo device and a 75mm long arm between the hinge and the piezo, while the SLAC present mover has about 400urad. The Y' mover is set on the Y mover. There would be two options for the Y mover. The one is mechanical mover using a triangle block whose horizontal move is transfered to vertical one. The another is to use a stacked piezo device, where a power supply stability must be less than 10-5 because of 100um dynamic range with nanometer stability.

Three V-blocks in the three mover systems shall be lapped on a basement plate with accuracy of less than 10um. Finally, the basement plate shall be set on a rigid girder which may be granite or steel.

He also showed a preliminary idea on how to monitor positions of three BPMs hopefully with nanometer resolutions. The positions are monitored by laser interferometers between the top-outer surface of BPMs and a reference block which is set above the heads of BPMs. So, the reference block must be "floating" in air.

Questions or issues to be investigated are summarized as;

"KEK design (2)", Y.Higashi

Higashi estimated vibrational properties of the KEK system which Honda explained. He considered a girder (granite) of 300mm hight, 500mm wide, about 1m long and 300kg weight which is supported by four legs. Q-values and frequencies(RF) of primary resonances were estimated as follows; Q=3130, RF=500Hz at the active Y-mover, Q=1346, RF=~216Hz at the Y' mover and Q=3000(571), RF=500(91)Hz transversely (longitudinally) at the girder. These frequencies seems to be high enough for isolation from ground motion and cultural noise.

He also showed a nanometer monitoring system based on a laser technology. His system consists of a laser beam intersecting three mirrors which branches off a laser to BPMs. Positions of laser shall be measured by Quadrant Photo Detectors (QPDs) at each BPM. Accuracy of QPD can be expressed by 0.2%/10nm with a laser spot size of 60um. Relevant issues are stabilities of laser and optical devices such as mirrors, which must be controlled at nanometer level.

"Some results from this beam test", S.Smith

(transparencies, Cavity BPM Status;12 pages, pdf ,930KB, Status, June 10,2003;5 pages, pdf,375KB, Status, June 3,2003;4 pages, pdf ,615KB) Daily progresses have been presented at the ATF evening and weekly meeting. At this meeting, Smith briefly reported present status dated in 12 June,2003.

Calibration run has started for BPM-X (from -50um to 50um) and X' (from -3 m rad to 3 m rad )movers on 6 June. X-position of BPM-1 have been fitted to a linear function of other BPMs. First, the 1X-positions were calculated from BPM2 X and BPM3 X. Resultant resolution was 604nm. Second, they were calculated from other BPMs including Ys and Y's. The result was 473nm. Residial distributions of 1X are uniform in BPM1X range of +/- 40um, The measured resolution is consistent with an estimated electronics; 13 ADC counts with noise reduction of 2.4 from "Processing gain", converting to X using measured calibration, which is 400 - 500nm.

There two major reasons for such large noise, which are (1) data taken with 20dB attenuator at front end and (2) many pulses saturate ADC, so short integration time (low processing gain with high noise bandwidth) and late evaluation of amplitude (lowers gain) have to be used. Improvements are (1) remove 20dB attenuator for 10 times signals, (2) avoid saturation by integrating longer and sampling earlier for 5 times signals at least, and (3) use IF bandpass filter to match the digitizer dynamic range.

Next calibration has been done on the BPM2Y mover. Fitting was also executed with 10 parameters in a linear function (Y1, Y3, X1,X2,X3 and their tilts). Best result was obtained to be 170nm, while one with only Y1 and Y3 was 230nm. There is an interesting systematics in a correlation plot in a plane of BPM2Y and its predicted values.

They found that calibration was very difficult with beam jitters of O(10um), and resolution of controlling magnet power supply was not enough to correct orbits. These issues must be improved.

The above results are so-called "online" analysis. So, more advanced analysis will be done, and some results are expected to be presented at the ISG10 in next week(6/17-20). We will also continue to have video conference in future.