Minutes of ATF2 weekly meeting, 13 July, 2011

July 13 15:00 - 16:00 (JST time), Webex and ATF Meeting Room, KEK

The new ATF wiki page (http://atf.kek.jp/twiki/bin/view) can be reached here, where the collaboration-page requests your login or registration.

LowQ-IPBPM progress, S.W. Jang (KNU)

First, SiWon showed a result of January beam test, i.e. the resolution of the low Q IPBPM was 70.44mm by using 3 signals of Y1I,Y2I, Y_REF together with Nakamura's 5.94nm and 19.7nm by using 10 signals of Y/X I/Q and Y/X_REF and the same 3 signals, respectively. All above values are normalized at the beam intensity of 1 x 1010/bunch, i.e. the published (Nakamura's) resolution is 8.72±0.28±0.35nm at the beam intensity of 0.68 x 1010/bunch.
Reference :
1. "Development of Beam-position Monitors with High Position Resolution", Master Thesis by T. Nakamura, Department of Physics, Graduate School of Science, The University of Tokyo, 3 February 2008, http://acfahep.kek.jp/subg/ir/nanoBPM/index.html#library
2. Y.Inoue et al., PR STAB 11,062801 (2008)

Above measurement was done with the KNU electronics of version 1. In this fall, the second version electronics will be tested. The major improvements in the second version are higher conversion gain of 70dB and isolation in the mixer of -40dB with the latency time of 25nsec.

KNU will fabricate one IPBPM with 2 cavities, one with single cavity, and two reference cavities. New IPBPMs will be made without brazing, so they will be installed in a vacuum-chamber. One of IPBPMs is put on a piezo-mover for fine alignment in vertical direction, at least.

Dimensions of the low-Q IPBPM cavity were shown. Also he showed results of the HFSS simulation such as the electric field mappings, parameters ( resonant frequencies, coupling, Q-values and decay times) and time structure of signals for two materials of copper and aluminum. Both materials have very similar results. Aluminum is preferable for the light weight, while it may need delicate fabrication.

The beam test will be setup on the granite table at the upstream test section, where the high-Q IPBPM test is also planned. Both setups are in-line and the test beam line can be swapped with a beam pipe at the ATF2 beam tuning.

Preliminary schedule is listed below.

Q : What is the angle correction in your table?
A : The source is the beam angular jitter, which appears in the Q-signals. The effect is estimated by analytic calculation.
Q : Does this mean that the angular jitter dominates in the resolution?
A : By Nakamura's Master thesis, the resolution was 28nm if we use only I signals of Y-cavities, and it becomes 15nm including X information.
Q: What is the 40urad in the table?
A : It is the measured angle between 2 BPMs.
Q : We still do not well understand a meaning of "Expectation for the beam angle correction" and ~12nm in Low-Q and 40urad in High-Q. From above answers, I guess that the resolution corrected the beam angular jitter is estimated to be ~12nm, where the angle jitter was measured to be 40urad in the beam test, January 2011. If so, the correction seems to be too large.
Q : You tested a setup with 1 low Q and 2 High Q IPBPMs. How did you calculate the resolution ?
A : I assumed the same resolution of all IPBPMs which have been calibrated by the movers on the post-supports.
Q : Did you use the same electronics as Nakamura's ?
A : Low Q IPBPM used the KNU electronics ( version-1), but high Q used KEK ones.
Q : What is difference in the sensor cavities between new and previous low-Q IPBPM ?
A : The dimension is slightly different for shorter decay time.
C : In general, Al is not so good for a slit for signal-antena.
A : Al must be OK since Shintake made the C-band RF-BPMs from Al at FFTB. However, we had better check to ask for Honda-san or Shintake-san if needed.
Q : High Q IPBPM has the decay time of 30ns. It looks OK for the multi-bunch beam. Why do you need the low Q IPBPM ?
A : No, the high Q IPBPM has 60ns since the loaded Q of the high Q is a few times larger than the low Q.

A short update of FONT, P.Burrows (Oxford univ.)

FONT5 system has been studied at the upstream section. It is phased-locked to the beam at 357MHz. Valencia's movers were used for calibration, which shows fantastic performance. After the earthquake, the hardware was inspected to check the status. We found 3 cables broken by the concrete shields, which are already replaced. So, the visual inspection is OK and the system is ready for the beam study. Indeed, our work in this June is good opportunity for above inspection.

As the ILC prototype system, the FONT5 verified experimentally the bunch-by-bunch feedback with latency of 140nsec and enough kicker drive power corresponding to more than ±300nm at the ILC IP, i.e. ±100um to be corrected at 1GeV corresponds to ±400nm at 100GeV by scaling power. So, Ben Constance successfully defended PhD thesis.

We improved the stripline BPM resolution. The best obtained one is 200nm, while the typical one is 1 um. The difference is investigated and it is a subject of Robert Apsimon PhD thesis.

The feedback performance is also improved with timing studies and monitoring downstream BPMs, where Javier Resta Lopez contributed so much who recently returned to Valencia university. Examples are listed below. First bunch with 13um jitter (RMS) was used for the feedback input, then the second and third bunches were corrected to have jitters of 5um and 3 um, respectively. Another results are described as follows. First bunch with 2.1um jiiter was corrected to 0.4um jitter in the second bunch. So, the FONT5 achieved factor of 5 jitter reduction. Propagating the 0.4um jitter in the perfect beam line, we expect 2.6nm jitter at IP with the feedback. However, the third bunch was only corrected to 0.8um of jitter in the same data. It is disappointed result which caused by white noises in the three bunches. Actually, the correlation between the second and third bunches is poor due to the white noises. Where does this noise come from ? Snapshots of the extraction kicker pulse show timing jitter in leading and trailing edges. From discussion with Terunuma-san and Naito-san, they were caused in the thyratrons.

For possible next steps with some recommendations,

For DR studies, a BPM of LW26 was installed with the FONT BPM processor. It can measure the 40K turns of data or 1 in n-turns. It also monitor an orbit correlation between DR and the extraction line. It has been commissioned.

IP FB loop consists of a new kicker, IPBPM/electronics, digital feedback and new drive amplifier. A short kicker will be installed in smaller space just downstream of QD0. Two pairs of IPBPM-cavities are good configuration at IP for both measurement of position and angle. The kicker with 15cm length could be fit there. We will continue to discuss with KNU and KEK for the optimization. The key issue is the KNU BPM electronics with nanometer resolution.

Our recommendation is the 2 bunch train operation for the IP feedback for high quality correlation in the train. This might remove requirement for multibunch cavityBPM signal processing. Obtaining the current IP configuration from Honda-san and discussing with Stewart, we like to try to demonstrate the IP FB with the present IPBPMs. There is a space for the kicker between the pre-IPBPM and IPBPM1,2 with 100nm resolution now. It could demonstrate it. So, it is worth to try.

Q : Since the jitter is scaled by the beam size, so it is large at the IPBPMs.
A : Yes, if the jitter is large, we can feedback anyway for demonstration and the verification in principle.
Q : It is interesting in propagating jitters to IP. Did you check what was the noise in the angles?
A : We just propagated the y position, no angle effect in the extrapolation now. We can measure position and angle, then we can estimate the angle. We are also interested in measuring the jitter at downstream BPMs.
Q : If it is possible to stabilize the beam by the upstream FONT , is it sufficient ? The jitter is dominant noise by the kicker.
A : Do not over-interpreted the results of the extrapolation. Of-course, if there is no large jitter, no IP feedback is needed.
Q : In your last slide, what do you check ? what is the demonstration ?
A : If the beam jitter is 500nm measured by 100nm BPM. It can improve the jitter. So, it is an operational demonstration and it may reduce the beam size.
Q Phil : to Stewart, Do you have comments on this test (of IP FB)?
A : Readout of the 3rd bunch is now fully prepared to understand the jitters and to compare with the upstream ones. It is nice, but it is a half way .
Q : Do you plan to install the short kicker ?
A Terunuma: not in this summer, but this winter may be possible.
Q :Can you prepare a drawing ? We need electode structure for preparation.
A : yes

Next meeting is 27th July, 2011.