PSB ME: Comparison Beamscope / Fast Blade Scanner

PS/AE Note 2002-43 (MD)

14 June 2002

PSB ME: Comparison Beamscope / Fast Blade Scanner

Date: 7 -21 November 2001, in parasite

Participants: H. Schönauer

Reported by: H. Schönauer

______

Abstract:

The Fast Blade Scanner, (FBS, "Guillotine") installed in the PSB in Ring 1 for vertical beam size measurements, was operated at 1.4 GeV, with beams of varying intensity and emittance. Beamscope measurements were performed for comparison. Results show satisfactory agreement between the two devices but are subject to fairly strong fluctuations. One also obtains some information on the FBS absolute blade position readout accuracy.

Objective:

Performance evaluation after recalibration of the Fast Blade Scanner (FBS) and comparison with Beamscope.

Procedure:

All measurements were made in Ring 1, in the VERTICAL plane, after synchronisation and just before rise of the extraction bump.

This avoids the complications due to the effect of dispersion and possible reactions of the beam control system to the change of orbit length due to the ejection bump.

The Beamscope application program handles both the FBS and Beamscope. It dumps all values, which can be read from the display, to a user-specific USER.res.html file.

Appending each new measurement results to a dump file permits a convenient (off-line) spreadsheet analysis.

The FBS can be operated at different crossing speeds, for which a control parameter RPM (for the drive motor) can be set to values: 200, 500, 700, 1000, 1300, 1500r.p.m. The available RPM settings span a range of crossing speeds from 4.5 through 9.5 m/s. The lowest value is not available from the application program. The highest value of 1500 RPM appears not to yield significantly higher average speed (cf. Fig. 1). For this setting, computed values of speeds and emittances show strong fluctuations, suggesting intense mechanical vibrations. For this reason, and because they are not useful in operation, the emittance results are not tabulated.

Fig.1:Measured crossing speeds as function of the "RPM" setting. Error bars show standard deviations. v, v denote upwards and downwards motion, respectively.

For Beamscope measurements, it is known that there is a weak increase of measured beam radii/emittance values with scraping speed. The effect is most likely due to the limited bandwidth of the beam transformer electronics. For this reason, Beamscope reference measurements were made with peak dipole currents of 40, 60, 80 and 100A, by this way covering the whole scraping velocity range of the FBS. Emittances are compared for matching scraping speeds.

The tests were planned both with beams typical for higher intensity and larger emittances ~61012p/p (user AD, 12 turns injected) and the nominal LHC beam (user TSTLHC, 2.5 turns injected). The latter is of particular interest as the effort of having ultimately three operational systems (wire scanners, Beamscope and the emittance measurement line) and one purely mechanical 'calibration' system (the FBS) is justified by the need to trust the emittances of the LHC beam measured there, where they are generated - in the PSB.

Unfortunately, in the course of the measurement series with the LHC beam, one bellow of the FBS developed a leak, ending abruptly all measurements of this year and leaving the series incomplete.

Prior to these two series, preparatory comparison series were made, one with low intensity and slightly 'shaved' beams (user SFTPRO, 3 turns injected) and two others during an exploratory machine experiment on the working point Qv = 4.28. During these series, Beamscope was not as carefully calibrated as in the final runs, also measurement timings were sometimes varied. Although less significative, these results are also presented in order to complement the incomplete final results.

Results:

The following table summarizes the results of all measurements. The series with user TSTLHC could not be completed.

USER
Intensitiy / Average Velocity / Average Emittance 95% / Stdev
Emittance 95% / Average
proj. Emittance / Rel. Diff. FBS  BSC / ML Emit.
m/s /  mm mrad /  mm mrad /  mm mrad / % / 
BSC / FBS / BSC / FBS / BSC / FBS / BSC / FBS / 95% / proj
AD
5.7E12 p/p / 3.2 / 4.3 / 11.15 / 11.44 / 0.38 / 0.35 / 10.17 / 10.01 / 2.6 / -1.6
5.8 / 5.8 / 11.55 / 11.70 / 0.50 / 0.44 / 10.39 / 10.11 / 1.3 / -2.7
7.9 / 7.5 / 11.81 / 11.49 / 0.52 / 0.54 / 10.41 / 10.06 / -2.7 / -3.4
9.7 / 9.1 / 11.96 / 11.86 / 0.53 / 0.58 / 10.55 / 10.04 / -0.8 / -4.8
TSTLHC
1.46E12 p/p / 3.1 / 4.3 / 3.22 / 3.53 / 0.15 / 0.23 / 2.62 / 2.75 / 9.6 / 5.0 / 3.04
5.5 / 5.7 / 3.52 / 3.65 / 0.15 / 0.27 / 2.72 / 2.80 / 3.7 / 2.9
SFTPRO
1.6E12 p/p / 3.0 / 4.5 / 3.45 / 3.74 / 0.14 / 0.27 / 2.99 / 3.16 / 8.4 / 5.7
5.6 / 5.7 / 3.49 / 3.53 / 0.15 / 0.10 / 2.97 / 2.94 / 1.1 / -1.0
7.6 / 7.3 / 3.57 / 3.99 / 0.11 / 0.20 / 3.12 / 3.34 / 11.8 / 7.1
9.2 / 8.8 / 3.76 / 3.48 / 0.12 / 0.24 / 2.93 / 2.94 / -7.4 / 0.3
2.2E12 p/p
ISOGPS QV=4.24
6.2E12 p/p / 3.2 / 4.5 / 6.38 / 7.24 / 0.56 / 0.32 / 5.17 / 5.57 / 13.5 / 7.7
5.6 / 6.61 / 0.31 / 5.18
7.4 / 7.58 / 0.20 / 6.19
8.9 / 6.56 / 0.58 / 4.99
3.3 / 4.5 / 10.82 / 11.37 / 0.75 / 0.71 / 10.01 / 9.82 / 5.1 / -1.9
5.4 / 10.86 / 0.38 / 9.29
7.0 / 11.80 / 0.49 / 10.36
8.9 / 10.90 / 0.65 / 9.46

Table 1:Emittances measured with FBS and Beamscope. Shaded part of table refer to earlier measurements where Beamscope was not exactly calibrated and measurement timing was sometimes varied.

Comments:

Apart from its final breakdown, the FSB worked reliably during all sessions. It no longer perturbs measuring equipment in the BOR, notably the Beam Loss Monitors and Beamsope itself. After some hours of use, the DSC needs to be rebooted.
There is a reasonable agreement of the average emittances between all the two devices. The largest difference of 9% between the low-speed TSTLHC 95%-emittances seems partially due to the aforementioned dependence of emittance values on the scraping speed – Beamscope scraping (3.7 m/s) was significantly slower than the FBS speed of 4.3 m/s. The small TSTLHC beam was also measured in the Emittance Line, which yields a 9 and 13 % larger projected emittance than the FBS and Beamscope, respectively. For such a small beam of ~3 this can be considered as satisfactory agreement.
The dependence of emittance values on the scraping speed observed with Beamscope seems to be less pronounced for the FBS, with the consequence that FBS emittances are somewhat higher than Beamscope ones at low speed, but are about equal at higher speed. Moreover, FBS emittances as a function of the four RPM settings show a zig-zag pattern: values for 4.5 and 7 m/s, and values for 5.4 and 9 m/s are about equal, respectively. This rather suggests systematic calibration (done separately for each RPM setting) errors than a trend.

Fig.2a:Trends of emittances measured with FBS and Beamscope vs. scraping speed. Top: Large beam (AD), 12 turns injected, Bottom: Small beam (SFTPRO), 3 turns injected; these trends are not very significant due to fluctuations .

The FBS measurements with small beams show ambiguous features: For the SFTPRO beam (Fig. 2a bottom), there is virtually no emittance increase with scraping speed. The (less numerous, but carefully calibrated) measurements on TSTLHC (Fig. 2b), however, show a weak slope comparable to that of Beamscope.

Fig.2b:Trends of emittances measured with FBS and Beamscope vs. scraping speed. Small beam (TSTLH), 2.5 turns injected; scatterplot of individual measurements for both scraping polarities. Note that in this reprsentation the slopes of the linear trendlines are about equal.

Another subtle difference between FBS and Beamscope is observable in the plots of Figs. 3 below. They suggest that for Beamscope the “+” (scraping upwards) emittances are slightly, but systematically, smaller than the ““ emittances. As the closed orbit at the Beamscope aperture is off-centre (this can be inferred from the bump amplitude data, but shows also up in a systematic difference in scraping speed), this may be explained by a minor non-linearity of the machine. The FBS results do not show this asymmetry.

Informations on the FBS:

For each measurement, the application software records the aquired positions of the FBS intercepting edge corresponding to 95% and 0% remaining circulation beam current. The latter position is supposed to represent the centre of the circulating beam and should thus be the same for both scraping directions. As Figure 4 shows, this is not the case: puzzling readout differences up to 20 mm between scraping directions have been observed. Beam radii fortunately vary much less, suggesting a displacement of the read-out values, possibly even in time. Moreover, results are inconsistent and confusing: different series of measurements (on different PLS users) show contradictory trends: the only apparent general feature seems to be an increase of the gap between beam centres measured at different scraping motions (up / down) with the speed (set in “RPM” values). But the amount of this trend depends on the PLS user (or on the day..?)

Conclusions:

No systematic differences between Beamscope and the FBS have been found. Hence one can conclude that the lattice is fairly well represented by the beam optics model used. Projected emittances – the ones to be compared with wire scanners and SEM grids – differ in general by less than 3 %.

It is thus suggested that the FBS should be repaired (new bellows of more resistant stainless steel), carefully calibrated (not necessarily in the ring) and reinstalled in horizontal measurement position.

It should be recalled that the real interest of the FBS is to have a measurement device to cross-check the principal operation workhorse, which will be the wire scanner. There is of course Beamscope, but the reliability of horizontal Beamscope measurements at 1.4 GeV may be compromised by the fact that the bumper dipoles are driven far into saturation, an effect that may be not completely compensated by software corrections applied. Also, a reaction of the beam control system cannot be excluded. In practical terms, the beam radii measured by Beamscope depend somewhat on the direction of scraping (in/out). These known limitations of Beamscope are in fact the raison d’être of the FBS, which should be exempt of these.

Acknowledgements

I would like to stress in this place the excellent work done on the FBS in terms of hardware installation, as well as control and applications software development, making the FBS as convenient to use as today’s Beamscope. My thanks go to Uli Raich, Enrico Bravin and Jean-Michel Nonglaton.

Average emittances measured with Beamscope and Fast Blade Scanner for Users AD and TSTLHC for varying scraping speed and for up (+) and down (-) scraping.

Ring 1, vertical plane; error bars represent standard deviations

Fig. 3a: 12 Turns injected, Average Intensity 5.7 1012 p/p / Fig. 3b: 2.5 Turns injected, Average Intensity 1.46 1012 p/p

Average emittances measured with Beamscope and Fast Blade Scanner for Users SFTPRO and ISOGPS for varying scraping speed and for top (+) and bottom (-) scraping. Ring 1, vertical plane; error bars represent standard deviations

Fig. 3c: 12 Turns injected, Average Intensity 5.7 1012 p/p / Fig. 3d: 4 and 12 Turns injected, Qv = 4.24
Average Intensity 2.2 1012 and 6.2 1012 p/p;

Information on the functionality of the FBS:

Recorded beam centre positions for the various series of measurements. Zones indicated are labelled with the “RPM” control values for the scraping speed. One observes a noticeable difference between scraping directions (up / down), which seems to increase with the scraping speed. Beam radii fortunately vary much less, suggesting a displacement of the read-out values, possibly even in time. All scales in mm.

Fig. 4a: First series of measurements, spread over hours. The beam radius (during MD, on an unusual working point) may have slightly changed during the time. One notices a systematic difference in beam radii with the direction of scraping. / Fig. 4b:.In this series (only) there appears a tendency of increasing beam radius with scraping speed
Note the general trend of an increasing gap between the beam centre values with increasing speed. / The gap increase between beam centres seems not too much dependent on speed: but there is larger number of outsiders
Fig. 4c: No trend for beam radii in this series, if not the fact that the values for 1000 RPM are systematically higher than the others. / Fig. 4d: Last few measurements on the small LHC beam before vacuum failure.
The general trend of an increasing gap between the beam centre values with increasing speed is most expressed in this series of measurements: At highest speed (1500 RPM) the gap reaches values of 20 mm ! / In this limited series no trend is recognizable.