ATF2 Project Meeting CERN, 14-15 March 2017 Status of IPBPM performance & resolution studies N. Blaskovic, T. Bromwich, P. Burrows, G. Christian, C. Perry, R. Ramjiawan FONT, John Adams Institute, University of Oxford Thanks to the ATF2 staff and collaborators, particularly T. Tauchi (ATF2), T. Naito (ATF2) S.Araki (ATF2), S. Wallon (LAL) and P. Bambade (LAL) Talitha Bromwich 14 March 2017

Outline Background FONT experimental regions at ATF ATF IP BPMs and associated electronics Operational Challenges Sample jumps Unwanted shapes in dipole cavity waveforms Timing concerns in driving the mixer Avoiding electronics saturation Results Latest resolution estimates New two-BPM IP feedback stabilisation results Summary Talitha Bromwich 14 March 2017

Background FONT experimental regions at ATF ATF IP BPMs and associated electronics Talitha Bromwich 14 March 2017

FONT experimental regions at ATF2 ATF2 goal 2: demonstration of nm level beam stabilisation at the IP. Upstream system IP system The FONT group now continues experiments with ATF2 collaborators to work towards the ATF2 goal 2 of achieving nm level beam stabilisation. This is much more challenging than the ILC requirements, as there is no beam-beam deflection, because there is only one beam at the test facility, so we rely on much higher resolution beam position monitors to directly measure beam misalignments. FONT works in two experimental regions at ATF: one upstream in the extraction line just after the damping ring, where there are three stripline BPMS: P1, P2 and P3, and two kickers, K1 and K2, so we can change the position and angle of the beam and stabilise the beam position all the way along the final focus. The second experimental region is around the interaction point, where there are three high resolution cavity beam position monitors IPA, IPB and IPC, with the interaction point nominally falling midway between IPB and IPC. Just upstream we have one stripline kicker, IPK, which can be used to correct the position of the beam. At ATF we typically work in 2-bunch mode for feedback experiments, with ILC-like bunch-to-bunch separations of a few hundred nanoseconds. Based on figure from G. White et al. (PRL, 2014) Talitha Bromwich 14 March 2017

Experimental focus in IP system ATF2 goal 2: demonstration of nm level beam stabilisation at the IP. Most of our work in the IP region is focused on: Developing new modes of IP feedback that could improve stabilisation capability (FONT) Our major limitation in stabilising the beam at the IP is the resolution of the IP BPM system, so much time is spent collaborating with ATF colleagues on: Optimising the IP set-up BPM alignment and mover system (LAL), cavity BPMs (KNU), characterising steps of the processing electronics etc. Quantifying the usable resolution of the IP BPMs Determining the operating dynamic range Troubleshooting unwanted features in signals IP system The FONT group now continues experiments with ATF2 collaborators to work towards the ATF2 goal 2 of achieving nm level beam stabilisation. This is much more challenging than the ILC requirements, as there is no beam-beam deflection, because there is only one beam at the test facility, so we rely on much higher resolution beam position monitors to directly measure beam misalignments. FONT works in two experimental regions at ATF: one upstream in the extraction line just after the damping ring, where there are three stripline BPMS: P1, P2 and P3, and two kickers, K1 and K2, so we can change the position and angle of the beam and stabilise the beam position all the way along the final focus. The second experimental region is around the interaction point, where there are three high resolution cavity beam position monitors IPA, IPB and IPC, with the interaction point nominally falling midway between IPB and IPC. Just upstream we have one stripline kicker, IPK, which can be used to correct the position of the beam. At ATF we typically work in 2-bunch mode for feedback experiments, with ILC-like bunch-to-bunch separations of a few hundred nanoseconds. Based on figure from G. White et al. (PRL, 2014) Talitha Bromwich 14 March 2017

IP BPMs & associated electronics The FONT group now continues experiments with ATF2 collaborators to work towards the ATF2 goal 2 of achieving nm level beam stabilisation. This is much more challenging than the ILC requirements, as there is no beam-beam deflection, because there is only one beam at the test facility, so we rely on much higher resolution beam position monitors to directly measure beam misalignments. FONT works in two experimental regions at ATF: one upstream in the extraction line just after the damping ring, where there are three stripline BPMS: P1, P2 and P3, and two kickers, K1 and K2, so we can change the position and angle of the beam and stabilise the beam position all the way along the final focus. The second experimental region is around the interaction point, where there are three high resolution cavity beam position monitors IPA, IPB and IPC, with the interaction point nominally falling midway between IPB and IPC. Just upstream we have one stripline kicker, IPK, which can be used to correct the position of the beam. At ATF we typically work in 2-bunch mode for feedback experiments, with ILC-like bunch-to-bunch separations of a few hundred nanoseconds. Talitha Bromwich 14 March 2017

Operational Challenges Sample jumps Unwanted parasitic shapes in dipole cavity waveforms Timing concerns in driving the mixer Avoiding electronics saturation Talitha Bromwich 14 March 2017

Sample jumps Visible shift of the beam pulse location within our digitised sampling window. Observed in FONT boards upstream and at the IP, and in the SIS digitiser at the IP. Large, permanent sample jumps (occurs on all digitiser boards, but not by the same numbers of samples) – requires new set-up Rapid back-and-forth between two locations (occurs on different boards and different banks) –fixed by power cycling. Single sample jumps (occur on all banks) - removed in analysis Tauchi-san has confirmed the stability of a raw stripline BPM signal with respect to the triggers used for the digitisers. Ongoing work to diagnose this problem. The FONT group now continues experiments with ATF2 collaborators to work towards the ATF2 goal 2 of achieving nm level beam stabilisation. This is much more challenging than the ILC requirements, as there is no beam-beam deflection, because there is only one beam at the test facility, so we rely on much higher resolution beam position monitors to directly measure beam misalignments. FONT works in two experimental regions at ATF: one upstream in the extraction line just after the damping ring, where there are three stripline BPMS: P1, P2 and P3, and two kickers, K1 and K2, so we can change the position and angle of the beam and stabilise the beam position all the way along the final focus. The second experimental region is around the interaction point, where there are three high resolution cavity beam position monitors IPA, IPB and IPC, with the interaction point nominally falling midway between IPB and IPC. Just upstream we have one stripline kicker, IPK, which can be used to correct the position of the beam. At ATF we typically work in 2-bunch mode for feedback experiments, with ILC-like bunch-to-bunch separations of a few hundred nanoseconds. Talitha Bromwich 14 March 2017

Unwanted parasitic waveforms Signals from all the IP cavity BPMs contain an unwanted ~60 MHz waveform of unknown origin in both I and Q. Amplitude of parasitic waveform increases linearly with charge. Seems identical frequency ~ 60 MHz on all channels. Not removed by placing +/- 100 MHz BPFs after the cavity. IP BPM port cables have been replaced and matched. Parasitic signal does not shift unexpectedly within the pulse on introduction of delay cables: rules out reflections. Operationally we smooth this out using BPFs between first and second stage of the processing electronics. Possible sources: Transient effects from strong unsuppressed modes. A signal from the reference propagated through the electronics. Off-frequency signals in the DR signal used for downmixing. The FONT group now continues experiments with ATF2 collaborators to work towards the ATF2 goal 2 of achieving nm level beam stabilisation. This is much more challenging than the ILC requirements, as there is no beam-beam deflection, because there is only one beam at the test facility, so we rely on much higher resolution beam position monitors to directly measure beam misalignments. FONT works in two experimental regions at ATF: one upstream in the extraction line just after the damping ring, where there are three stripline BPMS: P1, P2 and P3, and two kickers, K1 and K2, so we can change the position and angle of the beam and stabilise the beam position all the way along the final focus. The second experimental region is around the interaction point, where there are three high resolution cavity beam position monitors IPA, IPB and IPC, with the interaction point nominally falling midway between IPB and IPC. Just upstream we have one stripline kicker, IPK, which can be used to correct the position of the beam. At ATF we typically work in 2-bunch mode for feedback experiments, with ILC-like bunch-to-bunch separations of a few hundred nanoseconds. Talitha Bromwich 14 March 2017

Latest tests Checked for unwanted harmonics in the output of the frequency multiplier spectrum analyser with help from Naito-san. Found two peaks +/- 10 MHz of the Y (5.712GHz) and X (6.426 GHz) LO outputs. Unable to check the signal outputs from the cavities directly as we do not have a high enough frequency scope. Working on simulations of other unwanted modes from the reference and dipole cavities that could potentially mix down if not sufficiently well suppressed (Matlab and GdfidL simulations). The FONT group now continues experiments with ATF2 collaborators to work towards the ATF2 goal 2 of achieving nm level beam stabilisation. This is much more challenging than the ILC requirements, as there is no beam-beam deflection, because there is only one beam at the test facility, so we rely on much higher resolution beam position monitors to directly measure beam misalignments. FONT works in two experimental regions at ATF: one upstream in the extraction line just after the damping ring, where there are three stripline BPMS: P1, P2 and P3, and two kickers, K1 and K2, so we can change the position and angle of the beam and stabilise the beam position all the way along the final focus. The second experimental region is around the interaction point, where there are three high resolution cavity beam position monitors IPA, IPB and IPC, with the interaction point nominally falling midway between IPB and IPC. Just upstream we have one stripline kicker, IPK, which can be used to correct the position of the beam. At ATF we typically work in 2-bunch mode for feedback experiments, with ILC-like bunch-to-bunch separations of a few hundred nanoseconds. Talitha Bromwich 14 March 2017

Timing of IPBPM signals Overlay of timing of noise floor I and Q signal measurements with the position signal- dependent scale factor data from a 0dB calibration taken immediately after. Position signal delayed by the band pass filters. Reference charge monitor delayed intentionally using a delay cable. Need additional delay on the reference being sent to the mixer. The FONT group now continues experiments with ATF2 collaborators to work towards the ATF2 goal 2 of achieving nm level beam stabilisation. This is much more challenging than the ILC requirements, as there is no beam-beam deflection, because there is only one beam at the test facility, so we rely on much higher resolution beam position monitors to directly measure beam misalignments. FONT works in two experimental regions at ATF: one upstream in the extraction line just after the damping ring, where there are three stripline BPMS: P1, P2 and P3, and two kickers, K1 and K2, so we can change the position and angle of the beam and stabilise the beam position all the way along the final focus. The second experimental region is around the interaction point, where there are three high resolution cavity beam position monitors IPA, IPB and IPC, with the interaction point nominally falling midway between IPB and IPC. Just upstream we have one stripline kicker, IPK, which can be used to correct the position of the beam. At ATF we typically work in 2-bunch mode for feedback experiments, with ILC-like bunch-to-bunch separations of a few hundred nanoseconds. Talitha Bromwich 14 March 2017

New set-up for better timing Move the delay before the limiter. Or somehow distribute delays between the limiter outputs for optimal timing. CURRENT SET-UP PROPOSED SET-UP The FONT group now continues experiments with ATF2 collaborators to work towards the ATF2 goal 2 of achieving nm level beam stabilisation. This is much more challenging than the ILC requirements, as there is no beam-beam deflection, because there is only one beam at the test facility, so we rely on much higher resolution beam position monitors to directly measure beam misalignments. FONT works in two experimental regions at ATF: one upstream in the extraction line just after the damping ring, where there are three stripline BPMS: P1, P2 and P3, and two kickers, K1 and K2, so we can change the position and angle of the beam and stabilise the beam position all the way along the final focus. The second experimental region is around the interaction point, where there are three high resolution cavity beam position monitors IPA, IPB and IPC, with the interaction point nominally falling midway between IPB and IPC. Just upstream we have one stripline kicker, IPK, which can be used to correct the position of the beam. At ATF we typically work in 2-bunch mode for feedback experiments, with ILC-like bunch-to-bunch separations of a few hundred nanoseconds. Talitha Bromwich 14 March 2017

Linear operating range In the past we have achieved linear calibrations by remaining within ± 0.3 V for the dipole cavity signals (this corresponds to approximately ± 3mm at a charge of 1x1010, which is the recommended linear operating range for the electronics from Honda’s Inoue et al PRSTAB, 11, 062801 (2008)). 214 = 16384 possible ADC values on SIS digitiser. For the 2V and 5V ADC voltage range settings on the SIS digitser, this linear range corresponds to: To prevent satuaration, this allows approximately: ± 2,500 ADC counts (@ 2V) ± 1000 ADC counts (@ 5V) Note: the FONT board has a range of 8000 possible ADC values, and a voltage range of 1V, so the linear operating range in terms of ADCs is very similar to SIS 2V setting. The FONT group now continues experiments with ATF2 collaborators to work towards the ATF2 goal 2 of achieving nm level beam stabilisation. This is much more challenging than the ILC requirements, as there is no beam-beam deflection, because there is only one beam at the test facility, so we rely on much higher resolution beam position monitors to directly measure beam misalignments. FONT works in two experimental regions at ATF: one upstream in the extraction line just after the damping ring, where there are three stripline BPMS: P1, P2 and P3, and two kickers, K1 and K2, so we can change the position and angle of the beam and stabilise the beam position all the way along the final focus. The second experimental region is around the interaction point, where there are three high resolution cavity beam position monitors IPA, IPB and IPC, with the interaction point nominally falling midway between IPB and IPC. Just upstream we have one stripline kicker, IPK, which can be used to correct the position of the beam. At ATF we typically work in 2-bunch mode for feedback experiments, with ILC-like bunch-to-bunch separations of a few hundred nanoseconds. +0.3 V 2,457 ADCs 983 ADCs - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - = 4,915 ADCs @ 2 V = 1,966 ADCs @ 5 V 0 ADCs -0.3 V -2,457 ADCs -983 ADCs Talitha Bromwich 14 March 2017

Examples of saturated ADC levels ADC region within the limits of saturating the processing electronics is shaded grey. Everything outside this region is operating in a non-linear regime that saturates the electronics. This is the data set that gives ~12 nm resolution when you fit all parameters. jitRun33_0dB_0.95_ipbpm_160316 Charge ~ 0.9 x 1010 Att Y sig = 0 dB ADC voltage range setting in SIS DAQ = 5V The FONT group now continues experiments with ATF2 collaborators to work towards the ATF2 goal 2 of achieving nm level beam stabilisation. This is much more challenging than the ILC requirements, as there is no beam-beam deflection, because there is only one beam at the test facility, so we rely on much higher resolution beam position monitors to directly measure beam misalignments. FONT works in two experimental regions at ATF: one upstream in the extraction line just after the damping ring, where there are three stripline BPMS: P1, P2 and P3, and two kickers, K1 and K2, so we can change the position and angle of the beam and stabilise the beam position all the way along the final focus. The second experimental region is around the interaction point, where there are three high resolution cavity beam position monitors IPA, IPB and IPC, with the interaction point nominally falling midway between IPB and IPC. Just upstream we have one stripline kicker, IPK, which can be used to correct the position of the beam. At ATF we typically work in 2-bunch mode for feedback experiments, with ILC-like bunch-to-bunch separations of a few hundred nanoseconds. Talitha Bromwich 14 March 2017

Calibration quality with saturation Calibration with saturation. Charge ~ 0.9 x 1010  March 2016 The FONT group now continues experiments with ATF2 collaborators to work towards the ATF2 goal 2 of achieving nm level beam stabilisation. This is much more challenging than the ILC requirements, as there is no beam-beam deflection, because there is only one beam at the test facility, so we rely on much higher resolution beam position monitors to directly measure beam misalignments. FONT works in two experimental regions at ATF: one upstream in the extraction line just after the damping ring, where there are three stripline BPMS: P1, P2 and P3, and two kickers, K1 and K2, so we can change the position and angle of the beam and stabilise the beam position all the way along the final focus. The second experimental region is around the interaction point, where there are three high resolution cavity beam position monitors IPA, IPB and IPC, with the interaction point nominally falling midway between IPB and IPC. Just upstream we have one stripline kicker, IPK, which can be used to correct the position of the beam. At ATF we typically work in 2-bunch mode for feedback experiments, with ILC-like bunch-to-bunch separations of a few hundred nanoseconds. Calibration without saturation. Charge ~ 0.5 x 1010  February 2017 Talitha Bromwich 14 March 2017

New Results Latest resolution estimates New two-BPM IP feedback stabilisation results Talitha Bromwich 14 March 2017

Resolution of 3 BPM IP system Use signals at two BPMs to predict the position at the third. Compare prediction with what was measured at the third BPM  resolution of the 3BPM system. METHODS: These methods use the position signals the FONT feedback system utilises to stabilise the beam, so are a measure of the current useable resolution for IP stabilisation. If the system is behaving linearly they should agree, as the BPM separations are well known. Geometric: Use known separations of the BPMs to predict vertical position at the third BPM based on the position information at the other two. Fitting: Ignore known geometric constraints, and just apply the best fit to the vertical position data at two BPMs to predict the position at the third. Multi-parameter fitting: Instead of calculating positions, just fit to the raw digitised signals to predict vertical position at the third. Allows you to introduce many more parameters such as the charge, X-information, any residual position in Q’ signal etc. The FONT group now continues experiments with ATF2 collaborators to work towards the ATF2 goal 2 of achieving nm level beam stabilisation. This is much more challenging than the ILC requirements, as there is no beam-beam deflection, because there is only one beam at the test facility, so we rely on much higher resolution beam position monitors to directly measure beam misalignments. FONT works in two experimental regions at ATF: one upstream in the extraction line just after the damping ring, where there are three stripline BPMS: P1, P2 and P3, and two kickers, K1 and K2, so we can change the position and angle of the beam and stabilise the beam position all the way along the final focus. The second experimental region is around the interaction point, where there are three high resolution cavity beam position monitors IPA, IPB and IPC, with the interaction point nominally falling midway between IPB and IPC. Just upstream we have one stripline kicker, IPK, which can be used to correct the position of the beam. At ATF we typically work in 2-bunch mode for feedback experiments, with ILC-like bunch-to-bunch separations of a few hundred nanoseconds. Possible insight into which signals contribute vertical position information. Invariably gives a lower estimate of the resolution because you have so many free fitting parameters. However, vertical position is never calculated this way operationally. Talitha Bromwich 14 March 2017

Latest resolution measurements In February we repeated resolution studies with X and Y information at different attenuation settings, with improved BPM alignment, operating at a lower charge of 0.5 x 1010 to ensure we were not saturationg the electronics. Unfortunately the hardware delay cable set-up was still in the previously mentioned configuration we have now identified is poor for timing. Plot by Rebecca Ramjiawan The FONT group now continues experiments with ATF2 collaborators to work towards the ATF2 goal 2 of achieving nm level beam stabilisation. This is much more challenging than the ILC requirements, as there is no beam-beam deflection, because there is only one beam at the test facility, so we rely on much higher resolution beam position monitors to directly measure beam misalignments. FONT works in two experimental regions at ATF: one upstream in the extraction line just after the damping ring, where there are three stripline BPMS: P1, P2 and P3, and two kickers, K1 and K2, so we can change the position and angle of the beam and stabilise the beam position all the way along the final focus. The second experimental region is around the interaction point, where there are three high resolution cavity beam position monitors IPA, IPB and IPC, with the interaction point nominally falling midway between IPB and IPC. Just upstream we have one stripline kicker, IPK, which can be used to correct the position of the beam. At ATF we typically work in 2-bunch mode for feedback experiments, with ILC-like bunch-to-bunch separations of a few hundred nanoseconds. Talitha Bromwich 14 March 2017

Resolution results example 0dB attenuation at a charge of 0.55x1010 714 MHz BPFs in place. Single sample, as used for FB Integrating 5 samples Parameter Geometric Fitting Multi-parameter fits No. param 2 3 5 6 11 13 Parameters used to predict vertical position at 3rd BPM. Y1 Y2 Y1 Y2 + const. Y1I’ Y2I’ Y1Q’ Y2Q’ + const. Y1I’ Y2I’ Y1Q’ Y2Q’ + Y Ref charge + const. Y1I’ Y2I’ Y1Q’ Y2Q’ + Y Ref charge X1I’ X2I’ X1Q’ X2Q’ + X Ref charge + const Y1I’ Y2I’ Y1Q’ Y2Q’ + Y Ref charge X1I’ X2I’ X1Q’ X2Q’ X3I’ X3Q’ + X Ref charge + const IPA Res (nm)  69  47 46 41 39 IPB Res (nm)  46 38 36 IPC Res (nm) 69 49 23 22 21  63  42 40 35 34 32 30 63 45 19 18 The FONT group now continues experiments with ATF2 collaborators to work towards the ATF2 goal 2 of achieving nm level beam stabilisation. This is much more challenging than the ILC requirements, as there is no beam-beam deflection, because there is only one beam at the test facility, so we rely on much higher resolution beam position monitors to directly measure beam misalignments. FONT works in two experimental regions at ATF: one upstream in the extraction line just after the damping ring, where there are three stripline BPMS: P1, P2 and P3, and two kickers, K1 and K2, so we can change the position and angle of the beam and stabilise the beam position all the way along the final focus. The second experimental region is around the interaction point, where there are three high resolution cavity beam position monitors IPA, IPB and IPC, with the interaction point nominally falling midway between IPB and IPC. Just upstream we have one stripline kicker, IPK, which can be used to correct the position of the beam. At ATF we typically work in 2-bunch mode for feedback experiments, with ILC-like bunch-to-bunch separations of a few hundred nanoseconds. Talitha Bromwich 14 March 2017

FONT IP feedback – two-BPM The system has now been enhanced with feedback firmware that uses weighted positions from two BPMs to stabilise the beam at an intermediate location. Beam stabilisation to 133 nm was demonstrated with this feedback system in October 2016 (presented at LCWS2016, Morioka). In October we performed a new study using two BPMs as input to the feedback to correct at a third intermediate location. So as shown in the block diagram we used IPA and IPC as inputs to the feedback system. So, assuming a straight line trajectory through the feedback system, the position signals from these BPMs can be weighted by setting feedback gains in the firmware to correct the position at any pre-determined intermediate location between them. For this set-up we chose the correction location to be at IPB, so we could have a measure of the correction. So position information from IPB is digitised synchronously with IPA and IPC, but the information from that BPM is not used in the feedback algorithm, so it provides an independent witness of the correction. It is vital for this study that the beam passes through the linear operating range (5 μm) of all three dipole cavity BPMs simultaneously. This is achieved by carefully adjusting the BPMs in both the vertical and horizontal planes using the piezo mover system they are mounted on, until the beam signals are minimised, indicating the beam is passing through the centre of the cavities. This has never been possible in the past, as offsets in the relative vertical alignment of the three BPMs were larger than the extremes of the BPM mover ranges. Careful iterative alignment studies and physical adjustments to the IP chamber over the past year have now culminated in a successful alignment set-up that makes this simultaneous centring possible – so we thank our LAL and ATF2 colleagues for their work on this. It is also important that the beam jitters at the three BPMs are within the operating range. With nominal IP optics the jitter either side of the tuned waist is larger than the dynamic range at the BPMs furthest from the IP (i.e. IPA and IPC). With a single-loop study set-up, shown previously, the waist location is simply shifted very close to (or even directly at) the BPM being used for correction. For this study a simple longitudinal waist adjustment is not sufficient, as it is necessary to have useable beam jitters at all three BPMs, necessitating a change on the beam optics. Feedback Bunch 2 jitter (nm) Off 264 ± 19 On 133 ± 10 The two-BPM feedback configuration, where IPA and IPC are used as inputs to the feedback and IPB information is used as an independent witness to the correction. Talitha Bromwich 14 March 2017

High-β optics set-up Consequently the optics used for this study form a βy∗ that is 1000 times larger than nominal and is subsequently known as high-beta optics, allowing a much lower divergence of the beam around the IP. What discernible waist remains is then moved closer to IPB to produce similar sized jitters at IPA and IPC. The plots here from data on shift operating with high-beta optics. The plot on the left shows many triggers across the course of a data gathering run, where the measured positions at IPA and IPC have been interpolated to reconstruct the orbit across the intervening distance, including at IPB. This method is used again later to predict feedback performance at IPB, based on measurements at IPA and IPC. This plot shows there is a discernible waist around -35 mm from the nominal IP, but it has been significantly widened by the high-beta optics tuning. The right-hand side plot shows the position jitters interpolated between IPA and IPC, but also between IPA and IPB and between IPB and IPC, using the witness IPB position information. The interpolated IPB jitter will always be less than the measured IPB jitter, because two measurements are being combined, improving the effective resolution limitation on measuring the true beam jitter. This shows that for this study we had jtiters at around 0.4 um at IPA and IPC, and around 0.25 um at IPB. βy* 1000 times larger than nominal means lower IP beam divergence and enables beam position jitters that are within operating range at all three BPMs simultaneously. Talitha Bromwich 14 March 2017

Two-BPM jitter stabilisation result The best feedback run in October stabilised the beam at IPB to 133 nm with 0dB attenuation. The best feedback run in February stabilised the beam at IPB to 83 nm with 10dB attenuation, charge of ~0.85 x 1010. Geometric resolution was calculated to be 74 ± 4 nm for this data set. Bunch-to-bunch correlation 90% at IPA, 91% at IPC and 92% at IPB. The variable of interest in this experiment is the position jitter, which is best visualised in a histogram of vertical position distributions. Here bunch 1 always has a jitter of approximately 260 nm, but for bunch 2 the 267 nm jitter is reduced to 133 nm when the feedback is turned on. This result was a very positive step as a first demonstration of the two-BPM feedback mode operating successfully, achieving the required BPM alignment and useable jitters at all three simultaneously, however the 133 nm correction falls short of the 73 nm best beam stabilisation previously achieved with single-loop feedback at the IP, and the typical stablilisation on this shift was 150 to 250 nm. To understand why the performance was not as good, it is important to look at the bunch-to-bunch position correlation performance, which is vital for successful feedback, as we rely on bunch 1 being a good prediction of the position in bunch 2. Feedback Bunch 1 jitter (nm) Bunch 2 jitter (nm) Off 265 ± 20 253 ± 19 On 252 ± 19 83 ± 6 ?????? Talitha Bromwich 14 March 2017

Summary Operational Challenges Sample timing jumps – under investigation Unwanted parasitic signal source still unknown – currently soothed using BPFs. Delay cable set-up in IP region is not optimised for best timing. Optimise for next operation. Need to take care with signal levels to avoid non-linear calibrations and saturation effects. New Results Latest resolution results suggest a useable geometric resolution of ~70 nm (60 nm with integration). This result is reduced to 20/30 nm using multi-parameter fits. X information appears to do very little to improve the result. New two-BPM IP feedback mode has been used to stabilise the beam at the IP to 83 nm with 10dB attenuation at a charge of 0.85x1010. Talitha Bromwich 14 March 2017

Appendix Talitha Bromwich 14 March 2017

IP BPMs & associated electronics CAVITY BPMs (KNU) 6.4 GHz in Y, 5.7 GHz in X 3 Dipole Cavity BPMs: Aluminium Mounted on LAL piezo-mover system IPA/IPB/IPC installed Nov 2014 IPA/IPB QL ~459, decay time ~ 12 ns IPC QL ~270, decay time ~7 ns New IPC with indium sealing installed April 2016 New IPC QL ~700, decay time ~19 ns Reference Cavity BPM Stainless steel Tuning pin to adjust monopole mode frequency ±35 MHz to match dipole cavities The FONT group now continues experiments with ATF2 collaborators to work towards the ATF2 goal 2 of achieving nm level beam stabilisation. This is much more challenging than the ILC requirements, as there is no beam-beam deflection, because there is only one beam at the test facility, so we rely on much higher resolution beam position monitors to directly measure beam misalignments. FONT works in two experimental regions at ATF: one upstream in the extraction line just after the damping ring, where there are three stripline BPMS: P1, P2 and P3, and two kickers, K1 and K2, so we can change the position and angle of the beam and stabilise the beam position all the way along the final focus. The second experimental region is around the interaction point, where there are three high resolution cavity beam position monitors IPA, IPB and IPC, with the interaction point nominally falling midway between IPB and IPC. Just upstream we have one stripline kicker, IPK, which can be used to correct the position of the beam. At ATF we typically work in 2-bunch mode for feedback experiments, with ILC-like bunch-to-bunch separations of a few hundred nanoseconds. Talitha Bromwich 14 March 2017

IP BPMs & associated electronics FIRST STAGE MIXER Down-mixes the GHz frequency signals with frequency multiples of a common 714 MHz source to produce signals that are all at 714 MHz. The FONT group now continues experiments with ATF2 collaborators to work towards the ATF2 goal 2 of achieving nm level beam stabilisation. This is much more challenging than the ILC requirements, as there is no beam-beam deflection, because there is only one beam at the test facility, so we rely on much higher resolution beam position monitors to directly measure beam misalignments. FONT works in two experimental regions at ATF: one upstream in the extraction line just after the damping ring, where there are three stripline BPMS: P1, P2 and P3, and two kickers, K1 and K2, so we can change the position and angle of the beam and stabilise the beam position all the way along the final focus. The second experimental region is around the interaction point, where there are three high resolution cavity beam position monitors IPA, IPB and IPC, with the interaction point nominally falling midway between IPB and IPC. Just upstream we have one stripline kicker, IPK, which can be used to correct the position of the beam. At ATF we typically work in 2-bunch mode for feedback experiments, with ILC-like bunch-to-bunch separations of a few hundred nanoseconds. Talitha Bromwich 14 March 2017

IP BPMs & associated electronics 714 MHz BPFs Smooth unwanted parasitic signals in pulse shape The FONT group now continues experiments with ATF2 collaborators to work towards the ATF2 goal 2 of achieving nm level beam stabilisation. This is much more challenging than the ILC requirements, as there is no beam-beam deflection, because there is only one beam at the test facility, so we rely on much higher resolution beam position monitors to directly measure beam misalignments. FONT works in two experimental regions at ATF: one upstream in the extraction line just after the damping ring, where there are three stripline BPMS: P1, P2 and P3, and two kickers, K1 and K2, so we can change the position and angle of the beam and stabilise the beam position all the way along the final focus. The second experimental region is around the interaction point, where there are three high resolution cavity beam position monitors IPA, IPB and IPC, with the interaction point nominally falling midway between IPB and IPC. Just upstream we have one stripline kicker, IPK, which can be used to correct the position of the beam. At ATF we typically work in 2-bunch mode for feedback experiments, with ILC-like bunch-to-bunch separations of a few hundred nanoseconds. Talitha Bromwich 14 March 2017

IP BPMs & associated electronics REFERENCE LIMITER Reference signal limited to a constant voltage to remove charge dependence. A copy of the reference (unlimited) is also kept as a beam intensity monitor. The FONT group now continues experiments with ATF2 collaborators to work towards the ATF2 goal 2 of achieving nm level beam stabilisation. This is much more challenging than the ILC requirements, as there is no beam-beam deflection, because there is only one beam at the test facility, so we rely on much higher resolution beam position monitors to directly measure beam misalignments. FONT works in two experimental regions at ATF: one upstream in the extraction line just after the damping ring, where there are three stripline BPMS: P1, P2 and P3, and two kickers, K1 and K2, so we can change the position and angle of the beam and stabilise the beam position all the way along the final focus. The second experimental region is around the interaction point, where there are three high resolution cavity beam position monitors IPA, IPB and IPC, with the interaction point nominally falling midway between IPB and IPC. Just upstream we have one stripline kicker, IPK, which can be used to correct the position of the beam. At ATF we typically work in 2-bunch mode for feedback experiments, with ILC-like bunch-to-bunch separations of a few hundred nanoseconds. Talitha Bromwich 14 March 2017

IP BPMs & associated electronics SECOND STAGE MIXER Combines the reference and dipole signals in phase (I), and in quadrature phase (Q), to isolate phase dependence. The FONT group now continues experiments with ATF2 collaborators to work towards the ATF2 goal 2 of achieving nm level beam stabilisation. This is much more challenging than the ILC requirements, as there is no beam-beam deflection, because there is only one beam at the test facility, so we rely on much higher resolution beam position monitors to directly measure beam misalignments. FONT works in two experimental regions at ATF: one upstream in the extraction line just after the damping ring, where there are three stripline BPMS: P1, P2 and P3, and two kickers, K1 and K2, so we can change the position and angle of the beam and stabilise the beam position all the way along the final focus. The second experimental region is around the interaction point, where there are three high resolution cavity beam position monitors IPA, IPB and IPC, with the interaction point nominally falling midway between IPB and IPC. Just upstream we have one stripline kicker, IPK, which can be used to correct the position of the beam. At ATF we typically work in 2-bunch mode for feedback experiments, with ILC-like bunch-to-bunch separations of a few hundred nanoseconds. Talitha Bromwich 14 March 2017

IP BPMs & associated electronics DELAY CABLE The reference diode used for charge normalisation is delayed to account for the delay on the dipole signals introduced by the BPFs, so all signals arrive at the digitiser at around the same time. The FONT group now continues experiments with ATF2 collaborators to work towards the ATF2 goal 2 of achieving nm level beam stabilisation. This is much more challenging than the ILC requirements, as there is no beam-beam deflection, because there is only one beam at the test facility, so we rely on much higher resolution beam position monitors to directly measure beam misalignments. FONT works in two experimental regions at ATF: one upstream in the extraction line just after the damping ring, where there are three stripline BPMS: P1, P2 and P3, and two kickers, K1 and K2, so we can change the position and angle of the beam and stabilise the beam position all the way along the final focus. The second experimental region is around the interaction point, where there are three high resolution cavity beam position monitors IPA, IPB and IPC, with the interaction point nominally falling midway between IPB and IPC. Just upstream we have one stripline kicker, IPK, which can be used to correct the position of the beam. At ATF we typically work in 2-bunch mode for feedback experiments, with ILC-like bunch-to-bunch separations of a few hundred nanoseconds. Talitha Bromwich 14 March 2017

IP BPMs & associated electronics FONT5 BOARD / SIS Digitises all the signals. The FONT group now continues experiments with ATF2 collaborators to work towards the ATF2 goal 2 of achieving nm level beam stabilisation. This is much more challenging than the ILC requirements, as there is no beam-beam deflection, because there is only one beam at the test facility, so we rely on much higher resolution beam position monitors to directly measure beam misalignments. FONT works in two experimental regions at ATF: one upstream in the extraction line just after the damping ring, where there are three stripline BPMS: P1, P2 and P3, and two kickers, K1 and K2, so we can change the position and angle of the beam and stabilise the beam position all the way along the final focus. The second experimental region is around the interaction point, where there are three high resolution cavity beam position monitors IPA, IPB and IPC, with the interaction point nominally falling midway between IPB and IPC. Just upstream we have one stripline kicker, IPK, which can be used to correct the position of the beam. At ATF we typically work in 2-bunch mode for feedback experiments, with ILC-like bunch-to-bunch separations of a few hundred nanoseconds. Talitha Bromwich 14 March 2017

IP BPMs & associated electronics FONT5 BOARD  IP kicker In feedback mode the FONT board uses one sample point in the first bunch pulse signals to calculate the necessary correction and apply a corrective kick to the second bunch in the train via a fast rise-time amplifier and stripline kicker, IPK. The FONT group now continues experiments with ATF2 collaborators to work towards the ATF2 goal 2 of achieving nm level beam stabilisation. This is much more challenging than the ILC requirements, as there is no beam-beam deflection, because there is only one beam at the test facility, so we rely on much higher resolution beam position monitors to directly measure beam misalignments. FONT works in two experimental regions at ATF: one upstream in the extraction line just after the damping ring, where there are three stripline BPMS: P1, P2 and P3, and two kickers, K1 and K2, so we can change the position and angle of the beam and stabilise the beam position all the way along the final focus. The second experimental region is around the interaction point, where there are three high resolution cavity beam position monitors IPA, IPB and IPC, with the interaction point nominally falling midway between IPB and IPC. Just upstream we have one stripline kicker, IPK, which can be used to correct the position of the beam. At ATF we typically work in 2-bunch mode for feedback experiments, with ILC-like bunch-to-bunch separations of a few hundred nanoseconds. Talitha Bromwich 14 March 2017

Noise floor measurements Place 70 dB attenuation on the dipole signal so there is no position signal. Do not attenuate the reference signal, so the electronics are still being driven. Use the observed jitter on output signals to estimate the noise floor of the system. 70dB 70dB 70dB The FONT group now continues experiments with ATF2 collaborators to work towards the ATF2 goal 2 of achieving nm level beam stabilisation. This is much more challenging than the ILC requirements, as there is no beam-beam deflection, because there is only one beam at the test facility, so we rely on much higher resolution beam position monitors to directly measure beam misalignments. FONT works in two experimental regions at ATF: one upstream in the extraction line just after the damping ring, where there are three stripline BPMS: P1, P2 and P3, and two kickers, K1 and K2, so we can change the position and angle of the beam and stabilise the beam position all the way along the final focus. The second experimental region is around the interaction point, where there are three high resolution cavity beam position monitors IPA, IPB and IPC, with the interaction point nominally falling midway between IPB and IPC. Just upstream we have one stripline kicker, IPK, which can be used to correct the position of the beam. At ATF we typically work in 2-bunch mode for feedback experiments, with ILC-like bunch-to-bunch separations of a few hundred nanoseconds. Talitha Bromwich 14 March 2017

Noise floor measurements Example outputs for the I and Q signals from IPC at 70dB attenuation, as well as the reference (non attenuated). Use the observed jitter of this signal to estimate the noise floor of the system in the location where the pulse would normally be with 0dB attenuation. You can then use calibrations taken at different attenuations to convert the jitter of this file into IPC position jitter in nm. The FONT group now continues experiments with ATF2 collaborators to work towards the ATF2 goal 2 of achieving nm level beam stabilisation. This is much more challenging than the ILC requirements, as there is no beam-beam deflection, because there is only one beam at the test facility, so we rely on much higher resolution beam position monitors to directly measure beam misalignments. FONT works in two experimental regions at ATF: one upstream in the extraction line just after the damping ring, where there are three stripline BPMS: P1, P2 and P3, and two kickers, K1 and K2, so we can change the position and angle of the beam and stabilise the beam position all the way along the final focus. The second experimental region is around the interaction point, where there are three high resolution cavity beam position monitors IPA, IPB and IPC, with the interaction point nominally falling midway between IPB and IPC. Just upstream we have one stripline kicker, IPK, which can be used to correct the position of the beam. At ATF we typically work in 2-bunch mode for feedback experiments, with ILC-like bunch-to-bunch separations of a few hundred nanoseconds. Talitha Bromwich 14 March 2017

Noise floor measurements The FONT group now continues experiments with ATF2 collaborators to work towards the ATF2 goal 2 of achieving nm level beam stabilisation. This is much more challenging than the ILC requirements, as there is no beam-beam deflection, because there is only one beam at the test facility, so we rely on much higher resolution beam position monitors to directly measure beam misalignments. FONT works in two experimental regions at ATF: one upstream in the extraction line just after the damping ring, where there are three stripline BPMS: P1, P2 and P3, and two kickers, K1 and K2, so we can change the position and angle of the beam and stabilise the beam position all the way along the final focus. The second experimental region is around the interaction point, where there are three high resolution cavity beam position monitors IPA, IPB and IPC, with the interaction point nominally falling midway between IPB and IPC. Just upstream we have one stripline kicker, IPK, which can be used to correct the position of the beam. At ATF we typically work in 2-bunch mode for feedback experiments, with ILC-like bunch-to-bunch separations of a few hundred nanoseconds. Plot the jitter of the 70 dB data set in ADCs as a function of sample number along the pulse. Sample in the region of the pulse from the calibration file: maximum signal ~ 38. Corresponds to ~ 7 ADC counts in I, ~ 9 ADC counts in Q at a charge ~0.5 x 1010 Convert to position jitter using a 0dB calibration file: 45 ± 3 nm. Talitha Bromwich 14 March 2017

Calibration quality with saturation Calibration with saturation 0dB attenuation Charge ~ 0.9 x 1010 IPAyCal1_0dB_0.95_ipbpm_160315 Calibration without saturation 0dB attenuation Charge ~ 0.5 x 1010 AQDOFFyScan1_0dB_{103:107}um_Board1_170217 The FONT group now continues experiments with ATF2 collaborators to work towards the ATF2 goal 2 of achieving nm level beam stabilisation. This is much more challenging than the ILC requirements, as there is no beam-beam deflection, because there is only one beam at the test facility, so we rely on much higher resolution beam position monitors to directly measure beam misalignments. FONT works in two experimental regions at ATF: one upstream in the extraction line just after the damping ring, where there are three stripline BPMS: P1, P2 and P3, and two kickers, K1 and K2, so we can change the position and angle of the beam and stabilise the beam position all the way along the final focus. The second experimental region is around the interaction point, where there are three high resolution cavity beam position monitors IPA, IPB and IPC, with the interaction point nominally falling midway between IPB and IPC. Just upstream we have one stripline kicker, IPK, which can be used to correct the position of the beam. At ATF we typically work in 2-bunch mode for feedback experiments, with ILC-like bunch-to-bunch separations of a few hundred nanoseconds. Talitha Bromwich 14 March 2017