UPGRADED CMS FAST BEAM CONDITION MONITOR FOR LHC RUN 3 ONLINE LUMINOSITY AND BEAM INDUCED BACKGROUND MEASUREMENTS
Reference intro slides:
Laboratory measurements
Fig. 1: The Fast Beam Condition Monitor (BCM1F) sensor pad capacitance measured as a function of the applied high voltage bias. The plot is an example from the measurement campaign performed prior to the BCM1F detector assembly - only best quality components were used. The plateau is observed after full depletion voltage around -260 V. Example samples of sensor pads are shown, left and right denotes the side of the tested sample.
3
BCM1F measured beam structure in LHC orbit
Fig. 3: Example raw BCM1F RHU counts per-channel, per lumi-section (LS~22.5 s), (left) LHC full orbit signal, (right) zoomed into a part of the orbit including individual bunch followed by a bunch train. Channel 0 in fill 8094 is shown, data collected over 6 hours is used, normalized per lumi-section (LS). In the very beginning of the orbit low intensity signal is visible, which is caused by the beam induced background signal of the non-colliding bunches. Afterglow tails are visible after each set of bunch trains.
4
BCM1F measured beam structure in LHC orbit - log scale
Fig. 2: Example raw BCM1F VME counts per-channel, per lumi-section (LS~23.6 s), (left) LHC full orbit signal, (right) zoomed into a part of the orbit including individual bunch followed by a bunch train. BCM1F VME Channel 0 is shown, data was collected over 6 hours, and normalized per lumi-section (LS). In the very beginning of the orbit low intensity signal is visible, which is caused by the beam induced background signal of the non-colliding bunches. Afterglow tails are visible after each set of bunch trains.
5
Afterglow correction
Fig. 3: First Run 3 afterglow correction model based on the two well-separated single colliding bunches. Left and Right plots are the same with different Bunch IDs (BCIDs) range. The exponential function was fitted using to the BCM1F raw data (with full time granularity of 6.25 ns), aggregated over many orbits and normalized to the collision counts (thus factor for first BCID is equal to 1). The colliding and the subsequent BCIDs were used which include counts from the afterglow.
6
BCM1F measured background per beam
Fig. 4: BCM1F average measured background rate over the fill 7575 from single circulating nominal bunches. Beam 1 was filled a Bunch ID (BCID) 101 and Beam 2 at BCID 1001. BCM1F background signal measured correspondingly by the channels on +Z (blue) and -Z (orange) side of the CMS. Data is shown in LHC orbit bins to indicate differences in time of arrival - the smaller peak corresponds to the incoming beam signal, whereas the bigger one to the outgoing. Data is not normalized to the bunch current sum.
7
BCM1F uTCA raw signal
Fig. 5: BCM1F uTCA sampled hit signal (blue), overlapped with corresponding derivative (orange) calculated with the smooth noise-robust differentiator. The detected peak at derivative zero-crossing is marked with red dashed vertical line. BCM1F uTCA back-end electronics is planned as the replacement for the baseline VME RHU. The new peak finder algorithm uses the derivative based threshold, which is designed to differentiate the overlapping pulses to maintain the detection efficiency at high pile-up.
8
BCM1F uTCA baseline derivative
Fig. 6: Samples of the empty orbits BCM1F uTCA baseline were taken and derivatives were calculated with the smooth noise-robust differentiator. Distribution of maximum derivative per orbit are shown for (left) a good and (right) a noisy channel. This derivative corresponds to the maximum slope gradient of the raw pulse rising edge shown in Fig. 5. The noisy channel has a high frequency noise contribution which results in the higher derivative values.
9
BCM1F uTCA amplitude spectra
Fig. 7: BCM1F uTCA per-channel amplitude spectrum. The distribution is cut at the amplitude threshold of 11. Data was collected for 10 h during fill 8118. The signal amplitude spectrum was built using the derivative-based peak finding algorithm. The main peak at 22 ADCs corresponds to the minimally ionizing particle (1 MIP) signal and that around 44 ADCs to 2 MIPs.
10
BCM1F measured background
Fig. 8.1: BCM1F measured Beam 1 and 2 background (lower plot) over a LHC fill cycle, together with the CMS vacuum pressure measurement close to the BCM1F location. Hot cathode ionisation (Bayard Alpert) gauges (VGI) pressure measurements are shown at 183 and 220 cm from IP, on both ends. The BCM1F backgrounds is composed of the beam induced background (BIB) from the non-colliding bunches as well as the incoming signal prior to the well separated colliding bunches. Different beam modes indicated in the upper plot legend are used in both plots with the corresponding background shades. In the upper plot luminosity signal is overlaid with per beam energy over the same period. Beam 1 total intensity was lower than Beam 2 in the presented fill. The vacuum pressure degrades when collisions start (in Adjust), caused by the increase in beam gas interactions. At the end of the fill, there is an emittance scan - both vacuum and backgrounds change along the scan steps.
11
BCM1F measured background
12
Preliminary BCM1F uTCA per-channel calibration
Fig. 9: Calibration constants ππ£ππ for BCM1F ΞΌTCA channels. All 40 connected channels are plotted in groups, to separate the points. The channel number is indicated in the legend, the average value of all channels is also plotted for reference (in black). This preliminary calibration was obtained from an emittance scan. Gaussian fit was used to extract the calibration constant per bunch - the errors are defined from the fit parameters, no systematic error is included. The spread between good channels is within +/- 15% from the average ππ£ππ . The differences between bunches are visible, and can be only minimized in the especially designed vdM calibrations.
13
BCM1F measured orbit displacements
Fig. 10: BCM1F measured per-bunch orbit displacement in π₯ plane, extracted from the emittance scan in a single fill. The displacement is extracted from the Gaussian fit, as its mean. The full LHC orbit is shown, bunches are grouped by the number of collisions each of them undergoes, per beam (B1+B2). Legend starts with the symmetric case with ATLAS and CMS in collision (2+2), 3 collisions include additional IP either IP2 or IP8 (explicitly βwIP2β means βwith IP2β), for just B1 or B2 (depending on placement before or after β+β symbol), or both when indicated. Lastly, 4 means collisions at all IPs.
14
BCM1F measured orbit displacements
Fig. 11: BCM1F measured per-bunch orbit displacement, during an emittance scan. The displacement in transverse direction is extracted from the Gaussian fit to the separation scan in given direction, as its mean. The zoom into a group of trains is shown with regular per-train structure in both planes {π₯, π¦}, which is dependent on the number of Long-Range interactions per bunch.
15
BCM1F measured beam overlap
Fig. 12: Beam overlap profile for 9 bunches with with various emittances in fill 7915. Both {π₯, π¦} planes are shown (left, right, correspondingly), points were fitted with the double Gaussian function. In the legend the Bunch IDs (BCIDs) are indicated. The differences between bunches are more pronounced in the {π¦} plane - in this plane the overlap does not include the contribution from the crossing-angle and the bunch length at the CMS.
16
BCM1F online calibration measurement stability
Fig.13: Online calibration constant ππ£ππ measurement stability over an early Run 3 period at nominal conditions, obtained as the average over all bunches per fill. Each Scan on the horizontal axis represents a scan pair in both {π₯, π¦} planes, which are used to obtain ππ£ππ measurement. The double Gaussian (DG) fit function was used to extract the peak rate and overlap widths. BCM1F rate was corrected with the afterglow model and the beam induced background contribution based on the non-colliding bunch response as well as the prompt contribution just before the collision. The points were also color-coded depending on the average single bunch instantaneous luminosity (SBIL).
17
Beam profile
Fig. 14: Beam overlap profile for various π½β steps during luminosity levelling in fill 8007. An example bunch, namely bunch crossing ID (BCID) 5, is shown, for scans in both {x,y} planes. The levelling started with the largest π½β value, the beams are squeezed in steps over time. The scan ranges are set based on the nominal emittance of 3.5 um. Therefore, the scan points are measured at changing nominal separation steps, depending on the π½β value.
18
BCM1F linearity
Fig. 14: BCM1F per-bunch calibration constant ππ£ππ over a wide range of SBIL. Data was fitted with SG/DG and corrected for background. Three scans are shown at different beam conditions: first one (red) was performed at the beginning of the fill, second (orange) after all beta* levelling steps, and third 7 h after, at the end of the fill.
19