Speaker
Description
Timepix3 hybrid pixel detectors installed inside the CERN SPS beam pipe for a Beam Gas Ionisation profile monitor showed intermittent communication loss, resets and pixel-data corruption with intense multi-bunch beams. Initial investigations considered both radiation-induced single-event effects and electromagnetic interference; dedicated tests indicated an EMI origin. EMC laboratory studies identified plausible coupling paths but could not reproduce the full SPS failure mode. Mitigations showed that external cable shielding alone was insufficient, while strengthening the detector Faraday cage enabled stable operation. This contribution presents the observed effects, investigations, shielding evolution and lessons for operating digital electronics inside accelerator beam pipes.
Summary (500 words)
Beam Gas Ionisation (BGI) profile monitors provide continuous, non-destructive transverse beam profile measurements by detecting ionisation electrons produced when the beam interacts with residual gas. Historically, these electrons have been detected by amplification with a micro-channel plate, conversion to light with a phosphor screen, and transport through an optical vacuum feedthrough to an intensified camera. This chain suffers from non-uniform ageing, point-spread effects and analogue degradation, requiring cross-calibration and limiting its use as an independent diagnostic.
CERN has developed BGI monitors based on direct detection of ionisation electrons inside the accelerator beam pipe using Timepix3 hybrid pixel detectors. Removing the optical conversion chain enables high spatial and temporal resolution, opening the possibility of continuous bunch-by-bunch and turn-by-turn non-destructive profile measurements, relevant for bunch-resolved emittance measurements and injection matching studies.
A key challenge is that complex digital electronics must operate directly in the accelerator beam environment. The Timepix3 assemblies are ultra-high-vacuum compatible and enclosed in a Faraday cage, with a porous RF shield allowing 10 keV electrons to reach the sensor. Early prototypes operated successfully in the CERN PS. However, after installation in the SPS in January 2024, both SPS BGI instruments were affected by high-intensity AWAKE and LHC25ns beams. Symptoms included loss of synchronisation in the Timepix-to-GBTx communication, occasional Timepix resets and corruption of the pixel matrix data. The effect was not persistent, and operation could be recovered by restarting or reconfiguring the chip. Initially, both EMI and radiation-induced single-event effects were considered.
Investigations indicated that the problem was electromagnetic rather than radiation-induced: a Timepix detector adjacent to the beam pipe as a beam-loss monitor was unaffected, while an internal Timepix signal showed EMI coincident with the disturbance. EMC laboratory studies identified plausible coupling paths, including direct beam-to-detector coupling and RF emission from the high-voltage cable coupling into signal and low-voltage cables. However, the SPS symptoms could not be reproduced with the stretched-wire setup, limited by RF amplifier power.
During YETS 2024–2025, two SPS BGI instruments were modified: SPS BGI-H, which measures the horizontal beam profile, and SPS BGI-V, which measures the vertical profile. Both received external mitigations, including improved shielding and grounding of signal, control and high-voltage cables, with ferrites on the high-voltage cable. However, only SPS BGI-V also received a strengthened Faraday cage, consisting of a denser RF shield and cover below the instrument. This proved decisive: SPS BGI-H continued to suffer failures, whereas SPS BGI-V operated stably.
This demonstrates that digital electronics can operate inside the accelerator beam pipe, but also that wakefields generated by beam passage through an instrument with residual impedance can couple to Timepix electronics if shielding is insufficient. External cable mitigations alone were insufficient; the dominant vulnerability was leakage through, or around, the detector Faraday cage. Further laboratory measurements extending to 2 GHz are planned to reproduce the SPS beam-induced failure mode. If successful, this would provide a practical way to optimise the shielding design without long delays between vacuum interventions, and without risking changes that could make a working instrument non-functional.