Beam watch: fighting noise in the Beam Conditions Monitor

3 November 2008

In the center of the photo, four diamond detectors are mounted on a cruciform circle around the beam pipe (courtesy of Heinz Pernegger)

Along the four rods that help support the pixel detector along the beam pipe, seated on cruciform circles 1.8 metres on either side of the interaction point, eight diamond detectors keep an eye on any stray beam particles. “The real main purpose for this detector is to protect the ID from beam incidents,” says Daniel Dobos of the Beam Conditions Monitor (BCM) team.

The BCM takes time-of-flight measurements of the beam with a sampling rate 64 times higher than the bunch-crossing rate in ATLAS. While particles coming from the interaction point will strike all detectors roughly at the same time, particles from beam incidents will run through a detector on one side of the interaction point and then strike another at the opposite side of the pixel detector 12.5 nanoseconds later. 

The BCM decides whether so many stray particles are coming through that the Inner Detector could be damaged. If the ID is in danger, it sends an “abort beam” signal to the LHC so that the protons (or lead ions) are dumped.

The diamond detectors are made of polycrystalline diamond pads, cut from wafers grown from plasma through Chemical Vapour Deposition (CVD). Electron-hole pairs are generated in the material by charged particles or photons passing through. However, even when the detector is hidden from light, some electron-hole pairs are created and contribute to the leakage current. “That’s why they call them dark currents:  they are measured in the dark,” Daniel explains.

“The energy levels are a bit different at the boundaries than in the diamond bulk,” says Andrej Gorisek, also on the BCM. “At the boundaries, you typically need less energy to excite an electron to become a free electron in the conductive band, leaving a hole in valence band.” The diamond pads contain many boundaries, so they tested ways of controlling the currents as early as two years ago.

The currents aren’t such a problem in themselves. The average dark current is orders of magnitude smaller than current produced by particles traversing ATLAS. However, dark currents in pCVD diamonds can fluctuate – erratic dark currents – and these larger signals have the potential to disrupt the measurements of the BCM.

Well, they would, but luckily the erratic fluctuations disappear when a magnetic field is present, perpendicular to the electric field generated by the voltage across the detector. When the solenoid is on, as it will be when ATLAS runs, the BCM works fine. “Moreover, the BCM relies on excellent timing capabilities of its detectors, which allows them to ignore single detector noise hits by requiring coincidence of hits in more than one detector,” says Andrej.

At the end of August, the team discovered a new problem – unexpected noise spikes in the electronic readout system. They discovered the source of the noise just a week before first beam. “Whenever someone was telephoning close to our rack, we saw noise spikes,” says Daniel.

As USA15 will be accessible while ATLAS is running, the team set about shielding their cables and racks from mobile telephone signals the next day, finishing the evening of September 9th. The temporary shielding is to be replaced during the shut-down, according to Daniel.

The BCM team has solved their detector’s noise issues, and it will be ready to protect the ID once beam starts circulating again next spring.


Katie McAlpine

ATLAS e-News