Cold welds in cables of magnet 1005

Early May 2002, CERN. Cryomagnet 1005 is tested by MTA; quench field is limited at 6 T, with quenches in the straight part of Aperture 2, at 4 m from the connection side (see report).

Early May 2002, CERN. Warm magnetic measurements of 1005, both collared coil and cold mass are reviewed. Data show an anomalous coil waviness (70 microns against 15-25 microns measured in normal coils) in Aperture 2, with two main peaks at 4 m from connection side and non-connection side respectively. the worst one being on the non-connection side. A warning was given after the measurement of the collared coil to the project engineer (G. Spigo). Worse coil waviness was then observed on 1006, and an improvement was found from collared coil 1007. Nevertheless, the relation between the anomalous coil waviness and the poor quench performance of 1006 was considered doubtful.

May 2002, Manufacturer 1. The cold mass is disassembled, and the field quality of the collared coil is measured again by V. Remondino, with a scan every 150 mm to detect anomalies. Results show that the field quality is exactly the same as it was measured in the collared coil in July 2001 (before final assembly and cold test).

June 2002, Manufacturer 1. The cold mass is disassembled in presence of A. Devred: the outer shell is cut, the yoke and collars are removed, and the poles of the two apertures are separated. No defects are visible.

July 2002, Manufacturer 1. The two layers of the pole where the quench was localized are separated and unwound. Coil heads are visually inspected and show proper positioning of conductor.

Conductor insulation is removed and conductor is inspected. A section (corresponding to the outer layer pole turn) shows numerous places where the staybrite coating is absent.

Cable is sent back to CERN, laboratory ??? and cold welds are found in at least 28 strands of the cable

August 2002, CERN. Cable section with cold welds is cut and critical current is measured in MFRESCA. The measured voltage-current caracteristique is consistent with the observed quench current in the magnet.

The magnet disassembly confirmed the absence of relation between the anomalous coil waviness and the poor quench performance.