MAGNETOM Symphony

• Broken actuating lever on a table limit microswitch (error 134 case)
• Poor contact or bounce in the microswitches of the finger-pinch guard plate. Vibration on high-gradient sequences makes it show up
• Cold (cracked) solder joints in the joystick contacts
• Lift mechanism problem after users had turned the table by hand with the manual crank (cause not confirmed)

1. Inspect the table limit microswitches and their levers. In the error 134 case the system only flagged the fault after the table reached the point where it changes to vertical travel
2. Check the guard-plate microswitches for bounce or intermittent contact
3. Inspect and re-check the solder joints on the joystick contacts
4. For a lift stopping midway: take off the left cover to reach the gearbox near the front panel. The manual crank sits next to it. Ask whether the table was cranked by hand

Solved: re-soldered the joystick contacts and cured the bounce in the guard-plate microswitches. In the error 134 case, the broken microswitch lever was bent as a temporary fix so it reached its actuator. The recommended permanent fix was to move a microswitch over from another position. The cause of the lift stopping midway was not confirmed.

• Open contact on the pins of the joystick connector on the PTAB backplane
• Broken wires in the horizontal table cable chain shorting to the shield at a certain table position
• Blown or badly seated fuse F11. The black fuse holders often lose contact

1. In service mode, check that the switches respond
2. Swap board A512 for a known-good one to rule it out
3. Check the switch cable for continuity to the MRI chassis (look for shorts to ground)
4. Check the fuses in the black holders for seating and continuity

Solved: re-soldered the joystick connector on the PTAB backplane (communication error). Repaired or replaced the damaged wiring in the horizontal cable chain (repeated F11 failures). Replaced F11 (18 V error).

• Helium line not refastened to the filter box after a cold head replacement, so it was acting as an antenna
• PMU acting as the interference source
• Loose gradient cable fastening where the cables join the busbars on the magnet
• Unresolved case: possible interference through the power filters, especially on the table motor supply

1. Run the RF calibration and the spike test
2. Rule out the room lighting (LED lamps were swapped for halogen in one case), fans and other light sources
3. Check that the helium line is secured to and in good contact with the filter box, especially after cryogenic work
4. Disconnect the PMU to see if the interference goes away
5. Check the gradient cable fastening where the cables join the busbars
6. Check the power filters, especially on the table motor supply

Solved: refastening the helium line to the filter box removed the interference, and so did disconnecting the PMU. Probable: retightening the gradient cable fastening at the busbar joint. The dotted-line artifact case was not resolved.

• Faulty fiber-optic link or optical module to the reconstruction computer
• Faulty RX4 board. Unplugging the optical link to the reconstructor produces the same symptom
• Drive letter of the rawdata disk on the reconstructor had changed

1. Check the fiber-optic link between the system and the reconstructor
2. Review the logs (warnings and info messages) from the time of the hang
3. If more than one system is affected, look for anything they share
4. Ask whether a disk was added or an image reloaded, and check the rawdata drive letter on the reconstructor

Solved: replaced the optical link/module to the reconstructor (adjust passed after that). Replaced the RX4 board (adjust hangs). Set the rawdata drive letter back (errors at scan start).

• Output power of the fiber-optic transmitter dropping with age
• Fault in the optical channel for the gradient temperature signal

1. Swap in a spare D9 module and a spare optical cable
2. Try the spare fibers in the optical cable
3. Measure the transmitter output with a meter and compare it with the datasheet
4. Swap the optical transmitters on D9 with each other to see if the fault moves

Probable: replace the optical transmitter. A similar case was fixed by replacing the emitter. Swapping the emitters on D9 is a quick first test. Gradient temperature monitoring must be working again before clinical use.

• Leak at a filter box connection, draining into the partition between the equipment room and the magnet room
• Suspected crack in the RCA heat exchanger between the inner and outer circuits
• Contamination in the cooling circuit clogging the filters

1. Log every top-up
2. Inspect the cabinets and take off the magnet covers to look for leaks
3. Lay sheets of paper under the filter box connections and look for wet spots
4. Check or replace the RCA heat exchanger if the leak is not found

Solved: the leak was found from wet spots on paper laid under the filter box connections. The RCA heat exchanger was also recommended for checking or replacement. To prevent fouling, the field practice is to flush the circuit regularly (water, then a citric acid flush, then rinsing until clear, then a distilled water fill). Switch off the chiller before cleaning the filters on the ACS, or the cabinet can flood.

• Faulty MSUP shelf that was not original to this system (it came from another system)
• Alarm Box case (not confirmed): stuck System On button, open circuit, or key switch position

1. Review the MPCU logs to diagnose the MSUP
2. Alarm Box: press and work the System On button, check the circuit for an open, and check that the key switch is unlocked

Solved: replaced the whole MSUP shelf. The Alarm Box case was not resolved: the system was started from the CCA cabinet, after which the System On indicator appeared. The button and circuit were left to check.

• A phase lost upstream (possibly at the substation). The 100 V reading was voltage induced from the other phases
• Supply voltage rising when the substation is lightly loaded (nights, weekends)

1. Measure phase and line voltages at the transformer input
2. Monitor the stability of the 3 x 380 V supply over nights and weekends

Both causes were identified as problems with the building or utility supply. The supply side has to be fixed by the facility or utility. The cases do not record a repair on the scanner.

• Carbon contamination from a refurbished adsorber, leading to failure
• Ice under the cold head and in the cryostat channel after a cold head replacement took too long
• Power to the cryo system switched off overnight
• Cold head temperatures outside normal (single case, not confirmed)

1. Log and trend helium consumption
2. Check that the cryocompressor has continuous power (nobody switches it off at night)
3. Report the cryo temperatures and consumption to the manufacturer

Solved in the field by cryogenic service work: replaced the cold head, adsorber and lines, and in another case de-iced and refitted the cold head. Continuous power to the cryo system was restored. All cold head, cryostat, helium and magnet work must be done by the manufacturer or its authorised cryogenic service only.

• VA40A installer on an HP Z420 reconstructor needs three hard disks to create the correct partitions
• Disk names and drive letters not preserved when the image is restored

1. Make an image or backup before replacing any host disks
2. Check how many disks are fitted in the reconstructor before installing
3. Check the disk names and drive letters after restoring

Solved: the host was restored from an Acronis backup and the reconstructor was installed from scratch from its original CD. The VA40A install on the Z420 worked correctly once three disks were fitted. Probable: when restoring with Acronis, turn on the option to restore the disk name as well so the names and letters are kept. With no backup, use the original installation disk (BOS), which partitions the disks itself.