Aquilion

• Faulty 24 V power supply that feeds the HV multiplier
• Dried-out capacitors in the 24 V power supply mounted on the multiplier

1. Recalibrate the filament currents first, to rule out a calibration problem
2. Check the 24 V supply on the HV multiplier under load, not just at idle, and use the generator graphs
3. Inspect the HV cable, the silicone ring and the tube receptacle

Solved in two cases. One was fixed by replacing the 24 V power supply that feeds the HV multiplier. The other was fixed by replacing the dried-out capacitors in that supply, then verifying it under load. A spare supply is worth keeping. After the repair, run the manufacturer's generator and image checkout.

• Flat battery in the original timekeeper (measured 0.3 V)
• Fault in the starter (anode rotation drive). This is a probable cause only.

1. Measure the timekeeper battery voltage
2. Decode the generator register using the troubleshooting section of the service manual
3. Connect to the XC and test it
4. Check the starter that spins up the anode

Timekeeper case, solved: replaced the timekeeper, then recalibrated the filament currents. Starter case, probable: inspecting the starter (anode rotation drive) was recommended, and the cause was not confirmed.

• HV cable or X-ray tube problem (probable)
• Faulty temperature sensor on the tube (seen by one engineer)
• Generator power supply problem (suspected only)

1. Check the HV cable at the generator end
2. Review the generator graphs and match the error to the description in the service manual
3. Test systematically rather than guessing between the generator and the tube

Probable only. Start by checking the HV cable at the generator end; the fault is likely in the tube or the HV cable. For the intermittent error, one engineer found a faulty tube temperature sensor, but this was not confirmed for the case discussed. Only qualified staff should work on HV parts, and the manufacturer's checkout applies afterwards.

• Slipping electromagnetic brake between the motor and the gearbox, so the motor encoder and the gearbox-output encoder disagree (not confirmed)
• Stiff encoder bearing
• Inter-turn short in the motor windings, causing an overcurrent fault (mentioned as a similar case)
• Worn horizontal drive motor assembly

1. Remove, clean and refit the encoder, and check that it turns freely by hand
2. Remove mechanical obstructions; lubricate and adjust the rollers and bearings
3. Check the infrared sensors
4. Check the servo driver power supply and reset the KGSTM board
5. Connect a cable to the AC servo and read its logs, or read the error code from the LED blink pattern

All three cases are probable. In the stiff-bearing case, greasing the encoder bearing while warming it with a heat gun was followed by no repeat of the error. For the encoder-mismatch case, colleagues recommended replacing the whole drive (servo motor with gearbox). For a table that will not move during scans, replacing the horizontal drive motor assembly was recommended. After any table repair, check the table motion and positioning with the manufacturer's procedure.

• Damaged fibre-optic data cable (for the interference that starts mid-scan)
• Degraded detectors, or degraded boards/DAC in the signal chain before the detectors (for ring artifacts)
• Noise may be normal at high-resolution scan settings

1. Inspect the fibre-optic lines and cables
2. Repeat Air and NRA calibrations
3. Check the detector supply rails with an oscilloscope (one case measured 50 mV ripple against a 100 mV tolerance)
4. Scan the image quality phantom with the settings given in the maintenance manual and compare the standard deviation with the tolerance

All three cases are probable. For interference that starts mid-scan, find and repair the damaged fibre-optic cable. For persistent ring artifacts, colleagues recommended replacing the detectors first, with the detector electronics as an alternative suspect. For noise, run the phantom image quality check: the image may well be within tolerance.

• Incompatible disk types mixed in one RDD RAID array
• Wrong disk firmware revision
• The server needs a non-standard sector size and a specific disk firmware

1. Review the logs and the RAID array status
2. Confirm the disk type, interface and firmware revision before mixing disks
3. On the IDD server, check the disk revision and sector format

RDD (Aquilion 64), solved: use compatible Fibre Channel disks with firmware 0003. The 73 GB (BSX73-2232E) and 300 GB (BSX73-2396E) disks can be mixed in one array; the 36 GB disk (BSX73-1828E) cannot. IDD server, solved in one engineer's experience: an equivalent standard disk worked after being reformatted to 520 bytes per sector under Linux. In the RAID-crash case (probable), the disks were repaired or replaced. Keep the software media and manuals on site.

• Corrupted application software
• Damaged system disk on the D console
• Consoles cannot see each other on the internal network
• Badly installed sound card drivers
• Faulty reconstruction unit (for the HOLD transfer problem)

1. Confirm the reconstruction subsystem is fully booted (recon, boards, RDD and GCIFA all ready)
2. Check the HDD and memory, and run chkdsk and defragmentation
3. Check the internal network settings, the hub in the CPU, and the console network cards
4. Check the disks for errors and confirm the configuration files are readable

Solved: protocol selection came back after reinstalling the software (restoring a saved backup is an alternative), and voice commands came back after reinstalling the sound card drivers. Probable: for the D console, clone the disk from the S console and reconfigure it following the installation manual. For 'fsys can not run', check the inter-console network first. For the HOLD transfers, suspect the reconstruction unit.

• Fault on the PACS server side
• DICOM node not configured
• Lost node configuration

1. Run ping and a DICOM echo from the CT to the PACS
2. Send a test study to a different destination to isolate the fault
3. Check the disks and the DVD drive (in one case a disc was stuck in the drive)
4. Photograph the node configuration before any service work

Probable: the PACS-side fault was likely in the PACS server itself, so set up a second destination for testing. To add a workstation, register it as a DICOM node in the service settings and reboot; it then appears in the transfer list. Solved: the lost node links were re-entered from a photo of the previous configuration.

• Many back-to-back contrast studies with poor room cooling (office-grade air conditioners)
• After tube replacement: vibration from the tube insert, or a gantry imbalance

1. Monitor the tube heat percentage
2. Review the daily number of contrast studies and the room cooling
3. After tube replacement, find whether the vibration comes from the tube insert or from gantry balance, and confirm the tube is original

Probable. For overheating: avoid scheduling contrast studies back-to-back, let the tube cool (keeping it at about 40% heat or below), and reduce protocol phases where clinically acceptable. For vibration: if it comes from the tube insert, the insert must be scrapped. If it is a gantry balance problem, use the gantry balancing procedure (there is no separate rotor balancing procedure). Do not run the gantry at speed with abnormal vibration.