SIGNA (1.5T family)

• Faulty gradient control board
• Oxidised contacts in the gradient controller board connectors
• Wear in the gradient amplifier or its power module (in one case, suspected supercapacitor wear in the lower module of the gradient cabinet)
• Heavy load on the Z axis during diffusion, which shows up weak gradient hardware first

1. Rule out the coil by testing it on another scanner
2. Compare prescan amplitude with a known-good system
3. Run a different gradient-intensive protocol to see whether the fault is gradient-related
4. Write down the exact error text (which axis, or which diffusion direction)
5. Swap the suspect control board between the three gradient axes. If the error moves with the board, the board is faulty
6. Swap gradient axes (for example Z with another axis) to see whether the fault follows the amplifier channel
7. Inspect the connectors on the gradient controller boards for oxidation

In one case, replacing the gradient control board fixed the fault. This was confirmed by the error moving with the board when it was swapped between axes. In another case, reseating or swapping the gradient control boards brought diffusion back, probably by clearing oxidised contacts, but the fix may only be temporary. Where gradient amplifier or power module wear is suspected, diagnose by swapping axes and then repair or replace the affected module.

• Gradient amplifier over-temperature (one case logged OverTemperature on the Y gradient), probably because the cooling circuit was contaminated
• Dirty chiller (lint and dust), which loses capacity in hot weather
• Litron pump: the nylon shear bushing is sheared, so the pump shows a normal temperature but the transmit coil is not cooled
• Room air conditioning too small for the scanner's actual heat output

1. Search the logs for OverTemperature events and note which gradient axis
2. Check the front-panel temperature of the Litron pump and confirm the transmit coil is actually being cooled
3. Check whether the pump's nylon shear bushing is sheared
4. Check what coolant is in the gradient amplifier circuit (factory fluid or topped up with distilled water)
5. Inspect and clean the chiller
6. Use 'select viewing level' so that all error levels are shown
7. Check that room cooling capacity matches the scanner's heat output into the equipment room and the RF room

Probable: flush the gradient cooling circuit (planned for the next scheduled maintenance), clean the chiller and do full cooling-system maintenance. If the Litron pump bushing is sheared, the advice was to replace the pump. Opening the cabinet door was used only as a temporary workaround.

• Oxidised contact in the coil connector or port on the table carriage (in one case the contact had been sitting in liquid)
• A pin in the connector with no continuity
• Break in the table cable harness
• Broken contact tabs on the coil itself

1. Swap in another coil to rule out the coil
2. Inspect the carriage coil port (port A) for oxidation and signs of liquid
3. Check continuity on every pin of the connector
4. Check the table cable harness for breaks

Probable: repair or replace the oxidised or open connector contact, or repair the break in the table harness. Broken coil contact tabs have been re-soldered and covered with heat-shrink, but because the coil is safety-critical this is a reviewer item (see flags).

• Interference on the mains supply (in one case it started after a regional substation switchover)
• Damaged filter on the RF plate (a coil or capacitor had broken loose inside)
• Unshielded lighting wiring in the scan room

1. Repeat the phantom spike check
2. Check scan room humidity, ventilation and temperature
3. Move a small flex coil around the room in expert mode to find where the noise comes from
4. Inspect the RF plate filters

Solved by replacing the scan room lighting wiring with shielded wiring and repairing or replacing the damaged RF plate filter. When working on the filter, hold the lower nut so the filter body does not turn.

• Configuration or installation state on the replacement ICN that does not match the system
• ICN part number or software version that does not match the system software

1. Confirm the new ICN powers up and its iDRAC responds
2. Check the ICN part number and that it matches the software version
3. Connect a VGA monitor to the ICN and watch the output during boot or installation to see where it stops
4. Reinstall the software (VRE)

In one case, a full factory reset from the reconstructor's Lifecycle Controller let the ICN configuration complete. For the motherboard-swap case the outcome was not reported. Watching the VGA output during installation was the recommended way to diagnose it.

• Table power supply failure (reported as the most common failure)
• Belt tension out of adjustment
• Clutch out of adjustment
• Actuator fault versus control board fault (needs to be isolated)

1. Check that the control board receives 24 V from the external supply
2. Load-test the table power supply (about 42–48 V output, located top right in the cabinet) with an electronic load
3. Isolate the actuator by powering it directly with 24 V from a separate source
4. Check position readings, belt tension and clutch adjustment (in normal operation the clutch should slip for about 2–3 s)

Probable: replace the table power supply if it fails the load test. For drive-engagement overrun, adjust the belt tension and the clutch.

• Flat battery (NVRAM battery module) on a battery-backed board
• Faulty network switch between the boards and the host

1. Ping the board and watch its status LED
2. Use the COM port to see whether the board keeps sending requests to the host
3. Before pulling any battery, check all connectors, including the backplane connectors
4. Note that a software reload alone did not fix the APS case

In one case, replacing a flat board battery fixed the fault. For the AGP board, the likely fix is to replace the NVRAM battery module and then reprogram the board BOOT. Removing the battery means the board has to be reflashed. For APS not answering ping, the likely cause was the network switch, and replacing it was recommended.

• Failed PIN diodes in the body coil (a colleague's suggestion, not confirmed)

1. Review the system logs before replacing anything

Not confirmed. Damaged body coil PIN diodes were suggested as the likely cause; check the logs first.

• Protocol settings (b-value, NEX) that don't suit the body coil's SNR

1. Collect example images
2. Review the protocol with the applications/clinical staff

Probable: a protocol issue rather than a hardware fault. Lowering the b-value or increasing NEX was recommended to improve SNR.