How a CT scanner is built: what stands still and what rotates
Place any CT complaint on the chain from mains to PACS, say whether the faulty part sits on the stationary side or on the rotor, and decide whether the cause lies before or after the patient.
One chain from the wall socket to the radiologist
A CT scanner looks like a single machine, but it is a chain of systems, and every complaint you will ever hear sits somewhere on that chain. Learn it in order, because the order is also the order in which energy and information flow.
Mains power enters a distribution cabinet. From there it feeds the stationary parts of the gantry, the patient table, the console and the image reconstruction computer, and it crosses onto the rotating part of the gantry. On the rotor, a high-voltage generator turns that power into the kilovolts and milliamps the X-ray tube needs. The tube produces a fan or cone of X-rays, which is shaped by filters and a collimator, passes through the patient, and lands on the detector on the opposite side of the rotor. The detector turns X-rays into small electrical currents; the data acquisition system (DAS), mounted right next to it, turns those currents into numbers. The numbers cross back from the rotor to the stationary side, the reconstruction computer turns millions of projections into slices, and the console sends the images to the archive and to PACS.
Write this chain from memory before you go further. When a radiographer calls and says 'the scanner stopped' or 'the images look strange', your first job is to put the symptom on one link of this chain. Everything else in this course hangs on it.
- Mains supply and distribution cabinet
- Power transfer onto the rotor
- High-voltage generator (on the rotor)
- X-ray tube, filters, bowtie filter, collimator
- The patient
- Detector and data acquisition system (on the rotor)
- Data transfer off the rotor
- Reconstruction computer, console, archive, PACS
The rotation boundary: the most important line in the machine
The gantry has a stationary frame (the stator) and a large rotating ring (the rotor). On a modern scanner the rotor carries the generator, the tube, the collimator, the detector and the DAS, together several hundred kilograms of equipment, and spins it at a fraction of a second per revolution. At those speeds every component on the rotor is pushed outwards with a force many times its own weight. That is why bolts, connectors and balance weights on the rotor matter more than on any other medical device you will service.
No cable can follow a ring that turns continuously in one direction. Only two things cross the rotation boundary, and they cross in very different ways. Power goes onto the rotor through a slip ring: conductive tracks on one side and brushes sliding on them on the other, a mechanical rubbing contact. Data comes off the rotor through a contactless link, optical or capacitive or radio-frequency depending on the design, which never touches anything. Some designs also transfer power without contact; you will find out which type you have from the manufacturer's documentation.
Remember one link between these two parts, because it explains a whole family of faults. Where brushes are used, they wear, and their wear produces fine conductive dust inside the gantry. That dust settles on the slip ring tracks, on insulators and on the contactless data link. Worn brushes therefore cause sparking and sporadic power problems on the rotor, and the dust they shed can cause data-transfer errors and streaks on the image. One consumable can take down two systems, which is why brush inspection and cleaning of the data path belong together in planned maintenance.
What rides on the rotor
Each part on the rotor has a typical way of failing. You do not repair these parts at Level 1, but you must know what each one does so that you can describe a symptom correctly in your report.
- High-voltage generator: produces the tube voltage (kV) and tube current (mA). Faults show up as kV or mA errors, arcing messages, or an exposure that stops part-way.
- X-ray tube: the most expensive consumable in the machine. A heated cathode emits electrons, a rotating anode stops them and produces X-rays and a great deal of heat. Wear shows up as arcing, rising image noise at an unchanged protocol, longer cooling waits and a changed sound when the anode spins up. Lesson 4 is devoted to it.
- Filters, bowtie filter and collimator: shape the beam before it reaches the patient. The bowtie is thicker at the edges than in the centre, to even out the dose across a body that is thicker in the middle. The collimator sets the beam width along the patient axis, and with it the number and thickness of slices.
- Detector: an arc of modules, each made of a scintillator that turns X-rays into light and photodiodes that turn light into current. A module is normally the unit of replacement.
- DAS: sits beside the detector because the detector currents are tiny and would be lost in electrical noise if carried far. It integrates and digitises the signal.
- Rotor-side cooling: fans, and on many designs an oil-to-air or oil-to-water heat exchanger for the tube.
What stays on the ground
The stationary side is less glamorous and, in practice, causes a large share of downtime. The gantry frame carries the main rotation bearing, the drive (a belt or a direct drive motor), the tilt mechanism on scanners that tilt, the positioning lasers, the operator controls on the gantry covers and the emergency stop buttons. The patient table has its own drive, guides, position encoder and a manual release. The console and the reconstruction computer have disks, fans, network links and software licences. Outside the scanner there is the distribution cabinet, often a chiller or water cooling unit on larger systems, and the room air conditioning.
The main bearing and the drive are mechanical parts that give warning long before they fail: a new hum, vibration, a change in sound as the rotor speeds up. The table matters more than it seems, because a spiral scan is reconstructed from the table position. If the table does not move exactly as commanded, slice thickness and geometry are wrong even though every image looks clean. Treat table play or uneven movement as an image-quality problem, not only a mechanical one.
Before or after the patient
Once you have the chain in your head, ask one question about every image complaint: did the problem happen before the X-rays reached the patient, or after? The answer narrows the search to half the machine.
Problems before the patient (generator, tube, filters, collimator) change the amount or the quality of radiation. They tend to affect the whole image: more noise everywhere, a shift in CT numbers, uneven brightness from the centre to the edge, an exposure that aborts. Problems after the patient (detector, DAS, data transfer) leave the overall image reasonable but add local, geometrically regular defects: rings, streaks that follow the channels, missing views.
The ring is the classic example. Each detector channel looks at the patient from every angle during one rotation, always at the same distance from the centre of rotation. If one channel reads slightly wrong, the reconstruction spreads that error around a circle of that radius, and you see a ring. The radius of the ring tells you roughly where in the detector arc the bad channel sits. Rings are often caused by calibration drift or temperature and are removed by the routine air calibration (lesson 5); a ring that survives a correctly performed air calibration on a warm scanner points at a detector module or DAS channel. That difference is the boundary between a routine task and a repair.
Where downtime really comes from
Engineers new to CT expect the tube and the detector to dominate their call list. Field experience on mature scanners shows a different picture. The causes below are typical of an older installed base, roughly from most to least frequent; your own sites will differ, but the pattern is worth keeping in mind.
Notice that several of the most common causes are not the scanner at all, but the room, the power supply and the way the system is operated. That is your strongest argument with a customer, and it deserves its own line of recommendations in every service report.
- Room cooling: the air conditioner cannot keep up in summer, the system waits for cooling or stops on temperature, and everyone blames the scanner.
- The X-ray tube: wears with use, predictably, if you track its counters.
- Slip ring brushes, where fitted: a consumable that is easily forgotten.
- Full disks, network or PACS problems: the scanner 'stops' for a non-technical reason.
- Detector calibration drift: rings that a routine calibration removes.
- A failed detector module: rings that calibration does not remove.
- Power quality: voltage dips under load, unbalanced phases, poor earthing.
- Table drive or encoder faults.
- The main rotation bearing: rare, expensive, and it warns you early.
Your scope at Level 1
This course prepares you to keep a scanner safe, calibrated and maintained using the means the system provides to users and to authorised service staff, and your hands and instruments. That includes inspection and cleaning, the system's own self-tests, the routine calibrations and phantom checks, reading error logs and tube counters, and checking cooling, power, earthing and room climate.
It does not include entering password-protected service levels, changing service parameters, factory detector calibrations, system software updates, or any work that requires manufacturer authorisation you do not hold. Dosimetry, CT dose index measurements and the legal verification of measuring instruments are for a medical physicist or an accredited organisation. When you find something outside your scope, you record it in the report, explain it to the customer and arrange the right party. Knowing where your scope ends is part of the job, not a weakness.
| Part | What it does | How it fails | |
|---|---|---|---|
| 01 | Distribution cabinet | Feeds every consumer; holds breakers and emergency-off | Supply dips, loose earthing give odd faults everywhere |
| 02 | Slip ring and brushes | Carry power onto the rotor by rubbing contact | Brushes wear; sparking and sporadic power faults on the rotor |
| 03 | Contactless data link | Carries scan data off the rotor without touch | Brush dust on it causes data errors and streaks |
| 04 | High-voltage generator (rotor) | Makes the tube kV and mA | kV or mA errors, arcing messages, exposure stops part-way |
| 05 | X-ray tube | Produces X-rays; most expensive consumable | Arcing, rising noise, longer cooling waits, changed sound |
| 06 | Filters, bowtie, collimator | Shape the beam; set beam width and slices | Problems before the patient affect the whole image |
| 07 | Detector modules and DAS | Turn X-rays into current, then into numbers | Local defects: rings, channel streaks, missing views |
| 08 | Main bearing and drive | Carry and spin the rotor | Warn early: new hum, vibration, sound change at speed |
| 09 | Patient table | Moves the patient; position used in spiral recon | Play or uneven travel: wrong slice thickness, clean images |
| 10 | Console and reconstruction computer | Reconstruct slices, send to archive and PACS | Full disks, network faults, fans, licences |
The scanner that broke down every afternoon in July
A hospital reported that its CT scanner had become 'unreliable': from early afternoon it kept pausing between patients with cooling waits and twice stopped with temperature errors. The radiographers suspected the tube. The engineer first wrote down the times of the errors from the log: all after midday, none in the morning. He measured temperature and humidity in the equipment area at 9:00 and at 15:00 and found the afternoon reading well above the morning one and outside the range in the installation documentation. The room air conditioner had been sized years earlier for a smaller scanner, and its filters were clogged. The tube counters and the arcing count in the log showed nothing unusual. After the filters were cleaned and the hospital added cooling capacity, the afternoon stops ended.
Put the symptom on the chain before you blame the most expensive part. Time patterns, room climate and the error log often point at the infrastructure, not at the scanner.
On-site checklist
- Draw the chain from mains to PACS and mark which links are on the rotor.
- For an image complaint, decide first whether the cause is before or after the patient.
- Ask when the problem happens: time of day, after a weekend, after a heavy list.
- Read the error log for timing and repetition before opening covers.
- Measure room temperature and humidity and compare with the installation documentation.
- Check free disk space and the link to PACS when the complaint is 'the scanner stopped'.
- Write down anything outside your scope and who must handle it.