The Level 2 method for mammography: logs, QC history, the subsystem map and phantoms
Combine the error log, exposure records and QC history to place a mammography fault in the exposure sequence, choose a phantom test or swap that separates two suspects, and know what you are and are not allowed to change.
Where Level 1 ends
Mammography Service Level 1 gave you the system, compression safety, the exposure chain of permissions, detector care, the QC programme and the first steps of fault-finding. X-ray Service Level 1 gave you the generator, the tube and radiation safety, and the Digital X-ray Detectors course gave you offset, gain and defect maps. None of that is repeated here. Level 2 starts when the unit fails in spite of all that: an exposure is refused or cut short, the compression will not hold, the C-arm will not turn, the QC phantom score drops, a radiologist rejects images that look subtly wrong.
Your job now is to name the subsystem that has failed, prove it with a test, and either repair it within your authorisation or hand over a precise request to whoever is authorised. The most expensive mistake at this level is replacing the board that reported the fault. In mammography it is made constantly, because so many errors are reported by the generator or the acquisition computer while the cause sits in the mains supply, a connector, a sensor or a mechanism.
Three records before you touch anything
A mammography unit gives you more evidence than most X-ray rooms, because it lives inside a QC programme. Use all three sources before you form an opinion.
The error log tells you what the system noticed and when. Read it backwards from the moment of failure and look for the first event of the burst, not the last line on the screen. Check that the clock is right first: a flat backup battery on the acquisition PC or a controller can reset the date and scramble the order of events. Note whether two errors always appear together; that usually means a shared cause such as a supply rail or a cable.
The exposure records tell you what the unit was asked to do: view, paddle, compressed thickness and force, target, filter, kV, mAs, focal spot, AEC mode and the dose indicator the system calculates. A fault that appears only with the small focus, only at the highest kV, only with the magnification stand or only above a certain thickness already points to a group of suspects.
The QC history tells you how the unit has been drifting. The radiographers' daily and weekly results, flat-field or uniformity images, phantom scores, signal-to-noise or contrast-to-noise values and AEC constancy readings, together with the physicist's last report, are a time series. A phantom score falling slowly over two months tells a different story from a sudden drop on Tuesday. Ask for the records, look at the trend, and note the date the change began; then ask what happened on that date.
- Photograph or export the log before any restart; some systems overwrite it.
- Copy the exposure parameters of the failed or rejected images.
- Look at the QC trend for at least the last two or three months, not only the last entry.
- Ask what changed: power cut, building work, room move, software update, new paddle, new detector calibration, a new radiographer.
The exposure sequence is your subsystem map
A mammography examination is a fixed sequence, and each stage depends on a known group of subsystems. Placing the failure in the sequence narrows the suspects before a cover comes off.
Power-up and initialisation depend on the mains, the internal supplies, the backup batteries some generators carry, the controllers and their communication. Motion of the column and the C-arm depends on drives, brakes, clutches, encoders and collision sensors. Compression depends on its own motor, brake, force and thickness sensing and paddle recognition. Preparation depends on the generator becoming ready, the anode reaching speed, the filament heating, the filter and focus moving into place and the detector reporting ready. Exposure depends on kV and tube current regulation, the AEC and the exposure switch. After the exposure come detector readout, image processing, display, storage and transfer, which are the computer's domain.
The question from X-ray Service Level 1 still works on images: is the defect before the breast, in the tube, filter, collimator and paddle, or after it, in the grid, the detector and the processing? Defects before the breast tend to change with focus, filter or paddle; defects after it tend to stay fixed to the detector whatever you do upstream.
Phantoms and test images are your instruments
Never diagnose an image complaint on patient images alone. Reproduce it on objects whose properties you know. A flat-field exposure of a uniform block shows non-uniformity, lines, blotches, dead or noisy areas and dust. A set of PMMA blocks of different thicknesses shows how the AEC behaves across the range, and the dose indicator and signal values can be compared with the baseline taken at acceptance. The accreditation or QC phantom used in the department shows what the radiologist would lose: fibres, specks and masses. A contrast-detail phantom, where the physicist uses one, gives the most sensitive picture of image quality.
Use them as instruments, one change at a time. Repeat the failing exposure with the same view, paddle, target, filter and mode. Then change one thing: the other focus, the other filter, no paddle, magnification stand without grid, a different thickness. If the defect moves with what you changed, you have found its side of the chain. Keep the images, label them with date, parameters and what you changed, and keep a copy for the physicist.
Any test exposure is a radiation exposure. Clear the room, use the operator's shielded position, and follow the local rules and the radiation safety lessons of the Level 1 courses in full. Never put your hand or anyone's body in the beam to 'see' an effect.
Treat the mains as a suspect
Many mammography units run from a single-phase supply and draw a short, heavy current during the exposure. Weak wiring, a poor contact in the distribution board, an overloaded phase or an automatic voltage stabiliser that switches taps under load can produce errors that look like generator, filament or exposure-switch faults. The field cases behind this course include low anode current errors traced to a loose phase contact, filament errors that came and went with the season and the supply voltage, an exposure that worked in manual mode but not in automatic mode until the unit was moved to a better phase, and an 'exposure button released early' error that tracked the clicking of a stabiliser.
So measure the supply at the unit, not at the outlet down the corridor, and measure it during an exposure, with a meter that can catch a short dip, or with a recorder left for a working day. Compare with the supply range and impedance the manufacturer's pre-installation documentation requires; the GE Senographe Crystal Nova pre-installation manual, for example, states a maximum supply impedance as well as the power demand. A reading that is fine at rest proves nothing about the voltage during the exposure itself, which lasts from a fraction of a second to a few seconds.
Swap and isolate, with discipline
Swapping identical parts is a strong test. Many units carry pairs of identical controller boards, motor drivers or power supplies, and a board borrowed from a working unit of the same type will answer a question in minutes. Change one thing at a time and write it down. Use only the same part number and an allowed revision. Put a part back if the swap proved nothing. Think before you plug a good supply into a circuit that may have a short; the fault that killed the first part can kill the second. And remember that many replacements need a calibration: a new filament board, a new compression sensor, a new detector or new firmware is not finished until the required calibration and checks are done.
Be especially careful with firmware and configuration files. Loading the firmware or tube data of another unit, even of the same model, may make an error disappear while leaving filament, kV or AEC settings that do not belong to this tube and detector. If the manufacturer's procedure is to restore a backup or reload factory data, follow it exactly, and plan the calibrations and dose checks that must follow.
What you may change, and what you may not
Your authority is set by your employer, your contract, the manufacturer's training and the law of the country. In both Russia and Kazakhstan, work with sources of ionising radiation and technical service of medical equipment are regulated, and personnel must be trained and admitted under the radiation safety rules; check the current rules that apply to you before you work.
Within that authority the usual safe territory is reading logs and running the tests your access level provides, cleaning, inspecting and reseating connectors, mechanical repair of drives and compression by the manufacturer's procedure, replacing modules and running the calibrations you are trained for. Outside it, and outside this course: using service passwords or keys you were not given, activating options or licences, changing AEC tables, kV or tube current calibration or dose-related parameters without the acceptance or constancy testing and physicist involvement that must follow, and bypassing any interlock, emergency stop, compression limit or exposure switch circuit, even for one test. If the repair needs a step you are not authorised for, stop and document the evidence for the party who is.
| Part | What it does | How it fails | |
|---|---|---|---|
| 01 | Mains, UPS or stabiliser, internal supplies | Feed the generator, drives and computers | Dips under load, a weak phase, a tap-switching stabiliser, flat batteries |
| 02 | Generator and filament circuit | Make kV, tube current and filament heating | Regulation errors, low anode current, filament faults, inverter fuses |
| 03 | Tube head, target, filter and collimator | Produce and shape the beam | Arcing, rotor faults, filter or blade position errors, overheating |
| 04 | Compression system | Holds the breast at a measured force and thickness | Brake, belt, sensor, potentiometer and paddle-recognition faults |
| 05 | Column, C-arm and accessories | Position the beam and detector; carry biopsy or magnification devices | Slipping clutches, broken shafts, encoder errors, accessory not recognised |
| 06 | Grid, bucky and detector | Remove scatter and convert X-rays into an image | Grid motion faults, temperature errors, artefacts, calibration failures |
| 07 | AEC | Chooses kV, target, filter and mAs from a pre-exposure | Wrong thickness input, wrong mode or curve, detector-region problems |
| 08 | Acquisition and review workstations | Control the exam, process, store and send images | Disk, database, BIOS, network, DICOM and worklist faults |
The exposure button that was 'released early'
A digital mammography unit began to abort exposures with an error saying the exposure button had been released too early. The radiographers swore they held it. The button was checked and found sound, and the error was intermittent, so the engineer stayed in the room for a session and wrote down every occurrence. The pattern appeared quickly: each error coincided with an audible click from the automatic voltage stabiliser feeding the unit, which switched taps just as the exposure drew current. The brief disturbance was interpreted by the controller as a broken exposure request. The advice was to replace the stabiliser with a suitable online UPS; colleagues with similar units reported that this had cured the same error for them, though the result on this unit was not confirmed.
An error message names what the controller saw, not what caused it. Record when it happens and what else happens at the same moment; the mains supply is a suspect on every mammography unit.
Filament errors and a board found by substitution
A Mammo-4MT unit showed filament and tube errors that blocked the tube calibration. The filament board was replaced first, then the current and voltage protection calibration was repeated, and the errors stayed. Instead of ordering more parts, the engineer borrowed a CPU board of the same type and fitted it as a test. The errors disappeared and the calibration completed, which proved that the controller board, not the filament circuit it was reporting about, was at fault. The CPU board was replaced and the tube calibration finished by the procedure.
A board that reports a fault in another circuit may itself be the faulty part. A controlled swap with an identical board is often the quickest proof.
High voltage only with the compression pedal held
On one unit the high voltage would start only while the compression pedal was held down; with the pedal released, the exposure button worked only now and then. Colleagues advised pressing the button many times without the pedal to see whether it switched reliably, remembering that this console had two exposure buttons wired in parallel, one on each side. Their likely explanation was cheap micro-switches with unplated contacts that oxidise. The advice was to take the button apart and clean the contacts; on a similar unit such a repair was reported to last about a year before the switches were replaced.
Before suspecting the generator, prove the request signal itself. Exposure switches are mechanical parts and wear like any other.
On-site checklist
- Export or photograph the log before any restart.
- Check the system clock and the order of events.
- Copy exposure parameters of failed or rejected images.
- Read the QC trend for two to three months and find the date the change began.
- Measure the mains at the unit during an exposure.
- Reproduce on a phantom, changing one thing at a time, room cleared.
- Swap only identical parts, record every change, calibrate after replacement.
- Keep a fault diary: as-found, tests, changes, as-left.