Probe or system: the swap method and the symptom map
Separate a probe fault from a port fault and from a system fault with a fixed swap sequence, and map the symptom on the screen to the block most likely to have failed before opening a cover.
Where Level 1 stopped and Level 2 starts
At Level 1 you learned to inspect a probe, read an air image and run a phantom test, and you learned the first rule of ultrasound service: most complaints are a probe. Level 2 begins where that rule runs out. The probe passed inspection, a second probe shows the same fault, or the system misbehaves with every probe. Now you have to find which block inside the cart has failed, and you have to do it without replacing boards at random, because beamformer boards, power supplies and probes are among the most expensive spare parts in the department.
The whole course rests on one habit: before you touch a screwdriver, move the fault. A fault that moves with the probe is in the probe. A fault that stays on one port is in that port or the path behind it. A fault that appears on every probe and every port is in the shared system. A fault that changes with the room, the outlet or the time of day may not be in the system at all. Each swap costs a minute and removes half of the suspects. A board swap costs a day and may remove none.
The swap sequence, always in the same order
Write the sequence on a card and keep it in your kit. Doing it in the same order every time is what makes it fast and what makes your report credible to the next engineer.
First, reproduce the complaint yourself with the preset, probe and mode the user describes. Save an image or a short clip. If you cannot reproduce it, ask exactly when it happens: which exam, which depth, which mode, after how long, in which room. Many 'intermittent' faults are tied to one mode or one temperature, and the question saves an hour.
Second, move the same probe to every other working port. Third, put a second probe of the same type, ideally known good from another system, on the original port and then on the others. Fourth, if you can, take the suspect probe to another system of the same model and look again. Fifth, try a probe of a different type on the suspect port. Record each result in a small table: probe by serial number, port, mode, result. That table is the most useful thing you will produce on the visit, because it is the evidence on which every later decision rests.
- Fault follows the probe on every port and on another system: probe.
- Fault stays on one port with every probe: that port, its pins, its relays or switching path.
- Fault appears on every probe and every port, in the same place on the screen relative to the channel layout: shared transmit, receive or beamformer channels.
- Fault appears on every probe in one mode only (for example colour or CW): processing for that mode, or settings.
- Fault changes when you move the system, change the outlet or switch off nearby equipment: external interference or mains, not the system.
Why a channel fault looks like a probe fault
A probe with 128 or 192 elements is often connected to a system with fewer physical transmit and receive channels. Multiplexers and relays decide which elements are connected to which channels for each line. A dead system channel therefore does not produce a dead line. It removes one contributor from every line that uses it, and the image shows a darker vertical stripe exactly where a dead element would show it. To the user, and to an engineer in a hurry, the two are indistinguishable on one probe.
Two things separate them. The first is the swap: the stripe from a dead channel stays in the same channel position whatever probe you plug in. The second is geometry. A dead element sits at one physical place on the array, so its stripe sits at the same place across the image on every depth and every preset for that probe. A dead channel is mapped to different elements by different probes and sometimes by different apertures, so on a probe with a different element count or pitch the stripe moves to another place on the screen, and on a phased array, where every element contributes to every line, it may not appear as a stripe at all but as a general loss of sensitivity or a slightly noisier image. If the stripe moves to a different position when you change probe type, but appears on every probe, think channel, not element.
The symptom map
After the swap sequence you know whether the fault lives in the probe, a port or the shared system. The symptom map tells you which block inside the system to suspect first. It is a starting point for the lessons that follow, not a verdict; each line is a probability, and the logs and the built-in tests decide.
- Dark vertical stripe on all probes: receive or transmit channel group, a front-end sub-board or its seating (lessons 5 and 6).
- Dark stripe on one port only: port pins, relay or multiplexer path for that port (lesson 5).
- Noise, snow or fan-shaped interference on all probes: earth, mains or external interference first, then receiver boards or the low-voltage supply to the front end (lessons 2 and 7).
- No image on any probe, system otherwise alive: transmit high-voltage supply, front-end supply, or a probe loading the transmit supply (lessons 5 and 7).
- Colour, PW or CW missing or noisy, B-mode fine: Doppler processing, transmit supply at Doppler settings, or presets (lesson 6).
- Reboots or shuts down under load (Doppler, 4D, long exams): power supply, cooling, PC (lessons 7 and 8).
- Does not reach the application, hangs, freezes: PC, disk, memory, operating system, software (lesson 8).
- Probe not recognised: probe identification, connector pins, port, or a probe outside the system's licence or software version (lessons 3 and 8).
- Touch panel, keys, monitor, printer, PACS: peripherals and integration (lesson 9).
Read the system before you open it
Every modern ultrasound system keeps an error log and has built-in front-end tests. Their names and access routes are model-specific and are described in the manufacturer's service documentation; some need service access that only an authorised engineer has. Use what you are entitled to use. Export the log to a USB stick before any restart, because some systems overwrite or lose entries, and note the time of the user's complaint so that you can find the matching entries.
Read the log for three things: the first error, not the last; errors that repeat at the same moment as the symptom; and power or temperature events just before a reboot. A transmit-voltage error a second before a restart tells a different story from a disk error a second before a freeze. Built-in front-end tests that report per-board or per-channel results are the fastest way to turn a stripe on the screen into a board name. Remember their limit: a test that passes says the circuit responded during the test, not that it works under every load and temperature.
Rules that keep you out of trouble
Change one thing at a time and record it. When the fault disappears after you have re-seated three boards and swapped a cable, you do not know what fixed it, and you will not know where to look when it returns. Photograph cable routing and board positions before you pull anything; on many systems sub-boards must go back in the same slot.
Do not move a suspect probe from system to system until you have looked at its connector and housing. A probe with a short or with liquid inside can damage a port or trip the transmit supply of a healthy system, and you will then have two faults instead of one. Inspect the connector pins with a light and a magnifier first, and if there is any sign of liquid, burning or a cracked housing, do not plug it into anything.
Work inside the cart only when the system is switched off, disconnected from the mains and left to discharge as the manufacturer describes. The transmit circuits carry high voltage while the system is scanning, and the power supplies hold stored energy after switch-off. Measurements on a live, open system are done only by trained personnel, following the procedure and test points in the manufacturer's service documentation, with the precautions it specifies.
Three probes condemned, one board at fault
A women's health clinic had replaced two convex probes in three months because of 'dead elements'. The third new probe showed a faint dark stripe within a week. The engineer who came next did not look at the probe first. He built the swap table: the stripe appeared with all three convex probes on every port, and also, in a different position, with the linear probe. A phased array looked clean but slightly noisy. That pattern pointed away from the probes and towards the shared channels. The built-in front-end test reported a failing group on one receiver sub-board. The board was re-seated and the stripe faded but did not disappear; a replacement board removed it. The two probes that had been returned as faulty had been tested by the repair facility and found to be within specification.
When a 'probe fault' repeats on new probes, stop buying probes. The swap table across probes, types and ports is cheaper than any part and is the evidence the clinic needs.
The port that stayed silent
A cart system ignored any probe connected to one particular port, while the other ports worked with the same probes. The engineer used the swap method first: three different probes, all fine elsewhere, were all ignored on that port, so the probes were excluded. On this platform the probe-select relays click audibly when a port is activated at start-up. He connected a probe only to the suspect port and restarted: the click heard from the other ports was missing. With the system off and discharged, the port's pins were inspected and found clean and straight. The port relay board was replaced, after which every probe was tested on every port.
When one port fails with every probe, the probes are innocent; listen for the relay clicks, inspect the pins, then look at the port's own switching board.
On-site checklist
- Reproduce the complaint; save an image or clip with the preset and mode noted.
- Inspect the suspect probe's connector and housing before plugging it into any other port or system.
- Build the swap table: probe serial, port, mode, result.
- Export the error log before any restart; note the time of the complaint.
- Run the built-in front-end tests you are entitled to use and record per-board results.
- Photograph board positions and cable routing before removing anything.
- Change one thing at a time and record what changed.