The most expensive mistake in probe service is not a bad repair. It is shipping a healthy probe across a border because the system was at fault. It costs freight both ways, two weeks of downtime, and it ends with an engineer telling you the probe tested fine — which nobody believes, so the probe gets shipped somewhere else.
This is avoidable in about twenty minutes with the machine you already have. The whole method reduces to one question asked three ways: how many probes are affected?
Branch 1 — Every probe fails
If nothing is recognised on any port, stop thinking about probes. You are looking at the system: the socket board, the probe select relay, or the power rail that feeds them.
This branch is worth naming explicitly because it produces the most wasted shipments. “The probe is not recognised” reads like a probe fault, and the department ships the probe that happened to be plugged in when someone noticed. Test a second probe before you pack anything. If the second probe also fails to enumerate, the probe you were about to ship is probably fine.
One variant deserves its own note. If the machine loses power or refuses to boot with one particular probe connected, that is not a system fault and not a normal probe fault either — it is a shorted connector pin pulling down the supply. Unplug it. That probe does need to come to us, but the machine is not broken.
Branch 2 — One probe fails, everywhere
Move the suspect probe to a different port. If it still fails, and other probes work in that same port, the fault has followed the probe. This is the branch where shipping is justified.
Before you pack it, narrow the fault so the assessment is faster:
- Flex the cable while scanning. If dropout lines appear, disappear or shift with your hand, you have a conductor break — see cable harness faults. Dead elements do not change when you move the cable.
- Scan at maximum depth with gain compensation off. Persistent vertical stripes in the same lateral position are element dropout, not noise.
- Look at the face under a light. Lens lifting, bubbles or a soft spot is delamination, and it is visible before it is diagnosable.
- Check the connector shell for bent, recessed or discoloured pins, and for gel or fluid inside.
Branch 3 — One probe fails, on one machine only
This is the branch people skip, and it is the one that catches the most false alarms. If the probe works on machine B but not machine A, the probe is not the problem. You are looking at machine A’s port, its socket board, or a software or preset mismatch.
Practitioners run into this repeatedly with probes that work fine on an older system and behave erratically on newer ones — same probe, different host, different result. Compatibility and firmware are as capable of producing “the probe is broken” as a broken probe is.
If you only have one machine, borrow the test: ask a colleague at another site to plug it in for ten minutes. That favour is cheaper than a round trip to Asia.
Two impostors worth ruling out first
Surface contamination. Dried gel, disinfectant residue and adhesive film on the acoustic face produce dark, patchy, low-contrast images that look exactly like element loss. Clean the face per the manufacturer’s instructions and rescan. This costs two minutes and resolves a surprising fraction of “degraded image” reports.
The preset. A probe running under a wrong or corrupted preset can look broken. Load the factory preset for that probe and scan a phantom before you conclude anything.
What to record before the probe leaves
Whoever assesses the probe next — us or anyone else — works faster with these five things, and they take a minute to collect:
- A screen capture of the fault on your own system, at depth, gain compensation off.
- Whether the image changes when the cable is flexed, and where.
- Whether the fault follows the probe to another port and another machine.
- Photographs of the lens face, the strain relief and the connector pins.
- The model, the part number, and the system it runs on.
That set is also exactly what we ask for before you ship anything. We assess from it and tell you whether the probe is repairable — and if it is not, we tell you not to send it. What comes back afterwards is documented to the same level.
Frequently asked questions
The system says “probe not recognised” but the probe images fine on another machine. Whose fault is that?
The machine’s, in almost every case — most often the socket board or a dirty port. Recognition is a handshake between the connector and the system; a probe that images correctly elsewhere has proven its side of it.
Is it worth testing with a phantom, or is scanning my own arm enough?
For dropout and uniformity, a phantom is better because it is reproducible month to month. For a yes/no on whether a fault exists, your forearm is fine. The important part is scanning the same thing the same way each time so you can compare.
How many dead elements before the image is actually affected?
Fewer than most people expect. Isolated dead elements can hide behind spatial compounding; a contiguous group of eight or more usually produces a visible shadow. By the time it is obvious on a clinical image, the count is often well into double figures.