Ultrasound Probe Leakage Current: The Limit Depends on the Condition

It is abnormal if no leakage current is measured

One sentence in GE's service manual for the VIVID E9 and VIVID E7 reads like a misprint the first time through the section that defines the probe leakage test. The sentence is: It is abnormal if no leakage current is measured.

It is not a misprint. The same note explains why. A probe will pass some current through its insulating barrier into a patient who is grounded somewhere else, the amount depends on how the probe is built, and small differences from one probe to the next are normal. Line voltage and the placement of the test lead move the number as well. What the sentence is guarding is the far end of the scale: a flat zero usually means the test setup is wrong, not that the probe is exceptionally well insulated. The manual says it plainly — if no leakage current is detected, check the configuration of the test equipment.

That gives the reading its correct size. A leakage figure is not a health score for a probe. It measures one physical property: how much isolation stands between the patient and mains-powered electronics. Which means "did it pass?" is not answerable from the number alone, because the number arrives without two things that decide its meaning — the class of the probe, and the condition the measurement was taken under.

It is not an incident task

Neither manual treats the leakage figure as something produced when a probe misbehaves. It sits on a standing schedule.

GE's customer care schedule carries leakage current checks as line items of their own — one for the console, one for peripherals, and separate ones for surface, endocavity, transesophageal and surgical probes — repeated after any corrective maintenance, and the probe and connector functional check ends by running a leakage test on every connected probe before it is considered fit. Philips arrives at the same obligation from the other direction: the testing interval belongs to the facility's own procedures for operating-room equipment, and two of the triggers it names are not on a calendar at all. A transducer that has been dropped, and a transducer with a crack or a cut, are tested immediately.

The testing-industry guidance explains why the interval is written that way. IEC 62353 asks the manufacturer to publish a test interval and procedure based on risk, typical usage and device history, and the number it attaches to life-support and other critical equipment is a floor of 24 months rather than a target to wait for. A facility can therefore own a defensible schedule and still miss the only test that mattered, because the event that produced the breach is not the passage of time — it is a drop, a bite or a cut, or a probe cable left where a wheel could reach it, which the same GE manual names among the cable precautions in its probe and connector check.

What the meter is asking

The test does not reproduce scanning. It injects a fault.

GE defines sink leakage as the current resulting from the application of mains voltage to the applied part, and marks it a required test for Type CF applied parts. For ultrasound, the applied part is the probe: the manual's table of definitions lists transducers and ECG leads as the parts that contact the patient to perform their function.

Philips describes the same circuit from the meter's side. In the EPIQ 7 user manual, the safety analyzer is arranged so that current which would normally return through the third wire is forced down another path: out of the system chassis, through the metal parts of the transducer, through the impedance of the transducer's outer insulating layer, into a saline bath, and out through a test electrode. With that layer intact, the impedance is on the order of 850 kΩ. With a hole in it, the impedance falls to roughly 500 Ω.

Those two numbers are the reason the test exists in the form it does. What is being measured is not the array, the beam or the image. It is whether a conductive path exists through a barrier that is supposed to be an insulator — and how good that barrier is otherwise.

The limit is a set of conditions, not a value

Both GE service manuals print the acceptance limits as a table with more than one answer for the same test, because the acceptable value depends on the probe class and on the state of the grounding contact during the measurement.

Type BF covers the non-conductive probes marked with the "man in box" symbol, which the manual says includes all transducers. Type CF covers non-conductive intraoperative probes intended for direct cardiac contact and isolated ECG connections, marked with the "heart in box" symbol. The CF column is an order of magnitude tighter than the BF column, and both columns shift when the grounding contact is opened.

Leakage current test	100-120 V mains	230-240 V mains
Chassis / enclosure, ground open	0.3 mA	0.5 mA
Type BF applied parts, transmitting, ground closed	0.1 mA	0.1 mA
Type BF applied parts, transmitting, ground open	0.5 mA	0.5 mA
Type CF applied parts, transmitting, ground closed	0.01 mA	0.01 mA
Type CF applied parts, transmitting, ground open	0.05 mA	0.05 mA
Type BF sink leakage, mains on applied part	5 mA	5 mA
Type CF sink leakage, mains on applied part	0.05 mA	0.05 mA

Source: GE VIVID E9 / VIVID E7 BT'13 service manual, Direction GB091046 Rev 2, Tables 10-8 and 10-9. The LOGIQ E9 service manual carries the same values with the figures restated in microamps — 5,000 / 100 / 500 / 10 / 50.

Two details in that table matter more than the numbers inside it. The tighter applied-part limit sits beside the closed-ground condition, not the open one, so the two conditions are not interchangeable and a reading filed without its condition cannot be compared with anything, including last year's result on the same probe. And the manual is explicit that its limits are summarized from NFPA 99 and IEC 62353 and in some cases are lower than the standards specify. A limit printed in a service manual is one manufacturer's deliberately stricter reading of a standard, and it is the reading its own service network tests against.

Electrical safety testing is not a procedure to be learned from a web page. Both manuals open their electrical safety chapter with the same condition of use: only trained persons may perform these inspections, and the equipment under test must not be in contact with a patient while they run. What follows is about what a result means. The procedure belongs to the safety analyzer's instructions and to the probe manufacturer's own documentation.

A field test is not a type test

The table above holds field limits. A hospital programme that borrows its acceptance value from the wrong standard will judge the same probe differently, and the probe will not have changed.

IEC 60601-1 is a type-testing standard — it describes how a manufacturer proves a design. Testing-industry guidance exists to explain why a second standard was written at all: 60601-1 carries no risk-management criteria and is impractical to run inside a hospital, and IEC 62353 was created for recurrent testing of equipment in service, before first use on a patient, during periodic testing and after repair.

The two do not print the same limit for the same class. Under IEC 62353's own applied-part leakage test, the acceptance value is 500 µA for Class I equipment across B, BF and CF applied parts, and 100 µA for the touch current of Class II equipment. The method differs too: a current-limiting resistor of about 66 kΩ sits in the path, and the reading is scaled up to the equivalent of 230 V across the applied part.

Set that against a 10 µA figure from a type-test table and the trap is obvious once named: a programme can fail a sound transducer, or release an unsound one, purely by pairing one document's number with another document's method. The reading needs its method attached for the rest of its life.

Instruments that show only a verdict create a smaller version of the same problem. A review of dedicated transducer leakage testers describes a market that settled on red-light and green-light boxes whose pass and fail criteria were invisible to the user, while manufacturers specify their transducers with a test voltage, a test frequency, and both a lower and an upper leakage limit. Some probes are characterised with 300 to 400 µA of leakage by design. Some carry a validated minimum as low as 2 µA. A verdict without its criterion is not comparable with the next facility's — or with the same probe last year.

The image cannot tell you

The probe is working. Images are clean, the exam is on schedule, and the leakage test is the only item on the maintenance sheet that has never found anything. That line of reasoning is worth taking apart, because the two tests measure unrelated physical properties.

Image quality is a property of the acoustic stack: the elements, the matching layers, the lens and the beamformer behind them. Leakage is a property of the insulating barrier and the ground shield. Nothing about a good picture requires the barrier to be intact, and Philips quantifies the gap in its own manual — about 850 kΩ of impedance with the outer layer sound, about 500 Ω with a hole through it. The array and the software are identical in both cases. One is an insulator; the other is a conductor touching saline.

This is why both manuals treat inspection and testing as separate duties. Philips asks for a visual and tactile inspection of a TEE transducer before each exam, looking for bumps, cracks and cuts, and notes that a small bump on the shaft can be the sign of a broken strand from the ground shield beginning to puncture the outer layer. Where something is suspected, the electrical safety check follows. The same manual says the leakage test should be performed any time a transducer is dropped, or when cracks and cuts are found — that is, on suspicious probes, and between patients for the ones that stay in service. Inspection narrows down which probes to test. It does not replace the measurement, because a breach that is not visible to the eye or the finger is exactly what the measurement is for.

The TEE case: one reading judged against another

Transesophageal probes are where the absolute limits stop doing the work, and the manual for one family shows why.

The TEE shaft and tip described in the EPIQ 7 manual are Type BF. Distal to the handle there are no exposed conductive surfaces, and inside the flexible shaft every active conductor sits inside a shield that is grounded to the chassis and runs the length of the transducer. If the outer layer of the shaft is punctured, the esophagus is exposed to whatever that shield is carrying. The manual's own qualification is worth reading twice: this is not hazardous provided the third wire in the system's power cable is intact and connected to a properly grounded outlet, and even with the ground broken the leakage stays inside the IEC 60601-1 limits.

So the field test for this family does not ask whether the reading is under a fixed number. It asks whether the reading has changed its path. Chassis leakage is recorded with the ground lifted, patient leakage is recorded in the same condition, and the transducer fails if the patient reading exceeds 80 percent of the chassis reading. The two readings are being compared to each other, not to a table, because what is being hunted is a conductive route through the barrier rather than a value. The shaft itself is also the part that wears, since the articulation controls and the bending section carry the mechanical load of every exam — TEE probe articulation wear is a check with a published tip angle on one model and none on another, which is a different problem from a breached barrier and belongs to a different entry on the maintenance sheet.

The consequences the manual attaches to that failure explain the emphasis. A hole gives a conductive pathway to internal metal parts, which is a hazard during external defibrillation or electrosurgery. It also admits organic material that cannot be fully rinsed out, which is why a probe with a breached barrier has to be repaired before it is used on a patient. The same ingress logic decides the working life of a flexible endoscope, where a passing leak test is not the same as a dry instrument. The probe is judged against the system it is plugged into, and against what the next reprocessing cycle will do to it.

Probe, slot, or ground

A leakage result identifies a fault path, not a component. The section that follows the limits in the GE manual gives the order to work in, and it separates the three things that can carry the current.

1. Run the same probe in a different connector slot on the same system. If the excess moves with the probe, the probe is what to investigate next. If it stays with the slot, the probe is not.
2. Where the excess is slot-dependent, inspect the system connector for bent pins, poor connections and ground continuity before any probe is touched.
3. Check the protective earth path: the resistance from the third pin of the mains plug to exposed metal should be below 0.2 ohms, the cord and plug intact, and the outlet verified as grounded. An outlet tester cannot detect a reversed neutral and ground, which the manual names as a condition to suspect when leakage runs high.
4. If it persists, remove connected items one at a time — probes first, then peripherals — and watch the reading as each one comes off.
5. Compare the result with the previous test on the same unit. The manual's caution is aimed at abrupt unexplained change rather than at the absolute value.
6. If the excess follows the probe into every slot and stays with it, the probe is the item to take out of service. A probe whose insulating material has been punctured or otherwise compromised should not be used, and the manual's limit table will not tell you which of several probes that is — the probe type has to be entered correctly on the checklist, because a body-surface probe and an intra-cavity probe do not share a limit.

Read that sequence as four different conclusions rather than one failure:

What the reading does	What it implicates	First move
High in one slot, normal in another	The connector path on the system	Connector pins, contacts, ground continuity
High in every slot with the same probe	The probe's insulation or internal shield	Remove from service; re-check method and condition first
High with no probe connected	The chassis, mains cord or outlet	Protective earth path, cord and plug, outlet wiring
Normal on a probe with a bump, cut or bite mark	Nothing yet — the measurement is not an inspection	Follow the manufacturer's inspection route for that model

When the fault does follow the probe, the next question is the probe itself: repair, or replace. geprobe is one of the third-party suppliers and repairers in that market — probes and the parts that go into them — and we are not affiliated with GE, Philips or any other manufacturer named on this page. What we need in order to quote is the model, the class marking on the connector, the reading with its test condition, and whether the excess follows the probe or the slot; if it follows the slot, the probe is not the item to buy. We cannot promise that a particular probe is repairable, or what a repair would cost, and the facility that receives it still has to satisfy itself that the probe is safe and performs as expected. The technical standards review from the AIUM is blunt on why that last step is not optional: the FDA treats an ultrasound probe as a finished medical device, third-party probe repair in the United States is not FDA-regulated, and insulation integrity has to be re-verified after any repair, alongside phantom checks of element dropout, resolution and measurement accuracy.

The limits quoted here come from two manufacturers' service and user documentation and one testing-industry guide, and each of them is a summary of a standard revision rather than the standard itself. The number that applies to a specific probe is the one in that probe's own documentation and in the facility's programme — and when the two disagree, the documented one wins and the disagreement belongs in writing. This page is written for biomedical and clinical engineering teams who own that programme. It is not a repair procedure and not clinical guidance, and no part of it should be run as one: electrical safety testing is a trained-person task, performed on equipment with the patient out of the circuit.