Soapy water is standard practice on compressed air systems for a good reason: it works. Spray a soap solution on a fitting carrying a leak and the escaping air forms visible bubbles. It's cheap, it requires no training, and it gives an unambiguous yes-or-no answer on the specific joint being tested. None of that is in question.
What's in question is whether that method (repeated joint by joint across an entire plant) is a workable way to find leaks you don't already know about, across a network most of which sits overhead, behind machinery, or several metres up a structural steel frame. It isn't, and the reason has nothing to do with the soap. It's the method's fundamental requirement: you have to already suspect a joint, and you have to be able to physically reach it.
Two conditions that quietly exclude most of a plant
Soap testing needs two things to happen before it can find anything: someone has to already suspect a specific fitting is leaking, and someone has to be able to get a spray bottle to it. The first condition rules out the leaks nobody's noticed, which, on an unmanaged system, is most of them, since leaks are inaudible over normal plant noise and don't announce themselves. The second condition rules out overhead mains, drop legs into machine cells, anything behind guarding, and anything that needs a ladder or an elevated work platform to reach safely.
Put those two conditions together and soap testing is left covering the leaks someone already knows about, on joints already at floor level or already scheduled for other maintenance work. That's a real and useful thing to be able to do. It is not a survey of a plant's compressed air network, because it was never designed to find leaks: only to confirm ones already suspected.
| Method | Finds leaks you don't already suspect? | Covers overhead / hard-to-reach runs? | Works while production runs? |
|---|---|---|---|
| Soap solution | No | No | Yes |
| Handheld ultrasonic detector | Partially | Limited: needs close range | Yes |
| Pressure decay test | Yes, for that section | N/A: whole-section result | No: requires shutdown |
| Acoustic imaging | Yes | Yes: surveys from several metres away | Yes |
Why "sprayed on overhead pipework" is its own hazard
There's a second, more physical problem with trying to extend soap testing to a whole network: overhead compressed air runs are exactly where a maintenance technician would need to apply soap solution from height, and anything sprayed on overhead pipework runs downward. On a factory floor, that's a slip hazard directly beneath the work being done, one that scales with how much of the network someone is trying to test this way. The method that's perfectly safe for a floor-level fitting becomes a genuine safety consideration the moment it's extended to the runs where compressed air networks actually spend most of their length.
Soap testing was never designed to search. It was designed to confirm: a different job, on a different, much smaller set of joints.
What ultrasonic detection changes
Every leak generates noise well above the range of human hearing, typically strongest somewhere around 40 kHz, against an upper limit for human hearing of roughly 20 kHz. That's true whether or not anyone suspects the leak exists, and it's true whether the joint is at floor level or thirty feet up a structural frame. A device that can detect that ultrasonic signature doesn't need a prior suspicion to find something: it can scan a run and let the sound itself indicate where a leak is.
A handheld ultrasonic detector takes advantage of this, but it's still a single-point instrument: sweep an area, then narrow in with a focusing probe, one zone at a time, which means it still needs to get physically close. That solves the "already suspected" problem but not the "hard to reach" one; ladders, elevated platforms and crawling behind machinery are all still required.
An acoustic imager (an array of ultrasonic-sensitive microphones paired with a camera, rather than a single sensor) solves both problems at once. It scans a whole area from several metres away, so overhead mains and elevated runs are visible from the factory floor rather than requiring access to them directly. And because it's listening across a wide area rather than waiting for a suspicion to test, it finds leaks nobody on site knew were there.
Where soap solution still earns its place
None of this argues for retiring soap testing. Once a survey (acoustic or otherwise) has identified and located a leak, confirming it with soap solution before a repair, or verifying a repair actually closed the leak afterward, is quick, cheap, and requires no equipment beyond what's already in most maintenance kits. It's also useful for a technician already working on a specific piece of equipment who wants to check a joint they've just disturbed. The method is genuinely fit for that purpose. The mistake is asking it to do a survey's job: finding leaks nobody has flagged, across a network most of which was never designed to be reached from the floor with a spray bottle.
A sensible sequence on most sites is the reverse of how soap testing tends to get used today: run a proper survey first to find and prioritise every leak, including the ones nobody suspected, then use soap solution as the cheap, fast confirmation step on each specific joint before a repair crew commits time to it.
The method that actually surveys a running plant
None of this makes soap testing a bad tool: it remains the right choice for confirming a specific joint once a leak is suspected, and it requires no equipment beyond a spray bottle. What it can't do is function as a survey method for an entire compressed air network, because a survey's whole purpose is finding leaks nobody has flagged yet, in places nobody can easily reach. That's a different job, and it needs a method built for scanning rather than confirming.
See how acoustic imaging locates and sizes leaks across a whole plant without a shutdown, or see the 12 places a leak typically hides, most of which sit exactly where soap testing struggles to reach.
The test worth applying to any method
The useful question for any leak-detection method isn't whether it works: soap solution, handheld ultrasonic detectors and pressure decay tests are all legitimate tools that genuinely do what they claim. The useful question is what it needs to be true before the method can find something: does it need a prior suspicion, does it need close physical access, does it need the line shut down. Soap testing needs the first two. A pressure decay test needs the third. Acoustic imaging needs none of them, which is the specific gap it fills rather than a general claim of superiority over tools built for a different job.