Refrigerant Leak Detector vs Soap Bubble Test: Which Finds the Leak?

As an Amazon Associate, Akermin earns from qualifying purchases. Product links on this page are affiliate links.

What each test physically measures, and why that decides the job

A soap bubble test measures one thing: gas escaping through a liquid film at a pressure high enough to inflate a bubble. You brush or spray a solution on a joint, and if refrigerant is escaping, the film stretches into a bubble that grows and pops. The method is old, cheap, and reliable at the joint it touches. Its limit is geometric. The bubble only forms where the solution sits, and the solution only sits where you can reach. A leak on the back face of an evaporator coil, inside a line set buried in a wall, or at a schrader port behind a service panel will never see the brush.

An electronic refrigerant detector measures a change in the air. A heated diode or infrared sensor reacts to halogenated refrigerant molecules drawn across it by a small pump or by diffusion, and the instrument turns that reaction into a signal: a tone that speeds up, a bar that climbs, a number on a display. The detector does not measure mass flow. It measures concentration in the sample it is breathing, which is why the reading depends on how close the probe is, how long you dwell, and how the air moves in the room. That is not a defect. It is the nature of a sniffer, and it is why the good ones give you a graded signal rather than a yes or no.

The practical consequence is that the two methods answer different questions. The sniffer answers “where is the leak, roughly, in this machine or this room?” The bubble answers “is this specific joint leaking right now?” If you already know the joint, the bubble is faster and needs no calibration. If you do not know where to look, the bubble cannot help you, because you would have to wet every joint in the system, including the ones you cannot reach.

Sensitivity: why the sniffer finds leaks the bubble never shows

The numbers matter here, and they are not close. A soap bubble becomes visible when the leak rate is high enough to inflate a film against surface tension and the surrounding air pressure. In practice, a bubble test is good at gross leaks and poor at slow ones. A system that loses a few grams a year will not make a bubble you can see, even if you wet the joint perfectly. The refrigerant escapes too slowly to stretch the film.

An electronic detector is specified differently. The Inficon Compass listing states a sensitivity of 0.15 oz per year (4 g per year). That is a mass-loss rate, and it is the kind of figure that matters when a system is losing charge slowly enough that the owner notices performance drift over months rather than a puddle on the floor. A detector at that sensitivity can respond to a leak that a bubble test would never reveal, provided the probe reaches the leak and the sensor is working. The Robinair LD7 listing does not publish a sensitivity figure, but it does describe two display modes: a bar graph that shows the amount of refrigerant detected in the air and changes colour as the sensor gets closer, and a sweep mode that draws a line graph so you can trace back and forth to pinpoint the source. Those modes exist precisely because the raw signal is a concentration, not a location, and the display helps you turn a gradient into a point.

The Aprvtio WJL-6000PRO adds a second measurement scale: ppm for micro leak detection and %LEL for monitoring A2L refrigerant safety. That distinction is worth understanding. PPM is a concentration in air, useful for finding a small leak. %LEL is a fraction of the lower explosive limit, which is a safety scale, not a leak-location scale. If you are working with A2L refrigerants such as R32 or R454B, the %LEL reading tells you whether the atmosphere is approaching a flammable concentration. It does not tell you where the leak is. For more on why those two scales are not interchangeable, see why a percent LEL alarm is not a ppm leak locator.

The false positive problem, and why it is worse for bubbles than for sniffers

Electronic detectors are often criticised for false positives. The criticism is fair but incomplete. A heated diode sensor responds to any halogenated compound, which includes the refrigerant you are hunting and also some solvents, degreasers, and the residue on a new coil. A detector that chirps near a brazed joint may be reacting to flux or to a cleaning agent, not to a leak. The remedy is technique: let the instrument settle in clean air, move the probe slowly, and confirm any hit with a second method.

The soap bubble test has its own false positives, and they are less discussed. A bubble can form from air escaping a poorly sealed joint during a pressure test, from residual nitrogen, or from the soap solution itself foaming in a draft. A bubble tells you gas is escaping, not which gas. If you are testing with nitrogen alone, a bubble confirms a mechanical leak but says nothing about refrigerant loss rate. If you are testing with refrigerant, a bubble confirms a leak but not its size. The bubble is a binary indicator at a single point, and it is only as good as the surface you wet.

This is where the detector’s graded display earns its place. The bar graph on the Robinair LD7 changes colour as the sensor approaches the leak, which gives you a direction. The sweep mode draws a line you can watch as you move the probe back and forth. Neither display is a measurement of mass flow, but both turn a yes or no into a search pattern. A bubble cannot do that, because a bubble either forms or it does not.

What a soap bubble test still does better than any sniffer

There are jobs where the bubble is the right tool, and it is worth being honest about them. The first is confirmation. When a detector has pointed you at a joint, a bubble test at that joint gives you a visual, unambiguous confirmation that costs almost nothing and requires no calibration. The second is a system that is already open and pressurised with dry nitrogen. A bubble test at that moment is direct: you see the leak, you mark it, you repair it. The third is a joint you can reach and a leak you suspect is gross. In that case the bubble is faster than setting up a detector, warming it up, and zeroing it.

The bubble also has an advantage in documentation. A photograph of a growing bubble at a specific joint is a record that a leak existed there. A detector reading is a number on a display that depends on probe position and ambient conditions; it is harder to defend later. If you are handing a job to another technician or writing a report, the bubble photograph is concrete evidence.

What the bubble cannot do is survey. It cannot tell you which of forty joints is leaking without wetting all forty. It cannot find a leak inside a sealed circuit. It cannot tell you whether a system that has lost charge over a season is leaking now or was leaking before it was topped up. For those questions you need a detector, and for the last one you need a pressure test.

Using the two together: a sequence that works

The sensible workflow is detector first, bubble second, pressure test when the question is whether the system holds at all. Start with the detector in clean air, zeroed according to the manufacturer’s instructions. Move the probe slowly along the likely path: joints, schrader ports, flare nuts, brazed connections, the coil faces, and the line set where it is accessible. Watch the display rather than listening for a tone alone, because a graded bar or line tells you when you are getting warmer. When the signal peaks, stop and mark the spot.

Then confirm with soap solution at that spot. If a bubble forms, you have a leak and a location. If no bubble forms but the detector still peaks, you have a decision: the leak may be too slow for a bubble, or the detector may be reacting to something else. In that case, clean the area, re-zero the detector away from the joint, and try again. If the signal persists and no bubble appears, a pressure test with dry nitrogen and a gauge is the next step, because it measures the system’s ability to hold pressure rather than the concentration in the air.

The Aprvtio WJL-6000PRO and the Simbow 3-in-1 both offer ppm and %LEL modes, and both list A2L compatibility. That matters for newer systems using R32, R454B, or R1234yf. The %LEL mode is a safety check, not a leak locator, and the ppm mode is the one you use for finding the leak. The Simbow listing names R134a, R410A, R404A, R454B, R32, R1234yf, R290, and R600A as compatible refrigerants, which is a broad list, but the listing does not publish a sensitivity figure. If you need a published sensitivity number, the Inficon Compass is the one in this set that states it.

Where both methods stop being the right answer

Neither a sniffer nor a bubble proves a system is leak free. A detector samples the air it breathes; a leak inside an insulated line set or behind a panel may never reach the probe. A bubble test covers only the joints you wet. If the question is “does this system hold its charge over a season?”, the answer is a pressure test or a standing vacuum test, not a survey with either tool.

There is also a safety boundary. A refrigerant detector is not a combustible gas alarm and not a carbon monoxide detector. The %LEL mode on the Aprvtio and Simbow units is a refrigerant safety scale, not a general flammable gas alarm. If you are working in a confined space with a refrigerant that can displace oxygen or reach a flammable concentration, you need a confined space gas monitor with the appropriate sensors, and you need to understand what each sensor does and does not detect. The same logic applies to domestic gas work: a refrigerant sniffer will not tell you whether a propane line is leaking, and a propane detector will not find a refrigerant leak. For the domestic side of that distinction, see natural gas detector vs propane detector.

Finally, calibration and sensor life are the maintenance nobody mentions. Heated diode sensors degrade with exposure to contaminants and eventually need replacement. The listings here do not publish sensor life or calibration intervals, so the honest statement is that the manufacturer’s instructions are the only reliable source for those figures. A detector that has not been checked against a known leak is a detector you are trusting on faith. The soap bubble, for all its limits, has no sensor to drift.

The short version

A soap bubble test confirms a leak at a joint you can reach. An electronic refrigerant detector finds the leak you cannot see and gives you a gradient to follow. Use the detector to search, the bubble to confirm, and a pressure test when the question is whether the system holds at all. Neither tool is a substitute for the other, and neither one proves a system is leak free.

Frequently Asked Questions