How to Choose the Right Leak Testing Method in HVAC/R Manufacturing

One of the oldest—and still widely used—approaches is water immersion, also known as the water bath or bubble test. The principle is simple: the component is pressurized and submerged in water, and the operator checks whether bubbles rise to the surface. The logic is straightforward—similar to checking a bicycle tire for punctures.

In practice, water bath testing can detect leak rates down to approximately 5×10⁻² mbar·l/s, which appears as a slow but visible stream of bubbles. At a leak rate of 1×10⁻⁴ mbar·l/s, however, bubble formation becomes extremely slow—taking roughly 30 seconds for a single bubble to appear.

Although the theoretical detection limit is even lower, it deteriorates significantly in real-world conditions, depending on the geometry of the test piece. An air bubble that would rise freely in open water may detach only with difficulty from a component with a complex shape. Another drawback is that the method cannot be used for components that must not come into contact with moisture.

Finally, there is the human factor: whether bubbles are detected at all depends heavily on the operator’s attention and condition on a given day. As a result, the risk of false-negative results is relatively high when using the water bath method.

Curious what leak rates your water bath can detect? Our calculator helps you convert bubbles into flow rates—and vice versa.

This method is simple and relatively inexpensive, but its effectiveness depends heavily on the operator’s attention and skill. Theoretically, soap spray testing can detect leaks as small as 1×10⁻³ mbar·l/s. In practice, however, its sensitivity is lower than water bath testing, with a realistic detection limit around 5×10⁻² mbar·l/s.

Soap spray testing also has specific challenges. Large leaks can escape detection because air may blow the foam away, making a leaky part appear tight. Additionally, if the leak detection spray doesn’t stick to the surface and runs off, certain areas—like the underside of a component—become particularly difficult to inspect.

There are three main techniques that detect leaks by monitoring pressure changes: the pressure‑decay method, the differential‑pressure method, and the pressure‑rise method. All three are widely used for integral leak testing, with the pressure‑decay method being the most common in industrial settings.

Pressure decay test

In a pressure decay test, a component is filled to a set overpressure, allowed to stabilize, and then monitored for any drop in pressure. A decreasing pressure indicates a leak. While the theoretical detection limit is around 1×10⁻³ mbar·l/s, in practice the sensitivity is usually closer to 1×10⁻² mbar·l/s due to temperature variations. Temperature fluctuations can cause pressure changes that mimic a leak or hide a real one. 

Differential pressure testing

Differential pressure testing increases sensitivity by comparing the test part with a leak‑tight reference component. This approach allows even small pressure differences to be detected, giving a theoretical limit of 1×10⁻⁴ mbar·l/s and practical values around 1×10⁻³ mbar·l/s.

Pressure rise testing

Pressure rise testing is used for vacuum-tight components. A vacuum is created inside the part, and the rate at which pressure increases indicates the leak rate. Although the theoretical sensitivity is 1×10⁻⁴ mbar·l/s, in practice it usually detects leaks around 1×10⁻³ mbar·l/s.

Together, these three methods provide reliable options for integral leak testing in industrial applications, but practical limitations such as temperature effects, equipment setup, and part constraints must be considered when choosing the appropriate method.

Helium is the most widely used tracer gas. Chemically inert, non-toxic, non-flammable, and with a low molecular mass of 4, it is particularly well suited for leak testing. Another key advantage is its low background concentration: helium naturally occurs at only about 5 ppm in air.

In integral helium leak testing in a vacuum chamber, the test object is placed in the chamber, either manually or automatically, for example by a robotic arm. A pump creates a vacuum, and the test object is first evacuated and then filled with helium tracer gas. Although this method is comparatively complex due to strict tightness requirements and the cost of the vacuum pump, it provides important benefits.

Helium testing in a vacuum chamber is the most sensitive of all tracer gas methods. Mass spectrometers used for detection can measure leak rates down to 1×10⁻¹² mbar·l/s under optimal conditions. The vacuum method is especially well suited for in-line testing in automated production processes, where every manufactured part undergoes an integral leak test. Short test cycles and fast throughput times are major advantages, particularly in fully automated operations. In addition, the high sensitivity often allows the helium concentration to be reduced significantly, for example to as little as 10%, lowering test gas costs accordingly.

To learn more about our helium leak detection solutions and find the right system for your process, visit our product pages.

Using refrigerants as tracer gases

In HVAC/R production, components such as refrigerators and heat exchangers are often already filled with refrigerant during end-of-line testing. Instead of introducing a different tracer gas, manufacturers can take advantage of this and use the existing gaseous end media as the tracer gas for leak detection. In this scenario, a wide range of refrigerants can serve as tracer gases, including synthetic compounds like R1234yf (2,3,3,3-tetrafluoropropene) and naturally refrigerants such as R290 (propane), R600a (isobutane), and R744 (CO₂).

Modern leak detectors are designed to handle this variety. Some multi-gas systems can even detect multiple refrigerants simultaneously, making them ideal for mixed production lines where different gases are used. The purpose of refrigerant-based sniffing leak detection is often to verify the tightness of critical connection points on finished components. However, it can also be used to locate leaks in parts that failed testing during production and were removed from the line, allowing them to be repaired during rework.

A very sensitive multi-gas sniffing leak detector, such as the Ecotec 4000, can detect leaks as small as 0.05 g of R600a per year. This level of sensitivity ensures that even minor leaks are identified before products leave the factory.