By Grant Laidlaw

Many people ask for assistance in the understanding of theoretical and practical aspects of the industry. I will endeavour to enlighten. I am going back to basics as I have questions coming in which indicate that the basic understanding necessary to work in industry is not in place.

James sent us this question: “Grant, with regards to leak testing, what would be the preferred method for hydrocarbons? It seems that soap water will be the only safe way moving forward.”

Hi James, we all need to keep abreast of changing technology and this is certainly an area worth looking into. Having spoken recently to people in the industry and in particular rural areas, it became apparent that some of the basics skills are an issue. I even ran across some refrigeration mechanics using the old halide testers, explaining to me that they work well. Of course, leak testing on a system that uses a flammable refrigerant with a leak tester that utilises an open flame is really dangerous.

The vast amount of refrigerants in use globally coupled with fairly high leakage rates are creating environmental issues. Leaks can reduce system performance, lead to energy wastage, product and production losses. Leak testing is therefore an important aspect of system installation and maintenance. Leaks should be located and attended to as a matter of priority.

The vast majority of refrigerant systems circulate oil when operating, and therefore when generally inspecting a system that has been running for a period of time, oil residue anywhere on the equipment or piping indicates a possible leak.

As this will not be the case with all leaks or on newly installed systems, a thorough leak test must be conducted.

The first step is conducting a vacuum test. In this test there will be no refrigerant in the system. Using a vacuum pump capable of reaching a vacuum of at least 500 microns, evacuate the system until the vacuum is below 500 microns. Isolate the system using your manifold gauges and observe the micron gauge placed after the manifold gauges. If, after a minimum period of two hours, the system has maintained the vacuum, one can conduct a pressure test. Two hours is a minimum, on larger systems it is recommended that the system stand under vacuum for a longer period.

For the following methods of leak detection, the system must be under a positive pressure utilising dry nitrogen. Only a small amount of detectable gas is required in the system to detect the leaks when using the methods that rely on monitoring the gas leaking out of the system.

Most leak testing instruments are sensitive to commonly used refrigerants although care must be exercised when selecting the instrument because not all instruments are suitable for the hydrocarbon refrigerants.

To minimise the amount of refrigerant for leak testing, the system is pressurised to about 50 kPa (g) with the refrigerant, with which the system normally operates. To bring the test pressure up to the test pressure required, dry nitrogen is added (care must be taken that the pressure does not become excessive).

You must use a nitrogen regulator and I mean ‘must’. If you decide to disregard this, I may be attending your funeral and I do not like funerals. So please use a regulator. Under no circumstances should compressed air, oxygen or acetylene be used for pressure testing.

It must be remembered that pressure testing serves the purpose of locating leaks and not that of testing whether the components can, in fact, withstand the pressure. If the entire system is tested, care must be taken that the pressure applied to the system does not exceed the maximum operating pressure of the components.

For pressure testing the entire system, including the low side components, the pressures recommended depend on the normal working pressure of the refrigerant and are as follows:

  • For 134A, R600a and R290 used in domestic applications with pressed aluminium, evaporators between 600 and 800 kPa is recommended.
  • For R134a systems and other refrigerants which have similar working pressures the recommended test pressure is 1 050 to 1 400 kPa.
  • For R 22 systems and other refrigerants which have similar working pressures the recommended pressure is 1 750 to 2 100 kPa.
  • For R410a, 3 800 kPa is recommended.

When pressurising the system, the pressure must be increased in stages with a comprehensive leak test between each stage and the pressure only increased when no leaks are found. The nameplate on some systems indicates the recommended test pressure; in this case this is the pressure to be applied. After the test pressure has been reached the most suitable leak detection procedure can be applied.

The oil spot method

This is an observation of the system rather than an actual leak detection method.

Together with the refrigerant, oil is circulating through the system. If the refrigerant leaks out there is usually an amount of oil leaking out with it. As there is always air (with dust) flowing over the plant the areas where there is leakage will be oily and dusty. Thus, to find the leaks look for these spots.

The soap bubble method

This method is probably the simplest and safest method of leak testing and can even be used by inexperienced persons, because all the materials used are non-toxic and non-explosive. The disadvantage of the method is that small leaks are difficult to detect, and the testing can take a long time. Also, in some instances you cannot easily access the joints to be tested.

For this method, a strong water-soap solution is ‘painted’ on all areas suspected of leaking; areas such as joints, gaskets, seals, fittings, and welds. Where there is a leak, the solution will cause bubbles. Specially made solutions are available from refrigeration suppliers.

There are some that feel that this method is not accurate enough to detect very small leaks but in general the method is cheap and easily applied. It may be advisable to use this method in conjunction with an electronic leak detector.

Electronic leak detector

Electronic leak detector. Image supplied by ACRA.

Electronic leak testing

The electronic leak detector is the most sensitive of the commonly available leak testing instruments and is available for hydrocarbons. Note: It is important to ensure that the electronic tester is suitable for the type of refrigerant utilised by the system being tested.

I often hear the complaint that these types of leak detection devices do not really work as they are too sensitive. Some older electronic detectors were so sensitive that they could not be used if there was refrigerant gas present in the general atmosphere around the plant. The more modern detectors have a button which, when pressed, cancels out the concentration in the general atmosphere and will only react when the sample contains more gas than the atmosphere. This is the preferred method for detecting all leaks and in particular very small leaks. A disadvantage is that fans can disperse leaking refrigerant making a leak hard to locate utilising this method alone.

Ultraviolet leak detection

With the ultraviolet leak detection method, a small amount of fluorescent dye is added to the oil in the compressor and is typical to automotive systems. This method can be used in all systems where there is oil circulating through the system with the refrigerant. If a leak occurs anywhere in the system the refrigerant, together with the oil and the fluorescent dye, will leak out. Leak finding is done with an ultraviolet light. When this light is shone onto the plant the fluorescent dye will show up as a bright green-yellow or pink, fluorescent glow. Very small leaks can be found with this method.

Because it is so easy and fast, this method can be used as a routine method. If the plant is in a dark surrounding it does not even have to be switched off as it is easy to see the fluorescent glow in these circumstances. The dye can be added directly to the oil, or it can be added into the suction side of the compressor with the aid of a special mist infuser. After the dye has been added the plant must be run for some time in order to give the dye a chance to leak out – if there is a leak.

Large leaks will become apparent in a matter of hours; small leaks take much longer to locate. The fluorescent dye stays in the system until the oil is changed, even if all the refrigerant escapes or the system is evacuated.

Note: Some manufacturers do not guarantee compressors if dye is added. Having said that, most automotive manufacturers add dye to the air conditioning systems to facilitate leak detection.

Ultrasonic leak testing

Ultrasonic leak detectors operate on a completely different principle from all the other methods. This detector detects and amplifies the sound of the high velocity gas escaping from or going into the system. This means that it does not matter what gas there is in the plant; it does not even matter whether the leak is outward or inward as long as a pressure difference exists. This method works equally well if the plant is under a vacuum. This type of equipment is fairly expensive and is not in general use by the industry but rather in specialised applications.

Nitrogen cylinders

There are precautions to be taken when using nitrogen cylinders. A regulator must be used when pressurising with nitrogen. The pressure in a nitrogen cylinder can be up to 32 000 kPa and it is absolutely essential that a regulator with a pressure relief valve is used when adding the nitrogen to the system. Use of a nitrogen cylinder without a regulator will lead to accidents and possible deaths. The system can and will explode as the uncontrolled pressure increases rapidly.

The purposes of leak testing and the consequences of leaks in the system

No system can operate at maximum efficiency if there is leakage into or out of the refrigerant circuit or if the circuit is contaminated with air or moisture. If there is a leak in the system and the system is operating at a pressure below atmospheric pressure, air and moisture will enter the system. The air will cause the condensing pressure to rise above normal pressure and the moisture will cause acids, sludge and copper plating of the system. If the system operates at a pressure above atmospheric pressure the refrigerant will leak out which will soon cause a reduction in system capacity. If all the refrigerant leaks out and the compressor is kept running it may burn out because of lack of cooling by the refrigerant gas entering the compressor.

The correct leak testing procedure

Step 1: Introduce a detectable gas into the system, and pressurise the system to approximately 50 kPa (g). The gas used should preferably be the refrigerant with which the system normally operates.

Step 2: Increase the test pressure to 500 kPa(g) with nitrogen and test for large leaks.

Step 3: Increase the test pressure to 1 000kPa(g) with nitrogen and test again for leaks.

Step 4: Increase the test pressure to the appropriate test pressure and test the system for leaks, using the most suitable method(s) for the system.

Appropriate pressures are as mentioned previously, or the nameplate on some systems indicates the recommended test pressure. In this case this is the pressure to be applied. All manufacturers’ instructions must be followed.

In summary

  • The procedure for leak testing should be conducted in a methodical manner. Each joint which is a potential leak must be tested. One method is to start at the discharge valve of the compressor and follow the pipes to the inlet of the compressor.
  • When moving the electronic detectors, the sensor must be moved at a speed of approximately 2 cm per second. Moving faster than this will lead to smaller leaks not being detected.
  • When using the soap and water method, each joint must be painted with the soap solution.
  • When using the electronic method, the system must be pressurised first with the correct type of refrigerant (usually the type the system uses) to a pressure of about 50 kPa.
  • Leaks on flares may be fixed immediately by tightening (not over-tightening) the flare nut before carrying on with the leak test. Use the correct torque settings for R410A systems.
  • All leaks found must be marked for later identification and corrective action.
  • To repair a leak, the refrigerant or nitrogen must be removed from that part of the system – the entire system may have to be evacuated if there are no shut-off valves.
  • When using the ultraviolet method, the surrounding light, if too bright, makes it more difficult to see the ultraviolet dye where it has leaked out. Thus, this method must not be used in direct sunlight.
  • To evaluate the leak test, all joints must be checked, and the system pressure must be observed. Any reduction in pressure means that there is a leak in the system, which must be found.
  • In addition, a vacuum test should be conducted.

James, I hope that this helps with your refrigeration work. Leak testing is an important aspect, and all systems must be leak tested. After working on the pressure envelope of a system a new certificate of conformance must be issued with the pressure test results stated on the certificate.

Thanks to everybody for the overwhelming response. I receive on average over 60 questions each month and cannot publish all of them. But keep them coming, as I may answer you directly. Looking forward to hearing from you.

Grant Laidlaw


  1. SETA training
  2. Testo 


 About Grant Laidlaw

Grant Laidlaw

Grant Laidlaw is currently the owner of the Air Conditioning and Refrigeration Academy (ACRA) in Edenvale. He holds a Bachelor of Business Administration and an associate degree in educational administration. He has a National Technical Diploma and completed an apprenticeship with Transnet. He has dual-trades status: refrigeration and electrical. He has been involved with SAIRAC for over two decades and served on the Johannesburg committee as chairman and was also president between 2015 and 2018. Currently he is the SAIRAC national treasurer.

Register for free to gain access the digital library for RACA Journal publications