Why leak detection is more than a single tool

A refrigerant leak is rarely "one method fits all." Modern systems run with smaller charges, lower tolerances, and stricter EPA leak-rate thresholds — particularly under the 2024 revisions to §82.157, which tightened repair timelines for appliances with 50 or more pounds of charge. Picking the wrong detection method can leave you with a callback, an undetected 0.3 oz/year seep, or a contaminated recovery cylinder.

The four primary field methods are: electronic leak detectors, soap-bubble solution, ultraviolet dye with a UV lamp, and ultrasonic acoustic location. Each has strengths, blind spots, and specific refrigerant sensitivities. The best technicians layer them.

Electronic leak detectors

Electronic detectors sense refrigerant via either a heated diode sensor or an infrared (NDIR) cell. Heated-diode units react to halogen-bearing refrigerants — they are sensitive to R-22, R-410A, R-134a, and most CFC/HCFC/HFC molecules, but they can false-trigger on solvents, hand sanitizer, and certain cleaning fluids. NDIR units measure infrared absorption at a refrigerant-specific wavelength and have a far lower false-positive rate.

Best practices when using an electronic detector:

• Calibrate it to the refrigerant you are actually hunting. A detector calibrated for R-22 will respond to R-410A but with reduced sensitivity.
• Move slowly — the recommended probe speed is under 2 inches per second.
• Allow the sensor to recover between sample points; if you park the probe in a leak plume, the sensor will saturate and need to clear.
• Verify with a known reference leak at the start of every job. A 1 oz/year reference leak source is the field standard.

The exam treats electronic detectors as the primary detection method for systems containing 50 or more pounds of charge — you should know the 2024 rule's exact threshold and what "verified leak rate" means in that context.

Soap-bubble solution

Soap-bubble solution remains the most universally available method and the easiest to verify visually. It is excellent for medium-to-large leaks (above 0.5 oz/year) on accessible joints. The downside is that it requires a charged system — you cannot bubble-test an evacuated appliance.

When you bubble-test, use a professional-grade solution with a thin-walled applicator brush. Thicker solutions produce visible bubbles on smaller leaks but can also produce bubbles on rough castings that are not actually leaking. Confirm any marginal reading by re-pressurizing the system and re-testing, or by switching to an electronic detector to triangulate.

UV dye

Fluorescent dye is injected into the system, circulated by the refrigerant and oil, and then located at leaks with a UV-A lamp. It is excellent for pinpointing slow leaks in hard-to-access areas (heat exchanger coils, evaporator fittings) but slow — you must circulate the dye for at least 15 minutes under power before sweeping.

A critical exam point: not all dyes are compatible with all refrigerants and oils. Polyolester (POE) oils used with R-410A and HFCs dissolve some older dye formulations. Use dye that is specifically labeled for the refrigerant and oil in the system.

Ultrasonic detection

Ultrasonic detectors listen for the high-frequency turbulence generated when a pressurized gas escapes through a small orifice. They are refrigerant-agnostic, work on charged or uncharged systems, and excel at finding leaks in noisy environments where electronic detectors false-trigger or where the leak is on a vibration-prone joint.

The downside is that ultrasonic detection is line-of-sound: dense insulation, foam panels, and enclosures block the ultrasound. You typically need to remove service panels or use a contact probe to listen through a solid surface.

Putting it together: a layered approach

For a 5-ton rooftop unit with R-410A:

1. Walk the visual. Look for oil residue, ice formation on a warm day, and discoloration around fittings.
2. Electronic sweep. Start at the compressor service valves and work outward; the most common R-410A leak point is the liquid line braze joint at the condensing coil.
3. Soap the joint where the electronic detector chirps — confirm visually before quoting a repair.
4. UV dye follow-up for any slow-leak suspicion the electronic sweep did not conclusively pin down.
5. Document the leak rate using the §82.157 calculation method before opening the system for repair.

The regulatory hook

If the leak rate you calculate exceeds the §82.157 threshold (currently 10 percent for commercial refrigeration, 15 percent for comfort cooling, 20 percent for industrial process refrigeration, all measured over a 12-month period), the appliance must be repaired, leak-verified within 30 days, and recorded in the operator's service log. Skipping the verification step — or failing to log it — is the most common violation in EPA enforcement actions.

Leak detection is not just a service skill. It is the front door to compliance.