Leak Detection
Refrigerant does not evaporate into thin air legally — if a system is low, it leaked, and your job is to find where.
By Anthony LimpertASE Master Automobile Technician- UV dye or a certified electronic detector finds most leaks. R-1234yf detectors must meet SAE J2913 or J2791 specs and are more sensitive than older units.
- Dry nitrogen is inert, dry, and safe. Shop air introduces moisture. Oxygen mixed with A/C oil can explode. Never use anything but nitrogen for leak testing.
Leak detection is a skill of method and equipment, and the shortcuts people take here are some of the most dangerous mistakes in the trade.
Finding Leaks the Right Way
For small refrigerant leaks, the reliable tools are UV dye and electronic leak detectors — soap bubbles and visual inspection alone miss the small, slow leaks that cause most comebacks. UV dye circulates with the refrigerant and glows under a UV lamp wherever it seeps out, leaving a permanent witness mark at the leak point. Electronic sniffers detect refrigerant gas directly. For R-1234yf systems, the detector must be certified to SAE J2791 or J2913 specifications — these units are more sensitive than the older R-134a-era detectors, and yf's different chemistry demands it. Work the detector slowly along the low points of lines and connections, since refrigerant is heavier than air.
When a system is too empty to leak-test with refrigerant, or when UV inspection turns up nothing, pressurize the system with dry nitrogen and listen, soap, or gauge-watch for the loss. Nitrogen is inert, bone dry, and safe at test pressures. Never substitute shop air — it pumps moisture straight into the system you are trying to keep dry. And never, under any circumstances, use oxygen: oxygen in contact with A/C compressor oil can explode. Dry nitrogen is the only acceptable pressurizing gas for A/C leak testing.
| Method | Standard / spec | Best use |
|---|---|---|
| Electronic detector, R-134a | SAE J2791 certified | Small active leaks; probe low points slowly — refrigerant sinks |
| Electronic detector, R-1234yf | SAE J2913 certified | Required sensitivity for yf; older detectors can miss it |
| UV dye | About 1/4 oz of refrigerant-safe dye per system | Slow seepers; leaves a permanent witness mark at the leak |
| Dry nitrogen pressure test | Typically 100-150 psi through a regulator; never exceed the low-side rating | Empty systems; listen, soap, and watch the gauge over time |
| Never use | Shop air or oxygen | Shop air injects moisture; oxygen plus refrigerant oil can explode |
Nitrogen cylinders run over 2,000 psi — always work through a regulator, and confirm the system's low-side pressure rating before testing.
- Confirm the system is genuinely low with gauges and note how low. The loss rate frames the search: a system empty in a week has a leak you can hear; one that took three summers needs dye patience.
- Sweep with a UV lamp first — every joint, the condenser face, the compressor shaft-seal area, and the evaporator drain. If dye was in the system, the leak is already marked and you are done in minutes.
- Sniff with the electronic detector along the bottom of every fitting and component, moving about an inch per second. Refrigerant is heavier than air, so probe below joints, and check the evaporator by sampling at the condensate drain with the blower off.
- If there is no hit and the system is too empty to sniff, recover what is left and pressurize with regulated dry nitrogen to 100-150 psi. Listen, soap suspect joints, and watch the gauge over time — a held pressure is also your post-repair proof.
- After the repair, evacuate, recharge with dye added, and recheck the repair point with the sniffer. That proves the leak you fixed is fixed, and the dye stands watch for the one you have not met yet.
- Recharging a low system without finding the leak. Refrigerant does not wear out or seep through intact walls — it left through a hole that is still there, and the customer comes back in a month having paid twice.
- Pressure-testing with shop air. Compressed shop air is loaded with moisture — you just injected the one contaminant the entire drying system exists to keep out. Dry nitrogen only.
- Using oxygen for a leak test because the bottle was handy. Oxygen in contact with compressor oil can explode. This is not a judgment call; it is a hard never.
- Sniffing fast and over the top of fittings. Detectors need slow probe movement, and refrigerant sinks below joints — a quick pass across the top misses exactly the leaks you were sent to find.
The sniffer alarms at the evaporator condensate drain but everything under the hood is clean. What does that tell you?
The evaporator core is leaking inside the HVAC case. Refrigerant escaping the core pools in the case and exits through the drain, which is why the drain is the standard sampling point. Confirm with dye traces in the drain water or a borescope before quoting — evaporator replacement is usually a dash-out job.
Why do R-1234yf systems require a J2913-rated detector instead of the older R-134a unit on the shelf?
The yf chemistry responds differently and the standard demands higher sensitivity than the R-134a-era spec. An outdated detector can pass a leaking yf system — and with a mildly flammable A2L refrigerant, a missed leak carries a safety dimension R-134a never had.
A nitrogen test holds rock-steady for 30 minutes, but the customer's system arrived empty. What next?
A static nitrogen test does not stress everything. Some leaks only open under operating temperature, vibration, or rotation — a compressor shaft seal that seeps while spinning is the classic. Charge with refrigerant and dye, run the system through real heat cycles, and recheck; some leaks only exist on a hot, vibrating engine.
The sniffer alarms at the evaporator condensate drain but everything under the hood is clean. What does that tell you?
Missed one? The reasoning above comes straight from the HVA exam bank, so this is the standard you will be held to.