Brazing With a Nitrogen Purge Standard

Purpose

Copper brazed without a nitrogen purge grows a black scale on the inside of the tube. That scale does not stay put: it breaks loose, travels with the oil, and plugs the metering device or scores the compressor months later, on a job nobody connects back to the joint. The purge is the difference between a repair that lasts and a repair that returns as a warranty compressor with your name on the last ticket. This procedure also fixes the second half of the problem, which is that the purge is usually set and never verified, so a tech can run a full changeout believing gas is flowing through a line that is dead-headed behind a valve core.

Scope

Covers brazing of refrigerant-bearing copper on residential and light commercial split systems and packaged equipment: line set replacement, component changeout, and repair of a failed joint, including the nitrogen purge, hot work controls, filler selection, and the nitrogen pressure test that follows.

Does not cover brazing on systems still holding refrigerant, which is never permitted under this procedure. Does not cover welding or brazing of steel gas piping or hydronic steel, which is a different filler and a different hazard set. Does not cover recovery itself, which the recovery and evacuation SOPs own.

Roles and responsibilities

Procedure

1. Confirm the system is recovered and open to atmosphere before any flame comes near it. Recover refrigerant to the level required by 40 CFR Part 82, Subpart F, then break the system open so there is a vent path. Acceptance: a gauge reading at atmospheric with a physically open port, not a closed valve you believe is empty. Hazard in this step and it is the one that ends careers: R-410A, R-134a and other halocarbons decomposing over a flame produce hydrogen fluoride and carbonyl fluoride, which you inhale before you smell anything useful. Stop rule: any doubt that the system is empty, and the torch stays in the truck until you have proved it at an open port.

2. Set the hot work controls before you light anything. Clear combustibles from the work area or shield them, position an extinguisher within reach, and post a fire watch who is doing nothing else. Acceptance: a permit if the property requires one, a named fire watch, and nothing combustible you cannot see. NFPA 51B, in the edition your authority having jurisdiction or the property's insurer has adopted, sets the hot work program; OSHA's general industry rules are 29 CFR 1910.252, with 1926.352 as the construction counterpart. Stop rule: no fire watch available means no torch, and that is not a judgment call a tech makes alone in an occupied building.

3. Select the filler for the base metals, and know what it releases when heated. Copper-phosphorus (BCuP) alloys are self-fluxing on copper to copper and must never be used on ferrous or high-nickel base metals, where the phosphorus forms brittle phosphides. Silver (BAg) alloys with the correct flux handle copper to brass and copper to steel. Acceptance: the alloy on the rod matches the joint in front of you. Hazard in this step: cadmium-bearing BAg alloys release cadmium oxide fume, an inhalation hazard covered by 29 CFR 1910.1027, and gloves do nothing for it. Use cadmium-free filler wherever the joint allows, and where it does not, work under local exhaust with respiratory protection issued through a program meeting 29 CFR 1910.134.

4. Set the regulator and verify actual nitrogen flow with an exit path open. Regulate down to a few psig, open a downstream port so the gas has somewhere to go, and confirm flow at that open end by feel on a damp hand or with a flow meter. Manufacturers commonly specify a trickle on the order of 2 to 5 SCFH, enough to displace air without pressurizing the tube. Acceptance: measurable flow at the far end, not just a needle moving at the regulator. What wrong looks like: a Schrader core left in, a cap left on, or a closed service valve, all of which dead-head the purge while the regulator reads exactly the same. Hazard in this step: never use oxygen in place of nitrogen, never open a nitrogen cylinder without a regulator, since cylinder pressure is well over 2,000 psig, and remember nitrogen displaces oxygen in a closet or crawlspace, so ventilate the space you are purging into. Stop rule: no verified flow at the exit, no heat.

5. Clean and fit the joint to the depth and gap the filler needs. Ream the cut end, remove the burr, clean both surfaces to bright metal, and check insertion depth and clearance. Acceptance: bright metal on both mating surfaces, full insertion depth, and a capillary gap rather than a rattling fit or a driven interference fit. What wrong looks like: sanding cloth used on a coated or plated line set without knowing what the coating is; a coating that releases fume when heated is an inhalation hazard, so identify it or cut back past it. Stop rule: a joint that does not fit does not get fixed with more filler, it gets recut.

6. Heat the base metal, not the rod, and feed the filler into the joint. Play the flame on the tube and fitting until the base metal draws the filler in by capillary action. Acceptance: a complete fillet visible the full 360 degrees around the joint, filler drawn into the socket rather than sitting on top of it. What wrong looks like: filler balled up on the surface, which means the base metal was too cold and you melted the rod with the flame instead. That joint is not a joint. Hazard in this step: a filter lens of at least shade 3 for light torch brazing and shade 4 where the flame is heavier, per the selection table at 29 CFR 1910.133(a)(5), and no quenching the finished joint with water or a wet rag, because the thermal shock cracks the filler on a joint that is about to hold system pressure. Stop rule: a cold-looking fillet gets cut out and redone, never re-heated and topped up.

7. Hold the purge until the joint is below color, then shut the nitrogen down deliberately. Keep gas flowing through the whole cooling period, then close the cylinder valve, bleed the regulator, and remove the purge line. Acceptance: joint cooled to black with the purge still running, and the regulator at zero before the line comes off. What wrong looks like: shutting the nitrogen off the moment the torch comes away, which lets air back into a tube that is still hot enough to scale the last inch you just brazed. Hazard in this step: the joint and the surrounding tube are still hot enough to burn through a glove for several minutes, so it does not get handled or moved on feel.

8. Pressure test with dry nitrogen to the pressure stamped on the data plate, and hold it. Regulate up to the low-side test pressure the manufacturer stamps on that unit, not a number you remember from another one, and hold for the interval the manufacturer specifies. Acceptance: no measurable decay over the hold, with ambient temperature at the start and end written down, because a gauge falling on a cooling evening is temperature and not a leak. Hazard in this step: never test with oxygen or with a mixture, never exceed the lowest-rated component in the loop, use a regulator with a relief, and stand clear of the capped ends while pressure is coming up. Stop rule: a decay you cannot explain by temperature is a leak; find it and rebraze, do not proceed to evacuation.

9. Keep the fire watch on after the torch is down. The watch stays in place for the interval the adopted edition of NFPA 51B requires, which in the 2019 edition is at least 30 minutes after the work is finished, followed by a monitoring period, and the property's own permit may require longer. Acceptance: the watch reports the all clear and it goes on the ticket with a time. What wrong looks like: packing up the moment the joint is cool, when the risk is smouldering insulation or framing behind a wall you never saw. Stop rule: the watch does not leave early because the tech is finished.

The record this produces

One brazing block on the ticket per joint or joint group: recovery confirmation and the port where atmospheric was proved; permit reference and fire watch name; filler alloy and whether it was cadmium-free; nitrogen purge verified at which open port, with the regulator setting; joints brazed, with location; pressure test pressure, hold duration, start and end ambient temperature, and the result; and the fire watch all clear time.

Two readers use it. The next tech who opens this system, who needs to know a joint was purged before spending an afternoon chasing a metering device restriction. And the shop, on a warranty claim: a compressor failure attributed to debris will get asked whether the piping was brazed under nitrogen, and a ticket carrying a purge verification and a pressure test hold with ambient temperatures on it is a different conversation from one that says "brazed and tested."

Worked pass: residential split system line set replacement

Condenser changeout with a full line set replacement, 3 brazed joints at the outdoor unit and 2 at the evaporator, R-410A system recovered the previous day.

Recovery was confirmed at an open low-side port reading atmospheric, and the system was left open. Combustibles cleared from the closet and the exterior wall behind the outdoor joints, extinguisher set, homeowner's teenage son was not accepted as the fire watch and the helper took it instead. Filler selected: cadmium-free silver alloy for the copper to brass service valve joints and a copper-phosphorus alloy for the copper to copper line joints, with no flux on the latter.

Step 4 FAILED. The regulator was set to a low trickle and the needle held steady, which looks exactly right. The tech went to verify flow at the evaporator end before lighting up and found nothing at the open end: the low-side service valve had been left front-seated after recovery, and behind it the Schrader core was still installed. The purge was running into a closed volume, pressurizing it slightly and moving no air at all. Under this step's stop rule that is a hard no-heat condition. The valve was back-seated, the core removed with a core removal tool, and flow was re-verified as a clear stream on a damp hand at the evaporator end before the torch was lit.

That failure is worth the extra six minutes because of what the alternative looks like. A regulator needle reads identically whether gas is flowing or sitting, so the only honest verification is at the exit. Every joint brazed under a dead-headed purge would have scaled internally, and the failure would have surfaced as a restricted metering device or a compressor with debris in it, months out, on a system whose paperwork said it was brazed under nitrogen.

Joints then ran normally. Base metal brought up until the filler drew, complete fillets confirmed on all five, purge held until each joint went black, then the cylinder closed and the regulator bled. Pressure test to the low-side test pressure stamped on the outdoor unit's data plate, held for the manufacturer's interval with ambient 84 F at start and 82 F at end, gauge steady. The 2 F drop was noted on the ticket precisely so the small gauge movement was not later read as a leak. Fire watch held 30 minutes past the last joint per the adopted edition of NFPA 51B, all clear logged with the time.

References

• NFPA 51B, fire prevention during welding, cutting and other hot work, in the edition your authority having jurisdiction or the property's insurer has adopted, for the permit and fire watch requirements
• 29 CFR 1910.252, general industry welding, cutting and brazing; 29 CFR 1926.352 is the construction counterpart
• 29 CFR 1910.1027 for cadmium and 29 CFR 1910.134 for the respiratory protection program referenced in step 3
• 40 CFR Part 82, Subpart F, for refrigerant recovery requirements before the system is opened
• See related: Evacuation and Vacuum Decay Verification; Charge Verification by Weight and Superheat