Refrigerant line brazing and nitrogen purge field guide

Brazing with a nitrogen purge, and why the inside of the pipe is the part that matters

Brazing a refrigerant line is joining two pieces of copper tubing with a melted filler metal that flows into the joint and seals it for the pressures a refrigerant system runs. The nitrogen purge is a low flow of dry nitrogen pushed through the inside of that tubing while you heat it, so the brazing heat has no oxygen to react with. Without the purge, the heat oxidizes the copper from the inside and lays down a black, flaky scale that the system later sheds into the compressor and the metering device. The braze holds the pressure. The purge protects everything downstream of the joint.

The joint you can see is the easy half. A good fillet, a clean fit, the filler drawn all the way around: that is craft you can inspect with your eyes. The inside of the pipe is invisible once the joint cools, and that is exactly where a brazed-without-nitrogen line carries its defect, hidden, for months. The system cools fine on startup day and the installer drives away. The scale is already in there, waiting for the oil to carry it home.

This guide is the brazing and purge step. Sizing the suction, liquid, and discharge lines so the pipe you braze is the right diameter lives in the line-sizing guide. Pulling the deep vacuum that comes after the leak test, the microns and the decay test, lives in the evacuation guide. Here the work is the joint itself: the purge, the filler, the heat, the safety, and the nitrogen pressure test that proves the joints tight before any vacuum goes on the system.

Why do you flow nitrogen when brazing?

You flow nitrogen so the brazing heat does not build copper-oxide scale inside the pipe. Heat copper above brazing temperature with air in the tube and the oxygen in that air reacts with the hot copper to form cupric oxide, a black flaky crust on the inner wall. Nitrogen is inert. Push a slow stream of it through the line and it displaces the oxygen, so there is nothing for the copper to react with and no scale forms.

That scale is the single most common reason a system that was brazed wrong fails, and it fails downstream, never at the joint. The flakes do not stay put. They break loose, and the compressor oil acts like a solvent that lifts them off the wall and circulates them through the whole loop. They wedge into the thermostatic expansion valve or pack the inlet screen of a metering device and starve the evaporator. They score the compressor and clog the suction screen. A new compressor that fails in a season, with black grit in the oil, is very often telling you the line set was brazed dirty.

This is why old hands treat the purge as non-negotiable rather than optional. You cannot scrub the scale out after the fact. Once it is in the pipe it is in the system, and the only honest fix is to cut the joint out and start over with the nitrogen flowing. Prevent it at the torch or own it at the compressor.

The flowing-nitrogen technique

The method is a trickle, not a blast. You set the nitrogen so a gentle stream moves through the tube, enough to keep oxygen swept out of the joint zone without building pressure that fights the molten filler. Many techs run it on the order of a few cubic feet per hour, a flow you can just feel on the back of a wet hand at the open end. The number that matters is low and steady, and the equipment manufacturer's install instructions are the authority when they give one.

Start the purge before the torch touches the copper and keep it running until the joint has cooled past the point where it could oxidize. Light the nitrogen first, braze, then let it run while the joint loses its color. Shut it off too early and the still-hot copper grabs oxygen from the air that rushes back in, and you have made scale at the finish line after doing everything else right.

The other half of the technique is giving the gas somewhere to go. Nitrogen has to enter one end of the assembly and leave the other through an open path, or it just pressurizes the section and does nothing. On a closed system you crack a valve, pull a core, or leave a fitting loose downstream so the stream actually flows past the joint instead of sitting against it.

The nitrogen purge setup

The setup is a dry-nitrogen bottle, a regulator, and a way to read or meter the low flow. Use dry nitrogen, the standard bottle off the truck, not shop air and never oxygen. A two-stage or single-stage regulator drops the bottle's high pressure down to a low working pressure, and a flowmeter or a flow-control ball lets you dial the trickle instead of guessing. Some techs run a dedicated low-flow purge regulator made for this; others meter it carefully off a standard regulator. Either works if the flow stays low.

Plumb it so the gas enters one end of the line and exits the other through an open port, with the joint you are brazing in between. The classic rig is a hose from the regulator to a fitting at one end, the assembly open at the far end, and the braze in the middle of that path.

The error that ruins joints here is too much pressure. Crank the nitrogen up and the stream pushes against the molten filler as it tries to fill the joint. You get a blown joint, a pinhole, or filler blown clean through into the bore. The fix is the opposite of intuition: back the flow down until it is barely moving. Low pressure, open path, steady stream. That is the whole setup.

What is the difference between brazing and soldering refrigerant lines?

Brazing and soldering both join copper with a melted filler, but they happen at different temperatures and make joints of very different strength, and only one belongs on a refrigerant line. Brazing uses a filler that melts above roughly 840 degrees F, and in practice the copper joint is glowing as you work it. Soldering uses a filler that melts below that, the tin-based solders you run on plumbing water lines, and the joint never gets near red.

Refrigerant systems run high pressures and see vibration and thermal cycling, and a soft soldered joint is not made for that duty. Refrigerant work is brazed, with a high-temperature filler that flows into the joint and forms a connection strong enough for the operating and test pressures the system will see. The brazing fillers used on copper refrigerant lines melt high enough that the base metal carries a clear color before the filler ever flows.

If you came up running solder on water pipe, the temptation is to treat a line set the same way, and that is a rookie tell. A soldered refrigerant joint can pass a quick check and then weep under pressure or fail under vibration later. Braze it, and confirm the filler you are using is rated for the joint and the metals you are connecting against the manufacturer's and the filler maker's data.

What filler rod do you use for copper refrigerant lines?

For copper-to-copper joints the workhorse is a copper-phosphorus filler, the BCuP family, because the phosphorus makes it self-fluxing on copper. The phosphorus reacts with the copper-oxide at the joint and cleans it as the filler flows, so you do not need a separate flux for copper-to-copper. The common rods carry a little silver, often described by their silver content such as a 0 percent, a 2 percent, a 5 percent, or a 15 percent silver grade. More silver buys a lower flow temperature and a more forgiving, ductile joint that takes vibration better; less silver is cheaper and works fine on a clean, well-fit copper joint.

When the joint is copper to a dissimilar metal, copper to brass or copper to steel, you change fillers. There the choice is usually a silver filler in the BAg family, and that filler is not self-fluxing on those metals, so it needs flux. Brass service valves and steel accumulator stubs are the usual places this comes up.

Match the filler to the metals and the joint, and treat the specific alloy and percentage as something the equipment manufacturer and the filler maker specify, not something to default to out of habit. The one rule that holds across the trade: copper to copper with a BCuP rod needs no flux, and copper to brass or steel needs a silver filler with flux.

When flux is needed and when it is not

Flux is for the dissimilar-metal joints, not for plain copper-to-copper work with a phosphorus rod. On copper to copper with a BCuP filler, the phosphorus in the rod is the flux, and adding paste flux on top of it does nothing useful and can leave a residue inside the system you did not need to introduce. Keep the flux off the copper-to-copper joints.

When you braze copper to brass or copper to steel, the filler is a silver alloy that does not clean those oxides on its own, so you brush flux onto the cleaned surfaces before you heat them. The flux pulls the oxide and lets the filler wet and flow into the joint. Use the flux made for the filler and the metals, apply it thin to the joint surfaces, and keep it where it belongs.

Flux residue is corrosive, so on the joints that need it, the leftover should be cleaned off the outside after the joint cools rather than left to sit on the copper. The short rule techs carry: no flux on copper-to-copper with BCuP, flux on copper-to-brass and copper-to-steel with silver filler.

Joint preparation and the capillary gap

A brazed joint is only as good as the fit and the cleaning, and both happen before the torch comes out. Cut the tube square, ream and deburr the inside edge so a clean stream of nitrogen and refrigerant moves through it later, and wipe or lightly abrade the surfaces that will mate. Burrs left in the bore become turbulence and oil traps for the life of the system, and a tube cut on a slant never seats right in the fitting.

Brazing relies on capillary action to pull the molten filler into the joint, and capillary action only works across a small, even gap. The tube and the fitting socket are sized to leave that clearance when they are seated, so a swaged or belled end and a proper insertion depth give you the gap the filler needs. Too tight and the filler cannot draw in; too loose and the gap is past what capillary action will fill, and you get voids.

Seat the tube to the full depth of the socket so the joint has the overlap length to hold pressure, then check that it is square and not cocked. A joint that is clean, round, fully inserted, and left with the right gap is a joint that will pull filler all the way around. Skip the prep and you chase pinholes you could have avoided.

Heating the joint: technique and color

The skill in brazing is heating the base metal, the copper itself, and letting the hot copper melt the filler, rather than melting the rod with the flame and dripping it on. Filler flows toward heat and into the gap by capillary action, so you bring the copper up to temperature evenly around the joint, then touch the rod to the joint and let the metal pull it in and around. If you have to hold the rod in the flame to melt it, the base metal is not hot enough and the joint will be cold.

Heat the heavier or larger side a little more, work the flame around the joint so the whole circumference comes up together, and watch the color of the copper rather than chasing a number. The copper takes on a dull glow as it reaches brazing heat, and the filler flows freely at that point. Keep the flame moving and feed the rod around the joint so the filler chases the heat all the way around the back side you cannot see.

The two ways to ruin it are opposite. Too little heat and the filler balls up, sits on the surface, and never wets in, a cold joint that leaks. Too much heat and you overheat the copper, burn the filler, and can weaken the base metal or burn through a thin tube. The target is even, enough, and no more, with the rod following the heat around.

Torch and fuel gas

The torch is whatever puts enough concentrated heat into the joint to bring the copper up fast, and the common choices trade heat output for convenience. Oxy-acetylene runs the hottest and brings up large-diameter and thick joints quickly, which is why it shows up on bigger field-built piping and rack work. Air-acetylene and the bottled fuel gases such as MAPP-type gas are lighter to carry and have plenty of heat for typical line-set and residential split-system joints.

Match the tip and the flame to the joint size. A tip too small for a large fitting cannot get the copper up to temperature before the heat conducts away, so you stand there cooking the surrounding metal while the joint never reaches braze heat. A tip too large on small tubing overheats and burns through.

Whatever the torch, the heat is the same problem the rest of this guide is about. It is enough heat to oxidize the inside of the pipe if there is oxygen in there, enough to start a fire on what is around the joint, and enough to decompose any refrigerant left in the system. The torch is a tool and a hazard in the same hand.

Brazing safety, the ones that hurt people

Brazing puts an open flame and high heat into tight, often flammable spaces, next to copper that may have refrigerant or oil in it, while you handle compressed gas. The hazards stack, so the safety is not optional and it is not generic.

Use nitrogen for the purge and never oxygen. Pure oxygen pushed into a tube during brazing turns the heat and any oil into a fire or a violent reaction, and people have been badly hurt grabbing the wrong bottle. The purge gas is dry nitrogen, full stop. Keep a fire extinguisher within reach and a heat shield between the flame and anything that burns, because you will braze near framing, insulation, wiring, and finished surfaces, and the flame wanders heat farther than you think. Watch for the fire you started behind the joint, in the wall cavity, after you have moved on.

Ventilate the space. Brazing fumes, flux fumes, and the gases that come off heated refrigerant residue are not air you want to breathe, and a closed mechanical room or attic concentrates them. Move air through the space.

And recover the refrigerant before you braze, every time. That hazard is big enough that it gets its own section next, because brazing on a charged or pressurized system is the one that puts you in the hospital.

Can you braze a line with refrigerant in the system?

No. You recover or fully clear the refrigerant first, then braze. Brazing on a charged or even a pressurized system is the mistake that turns a routine repair into an emergency, and it does it two ways. First, a sealed system full of refrigerant under heat builds pressure, and a joint or a component can rupture and throw metal and hot refrigerant. Second, and worse, an open flame on refrigerant breaks it down into toxic decomposition products, including acid gases and phosgene, a gas that is dangerous in tiny amounts and that has hospitalized and killed people in the trade.

You cannot remove the flame, because the flame is how you braze. So you remove the refrigerant from the path of the flame instead. Recover the charge into a recovery machine and cylinder, then make sure the section you are about to open is clear, because a trapped pocket of refrigerant behind a closed valve is still refrigerant the flame can find.

This is the bluntest rule in refrigerant brazing. If there is refrigerant in the system, the torch does not come out. Recover it, confirm the section is clear, then braze. There is no deadline and no shortcut that is worth the gas coming off a flame in a closed room.

The nitrogen pressure test

After the joints are brazed and cooled, you pressure-test the piping with dry nitrogen to find leaks before any vacuum goes on the system. Nitrogen is the right test gas because it is inert, dry, and cheap, and because pressurizing with refrigerant to find a leak is wasteful and, with some refrigerants, against the rules. You bring the piping up to a test pressure and look for the joints that do not hold.

The test pressure is the system's design pressure, and the number to use is the one on the equipment nameplate or in the manufacturer's instructions, not a rule of thumb off the truck. High-side and low-side test pressures differ, and they vary by refrigerant, so the nameplate is the authority. Bring the pressure up in steps rather than slamming it to full, and keep yourself and others clear of the gauge and the joints while it climbs, because a joint that is going to fail under pressure is a thing you do not want to be leaning over.

This is the step that catches the cold joint and the pinhole while they are still cheap to fix. Find a leak now, with nitrogen, and you cut and rebraze a dry, empty joint. Miss it now and you find it after the vacuum and the charge, with refrigerant escaping and the whole job to redo.

How do you pressure-test brazed refrigerant joints?

Bring the piping up to the manufacturer's test pressure with dry nitrogen, then find leaks and hold a standing pressure to prove the system tight. To find the leak, brush a bubble solution on every joint and watch for growing bubbles, or sweep the joints with an electronic leak detector suited to the gas. Bubbles are slow but they show you exactly which joint, which is what you want before you cut.

The standing-pressure test is the proof. You bring the system to test pressure, close it off, note the pressure and the time, and let it sit, often for a stretch of hours or longer on a job that calls for it. A pressure that holds steady, after you account for temperature, says the system is tight. A pressure that drifts down says there is a leak to find.

Temperature is the catch that fools people. Nitrogen pressure rises and falls with the temperature of the gas, so a system pressurized in a warm afternoon will read lower when the night cools it, with no leak at all. Read the ambient temperature along with the pressure at the start and the end, and correct for it before you call a drop a leak. A pressure change that tracks the temperature swing is the gas behaving, not a hole in your work.

After the leak test: pulling the vacuum

Once the piping holds the nitrogen pressure test, you bleed the nitrogen out and pull a deep vacuum to remove the air and the moisture before you charge. The leak test proves the system tight; the vacuum proves it dry and clean. They are two different jobs and you do not skip from a good braze straight to charging.

On a system that got wet, was open a long time, or sat with moisture in it, the practical move is a triple evacuation: pull down, break the vacuum with dry nitrogen to absorb the remaining moisture, and repeat, so each pull-down carries out more water than a single vacuum could. That whole subject, the micron target, the gauge, the pump and hose setup, and the decay test that proves the vacuum held, is covered in the evacuation guide. Run the leak test here, then go there for the vacuum.

The order: braze, test, evacuate, charge

The sequence is fixed, and each step protects the one after it. Recover any refrigerant first so the torch is safe. Braze the joints with the nitrogen flowing so no scale forms. Pressure-test with nitrogen to the design pressure to find leaks while the system is empty. Pull the deep vacuum to remove air and moisture. Then weigh in the charge. Do them out of order and you undo the protection.

The reason the test gas and the purge gas are nitrogen and not oxygen or air comes back to the same physics the whole guide turns on. Oxygen feeds the oxidation and the fire risk; air carries moisture into a system you are trying to dry. Nitrogen is dry and inert, so it purges without scaling and tests without wetting. There is no place in this sequence for shop air or oxygen.

Skip a step and the failure shows up later, somewhere else. Skip the purge and the compressor eats scale. Skip the leak test and the charge leaks out. Skip the vacuum and moisture and acid attack the system from inside. The order is the discipline that keeps each of those from happening.

Keeping the system clean

The nitrogen purge handles the scale, but it is one part of keeping junk out of a system that has to run sealed for years. Tube and fittings come capped for a reason: an open end of copper sitting in a damp mechanical room or a muddy trench takes on moisture and dirt that you then seal inside when you braze it. Keep the ends capped until the moment you join them, and do not braze a line that has been open to the weather without cleaning and clearing it first.

Moisture is the quiet enemy alongside scale. Water in the system combines with the refrigerant and the oil to form acids that attack the windings and the metal from inside, which is why the vacuum step exists and why you keep the open pipe dry. A line set that took on rain through an uncapped end carries that water straight into the charge.

The filter-drier is the last line, not the first. It catches residual moisture and fine debris the rest of the work missed, and it is sized and placed by the manufacturer for that job. It is there to handle the small stuff, not to clean up after a line that was brazed dirty or left open. Do the clean work upstream and let the drier do the polishing it was made for.

A2L and mildly flammable refrigerants

The lower-GWP A2L refrigerants now common in new equipment are classified as mildly flammable, which adds a layer to brazing work and changes what carelessness costs. The core rule does not change: you recover and clear the refrigerant before you braze, and you purge with nitrogen while you braze. With an A2L the recovery and clearing is also a fire-safety step, because the gas itself can ignite as well as decompose.

Ventilation matters more with these refrigerants because a leak or a residue pocket is flammable as well as something you should not breathe. Move air through the space, watch for low spots where vapor can collect, and follow the equipment manufacturer's A2L install requirements, which can include room-size and detection provisions that go beyond older refrigerants.

The standards and best-practice guidance for A2L work are still settling as the equipment rolls out, so treat the manufacturer's instructions and the current industry guidance as the authority rather than older habits. The safe defaults carry over and get stricter: recover and clear first, purge with nitrogen, ventilate, and confirm the A2L-specific steps for the equipment in front of you before you light the torch.

Common braze defects and what causes them

Most brazing failures fall into a short list, and each one points back to a specific thing that went wrong at the joint. Knowing the symptom tells you the cause, and the cause tells you what to fix on the next joint.

Internal scale is the no-purge defect: black flakes inside the pipe because oxygen was in the tube during heating. You do not see it from outside, you see it as a plugged metering device or a fouled compressor later. A leaky or cold joint is a heat and fit defect: the filler balled up and never wet in because the base metal was too cool, or the gap was wrong, and it leaks under the pressure test or in service. A burned or overheated joint comes from too much heat for too long, which can burn the filler, weaken the copper, or melt through a thin tube. A blown joint, with filler blown through into the bore or a pinhole at the seam, comes from too much nitrogen pressure pushing on the molten filler.

The pressure test catches the leakers and the blown joints while they are still cheap. The scale it cannot catch, which is the whole argument for the purge: there is no test on a finished system that shows you the oxide inside, so you have to have prevented it at the torch.

Workmanship and quality check

A quality brazed joint reads the same to the eye every time: a clean joint with a continuous fillet of filler drawn all the way around, no balled-up rod sitting on the surface, no scorched or burned copper, and a tube that went in square and to full depth. The filler should look like it flowed into the joint, not like it was puddled on top. On the joints that took flux, the residue should be cleaned off rather than left to corrode the copper.

The check is part visual and part test. You look at every joint for the fillet, the color of the copper, and any sign of a starved or overheated spot, and you back the look with the nitrogen pressure test that proves the joints hold. The one thing no inspection of a finished system can verify is the inside of the pipe, so the proof that the joints were purged is in the procedure, not in an after-the-fact check.

A commissioning agent or a careful service manager will ask how the joints were brazed, whether nitrogen was flowing, and whether the system held its nitrogen pressure test before evacuation. Those questions are the QC, because they get at the two failures, scale and leaks, that a glance at the outside of a clean-looking joint will never reveal.

Critical cooling and data-center refrigerant brazing

On critical-cooling work, computer-room units, close-coupled cooling, and the refrigerant piping that keeps a data hall in its temperature band, the cost of a brazing defect is not a callback, it is downtime on equipment that cannot afford it. The procedure does not change, but the discipline tightens, because a plugged metering device or a scale-fouled compressor on a critical unit takes cooling offline where there is the least slack to absorb it.

Everything in this guide gets done by the book on that work: the nitrogen purge on every joint with no exceptions, the filler matched to the joint, the full nitrogen pressure test held and temperature-corrected, and the deep evacuation after. The redundancy in a data hall protects against equipment failure, not against a line set that was brazed dirty and seeds scale into every compressor on the loop.

The thermal guidelines and design references for these spaces set the conditions the cooling has to hold, but the brazing itself is the same craft, executed without the shortcuts that a comfort-cooling job might survive. On critical cooling, the purge and the leak test are not where you save ten minutes.

What to document

A brazed system that fails six months out raises one question: was the work done right. The record is how you answer it, and on refrigerant work the things worth recording are the things you cannot see in the finished pipe. Capture that nitrogen flowed during brazing, the filler used, the test pressure and how long the system held it, the temperature correction on the standing test, and that the refrigerant was recovered before any torch work.

Write down the filler family and silver content, the test gas and the design test pressure off the nameplate, the start and end pressure and temperature of the standing test, and who did the work. If you reused or modified existing piping, record that it was cleared and purged. The point is the same as the leak test itself: make the invisible part of the job checkable later, so the record defends the work.

Common mistakes

• Brazing with no nitrogen purge, so internal oxide scale forms and later plugs the metering device and damages the compressor.
• Brazing on a charged or pressurized system instead of recovering and clearing the refrigerant first.
• Running the nitrogen too high, so the pressure blows the joint or drives filler into the bore.
• Using the wrong filler or skipping flux on copper-to-brass and copper-to-steel joints, or adding flux on copper-to-copper.
• Melting the filler in the flame instead of heating the base metal, leaving a cold joint that leaks.
• Overheating the copper until the filler burns or the thin tube melts through.
• Skipping the nitrogen pressure test and going straight to the vacuum, so leaks surface after the charge.
• Leaving tube ends open to weather or shop dirt, sealing moisture and debris into the system.
• Calling a temperature-driven pressure change a leak because the standing test was not temperature-corrected.

Field checklist

Standards and references

The equipment manufacturer's installation instructions are the first authority on a refrigerant brazing job, and they govern the specifics: the filler to use, the nitrogen test pressure for the high and low side, the line-set requirements, and the A2L provisions for the equipment. Where this guide gives a temperature, a filler, or a flow, treat it as the working practice and confirm the actual number against the manufacturer and the filler maker's data.

For the brazing itself, the AWS brazing classifications name the fillers, the BCuP copper-phosphorus group for copper-to-copper and the BAg silver group for dissimilar metals, and AWS brazing practice covers joint design, the capillary gap, and technique. Refrigerant handling, including the requirement to recover rather than vent and to keep refrigerant out of the path of the flame, falls under the EPA Section 608 rules in the United States, with recovery and certification requirements that apply to the work. For the A2L refrigerants, ASHRAE safety classification and the developing industry best-practice guidance set the flammability handling, and that guidance is still maturing, so follow the current edition and the manufacturer.

Two things are not negotiable regardless of which document you reach for. You flow dry nitrogen through the line while you braze, and you never braze on a system that still has refrigerant in it. Cite the standard that governs the point, and let the manufacturer's instructions and the adopted rules override any rule of thumb.

Units and terms

Refrigerant brazing carries its own vocabulary, and the same idea shows up under different names across a manufacturer sheet, a filler label, and a spec.

Nitrogen flow is given in cubic feet per hour, abbreviated CFH, on a flowmeter. Test and design pressures are in pounds per square inch gauge, psig, sometimes shown in kPa or bar on metric equipment. Filler rods are named by their AWS class and often by silver content as a percentage. The vacuum that follows the leak test is measured in microns, covered in the evacuation guide, not in psig.

Nitrogen purge

A low flow of dry nitrogen through the tube during brazing to displace oxygen and prevent internal oxide scale

Cupric oxide scale

Black flaky copper-oxide formed inside the pipe when copper is brazed with oxygen present; breaks loose and plugs the system

BCuP

Copper-phosphorus filler family, self-fluxing on copper-to-copper joints; common in refrigeration

BAg

Silver filler family for copper-to-brass and copper-to-steel joints; requires flux

Capillary gap

The small, even clearance between tube and fitting that lets molten filler draw into the joint

Standing pressure test

Holding a sealed nitrogen pressure over time, temperature-corrected, to prove the joints are tight

A2L

ASHRAE safety class for mildly flammable, lower-GWP refrigerants, with added handling and ventilation requirements

Related tools

Calculators and readiness checks for this work

Compare your options

FAQ

People also ask

• How do you size refrigerant lines?
• Why do you evacuate an AC system?
• What is a VRF system?
• Who can install medical gas piping?
• What is oil return in refrigeration?