How to Make and Torque a Flare on a Mini-Split Line Set

Why this matters

A mini-split has four flare joints on a single-zone install and every one of them is a potential slow leak. The classic ductless callback is the unit that cooled fine at start-up and is blowing room-temperature air six weeks later, with a quarter of its charge gone through a 1/4 inch flare that looked good and was never quite right. The tech who comes back finds nothing obvious, because the leak is a few ounces a week at a joint that was soaped at 150 psi of nitrogen and never again.

Almost every one of those leaks traces to the same short list: a cut that was not square, a burr left in, a flare a hair short, a scratch on the face, or a nut tightened by feel. Each one is decided in the minute it takes to make the joint, which is why this is a procedure and not a judgment call.

Why a flare seals at all, and what each failure does to the seal, is in What a Flare Joint Depends On. The torque values and flare dimensions are in Mini-Split Flare Torque and Flare Size Field Card. This article is the hands-on sequence.

Start state

• The maker's installation manual open to the flare section. It sets the flare dimension, the torque, and whether to oil the flare and where. Those three vary between makers, and the manual governs.
• Tools: a tube cutter with a sharp wheel, a reamer or deburring tool, an eccentric-cone flaring tool made for R-410A and later refrigerants (with its height gauge), a torque wrench with flare-nut crowfoot heads or a flare torque wrench set, a backup wrench, and the flare nuts the unit shipped with.
• The system is not charged into the line set. The outdoor unit's refrigerant stays behind closed service valves until the joints pass a pressure test and evacuation. The indoor unit usually ships with a holding charge of nitrogen or dry air; when you remove its stub caps, point the stubs away from your face, because the gas comes out with a pop.
• For an A2L unit, read the maker's joint requirements and A2L Equipment Install Requirements Field Card before making any flare, especially one that will end up inside the room. Some makers restrict which joint types may sit in the occupied space.

The sequence

1. Cut square, with a light feed. Tighten the cutter a quarter turn per revolution, not a half. A heavy feed pushes the tube wall inward and leaves a lip; a cutter that walks leaves the end at an angle. An angled end flares off-center: one side of the sealing ring wide, the other side thin, and the thin side is where it leaks. If the end of the tube was dented or crushed in the box, cut back past the damage first.

2. Ream with the tube end pointing down. Remove the inside burr with a few light turns so chips fall out of the tube, not into it. A burr left inside becomes a ridge on the flare face. Ream too hard and you thin the wall at the end, which makes a short, weak flare. You want the inside edge smooth, not chamfered.

3. Put the nut on before you flare. Every tech has flared a perfect end and then looked at the nut on the bench. Check its orientation, threads toward the end of the tube, and use the nut supplied with this unit; nuts from an old system or a different pressure class may thread on and still be the wrong cone and wall for this flare.

4. Set the height and flare. Clamp the tube in the bar with the end at the height the tool's gauge shows; too low makes a short flare, too high makes an oversized one that may split or not fit the nut. Turn the cone until the clutch slips or the tool says the flare is done. The eccentric cone burnishes the face as it forms, which is why a modern tool makes a better seal than an old ram-style tool.

5. Inspect it before anything else touches it. Look into the flare under good light, and measure it.

 LOOKING INTO THE FLARE (end-on)

   GOOD                    OFF-CENTER
    .------.                .------.
   /  .--.  \              / .--.   \
  |  |    |  |            | |    |   |
   \  '--'  /              \ '--'   /
    '------'                '------'
   sealing ring even       ring wide on one side,
   all the way round       thin on the other

   SHORT                   CRACKED
      .--.                  .------.
     /.--.\                /  .--.  \
    ||    ||              |  |    | /|
     \'--'/                \  '--' \/
      '--'                  '------'
   ring narrow, outer      split running out
   diameter under gauge    through the rim

A good flare has a smooth, shiny face with no scratches or tool marks, a sealing ring of even width all the way round, a rim with no splits, and an outer diameter at the face (the A dimension) matching the manual. Check the outer diameter with a flare gauge or calipers against the card's table: about 9.1 mm for 1/4 inch and about 13.2 mm for 3/8 inch, with the manual's figure and tolerance governing. Then slide the nut up over the flare: it should cover the face fully and the flare should sit in the nut's cone without catching.

Any of the three defects in the drawing is a cut-off-and-redo. A scratch across the face is too. Do not try to rescue a flare by re-running the tool over it; the face has already formed and a second pass thins it.

6. Oil only as the manual says. Some makers call for a drop of refrigeration oil of the specified type on the back of the flare and the nut threads, so the nut turns smoothly and the torque reaches the face. Some say keep oil off the sealing face. Some say no oil at all. Use the type and the place they name; mineral oil on a POE system is a contaminant, and oil on a face some makers require dry can lubricate a leak path.

7. Align and start the nut by hand. Bring the tube straight into the fitting so the flare meets the cone square, and turn the nut by hand until it seats. If it binds, back it off and find out why: a crooked tube, a crossed thread or a flare that is too big. A nut started with a wrench on a crossed thread will reach its torque and not seal.

8. Torque with two wrenches. Hold the valve body or union with a backup wrench so the force goes into the flare, not into twisting the service valve or the indoor stub. Set the torque wrench to the middle of the range the card and the manual give. If you use a crowfoot head, set it at 90 degrees to the wrench handle; in line, it lengthens the wrench, and the nut gets more torque than the dial reads. Pull smoothly until the wrench clicks, once. Do not "add a little for luck"; an over-torqued flare thins at the root and cracks after a season of heating and cooling, and the nut is scrap.

9. Pressure test, then evacuate. Every flare is soaped at each stage of a standing nitrogen test as described in How to Run a Standing Nitrogen Pressure Test on a New Line Set, and only then does the system get evacuated per Evacuation and Vacuum Decay Verification. Release nitrogen to 0 psig outdoors before pulling cores. After the service valves are opened and the system runs, check the flares once more with an electronic leak detector suited to the refrigerant.

Worked example: a 12,000 Btu/h single-zone

A 12,000 Btu/h wall unit with 25 feet of 1/4 inch liquid and 3/8 inch gas line. Four flares: two at the indoor unit's stubs, two at the outdoor service valves. The manual calls for refrigeration oil on the back of the flare only, and torque ranges that match the card: 1/4 inch at 14 to 18 N-m, 3/8 inch at 34 to 42 N-m.

Wrench settings. Middle of each range: 16 N-m for 1/4 inch, which is 16 x 0.7376 = 11.8 ft-lb, and 38 N-m for 3/8 inch, which is 38 x 0.7376 = 28.0 ft-lb. The crew's wrench reads in N-m, so they set it in N-m and skipped the conversion.

Indoor flares. Both made, inspected and measured: 1/4 inch at 9.1 mm, 3/8 inch at 13.2 mm, faces clean. Torqued with a backup wrench on the indoor stub's flare union.

Outdoor 3/8 flare, first attempt. The end of the 3/8 tube had been flattened slightly in the box. The tech cut about 1 inch back, but the cutter walked on the flattened section and left the end slightly angled. After flaring, the ring was visibly wide on one side and thin on the other: off-center, the second drawing. The nut fit. The tech cut it off anyway, an inch back on round tube, reamed, flared again, and got an even ring at 13.2 mm. The redo cost about another inch of line set and four minutes. Left in, it was the leak.

Outdoor 1/4 flare. Clean on the first try, 9.1 mm.

Torque. All four pulled to the wrench's click at 16 and 38 N-m, backup wrench on every one.

Pressure test. The manual's stated test pressure for this unit is 450 psig (illustrative; read yours). Stages at 150 and 300 psig, soap on all four flares and both service valve stems, a full minute each: dry. Held at 450 psig, corrected for temperature, and passed. Vented outdoors, evacuated, decay test passed, valves opened.

Six weeks later. No callback. That is the only finish test that really matters, and it is the one the next five checks are designed to predict.

Finish test

• Each flare inspected and measured before joining: even ring, smooth face, no splits, outer diameter to the manual.
• Each nut torqued with a calibrated torque wrench to a value inside the manual's range, with a backup wrench, recorded as the value set.
• Each flare soap-dry at every stage of the nitrogen test, and the test passed after temperature correction.
• Evacuation and decay test passed before the valves were opened.
• An electronic leak check at each flare with the system running, using a detector suited to that refrigerant.

Write the torque value used for each size on the job record. A leak call later goes very differently when the ticket shows the joint was made to a number.

References

• Indoor and outdoor unit installation manuals: flare dimension, torque range, oil instructions, test pressure
• See related: Mini-Split Flare Torque and Flare Size Field Card; What a Flare Joint Depends On
• See related: How to Run a Standing Nitrogen Pressure Test on a New Line Set; Evacuation and Vacuum Decay Verification
• See related: A2L Equipment Install Requirements Field Card