Service Valve and Schrader Core Replacement

Purpose

The system that has been topped up twice in two years is usually leaking at a service port, and the service port has three separate sealing surfaces that fail for three different reasons. A core weeps because it has been depressed by a hose a hundred times. A cap leaks because someone left it off, or fitted one with no gasket, and the cap is the primary seal on that port rather than a dust cover. A valve stem leaks through its packing after being cranked. Replacing the wrong one, or topping the system up again without finding out which, is how a shop ends up funding a customer's refrigerant habit.

Scope

Covers Schrader core replacement, service port cap sealing, and service valve stem and body repair on residential and light commercial split systems, heat pumps and packaged units, on both the front and back seating valves at the condensing unit and any line-mounted access fittings.

Does not cover system leak searching beyond the port itself, which the refrigerant leak response SOP owns; recovery method and cylinder handling; brazing and nitrogen purge; or evacuation and decay targets. Where a valve body must be cut out, this procedure hands off to those three siblings rather than re-deriving them.

Roles and responsibilities

Procedure

1. Localize the leak to one of the three surfaces before anything is opened. With the system running and the caps still on, sweep the port with an electronic detector, then shut down, remove the cap and re-check the core seat, the stem packing and the valve body separately, confirming any hit with bubble solution. Acceptance: a hit attributed to exactly one of core, cap seat or stem and body, confirmed a second way. Wrong looks like a general indication around the whole valve with no surface named, usually from wind or from oil that migrated. Stop rule: an unlocalized indication goes back to the leak response SOP rather than becoming a core replacement on a hunch. Hazard: you are working at high side pressure with eye protection on, and a detector probe never goes into the fan discharge.

2. Shut down, isolate, and let the pressures equalize before a tool touches a port. Stop the system at the thermostat, open the disconnect, prove dead with the meter proved on a known live source before and after per NFPA 70E-2021, 120.5 and work practices at 29 CFR 1910.333(b)(2), then let the standing pressure settle. Acceptance: under 1 volt at the line terminals, and a standing pressure consistent with the refrigerant's saturation pressure at the outdoor ambient, which tells you the system still holds a charge. Wrong looks like a core pulled with the compressor able to restart on a thermostat call. Stop rule: no core comes out until the disconnect is open, because a compressor starting with a core removal tool on the liquid port puts high side pressure through the tool. Hazard: liquid refrigerant freezes skin instantly and the port is at system pressure, so gloves and eye protection stay on from here to step 7.

3. Choose the repair path on the evidence, not on convenience. Decide between a core change under pressure, a stem packing repair, and a valve body replacement that requires recovery. Acceptance: a path chosen against three facts written down, which surface leaked, whether the port accepts a core removal tool with full thread engagement, and whether the valve is a serviceable design with a replaceable packing gland. Wrong looks like a core change ordered for a leak that was actually at the stem. Stop rule: a valve body or a brazed joint leak is not a core job, so it goes to recovery and the service manager is called before the charge comes out. Hazard: none at this step, it is a decision made at the truck with the system still closed; its risk is a system opened for the wrong reason.

4. Change the core with a core removal tool, fully seated before anything is broken loose. Thread the tool fully onto the port, close the tool's valve, then break the core loose, withdraw it, fit the new core and drive it home. Acceptance: tool threaded to full engagement with no movement at its own gasket, venting limited to the small volume trapped inside the tool, and the new core seated to the valve manufacturer's torque with a core driver rather than by feel. Wrong looks like a tool that will not thread down, or one that hisses at its base while the core is loose. Stop rule: never break a core loose on a partly seated tool, because that vents the system into your hands and is also a knowing release under 40 CFR 82.154. Hazard: this is a pressurized port with the core about to come out, so the tool body points away from your face and the hand behind it, not around it.

5. Repair the stem and packing, or condemn the valve body. Where the leak was at the stem, back the valve to the position the manufacturer specifies and tighten or replace the packing gland if the valve is a serviceable design. Acceptance: no indication at the stem after the repair with the system at operating pressure, and a stem that turns without the packing weeping. Wrong looks like a stem sealed by cranking it hard against its seat, which distorts the seat and shows up as a valve that no longer closes. Stop rule: a non-serviceable valve or a leaking valve body is a recovery and a brazed replacement, following the recovery and brazing SOPs, not a tightening exercise. Hazard: the system is still at pressure through this step, and over-torquing a stem can part it, so the wrench stays on the manufacturer's flats and the body is supported against twisting the line set.

6. Fit the correct cap with its seal, and set the stem position deliberately. Install a cap with an intact gasket or O-ring on every port, and leave each valve stem in the position the manufacturer requires for normal operation. Acceptance: every port capped with a sealing cap torqued to the manufacturer's figure, and each stem fully in its specified running position rather than partway. Wrong looks like a bare cap with no gasket, or a valve left cracked open one turn, which throttles the line and shows up later as a charge nobody can make sense of. Stop rule: fit the correct cap rather than reuse a gasketless one, because the cap is the sealing surface the core is backing up, not the other way round. Hazard: the ports are still at system pressure, so caps go on hand tight first and the wrench finishes them.

7. Restore, verify at operating pressure, and re-prove the protection you opened. Close the disconnect from the hinge side, run the system, and re-check all three surfaces at operating pressure, then again after a stated interval. Acceptance, all four: no indication at core, cap or stem with the detector at operating pressure; no indication on a repeat check at least 15 minutes later; suction and liquid pressures and the resulting superheat and subcooling inside the manufacturer's chart; and the disconnect pulled with the unit running to confirm it stops the unit, which is the protection you opened at step 2. Wrong looks like a clean first check and a hit on the repeat, which is a core seated crooked. Stop rule: a repeat-check indication is re-done now rather than noted as a small leak. Hazard: the cabinet is energized with the fan turning and hoses under system pressure, so panels go on before power does, hoses route clear of the blade, and the customer stands clear before the disconnect closes.

Exception path. A customer who wants a top-up rather than a repair gets a written note of what leaked and where, and the shop's position stated plainly: adding refrigerant to a located leak is a decision they are making, not a service being recommended. The federal leak-repair requirements at 40 CFR 82.157 apply to appliances with a full charge of 50 pounds or more, so most residential systems sit outside them and the standard here is the shop's own, while the venting prohibition at 82.154 applies to every system regardless of size. Where the port threads are damaged past a tool's engagement, the visit becomes a recovery and a valve replacement, not an improvised adapter.

The record this produces

One line per port, filled where it is read: which of the three surfaces leaked and the two methods that confirmed it; standing pressure and outdoor ambient at step 2; the repair path chosen and why; core torque method; cap type and whether the old one had a gasket; recovered and weighed-in amounts where the charge came out; and both step 7 detector checks with their times.

These land against the unit serial. Naming the surface is what stops the next tech repeating the work, because "leak at the service valve" sends someone back to the same three surfaces while "weeping core on the suction port, cap missing on arrival" tells them what happened and who left it that way.

Worked pass: 3 ton R-410A heat pump, topped up twice in fourteen months

Step 1: with the system running and caps on, the detector hit at the suction service port. Shut down, cap removed, and the hit localized to the core seat with bubble solution showing a slow riser there, nothing at the stem packing and nothing at the valve body braze. The suction cap was present but had no gasket in it.

Step 2: system stopped, disconnect open, 0.2 volts at the line terminals with the meter proved live either side. Standing pressure settled at 198 psig against a 68 F outdoor ambient, which is close to the saturation pressure for R-410A at that temperature, so the system was still holding most of a charge and a core change under pressure was on the table.

Step 3: surface named as the core, valve confirmed non-serviceable at the packing gland but that did not matter because the stem was tight, and the remaining question was thread engagement for the core tool.

Step 4 FAILED. The core removal tool would not thread past about half its engagement, and the port threads showed the flattening a cross-threaded hose leaves. Under this step's stop rule the core does not get broken loose on a partly seated tool, so the job changed shape rather than being forced. The service manager approved recovery. Nameplate charge 7 lb 2 oz, which is 114 ounces, and 6 lb 9 oz came out, which is 105 ounces, so 114 minus 105 leaves 9 ounces gone, and 9 divided by 114 is 7.9 percent of the charge lost across fourteen months, consistent with a weeping core rather than a hole.

The valve body was cut out and a new service valve brazed in with nitrogen flowing per the brazing SOP, then nitrogen pressure tested to the manufacturer's figure with no measurable loss, evacuated to 390 microns and held at 445 microns after 10 minutes, a 55 micron rise and inside the 500 micron gate. 7 lb 2 oz was weighed back in.

Steps 5 and 6: stem checked at operating position and dry, both ports capped with gasketed caps torqued to the manufacturer's figure, and both stems left fully in their specified running position. Step 7: detector clean at core, cap and stem at operating pressure, clean again on the repeat check at 22 minutes, superheat 11 F and subcooling 10 F against the chart, and the disconnect pulled with the unit running to confirm it stopped the unit.

What the failure teaches: everything before step 4 pointed at a five minute repair, and the correct call was to abandon it. A tech who leans on the tool gets it another turn or two, breaks the core loose, and discharges a system into the yard at the exact moment their hand is closest to the port. The damaged thread was not visible until a tool was on it, which is why step 4's acceptance is full thread engagement checked before the core is disturbed rather than after.

References

• 40 CFR Part 82 Subpart F, the venting prohibition at 82.154 and the leak repair requirements at 82.157, which apply at a full charge of 50 pounds or more
• 29 CFR 1910.333(b)(2) with 1910.332, and NFPA 70E-2021, 120.5, for the isolation and proving in steps 2 and 7
• Equipment manufacturer's service literature for valve stem running position, cap torque and core torque
• See related: Refrigerant Leak Response and Repair Authorization; Brazing With a Nitrogen Purge Standard; Evacuation and Vacuum Decay Verification; Charge Verification by Weight and Superheat