Duct Sealing and Mastic Application Standard

Purpose

This standing instruction sets the workmanship standard for sealing duct joints so the seal is still there in ten years, and it makes the crew prove the thickness they applied rather than describe it.

Duct sealing fails on application, almost never on product. It fails three ways: mastic used as a structural connection on a joint that was never fastened, a coat spread so thin it cracks when the metal moves, and a gap bridged with unsupported sealant instead of embedded mesh. All three look identical the day they are done. The difference shows up at the next leakage test, on somebody else's ticket.

Safety actions that gate this work

• Solvent-based sealants release flammable and irritating vapor, and a crawlspace concentrates it. Read the SDS before the lid comes off, ventilate the space mechanically, remove ignition sources including any appliance that can cycle, and use respiratory protection selected under a written program meeting 29 CFR 1910.134. Water-based products still require eye protection and skin cover.
• The system is off and isolated before any product is opened, not only for the electrical reason but because a running blower distributes solvent vapor and disturbed dust into the rooms where the family is sitting.
• Fibrous glass duct board and old liner release fibers when cut, trimmed or abraded. Cut with a sharp blade rather than tearing, bag offcuts as you go, and keep the same respiratory protection on for that work.
• Sealing a return in a space holding an atmospherically vented appliance changes the pressure that appliance vents against. Verify draft on every such appliance under worst-case conditions before you leave.

Scope

Covers sealing of accessible duct joints, seams, takeoffs, boots and plenums on residential and small light-commercial forced-air systems: materials, surface prep, mechanical fastening, application thickness, mesh reinforcement, cure and verification.

Does not cover duct leakage testing and reporting, which the duct leakage inspection SOP owns and which this procedure consumes and re-runs. Does not cover flexible duct geometry, owned by the flex duct inspection SOP, or insulation and vapor barrier repair, owned by the attic and crawlspace duct insulation SOP. Aerosolized interior sealing is out of scope.

The materials rule, stated as a spec

Two rules sit above all of that. Mechanical first: a joint is fastened to the duct construction standard your jurisdiction's mechanical code references before any sealant touches it, because mastic is a gasket, not a connection. Repair before seal: a disconnected boot, a rusted-through plenum wall or a collapsed platform is a repair, and sealing around it buries the defect where the next tech will not find it.

Roles and handoffs

Procedure

1. Split the findings into repairs and seals before anything is opened. Acceptance: a written list where every item is marked repair or seal, with disconnected fittings, rusted-through metal and failed platforms on the repair side. Wrong looks like mastic troweled over a boot that is hanging off its collar. Stop rule: a repair item is not sealed today; it is scoped and done first. Hazard: crawlspace and attic entry, so two lights, structure only, and nobody under the house alone.

2. Shut down and isolate before the first lid comes off. Acceptance: disconnect open with your lock and tag on it, dead proven live-dead-live on a known live source before and after per NFPA 70E-2021, 120.5, with work practices at 29 CFR 1910.333(b)(2), and the space ventilated. Stop rule: no work starts with the blower able to run, because it will carry vapor and dust into the house. Hazard: solvent vapor plus any ignition source is the fire risk here, so pilots and cycling appliances in the same space are shut down first.

3. Fasten every joint mechanically before it is sealed. Acceptance: fasteners at the count and spacing the duct construction standard referenced by your adopted mechanical code requires, flex cores under a positive mechanical clamp, and nothing relying on old tape for its connection. Wrong looks like three takeoff collars held on by cloth tape and about to be sealed in place. Stop rule: an unfastened joint is fastened first, without exception. Hazard: sheet metal edges and driven fasteners, so cut-resistant gloves and eye protection.

4. Prepare the surface and confirm the conditions the product needs. Acceptance: joints wiped free of dust, oil and loose rust, surfaces dry, and ambient and surface temperature inside the range on the product's data sheet, with the product name and both temperatures written down. Wrong looks like water-based mastic applied below its stated minimum temperature, which never develops strength. Stop rule: out of range means the work waits or the product changes. Hazard: solvent wipes carry their own vapor and flammability, so keep them out of a confined space.

5. Apply the listed closure to the stated thickness, with mesh at every gap. Acceptance: wet film thickness per the data sheet, verified by a wet film gauge or by comparing area covered against the product's rated coverage; mesh embedded wherever the gap exceeds the product's stated limit; coverage carried onto both sides of the joint by the distance the manufacturer states. Stop rule: a coverage check showing the container went materially further than its rated area means the coat is thin and gets a second pass before it skins over. Hazard: respiratory protection stays on for the whole application, not just the first joint.

6. Let it cure, then re-test rather than declaring it done. Acceptance: the data sheet's cure time elapsed before the system runs or before insulation goes on, then a repeat of whichever test produced the original finding, reported as before and after with the same basis. Wrong looks like insulation wrapped over uncured mastic, which traps solvent and never sets properly. Stop rule: a target not met is written as a target not met, with what is left named. Hazard: re-entry to the space after cure still needs the ventilation running.

7. Restore, verify draft and pressure, and record. Acceptance: panels on, your lock removed by you, power restored at the disconnect with the door closed, TESP re-measured and expected to rise slightly, system airflow re-measured, blower-door interlock proven by opening the door on a running blower, and no spillage at the draft hood of any atmospherically vented appliance within 60 seconds of firing under worst-case depressurization. Stop rule: spillage means the appliance does not go back in service and the combustion air situation is corrected first. Hazard: this is the step that restores power, air and combustion with the family in the house, so the CO monitor is on you for the draft checks.

The record this produces

• The repair-versus-seal split, with what was repaired and by whom
• Product name, listing mark, and the ambient and surface temperature at application
• Joint area covered, quantity of product used, the product's rated coverage, and the resulting thickness ratio
• Where mesh was embedded and the gap width that called for it
• Cure time observed before the system ran
• Before and after test results on the same basis, with any remaining target named
• TESP and system airflow before and after, and the draft verification result for each vented appliance

The test tech reads the thickness ratio when the next leakage number disappoints, and the office reads the repair list when the customer asks why the same duct is on a second invoice. The draft verification is the line that matters most if anyone asks what changed in that crawlspace.

Worked pass: crawlspace return platform and supply plenum, 3-ton system

Step 1: the boot at the back bedroom is hanging off its collar and goes on the repair list; the platform seams, the plenum seams and eight takeoff collars go on the seal list.

Step 2: disconnect open, lock and tag on, dead proven before and after on a known live source, crawlspace ventilated with a fan, and the gas water heater in that space shut down while the product is open.

Step 3: three of the eight collars are found held by cloth tape with no fasteners. They are fastened to the spacing the adopted duct construction standard requires before any mastic goes on them.

Step 4: surfaces wiped, ambient 58 F, surface 56 F, both inside the product's stated range, and the product name and both temperatures written on the ticket.

Step 5 FAILS. The joint area sealed measures 96 sq ft, and the crew used one container whose data sheet rates it at 55 sq ft at the specified wet thickness. The applied thickness is therefore 55 divided by 96, about 0.57 of specification, which is 43 percent short. Stop rule taken: a second pass goes on before the first skins over. Total product used is then two containers over 96 sq ft, which is 48 sq ft per container against a rated 55, so the thickness lands at 55 divided by 48, about 1.15 times specification. Mesh was embedded at four seams where the gap exceeded the product's stated limit.

Step 6: cure time observed per the data sheet before the system was allowed to run. Re-test on the same basis as the original: total duct leakage falls from 214 CFM at 25 Pa to 88, a reduction of 126, or 126 divided by 214, about 59 percent. The pressure pan at the return platform falls from 4.2 Pa to 0.4 Pa.

Step 7: TESP rises from 0.41 to 0.46 in w.c. against a 0.50 rating, the expected direction, because the leaks were an easy path and closing them raises the pressure the ducts see. System airflow falls from 1,180 to 1,150 CFM, about 2.5 percent, while summed register readings rise from 940 to 1,065 CFM. Those are not the same quantity as the leakage figures and are not compared to them: the leakage values are flow at a 25 Pa test pressure, these are delivered air at operating condition. Draft verified at the water heater with no spillage at the draft hood within 60 seconds of firing under worst-case depressurization, CO monitor on.

Checking the pass against the sections above: the repair item was separated in step 1 and not sealed over, which is the rule stated above the materials table; mechanical-first was enforced on three collars before any product went on them; the failing step is step 5 and its stop rule is taken on a measured coverage ratio rather than an impression of thickness; and the leakage figures are reported on the same test basis before and after, with delivered-air figures kept in a different unit.

References

• UL 181A for closure systems used with rigid fibrous glass duct and UL 181B for closure systems used with flexible duct and air connectors, with the marking on the product itself as the field evidence
• The mechanical code your jurisdiction has adopted, in its adopted edition, and the duct construction standard it references, for joint fastening before sealing
• The sealant manufacturer's data sheet for wet film thickness, rated coverage, application temperature range, gap limit and cure time, plus its SDS for the vapor hazard
• 29 CFR 1910.134 for the written respirator program, and NFPA 70E-2021, 120.5, for the live-dead-live sequence with work practices at 29 CFR 1910.333(b)(2)
• See related: the duct leakage inspection and report SOP and the flex duct inspection and correction SOP