What a Penetration Firestop Is Actually Restoring

Why this matters

The common belief on a job site is that a firestop has to "match the wall" - a 2-hour wall gets a 2-hour firestop and the hole is closed. That belief is half a specification. A rated assembly earned its hourly number by satisfying several separate criteria at once, and a firestop system is rated on those criteria separately too. Buy only the first one and you have restored flame passage and left the criterion that actually decides whether the stack of cardboard on the far side lights off. The difference is not academic: it is the difference between two published systems that both say "2 hour" on the label.

One gate decides most of it

The gate is this: what is on the unexposed side of this penetration, and can it touch the penetrating item?

Everything downstream follows. A steel pipe is a heat pipe. In a furnace test the pipe on the exposed side gets to fire temperature and carries that heat straight through the barrier along the metal, past whatever sealant is packed around it. Flame never passes. The pipe on the far side still gets hot enough to ignite what is stacked against it. That is the whole reason firestop systems carry more than one rating.

The ratings, each with what it measures

Through-penetration firestop systems are tested to ASTM E814, published in parallel as UL 1479, in the edition referenced by the model code your jurisdiction adopted. The test produces up to four ratings, and a published system may carry some and not others.

Three things fall out of that table immediately. The T rating is measured on the penetrant, not on the wall, which is why a bare metal pipe can hold an F rating equal to the barrier and a T rating of zero. The L rating is an air leakage number with its own units and its own reference pressure, so it is not derivable from the F rating and does not come free with it. And the hose stream is already inside the F rating - it is not something you add on top.

What the system is, and why substitution is not allowed

A firestop system is a listed assembly in the same sense a rated wall is. The listing names, at minimum:

• The rated assembly it may be used in, by construction type and minimum thickness.
• The penetrant: material, maximum diameter, and often wall thickness or gauge.
• The annular space: a minimum and a maximum, point-to-point. Both ends matter. Too tight and there is no room for the fill; too loose and the tested condition no longer exists.
• The fill material, by generic type and by depth, plus any packing or backing material and its density and compression.
• Whether the penetrant is continuously supported, and how close the nearest hanger sits.

Change any of those and you are not on that system. There is no partial credit and no interpolation between two published systems.

        exposed side        |        unexposed side
                            |
   wall thickness           |
  |<---------------------->||
   ####################### || #######################
   ####   barrier    ##### || ####   barrier   ######
   #######################  ||#######################
        packing material  ->|<- fill depth
             ================================     <- penetrant
        packing material  ->|<- fill depth
   #######################  ||#######################
   ####   barrier    ##### || ####   barrier   ######
   ####################### || #######################
                            |
   annular space is measured point to point,
   penetrant surface to opening edge, all around

Annular space is where field arguments happen. It is measured at the tightest point and at the widest point, not averaged, because a system with a stated maximum has to hold at the widest point and a system with a stated minimum has to hold at the tightest.

Through penetrations and membrane penetrations are different problems

A through penetration goes in one face and out the other. A membrane penetration stops inside the assembly: a recessed box, a can light, an access panel. Membrane penetrations are governed by their own limits in the adopted code and by the assembly listing, usually as a maximum opening area per unit of wall area plus a minimum separation between openings on opposite faces, or a listed protective treatment in place of the separation. Nobody thinks of an outlet box as a penetration, which is exactly why the back-to-back box is one of the most common defects in the field.

Joints are a third category with a third test: head of wall, wall-to-wall and floor-to-wall movement joints are tested under UL 2079 or ASTM E1966, and their listings carry a movement class the field detail has to respect. Ducts are a fourth category and not a firestop problem at all: a duct through a rated assembly is protected by a listed fire damper closing on heat, and through a smoke barrier by a smoke damper closing on a smoke detection signal, with combination units where the barrier is both. The barrier type selects the device, and no through-penetration system number covers either. Perimeter fire containment at the slab edge is tested under ASTM E2307. None of these is a firestop system in the E814 sense, and a through-penetration system number does not cover them.

One gate, two outcomes: the same pipe in the same wall

A 2-hour fire barrier separates a corridor from a back-of-house area on a single floor. Two identical penetrations, each a 4 in nominal steel pipe with roughly a 3/4 in annulus. Both need an F rating of at least 2 hours because the barrier requires 2 hours. That much they share.

Outcome A: the pipe lands in a storage room where pallets are stacked to the wall. The gate answers yes: combustibles on the unexposed side can contact the penetrant. Whether the adopted code REQUIRES a T rating here is a jurisdictional question rather than a settled one: in the model building code as commonly adopted the T rating attaches to through-penetrations of rated floors rather than rated walls, and any wall trigger comes from that jurisdiction's own text. Read it off the permit set. On the physics the answer is yes, so the system must carry a T rating as well as an F rating, and a bare-pipe system with F 2 hours and T 0 hours does not qualify. The listing that works here wraps the pipe in a defined insulation for a stated distance either side of the barrier, and that distance is a field of the listing, not a rule of thumb.

Outcome B: the same pipe lands inside a concealed chase with no storage and no access. The gate answers no, and depending on the adopted code's exception the F rating alone may be sufficient. That is a code-text question with a jurisdictional answer: NFPA 101, the model building code, and any local amendment all sit on it, and the authority having jurisdiction, which is a named role with legal authority under the adopted code rather than a synonym for whoever inspects, owns the interpretation.

Qualifier lines, printed:

• Annular space, both bounds: measured 3/4 in at the widest point and just under 1/2 in where the pipe sits low in the opening. The selected system's stated range has to contain both, or the opening gets adjusted before anything is installed.
• F rating basis: 2 hours, matching the barrier's required rating, hose stream already included in that number rather than added to it.
• T rating basis: 2 hours where required, measured on the penetrant and the fill at 325 degrees F above starting temperature, not on the wall face.
• L rating: not carried by either selected system, and not required here. If this barrier were also a smoke barrier, the air leakage criterion would be a separate specification line in cfm per square foot at its stated pressure differential, not something the F rating supplies.
• Correction character: selecting the insulated system in Outcome A is a re-basing, not an addition. The bare-pipe system was never partially compliant that got topped up; it was the wrong system.
• Adoption gate: the T-rating trigger quoted is the one in the code edition this jurisdiction adopted and amended, read off the permit set.

What it cost, in effort rather than currency: the survey of this barrier found 22 penetrations. Three sat in the Outcome A condition. Walking and documenting all 22 took about 2 hours for one person. Correcting the 3 took roughly 3 times that, because two of them had already been closed with the wrong system and the fill had to come out before the right one went in. The ratio is the argument for surveying before installing rather than after.

Sibling-rule check, printed: the barrier in this example is continuous deck to deck, which is the continuity rule the rated-assembly article states, so the penetrations are the only openings in question. No rounding here runs in the flattering direction: the annulus is judged at the widest point, and the ambiguous pipe in Outcome B is routed to the AHJ rather than assumed compliant.

How to verify a firestop before you accept it

Ask for the system number first, then check the four things the number promises: assembly type and thickness, penetrant type and size, annular space at its tightest and widest, and fill depth. If the fill is already cured, depth is only knowable from the installer's record and the photo taken before close-in, which is why an installation with no photograph is an open finding rather than a pass. Then ask whether a T rating was required at this location, and whether the installed system carries one. Finally check for the identification label the adopted code requires at the penetration, legible after installation, because a correct system with no way to identify it turns every future survey into a demolition.

Probe an annulus with a rigid gauge and not a finger: a packed opening can contain cut metal edges and mineral fibre, and disturbing mineral wool releases respirable fibre, so wear eye protection and a respirator selected under a written program meeting 29 CFR 1910.134 if you are opening or repacking rather than looking.

References

• ASTM E814 and UL 1479 (through-penetration firestop systems); UL 2079 and ASTM E1966 (joint systems); ASTM E2307 (perimeter fire containment), each in the edition referenced by the adopted code
• The model building code and fire code as adopted and amended by your authority having jurisdiction, which owns the T-rating trigger and the identification requirement
• 29 CFR 1910.134 (respiratory protection program) where fibre or dust exposure is created by the work
• See related: What a Fire-Rated Assembly Is, and How It Gets Broken; How to Firestop a Penetration So the Rating Survives