Vapour Barrier Failures in Chilled Water Insulation

Vapour Barrier Failures in Chilled Water Insulation: Where the System Actually Gets Wet

Our chilled water piping guide says once, in passing, that vapour barriers are the part that actually fails. Then it moves on to thickness tables. This article is what got left out: how vapour transmission is measured, why a datasheet number is valid at only one set of test conditions, and where on a real pipe run the barrier stops being continuous. Insulation thickness is a heat-transfer decision. The vapour barrier is a survival decision, decided by workmanship at roughly a dozen locations per hundred metres of pipe.

Why Below-Ambient Is a Different Problem, Not a Colder Version of the Same One

On a hot water line, vapour drives outward. Moisture that gets into the insulation is driven back out. On a chilled water line the direction reverses permanently. Moisture vapour “in conditions with high water vapour pressure will travel toward conditions having low water vapour pressure,” and vapour drive “is influenced most significantly by humidity and temperature” (Johns Manville, The Importance of Vapor Retarders on Chilled Water Insulation Systems). The condition test is simple: vapour pressure drives moisture inward whenever “the dew point temperature is above the operating temperature of a system” (Owens Corning FOAMGLAS, Moisture intrusion in chilled water applications). With a typical chilled water operating band of 4.4 °C to 7.2 °C (40 °F to 45 °F) (same source), that condition is met most of the year in any humid climate, and effectively continuously in the Gulf.

The mechanism is more aggressive than slow ingress. The chair of the ASTM E96 task group describes the cold pipe as behaving like the desiccant in the laboratory test itself. In the field, the cold source on the far side of the insulation “effectively acts as the desiccant, pulling the moisture out of the air/insulation under the vapor retarder and thereby reducing the relative humidity on the inside of the vapor retarder” (NIA, Introduction to ASTM E96, Charlie Petty, Lamtec Corp., ASTM E96 Task Group Chair). The worked field case in that source: ambient at 90 °F/90 % RH (32.2 °C), 0 % RH under the vapour retarder, and a pipe at 36 °F (2.2 °C) labelled a “vapor magnet (like desiccant in test).” Where closed-cell foam is used with no separate sheet retarder, the same source states the case as ambient 90 °F/90 % RH (32.2 °C) against 36 °F/0 % RH (2.2 °C) under the insulation.

That distinction changes how you read a datasheet. With a porous insulation and a sheet retarder, the retarder sits at close to a single temperature — nearly isothermal across its own thickness, which is the condition the laboratory test replicates. With foam and no separate retarder, the temperature gradient falls across the insulation body itself, so the material is doing a job the cup test was never set up to describe.

The design objective follows, and it is a rate objective rather than an absolute one. NIA states it plainly: “In below-ambient applications, it is important to minimize the rate of water vapor flow to the cold surface. This is normally accomplished by using vapor retarders with low permeance, insulation materials with low permeability, or both” (NIA, Insulation Design Guide — Materials and Systems). Note the term of art. The correct name is vapour retarder. Nothing on the market is a vapour barrier in the literal sense; every product transmits some vapour, and the design question is how much and for how long. We use “barrier” in the title because that is what specifiers search for. Every clause below is written against retarder logic.

The Measurement Chain: WVTR, Permeance, Permeability — and Why Two of the Three Get Misused

Three quantities are involved and only one is measured. The measurement is water vapour transmission rate (WVTR): the steady water vapour flow in unit time through unit area, normal to specific parallel surfaces, under specific conditions of temperature and humidity at each surface. Water vapour permeance (WVP) is calculated from WVTR by dividing by the partial vapour pressure differential — the time rate of water vapour transmission through unit area of flat material or construction induced by unit vapour pressure difference between two specific surfaces. Water vapour permeability is the same quantity through unit area of flat material of unit thickness, calculated from permeance by multiplying by thickness, which is only legitimate for homogeneous materials. The chain runs WVTR → permeance → permeability (all definitions: NIA, Introduction to ASTM E96).

The units, with the symbol meanings written out, because most specifications quote them without:

Reference conversion: 1 grain = 0.065 gram = 0.0023 ounce (NIA, Introduction to ASTM E96).

The right-hand column is the most commonly abused item in this subject. Permeance is valid for a laminate, a facing or a whole assembly. Permeability is not — the NIA source, written by the E96 task group chair, states that permeability “is an inappropriate term for use with facings.” A facing is a laminate of film, scrim and adhesive; dividing its performance by its thickness produces a number with no physical meaning, because there is no homogeneous material to divide. Yet “perm-inch” appears on facing datasheets and in installation guides routinely. When a facing or tape specification quotes perm-inch, read it as perm and treat the thickness normalisation as noise. Our reading of the field literature is that this is a copy-and-paste artefact rather than a claim anyone defends — but it means two datasheets can look comparable when they are not.

US perm, metric perm and SI are three different numbers

European datasheets built on EN/ISO methods and North American datasheets built on ASTM are not directly comparable without conversion, and the two “perm” units differ by more than 50 %:

1 US perm = 0.659045 metric perm; 1 metric perm = 1.51735 US perm. The SI base unit is kg·s⁻¹·m⁻²·Pa⁻¹, and 1 kg·s⁻¹·m⁻²·Pa⁻¹ = 1.74784 × 10¹⁰ US perm (Wikipedia, Perm (unit)). A submittal listing “0.02 perm” without stating which perm has left a 52 % ambiguity in the governing performance number. Require the unit system on the submittal, not just the digits.

ASTM E96: What the Test Actually Does, and the Arithmetic Behind the Driving Force

The governing test method is ASTM E96/E96M, Standard Test Methods for Gravimetric Determination of Water Vapor Transmission Rate of Materials. It sits under ASTM Committee C16 on Thermal Insulation, Subcommittee C16.33 on Insulation Finishes and Moisture. It has existed as a standard since 1953, with latest revisions approved December 2021; UFGS cites the current edition as ASTM E96/E96M (2024a). “Methods” is plural because the standard contains both a Desiccant method and a Water method. E96 is the gravimetric or “cup” method, distinct from ASTM F1249, which uses an automated analyser (NIA, Introduction to ASTM E96).

Gravimetric means the result is a weight change. A specimen is sealed over a cup; the cup contains either desiccant (dry cup) or water (wet cup); the assembly sits in a controlled chamber and is weighed over time. The mass gained or lost per unit area per unit time is the WVTR. Everything else is calculated from it.

The driving force, calculated

The partial vapour pressure differential is what makes permeance a portable number, and it comes from one line of arithmetic:

Δp = p_sat(T) × (RH_high − RH_low), where p_sat is the saturation vapour pressure at the test temperature and RH is expressed as a decimal.

Worked for Procedure A, both sides at 73 °F (22.8 °C), the condition behind most published data:

• p_sat at 73 °F = 0.82 inHg
• Ambient side: 0.82 × 0.50 = 0.41 inHg
• Under the specimen, over desiccant at 0 % RH: 0.82 × 0.00 = 0.00 inHg
• Δp = 0.41 − 0.00 = 0.41 inHg (0.20 psi)

(All values: NIA, Introduction to ASTM E96.) That 0.20 psi is the entire driving force behind a number your whole system depends on. It is a small pressure, which is exactly why the failure mode is slow, invisible and cumulative rather than dramatic.

The four condition sets used in this industry — and the trap inside two of them

(All four rows: NIA, Introduction to ASTM E96.)

Read the first two rows again. Procedures A and B impart the same driving force — 0.41 inHg in both. What differs is that Procedure B additionally exposes the specimen to 100 % RH. A facing can therefore pass A and fail B, not because the vapour push got harder but because liquid-adjacent conditions change how the laminate behaves. This is a real and citable failure mechanism, and it is why the most technically precise clause in the US federal specification pins its sealant requirement to Procedure B (see the UFGS clauses below). If your service is a chilled line in a humid plant room, the wet-cup case is the one that resembles your site.

You may not extrapolate between condition sets

E96 states it directly: “Results of tests performed at a given set of conditions are not to be used to extrapolate results at other conditions. This is because the structural characteristics of a material may change at different conditions and impact the water vapor transmission properties” (NIA, Introduction to ASTM E96). E96 permits testing at any conditions the user chooses, and any report of results must state both the test type (dry or wet cup) and the conditions.

The consequence for a specifier: a “0.02 perm” value measured at Procedure A cannot be assumed to hold at 90 °F/90 % RH. For a Gulf plant room or an outdoor rack, the published number and the service condition are not the same case, and the standard forbids bridging them by calculation. If the higher-drive condition matters to your project, it has to be tested at that condition, not derived from the Procedure A figure. A submittal that gives a perm value with no procedure and no conditions is, per the standard’s own reporting requirement, incomplete.

What a Perm Number Is Worth in Litres

Perm values are abstract enough that the difference between 1 perm and 0.02 perm reads as small. Converted to water it does not. The calculation below is ours, worked from the definitions already given above so you can check every step: permeance in perm is grains/(hr·ft²) per inHg, so mass = permeance × Δp × area × time × 0.0648 g per grain. Take 100 ft² (9.29 m²) of faced insulation at the Procedure A driving force of Δp = 0.41 inHg established earlier, over one year (8,760 hours):

(Arithmetic ours, using the perm definition and the 1 grain = 0.0648 g conversion; the Δp = 0.41 inHg Procedure A driving force is from NIA, Introduction to ASTM E96. 359,160 grains × 0.0648 g = 23,274 g = 23.3 litres. We previously carried a higher published pair of figures here; we could not reproduce them from the Procedure A driving force this article uses, so we have replaced them with the calculation shown. Any such figure is only as good as the Δp behind it — at a hotter, wetter condition the driving force is several times larger and so is the water.)

Two qualifications on that arithmetic, both mattering more than the headline. First, this is a theoretical diffusion-only calculation — what passes through the intact material. Second, it excludes leakage at laps and penetrations entirely. That exclusion is the subject of the second half of this article. A 0.02 perm facing installed with an open lap is not a 0.02 perm system; it is a 0.02 perm sheet with a hole in it, and the hole appears nowhere in the arithmetic above. The difference in the table is exactly 50-fold — it has to be, because permeance enters the calculation linearly and 1 ÷ 0.02 = 50. That factor of 50 is the prize for specifying correctly. Losing it takes one unsealed seam.

The Three Standards You Are Actually Buying Against

Three documents do three different jobs, and specifications routinely cite the wrong one for the job in hand.

C755 contains the single most important sentence in this field for anyone writing a specification. It states that “uncontrolled water entry into thermal insulation is the most serious factor causing impaired performance,” and — critically — that “it is impractical to specify an as installed permeance value” (ASTM C755-25, Significance and Use). C755 identifies the design as being affected by retarder thickness, substrate, number of joints, sheet length and width, useful life, and inspection procedures.

Read what that concedes. The governing selection standard is telling you that installed performance cannot be specified as a number. You specify materials plus an installation procedure, and the outcome is decided on site. That is not a loophole; it is the honest position, and it is why this article is organised around failure locations rather than product selection. It also explains why post-installation verification is so weak in practice — no accepted field measurement returns an as-installed perm value for a completed pipe run.

ASTM C1136 Types, as far as they are publicly documented

ASTM C1136 (2023) classifies flexible low-permeance vapour retarders into Types I–VI. We have verified two of them, and only through the US federal specification that invokes them: Type I at a maximum 0.02 perms with 50 Beach units minimum puncture resistance, and Type II at a maximum 0.02 perms with 25 Beach units (UFGS 23 07 00 Thermal Insulation §2.2.9.3). Both documented types cap permeance at the same 0.02 perms; the difference between them is mechanical, not vapour-related.

Three limits on how far you should carry those two rows. First, our source is UFGS 23 07 00 §2.2.9.3 invoking C1136 — we have not read the C1136 text itself, so these are the federal specification’s invoked values, not a direct reading of the standard. Second, the perm figures inherit the reporting problem this article has already set out: UFGS pins its sealant clause to ASTM E96/E96M Procedure B, and “0.02 perm” is ambiguous by 52 % between US perm and metric perm. A type designation alone does not tell you the procedure or the unit system, so a submittal citing “C1136 Type I” still owes you both. Third, we do not have Types V and VI and we will not guess at them.

Types III and IV are documented outside the standard text as well, in the C1136 requirements table published by Lamtec Corporation: Type III at a maximum 0.10 perms with 50 Beach units, and Type IV at a maximum 0.10 perms with 25 Beach units. The per-type permeance and Beach-unit values for Types V and VI: Coming soon — the C1136 standard text is paywalled and we do not publish standard values we have not read. Note what the four documented types tell you: within each permeance tier the difference is mechanical — 50 versus 25 Beach units of puncture resistance — but the tiers themselves differ by five-fold in permeance, so type selection is both a durability and a vapour decision. Exposed pipe in a plant room where ladders, trolleys and other trades pass is a Type I location for reasons that have nothing to do with permeance.

The 0.02 Perm Threshold Is a Specification Limit, Not a Rule of Thumb

The figure 0.02 perms circulates as though it were folklore. It is not. It appears as a hard limit across multiple product classes in UFGS 23 07 00, the US federal thermal insulation specification:

Vapour retarders must additionally meet a maximum flame spread index of 25 and a maximum smoke developed index of 50 per ASTM E84 (UFGS 23 07 00).

The sealant clause is the most technically precise citation in this subject, and it is worth understanding why. It does not just say 0.02 perms — it names Procedure B, the wet-cup case that exposes the specimen to 100 % RH. Whoever wrote that clause understood that a sealant at a lap joint on a cold line is the component most likely to sit against liquid water, and specified the test condition that reproduces it. Most private specifications quote the perm value and omit the procedure, which per E96’s own reporting rule leaves the requirement unenforceable.

A note on the PVDC row: its 150 Beach unit puncture requirement is three times the C1136 Type I minimum, and it carries a tensile requirement the laminated films do not. That is a materially different product class, not a substitution.

Permeance classes by film thickness

For PVDC films, NIA ties thickness directly to a permeance class:

(NIA, Insulation Design Guide — Materials and Systems.)

Two typical laboratory values from NIA’s Insulation Outlook article “Chill Out!”, worth holding in mind when reading a datasheet: facings and jacketing achieve values as low as 0.02 perm under laboratory conditions, while insulation materials themselves can reach less than 0.005 perm per inch (perm-inch). Read these as representative achieved values, not as physical limits — the 6 mil PVDC Class A row above is specified at 0.01 perms maximum, which is below the 0.02 perm figure, so 0.02 perm is plainly not a floor for what a facing can reach. The second figure is correctly expressed in perm-inch because it describes a homogeneous insulation body; the first is correctly expressed in perm because a facing is an assembly. The units are doing real work there.

Where the System Actually Gets Wet

Everything above concerns the intact sheet. Field failures are almost never in the intact sheet. Below are the locations, with the published installation requirements governing each, drawn from a chilled water installation guide for rigid polyiso pipe insulation (Dyplast Products, ISO-C1 Chilled Water Installation Guidelines, Rev 0619) and from UFGS 23 07 00.

1. Aligned longitudinal joints in single-layer or badly stacked double-layer work

For double-layer systems the geometry rule is specific. Stagger all longitudinal joints between inner and outer layers, and install inner and outer longitudinal joints 90° to each other — inner layer joints at the 12 and 6 o’clock positions, outer layer joints at the 3 and 9 o’clock positions. All butt joints between inner and outer layers shall be staggered between 6 and 18 inches (152–457 mm) (Dyplast §4.1.2, §3.2). The reason is direct: an aligned joint through both layers is a straight radial path from ambient to pipe wall. The clock-position rule exists so no such path exists anywhere on the run.

It also explains a structural weakness of single-layer systems that no amount of thickness fixes. A single layer has exactly one longitudinal joint line, continuous along the whole pipe. NIA’s laboratory work on self-drying wick systems gives a sense of scale for what a continuous longitudinal opening does: there a deliberate slot cut through the insulation to carry a drainage wick, “generally less than 0.1 inches” (2 mm) wide but running the length of the pipe run, leaves “a relatively open path for water vapor to diffuse into the system” until the wick saturates (NIA Insulation Outlook, “Chill Out!”). A 2 mm opening the length of a 30 m run is not a diffusion problem; it is an opening, and none of the perm arithmetic above applies to it.

2. The vapour retarder film lap, wrapped in the wrong orientation

This is the detail most often violated and the easiest to inspect. “Vapor retarder film should be cut to length longitudinally and wrapped around the circumference of the pipe with lap joint and installed facing downward avoiding the placement of the joint at the top or bottom of the pipe. Lap joint to be sealed using liquid adhesive. Butt joints shall be covered with vapor retarder tape” (Dyplast §4.1.11). Where the film is factory-applied, the lap joint is sealed with SSL (self-sealing lap) tape (Dyplast §4.1.13).

Two rules are stacked in that clause and they are frequently confused. The lap must face downward, so water tracking on the surface runs across the lap rather than into it. And it must not sit at 12 or 6 o’clock — 12 o’clock collects everything falling from above, 6 o’clock is where condensate and washdown water collects and sits. The correct position is the lower flank. In our experience reading these clauses against site photographs, the failure is usually not ignorance of the rule but installation convenience: the top of the pipe is where the installer’s hands are.

3. Elbows, tees and fittings

Curved geometry is where a flat film cannot be made continuous by wrapping alone, and where the published requirement becomes quantitative. “Elbows and fittings shall be wrapped with vapor retarder tape with a 50 % overlap,” applied in a spiral configuration; the alternative is a mastic-type vapour retarder. The tape is held to the same limit as the film: “Vapor retarder film and tape shall have a maximum permeance of 0.020 perm-in” (Dyplast §3.4.3, §3.4.4, §4.1.12).

The 50 % overlap is not a generosity margin. Spiralling tape around a curve stretches the tape on the outer radius and bunches it on the inner; at less than 50 % overlap the effective coverage on the outer radius of an elbow drops below one full layer. On the units point: that clause reads “perm-in” for a tape, which per the E96 task group chair is the inappropriate term for a facing — read it as the 0.02 perm requirement it plainly intends.

4. Valves, and anything that will be removed later

Vapour stops are the most under-specified item in chilled water insulation, and the only element on this list whose purpose is to limit damage rather than prevent it. “Vapor stops should be used on either side of valves frequently removed for servicing, valve stations left exposed, or odd fittings, elbows, tees, etc. where the chance of moisture infiltration is high” (Dyplast §4.1.10). UFGS states the same principle at specification level: use “properly installed protective vapor retarder/barriers and vapor stops on high relative humidity and below ambient temperature applications to reduce movement of moisture through or around the insulation to the colder interior surface” (UFGS 23 07 00).

The logic is compartmentalisation. Without vapour stops, one breach at a valve wets the insulation annulus and the moisture migrates along the run — the failure is unbounded in length. With vapour stops, the same breach wets one compartment. This is the closest thing this discipline has to a bulkhead, and it costs almost nothing at installation. Our view, stated as opinion rather than sourced fact: any valve on a chilled line that maintenance will realistically open should be treated as a future breach and bracketed accordingly at first fix.

5. Every mechanical fastener anyone drives through the system

This is the highest-value rule in the subject and the most violated. Verbatim from the installation guide: “Aluminum protective jacketing shall not be considered a vapor retarder.” “Stainless steel protective jacketing shall not be considered a vapor retarder.” “No fastener capable of penetrating the underlying vapor retarder shall be used.” And, separately: “Do not use screws, staples or other fasteners on lines containing a vapor retarder system.” Banding and seals are used instead, at a maximum spacing of 9 inches (229 mm) on centre for straight sections (Dyplast §4.2.5, §4.2.6, jacketing table).

Two things are being said. First, metal jacketing is weather protection, not vapour protection — a metal-jacketed line with a compromised retarder underneath looks finished and is not. Second, every screw and staple through the retarder is a permanent, unrepairable vapour path, and there are typically hundreds of them on a jacketed run. This is the failure mode that produces “we specified 0.02 perm and it still failed”: the specification was met by the material and destroyed by the fixing method.

6. The jacketing sequence, and why post-hoc inspection is impossible

Outdoor jacketing overlap “shall be a minimum of 2 in (51 mm) at butt joints and a minimum of 2 in (51 mm) at longitudinal joints. Jacketing shall be caulked before closing and banding and positioned in an orientation to avoid water ingress.” The outer layer is secured with fibre-reinforced tape at 25 % circumferential overlap on 12 in (305 mm) centres (Dyplast §4.1.7, §4.1.14, §4.2.5, §4.2.6).

The governing clause is the sequencing one: “before jacketing can be installed on a portion of the piping, the vapor retarder system on that portion must be complete and continuous” (Dyplast §4.1.14). Once the jacket is on, the retarder is invisible. No non-destructive field test returns an as-installed permeance — which is precisely what ASTM C755 concedes. Vapour retarder inspection is therefore a hold-point discipline before jacketing, or it does not happen at all. That single sequencing fact is why workmanship dominates outcomes here more than in any other part of a chilled water installation.

7. Sealants that stop being flexible

A slower failure, showing up years after handover. “Mastics and sealants shall remain flexible across the expected temperature range of the application,” and joint sealers “shall be preferably a low or zero-perm vapor retarder type” (Dyplast §3.3.1, §3.3.2). Chilled lines cycle thermally and contract; a sealant that hardens cracks at exactly the joints it was sealing, and the cracks open in service, not at handover.

Contraction and expansion joints are a required design element for the same reason. They are filled with resilient mineral fibre, fibres oriented parallel to the pipe, and the filler shall be twice the thickness of the joint so it is compressed to half thickness on installation (Dyplast §3.5.2). The 2:1 compression keeps the joint filled when the pipe contracts and the gap opens.

What Continuity Failure Actually Costs

The requirement is unforgiving because it is binary: the vapour retarder system must “be continuous at all the joints, elbows, valves and fittings present in real installations,” and success depends on installer workmanship and post-installation maintenance (NIA Insulation Outlook, “Chill Out!”). The same source documents a failure photograph showing “dripping, corrosion and mold growth” — despite vapour retarder jacketing being present over closed-cell insulation. Present, but not continuous.

The thermal consequence is quantified and traceable to a named journal paper rather than a vendor claim: a 1 % increase in moisture content can result in a 23 % increase in thermal conductivity in open-cell materials. The underlying study is A. M. Gusyachkin et al 2019, IOP Conf. Ser.: Mater. Sci. Eng. 570 012029, “Effects of moisture content on thermal conductivity of thermal insulation materials,” testing rock wool from different manufacturers by the standard guarded hot plate (GHP) method, and reporting that “higher thermal conductivity is always associated with higher moisture content for all investigated samples” (as reported by Owens Corning FOAMGLAS; the paper itself is open access under a Creative Commons Attribution 3.0 licence).

The framing statistic for the discipline is old but still the most-cited: “It has been estimated that 98 % of the problems with insulation systems are due to moisture” — Adams, Ludwig, “Thermal Conductivity of Wet Insulations,” ASHRAE Journal, pp. 61–62, October 1974 (via Owens Corning FOAMGLAS). It dates from 1974 and we cite it as the historical framing figure it is, not as a current measurement.

Put the three together and the economics are clear. The insulation is a heat-transfer component with a design life measured in decades. Wet, it degrades continuously and does not recover, because the vapour magnet effect keeps driving inward as long as the pipe is cold. And the corrosion under insulation that accompanies it attacks the pipe, not just the insulation — which converts an insulation defect into a piping replacement.

If Your Specification Is Written to European Rules

The EN/ISO counterparts to ASTM E96 are:

• EN 12086 (current: EN 12086:2013), Thermal insulating products for building applications — Determination of water vapour transmission properties. Specifies equipment and procedures for determining water vapour transmission rate, water vapour permeance and water vapour permeability in the steady state under different specified test condition sets; applicable to homogeneous materials and to products with integral skins or facings.
• EN ISO 12572 (current: EN ISO 12572:2016), the cup-test method for water vapour permeance of building products and permeability of building materials under isothermal conditions. ISO 12572 was prepared by CEN in collaboration with ISO/TC 163 (iTeh Standards listing; EN 12086 scope per standards catalogue listings).

The specific lettered condition sets in EN 12086 and EN ISO 12572 — the temperature and RH pairs analogous to ASTM’s Procedure A and B — are Coming soon. Both standards are paywalled and we were unable to read the condition tables. We will not publish EN condition-set numbers we have not verified, and neither should your submittal reviewer accept them without the standard in hand. The same applies to the µ (water vapour diffusion resistance factor) and s_d (diffusion-equivalent air layer thickness) formulae that European datasheets quote: they are widely reproduced, but we could not open a primary source for them and therefore do not state them here.

The practical warning stands without those numbers: an EN-tested µ or s_d value and an ASTM perm value are results from different methods at different, unstated conditions. Given that ASTM E96 forbids extrapolation between its own condition sets, extrapolating across standards families is worse. If a European product is submitted against a perm-based specification, ask for E96 data at the specified procedure, not a conversion.

A Specifier’s Checklist

• State the unit system. “0.02 perm” is ambiguous by 52 % between US and metric perm.
• State the E96 procedure and the conditions. A value without them is incomplete per the standard’s own reporting requirement, and cannot be extrapolated to your service conditions.
• Consider requiring Procedure B for anything that will sit against liquid water — sealants, laps, low-flank details. UFGS does exactly this for sealants.
• Specify permeance for facings, tapes and assemblies; permeability (perm-inch) only for homogeneous unfaced insulation. Correct any submittal that reverses this.
• Pick the C1136 Type on both grounds. Types I and II cap permeance at 0.02 perms and Types III and IV at 0.10 perms; within each pair the difference is puncture resistance (50 vs 25 Beach units).
• Write the installation procedure into the specification, because ASTM C755 tells you an as-installed permeance value cannot be specified. Include the clock positions, the 50 % tape overlap at fittings, the lap orientation, and the fastener prohibition.
• Make vapour retarder inspection a hold point before jacketing. After jacketing no test can find the defect.
• Mark vapour stop locations on the drawings at every valve maintenance will open, and at every complex fitting group.
• Ban screws and staples on the drawing, not just in the specification. Banding at max 9 in (229 mm) centres on straight runs.

Where IFANNova Fits, Stated Honestly

IFANNova is a French brand; products are manufactured by Zhuji Fengfan Piping in Zhuji, Zhejiang — 30+ years of manufacturing, 1000+ employees, 118+ countries served, a 120,000 ㎡ facility. Certifications on record: SKZ, CE, WRAS, DVGW, SGS, ISO 9001, ISO 14001 (certificate numbers: Coming soon).

What we supply is the piping, and the ceiling is real. Our pressure-pipe ceiling is Φ110 — UPVC 806 PN16 and HDPE PN16 marked to DIN 8077/8078 (per our catalogue), both Φ20–Φ110, with 203 fittings and the 603/604 no-weld compression fittings respectively; PPR PN20 (1103 pipe) stops at 32 mm, available only in 20/25/32 mm with 75 fittings in the 1138 series. Our PVC 902 line reaches Φ110 in pipe and Φ160 in 1902 fittings, but 902 is non-pressure drainage and is excluded from chilled water pressure duty. Also in range: PEX 2114/2121 at 16–32 mm; brass 2405 (all per our catalogue). We cannot supply DN150–400 chilled water mains. If your project is main-header work, this range is not the answer for that section — we are a terminal branch supplier and say so.

On insulation: we can supply insulation alongside the pipe, but insulation material, thickness, facing type, vapour retarder specification and permeance data are Coming soon. Every perm value, test condition and installation clause in this article is published third-party data attributed to its source, and none of it describes an IFANNova-supplied insulation product. We have no vapour retarder test data of our own to offer, and we will not imply otherwise.

One thing worth noting about the branch scale we do serve: at Φ20–110 the insulation annulus is small, the fitting density per metre is high, and the ratio of elbows, tees and valves to straight pipe is far worse than on a main. Every failure location catalogued above is more frequent per metre on terminal branch work than on the header. If you are insulating branch runs, the workmanship argument in this article applies to you more, not less.

Frequently Asked Questions

Is a vapour barrier more important than insulation thickness on a chilled water line? They answer different questions, but the barrier decides whether the thickness keeps working. Thickness sets the surface temperature relative to dew point — the condensation calculation covered on our pillar page. The barrier sets whether the insulation stays dry, and wet insulation loses conductivity performance continuously and without recovery (23 % per 1 % moisture in open-cell materials, per Gusyachkin et al 2019 as reported by Owens Corning FOAMGLAS). A correctly sized thickness with a discontinuous barrier degrades toward failure; no thickness compensates for a breach.

Why is 0.02 perm the number everyone quotes? Because it is a hard specification limit, not a convention. UFGS 23 07 00 applies it to laminated film retarders, PVDC films, sealants (specifically at E96 Procedure B) and self-adhesive vapour barrier jackets. ASTM C1136 Types I and II both cap at 0.02 perms as invoked by UFGS.

Can I use a datasheet value tested at 73 °F/50 % RH for a Gulf plant room? No. ASTM E96 states that results at one set of conditions are not to be used to extrapolate results at other conditions, because a material’s structural characteristics can change between conditions. The 90 °F/90 % RH desiccant condition imparts 1.28 inHg of vapour drive against Procedure A’s 0.41 inHg — a different test, requiring a different result.

Does metal jacketing act as a vapour barrier? No. Both aluminium and stainless steel protective jacketing are explicitly stated not to be vapour retarders (Dyplast ISO-C1 guidelines). Jacketing is weather and mechanical protection installed over a completed vapour retarder system.

Can I screw the jacket on? No fastener capable of penetrating the underlying vapour retarder may be used, and screws, staples and similar fasteners are prohibited on lines containing a vapour retarder system. Use banding at a maximum of 9 in (229 mm) centres on straight sections (Dyplast ISO-C1 guidelines).

Where should the vapour retarder film lap sit on the pipe? Facing downward, and not at the 12 or 6 o’clock position — on the lower flank, with the lap sealed using liquid adhesive and butt joints covered with vapour retarder tape (Dyplast §4.1.11).

What is a vapour stop and do I need one? It is a discontinuity-limiting element installed on either side of valves that are frequently removed for servicing, exposed valve stations, and complex fitting groups where the chance of moisture infiltration is high (Dyplast §4.1.10; UFGS 23 07 00). It does not prevent a breach; it confines one so a single failure does not wet an entire run.

Does IFANNova supply vapour retarder materials? Insulation can be supplied alongside the pipe. Material, thickness, facing and vapour retarder permeance data: Coming soon. We hold no vapour retarder test data of our own.