Roofing
Henry vs GAF roof coating adhesion tests: which result passes?
Match the test result to the exact coating, substrate and manufacturer instruction, with side-by-side methods, preparation records and failure-plane decisions.
Direct answer
Henry's October 2026 Pro-Grade silicone system guide requires at least 2.0 pounds per linear inch for each adhesion test; its February 2023 Pro-Grade 280 acrylic instructions require at least seven days of cure and no delamination at 2 PLI. GAF's Liquid-Applied Roofing Manual Version 2.2 calls for an average above 2 PLI in its field-peel procedure, while its high-solids silicone quick specs for metal and aged TPO specify a 2.0 PLI minimum. At exactly 2.0 PLI, record the applicable GAF project instruction before accepting the result. Acceptance belongs to the tested coating, substrate, preparation, cure and method; peel in PLI cannot substitute for pull-off in psi.
On this page
Which Henry and GAF adhesion results pass?Is 2 PLI the same passing rule in every document?Roof coating peel versus pull-off: what does each measure?How many test areas does the roof need?Test-area preparation recordWhich substrates need primer before silicone?How long must the test patch cure?Peel-force reporting rulesFailure-plane atlas: where did the test break?What happens after a low, invalid or passing result?Worked examples: strip width, boundary readings and roof coverageWhat should an estimator allow for testing and retesting?System acceptance checklist and closeout recordOpen the related tool for the calculation, checklist, or packet.
Which Henry and GAF adhesion results pass?
A passing result needs three decisions: the specimen and test are valid, the measured bond meets the named instruction, and the roof meets the remaining system requirements. Keep separate fields for those decisions.
| System | Authority / edition | Specimen / preparation | Cure / conditioning | Acceptance | Sampling / documentation |
|---|---|---|---|---|---|
| Henry Pro-Grade 988 silicone system | E059, 10/26, Step 1.II; 988 single-ply guide spec, undated | Prepared substrate and exact silicone system; obtain issued fabric construction instruction | Single-ply specification: at least 72 h | Each test ≥2.0 PLI for warranty eligibility; primer/base and retest below | At least 2 field tests, 1 per 10,000 sf, plus wear, changed substrate, installation phases, shade, ponding indications; narrow uncoated-granulated-MB exception |
| Henry Pro-Grade 280 acrylic | E047, 02/23 | Three test areas, three 1 x 6 in strips in each; 2 in free; straight-up pull with handheld scale; Henry 195 or 196 fabric; two coats each stated approximately 1.5 gal/100 sf, nominal 20 wet mils; EPDM-only 825 rinseable primer sequence | At least 7 days | No delamination at 2 PLI; minimum 2 PLI, less fails | Area/condition coverage plus three-area construction; submit warranty confirmation |
| GAF liquid-applied field peel | COMCO180, v2.2, p.31; p.34 LAR-1 boundary comparison | Intended system mock-up; specified primer; 1 x 12 in fabric, about 6 in embedded and 6 in free; encapsulate fabric | Standard practice 1 week; stated range 8 h to 2 weeks; silicone/PVDF may need longer; covered wet soak usually 1 h is best practice | p.31 average above 2 PLI; LAR-1 p.34 minimum 2.0 PLI | 3 patches first 100 squares plus 1 each additional 100 squares for questionable substrates; label, photograph, retain documentation |
| GAF High Solids Silicone over metal / aged TPO | COMCO180 v2.2, UH-1 p.67 / UH-10 p.75, read with p.31 | Named substrate system; specified application rates; p.31 fabric procedure | p.31 conditioning; silicone may need extended cure | Quick specs: minimum 2.0 PLI; retain p.31 average-above wording | UH-10 directs new TPO or results below 2 PLI to GAF Design Services |
| Polyglass elastomeric field procedure | TB 2015-02, Rev 013025 | At least 3 typical roof sites, about 4 sf each; 24 wet mil base, 1 x 12 in fabric, free tab, 16 wet mil overcoat | Complete cure, typically 7 days | Approximately 10 lbf with predominantly cohesive failure; without gauge, completely cohesive failure | Polyglass materials and conditions only |
| EPDM Coatings field procedure | Undated manufacturer PDF, checked 2026-11-05 | 2 x 12 in fabric, 6 in free; 16 wet mils below and 16 above | At least 4 to 7 days, as published | At least 4 lbf pull at 90 degrees; equivalent width arithmetic is 2 PLI | Clean, dry, remove loose rust/dust |
Source basis: S1 S2 S3 S4 S6 S11 S15.
Obtain Henry's issued silicone patch-construction instruction before placing a 988 test: E059 establishes acceptance and frequency without a complete fabric recipe. Record fabric, dimensions, film build and pull direction from that instruction; the 280 acrylic recipe does not establish the silicone method.
Henry's 280 sheet E047 (02/23) pairs approximately 1.5 gallons per 100 square feet with 20 wet mils per coat; Anvilfield estimates about 24 wet mils from that rate, dividing 346.5 cubic inches by 14,400 square inches and converting inches to mils. Retain the printed pairing and resolve the intended test build in the project instruction. The wet-film thickness and coating yield guide explains the area-volume arithmetic. [S3]
Use the coating restoration eligibility and chemistry guide to choose the restoration scope.
Is 2 PLI the same passing rule in every document?
GAF's Liquid-Applied Roofing Manual Version 2.2 says the field-peel average should be above 2 PLI on page 31, while its LAR-1 acrylic metal quick specification on page 34 and UH-1/UH-10 high-solids silicone quick specifications on pages 67 and 75 state a minimum of 2.0 PLI. [S6]
Which instruction controls the proposed system?
GAF describes its quick specs as abbreviated instructions, so retain the complete project specification as well as both relevant manual passages. Obtain GAF's written project disposition for a boundary result instead of inventing a universal hierarchy between those passages. [S6]
Anvilfield's reporting rule is to compare the underlying reading before rounding: 1.96 PLI stays below a 2.0 PLI minimum. Instrument resolution describes the display increment, not measurement accuracy. Retain the scale's accuracy and check status; resolve a boundary reading that the instrument cannot distinguish reliably, before accepting it.
These are system acceptance rules, not a ranking of coating brands. A larger force under different geometry, cure or failure criteria does not establish better adhesion on your roof or longer service life.
Roof coating peel versus pull-off: what does each measure?
Fabric peel measures force while a flexible strip progressively separates; report that force per unit strip width in PLI. Dolly pull-off loads a defined area perpendicular to the coating and reports tensile stress, commonly psi, with the fracture location. No universal numerical conversion exists between PLI and psi because their denominators and loading mechanisms differ. [S7 S8 S9]
| Method | Result and loading | Appropriate scope | Acceptance source | Comparison rule |
|---|---|---|---|---|
| OEM fabric field peel | lbf and strip width, normalized as PLI; progressive peeling | Exact roof coating/primer/substrate tested | Named OEM procedure and project system | Compare only matching geometry, conditioning, cure and procedure |
| ASTM D903-98(2025) | Comparative peel/stripping strength at controlled specimen and testing conditions | Adhesive bond specimens meeting full method | Specified material/system criterion | Hand pull is not automatically full ASTM conformity |
| ASTM D4541-22 | Perpendicular pull-off stress; methods B-F; protocol 1 fracture or 2 load criterion | Coatings on metal; other rigid substrates within method limitations | Project criterion and designated procedure/protocol | Fixture, glue, scoring, cure, instrument and rate affect results |
| ASTM D7234-22 | Perpendicular pull-off strength plus fracture plane | Coatings on concrete | Project criterion and named apparatus/fixture | Use concrete method; substrate fracture is not the coating-interface strength |
| ASTM D6083/D6083M-24 and D6694/D6694M-25 | Material properties, including adhesion | Acrylic roof coatings; silicone for SPF, respectively | Material specification | Public scopes do not give field application guidance |
Source basis: S7 S8 S9 S10 S12.
Call a rooftop hand pull an OEM field-peel test unless the complete ASTM D903-98(2025) method was followed. A flexible roof membrane needs its specified OEM procedure; a rigid-dolly test does not become applicable merely because the instrument can be attached. [S7 S8 S9]
Material qualification is different from roof acceptance
ASTM D6083/D6083M-24 covers acrylic coating material properties, while ASTM D6694/D6694M-25 covers solvent-dispersed silicone coating used in spray polyurethane foam roofing systems; both public scopes exclude application guidance. D6694's principal-polymer requirement of more than 95% silicone describes polymer composition, not coating solids percentage or adhesion strength. [S10 S12]
How many test areas does the roof need?
Plan Henry silicone testing at a minimum of two field locations and one per 10,000 square feet, then cover wear, changed substrates, installation phases, shade and ponding indications under E059 (10/26). GAF Version 2.2 page 31 specifies three patches for the first 100 squares plus one for each additional 100 squares for questionable substrates. One roofing square is 100 square feet; use the roof-area takeoff guide to establish that denominator. [S1 S6]
A location is a place on the roof, a patch is a prepared test assembly, and a strip is a specimen within it. Henry 280 separately directs three test areas with three strips per area; those replicates do not replace the sheet's coverage of roof size and distinct conditions. Moisture-survey cores are a separate count and cannot substitute for adhesion patches. [S3]
Map previous coatings, repairs, shaded zones and ponding indications before selecting locations. Anvilfield's quality-control rule is to repeat matching qualification when the production cleaner, primer, coating or preparation changes. Record partial area increments conservatively in the test plan; GAF's wording does not supply a rounding rule.
Where does the Henry granulated-roof exception stop?
Henry E059 (10/26) exempts uncoated granulated modified bitumen from adhesion testing but requires tests at previous repairs and worn areas or heavy granule loss. Preserve the exception's scope and do not transfer it to another manufacturer's system. [S1]
Test-area preparation record
Prepare the patch with the production cleaning, rinsing, drying, repair and primer sequence. A specially scrubbed island cannot qualify a field that receives weaker preparation. Henry's 988 data sheet places the primer or base sealer before silicone defect repairs, so the system's actual sequence must survive the transfer into the field work order. [S1 S2 S3]
Henry E059 (10/26) specifies washing at 1,500 to 2,000 psi for modified bitumen, built-up roofing, single-ply and previously coated surfaces, versus 3,000 to 4,000 psi for metal, without damaging surfaces or forcing water into the assembly. The September 2023 988 data sheet instead gives a general minimum of 2,000 psi. [S1 S2]
| Field | Filled planning entry / required field | Why retained / source | Job observation / completion |
|---|---|---|---|
| System and instruction | Henry Pro-Grade 280; E047, 02/23; reviewed 2026-11-05 | Exact recipe, S3 | ________________ |
| Roof location and condition | Assign A, B, C to mapped roof areas; record substrate, old coat and repairs individually | E047; Anvilfield form | ________________ |
| Cleaning and final state | Non-filming detergent, appropriate washing pressure, scrub contamination, treat growth, rinse, dry; enter actual completion time | Production prep match, S3 | ________________ |
| EPDM preparation if applicable | 825 rinseable primer at 500 sf/gal; wash off after 15 min | Acrylic-on-EPDM instruction only, S3 | ________________ |
| Patch construction | 3 areas; 3 Henry 195/196 strips per area; each 1 x 6 in, 2 in free; record actual wet build and printed rate discrepancy | E047; see matrix rate discrepancy | ________________ |
| Cure | Earliest evaluation 7 days after placement; actual timestamps required | S3 | ________________ |
| Conditions | Air/surface temperature, RH, dew point, precipitation during cure; record actual values | Anvilfield record; system conditions govern | ________________ |
| Force record | Instrument ID, units, width, pull description and retained readings for each strip | E047 scale; Anvilfield supporting fields | ________________ |
| Fracture and disposition | Layer on roof; layer on removed strip; photos; force criterion; pass/correct/retest | S3/S8/S9; Anvilfield synthesis | ________________ |
| Closeout | Compatible patch repair, material and date; warranty application confirmation | E047; GAF repair framework | ________________ |
The filled entries above plan Henry's acrylic procedure; complete the blank observation column on the job. Use the inspection and maintenance guide, moisture-survey definition and leak-diagnosis guide for the separate dry-roof qualification. A dry test surface does not establish dry insulation beneath it.
Record weather during installation and cure
Henry E059 (10/26) requires silicone application with substrate temperature above 35°F and rising, at least 6°F above dew point, and dry conditions for 15 minutes after application. The 988 data sheet gives an ambient range of 35°F to 120°F and slower cure at temperatures below 50°F and relative humidity below 35%. Record air and surface temperatures separately, together with humidity, dew point, rain, dew exposure and shelter. [S1 S2]
For Pro-Grade 294, E059 (10/26) requires substrate temperature above 50°F and rising, at least 6°F above dew point, and 24 hours dry after application. Its printed 2°C equivalent is incorrect: 50°F is 10°C. The base must dry fully for at least 48 hours before silicone; humidity or rain extends the cure. [S1]
Henry's acrylic system guide E419 (04/24) sets a 50°F minimum substrate temperature for Pro-Grade 280 and requires the substrate to remain dry for 12 hours after installation. Use those acrylic conditions when constructing the preparation-record example. [S5]
Which substrates need primer before silicone?
Qualify the complete substrate, cleaner, primer and coating assembly. RCMA's Coatings Tech Note 2, revised March 2025, separates adhesion promotion, rust inhibition and bleed blocking, and states that primer cannot replace surface preparation. [S13]
| Existing surface | Verified Henry instruction | Other primary guidance / disagreement | Field disposition |
|---|---|---|---|
| Aged TPO/PVC/EPDM under 988 | Test; 941 for insufficient adhesion in TDS; current guide lists 941 and Prime-Tek TPO primer | RCMA says membrane types have distinct preparation/primer needs | Qualify complete system |
| Aged TPO under GAF High Solids Silicone | Henry primer instructions apply only to the Henry system | UH-10, v2.2 p.75: new TPO or results below 2 PLI go to GAF Design Services | Follow GAF system disposition |
| EPDM under 280 acrylic | 825 rinseable primer at 500 sf/gal, wash off at 15 min | Acrylic system only | Keep acrylic scope |
| Asphaltic membrane under 988 | Pro-Grade 294 bleed blocker; current guide requires full drying minimum 48 h before silicone | Bleed prevention and adhesion promotion are separate functions | Do not omit specified base solely because a pull holds |
| Sound uncoated granulated MB | E059 testing exception | Prior repairs and worn/heavy-granule-loss areas require tests | Map exception and required tests |
| Existing acrylic coating | 988 lists well-adhered acrylic as a suitable existing surface | RCMA says primer need depends on age and condition | Identify any failure within or beneath existing acrylic |
| Existing silicone | E059 prohibits applying 294 over silicone coatings/mastics | RCMA favors silicone-based assemblies | Obtain coating-specific sequence and test it |
| Rusted metal | Remove rust/scale; rust-inhibiting primer per system | RCMA says primer cannot replace preparation | Repair deficient panels and test the prepared finish |
| Kynar/Hylar metal | 988 TDS and current guide exclude this use | RCMA discusses differing factory finishes generally | Select an approved system; passing pull does not override exclusion |
| Concrete | 988 TDS lists concrete, while current E059 system introduction lists other substrate groups | D7234 is the concrete pull-off method | Use issued concrete system specification and criterion |
Source basis: S1 S2 S3 S6 S9 S13.
EPDM, TPO and existing silicone are different test surfaces
Record the membrane and existing coating by product, rather than writing 'rubber roof.' Test the entire remaining coating stack, including repaired areas. [S2 S6 S13]
Henry's September 2023 988 data sheet excludes Kynar/Hylar finishes, gravel, degraded SPF, plywood, walking decks and shingles. Use the metal restoration guide for rust, fastener and seam repairs before qualifying the approved metal finish. [S2]
How long must the test patch cure?
Schedule at least seven days for Henry 280 under E047 (02/23), and at least 72 hours before adhesion testing under Henry's undated 988 single-ply guide specification. Record placement and evaluation timestamps so actual elapsed cure is visible. [S3 S4]
GAF Version 2.2 page 31 describes one week as standard practice within an eight-hour to two-week range, with silicone and PVDF potentially needing longer. Its weather-dependent drying examples do not guarantee full cure. [S6]
GAF's covered wet conditioning, usually a one-hour soak after cure, is a recommended best practice. Use separate labeled specimens for dry and wet results and record the soak duration. A dry pass cannot qualify a required wet test that fails; retain both results for the system decision. [S6]
Why does a 15-minute rain-safe claim not authorize a pull test?
Henry's September 2023 988 data sheet separately lists rain safety at 15 minutes, tack-free time at one to two hours, and weather-dependent full cure/recoat discussion at two to six hours. Those milestones do not replace the single-ply specification's adhesion-test interval. Progressive Materials' August 23, 2023 procedure uses at least 24 hours for its qualitative fabric screening, a separate manufacturer's procedure. [S2 S4 S14]
Peel-force reporting rules
Divide the measured pull force by the actual strip width to report PLI. Anvilfield's normalization rule is: peel force (PLI) = force (lbf) / width (in).
Use the named method's reinforcement fabric, embedded length, free tab and film build. Henry 280 directs a straight-up scale pull; EPDM Coatings and Polyglass explicitly use 90 degrees. GAF page 31 calls for a slow pull without specifying a numerical angle or speed. [S3 S6 S11 S15]
| Record element | Rule | Error it prevents |
|---|---|---|
| Force and width | Retain lbf and measured inches; report PLI as force divided by width | Treating 4 lbf on 2 in as 4 PLI |
| Method/edition | Name OEM, product, instruction/date and test construction | Calling every fabric pull ASTM D903 |
| Geometry and pull | Keep width, embedded length, free tab, wet build, pull direction and rate description | Comparing unlike angles, films or fabric construction |
| Instrument | Record range, display resolution, stated accuracy, zero/check status and instrument ID | Confusing display precision with accuracy |
| Cure and conditioning | Record age and dry/wet condition, soak duration if used | Equating GAF one-hour wet conditioning with laboratory immersion |
| Readings | Retain readings and averaging method | Peak substituted for average |
| Failure location | Photograph and label both faces | New coating's cohesive tear concealing old-coating delamination |
| Precision and threshold | Compare underlying reading to applicable boundary before display rounding | 1.96 PLI promoted to a 2.0 PLI minimum |
| Grouping | Separate substrate/preparation groups | A good group masking an unqualified condition |
| Repair | Record compatible repair at disturbed test sites | Open test damage left in waterproofing |
A peak-hold maximum is not the average requested by GAF. Agree on the collection method, retain the readings or trace, and say whether the mean describes one strip's progressive peel or several specimens. GAF page 31 does not define a numerical hand-pull rate or a cross-roof pooling rule. [S6]
Pack the specified materials, ruler, wet-film gauge, map, scale, weather instruments, camera and repair kit. Zero and check the scale according to its instructions, using the instrument fields above.
What belongs in a pull-off report?
Record instrument method, fixture size, glue, scoring status, loading protocol, conditions, strength and fracture location. ASTM D4541-22 identifies methods B through F and protocols 1 and 2; protocol 1 applies when the protocol or pass/fail criterion was not designated. ASTM D7234-22 recommends agreement on instrument and fixture and sufficient coating cure, extended for weather where necessary. Separate glue failure, equipment limits and substrate fracture in the report. [S8 S9]
Failure-plane atlas: where did the test break?
Photograph both fracture faces and name the limiting layer. ASTM D4541-22 and D7234-22 describe failure through the weakest plane in the fixture, glue, coating and substrate stack. [S8 S9]
| Plane / visual signature | What the load established | Action | Authority basis |
|---|---|---|---|
| New coat lifts cleanly from exposed substrate | Interface detached at recorded force | Check preparation, cure, primer and compatibility; correct and retest | S1/S2, S8/S9 |
| New coating tears within itself; coating remains on roof and strip | Coating cohesion limited the pull at that load | Apply OEM force and failure-mode criteria; no automatic pass | S1/S3/S11 |
| Primer remains on roof, topcoat on strip | Primer/topcoat interface was limiting | Check compatibility and recoat sequence; retest | S8/S9; diagnosis synthesis |
| New coat and old coating lift together | Remaining existing coating bond failed | Remove unsound old system or change restoration scope, then qualify | S1/S2; weakest-plane principle S8/S9 |
| Foam, granules, facer or concrete travels with coating | Substrate limited the load | Determine substrate repairs; report load and limiting plane | S1/S8/S9 |
| Fabric slips or coating detaches from reinforcement | Pulling assembly failed | Rebuild valid specimen; do not issue substrate pass from this result | S7 and OEM encapsulation rules |
| Dolly glue separates from coating or fixture | Glue connection limited test | Correct glue preparation/cure; repeat named method | S8/S9 |
| Several planes visible | Mixed fracture | Record visually estimated fractions and locations; apply actual job criterion | S8/S9; Anvilfield reporting design |
Source basis: S1 S2 S3 S7 S8 S9 S11.
Cohesive tearing at low force does not establish that the specified force was reached. Substrate fracture above a target establishes the load reached with that limiting plane, not unlimited coating-interface strength. Record mixed fracture as visually estimated areas and locations without inventing universal passing percentages. Use the deck and substrate guide where the supporting material is failing.
Is a clean peel always a preparation failure?
Clean separation identifies an interface, not a unique cause. Contamination, an incompatible primer, incomplete cure and an unsuitable old coating can produce an interface failure. Compare corrected preparation or approved primer trials within the same mapped condition, retain both outcomes, and identify what changed.
What happens after a low, invalid or passing result?
A low valid result leaves its represented condition unqualified: identify the failed layer, correct the preparation or approved system and retest. Henry E059 (10/26) calls for an appropriate primer or bleed blocker and retesting below its minimum; the 988 data sheet names 941 for insufficient adhesion. If the repeat silicone test remains below 2.0 PLI, E059 directs the result to Henry Product Support or its sales representative; keep that condition unqualified pending a revised system decision. [S1 S2]
An invalid or premature result needs a corrected test. A passing valid result releases only the matching preparation and coating system for the represented condition under Anvilfield's field disposition framework. Complete the roof repairs, thickness checks and warranty work separately, using the warranty coverage guide to distinguish eligibility from issued coverage.
Can a missed test be done after installation?
GAF Version 2.2 page 31 provides a one-inch butyl-strip fallback when the standard pre-installation test was omitted, followed by compatible coating repair. Identify this construction in the report and retain the failure plane so tape separation is not mistaken for coating detachment. Henry 280 and the 988 single-ply specification require testing before coating installation. Photograph compatible repairs at disturbed test sites. [S3 S4 S6]
Adhesion acceptance does not establish a dry or structurally sound roof
Henry E059 (10/26) requires verification of a dry, leak-free assembly for warranty eligibility and prohibits coating saturated insulation or substrates. For GAF Liquid-Applied NDL Diamond Pledge eligibility, Version 2.2 page 18 requires a moisture survey and tear-off where moisture covers 25% or more of the area; page 5's general discussion uses softer language about considering tear-off. Preserve the guarantee-specific instruction when selecting the tear-off or recover scope. [S1 S6]
Worked examples: strip width, boundary readings and roof coverage
| Calculation | Real published input | Result | Practical reading |
|---|---|---|---|
| EPDM Coatings' minimum normalized by width | 4 lbf; 2 in strip; 90-degree pull; 16 + 16 wet mils; 4 to 7 day cure [S15] | 4 / 2 = 2 PLI | Its four-pound number is force on a two-inch strip |
| Henry 280 minimum expressed as force | 2 PLI minimum; 1 in strip [S3] | 2 x 1 = 2 lbf | Required force for that width; follow its no-delamination rule |
| GAF p.31 boundary expressed as force | Average above 2 PLI; 1 in fabric [S6] | Average above 2 lbf on the specified width | Exactly 2 lbf does not meet 'above' wording; quick-spec minimum wording differs |
| Polyglass force normalized by width | Approximately 10 lbf; 1 in fabric [S11] | Approximately 10 PLI | Includes predominantly cohesive failure; do not use Henry's value for it |
| Henry acrylic preparation quantity | 3 areas x 3 strips; 6 in total minus 2 in free [S3] | 9 strips; 4 in embedded per strip | Three locations with replicates, not nine locations |
Method note: Anvilfield estimates the normalized forces and strip quantities above from manufacturer criteria and dimensions checked November 5, 2026, using the width formula in the reporting section and multiplying areas by strips per area.
What happens at the numerical boundary?
The worksheet uses valid, cured one-inch specimens; GAF's input is the agreed average. Henry acrylic also needs its no-delamination condition. Exactly 2.00 PLI meets the numerical minimum but does not exceed GAF page 31's boundary; the project instruction resolves the quick-spec wording. [S1 S3 S6]
| Force input on 1.00 in fabric | Calculated PLI | Henry E059 minimum, and E047 force threshold | GAF v2.2 p.31 average | GAF UH-1 / UH-10 minimum wording |
|---|---|---|---|---|
| 1.96 lbf | 1.96 | Below minimum; correct and retest | Below criterion | Below minimum |
| 2.00 lbf | 2.00 | Meets numerical threshold | Does not exceed boundary | Meets numerical wording; resolve p.31 project instruction |
| 2.25 lbf | 2.25 | Meets numerical threshold | Exceeds boundary | Meets numerical threshold |
How should a peel average be reported?
In a worked example using five readings from one one-inch strip, 1.90, 2.10, 2.25, 2.20 and 2.80 lbf total 11.25 lbf. Divide by five readings to obtain a mean of 2.25 lbf, then by the one-inch width to report 2.25 PLI; the peak is 2.80 lbf.
Anvilfield estimates that mean from the stated inputs, assuming the project has agreed to this sampling of the peel. Retain the low reading, peak and failure plane with it; the five-reading scheme is a worked reporting choice, not GAF's prescribed sampling algorithm.
How does the sampling count work on a 20,000-square-foot roof?
Anvilfield estimates three Henry silicone locations or five GAF patches for a 20,000-square-foot aged-single-ply roof with one separately mapped shaded condition. [S1 S6]
| Planning step | Henry Pro-Grade silicone E059 (10/26) | GAF field peel, COMCO180 v2.2 p.31 |
|---|---|---|
| Roof area | 20,000 sf | 20,000 sf / 100 = 200 roofing squares |
| Area-based field count | 20,000 / 10,000 = 2, meeting the minimum of 2 | 3 for first 100 squares + 1 for next 100 = 4 |
| Separately mapped shaded condition | Add 1 for shade outside the two field locations | Add 1 under this Anvilfield layout for the distinct questionable condition |
| Total selected locations / patches | 3 | 5 |
| Record kept at each location | Assembly, preparation, cure, force, failed layer | Assembly, preparation, cure, conditioning, force, failed layer |
Method note: apply the published area frequencies to 20,000 square feet and add one separately located condition patch. The two Henry field locations represent the main exposure, so shade needs its own location. GAF's four area-based patches likewise miss that condition in this layout; its fifth patch is Anvilfield's addition. These totals describe the selected map, not automatic additions when the base locations already cover shade. Add locations for further substrates, phases or repairs; record primer-comparison strips within their locations.
What should an estimator allow for testing and retesting?
Anvilfield estimates three labor-hours across two visits for the accessible-roof, three-area Henry 280 test allowance below, with a minimum seven-day cure between placement and evaluation. Reserve five labor-hours when cleaning, access or documentation needs two additional hours. The elapsed cure period is separate from paid labor. [S3]
| Task | Selected input | Schedule or quantity |
|---|---|---|
| Place and label patches | Local preparation and three Henry 280 test areas | 1.5 labor-hours on first visit |
| Evaluate and repair test sites | Nine strips, measurements and compatible repairs | 1 labor-hour on return visit |
| Complete records | Photos, results and warranty submission | 0.5 labor-hour |
| Base labor total | Placement + evaluation + records | 3 labor-hours |
| Additional-work allowance | 2 extra labor-hours for site demands | 5 labor-hours total |
| Earliest evaluation | Henry E047 (02/23) minimum cure | 7 days after placement |
| Separate quantities | Coating, fabric, cleaner, primer, repairs, travel and scale use | Enter actual job allowances |
Method note for the testing allowance
The November 5, 2026 labor model adds 1.5 hours for placement, one hour for evaluation and repair, and 0.5 hour for records. It assumes existing roof access, an available scale and local patch preparation; these are selected estimating inputs. Henry supplies the three-area construction and cure requirement. [S3]
Build the job allowance from labor-hours at your loaded labor rate, then add consumables, travel, access and equipment. Full-roof washing, moisture surveys, destructive cores, permanent access, engineer or third-party fees and restoration work need separate scope items.
A failed condition adds another preparation, cure and evaluation cycle. In a Henry 280 schedule, replacement patches placed on day seven cannot be evaluated before day fourteen under E047's minimum cure; later preparation shifts that date. Keep production coating behind the accepted test result.
Group inspection visits by service area and include return travel in the estimate. Estimators scheduling adhesion-test visits receive building lists in CSV or Excel within 48 hours through PorchRocket's Ground Control.
System acceptance checklist and closeout record
Assign each checklist item a status of pass, correct/retest, or not applicable with a reason. Keep the specimen's measured result separate from the system disposition. Use the checklist to connect each accepted roof condition to its test evidence.
Henry 280 E047 (02/23) requires confirmation of passing results on the warranty application sent to its warranty department; the undated 988 single-ply guide specification also requires passing-result submission. GAF Version 2.2 page 31 requires testing before guarantee registration, labeled and photographed patches, and retained documentation available on request. A test report does not itself issue a guarantee. [S3 S4 S6]
Field checklist
- Exact system and current documents named; record the applicable acceptance provision and any written boundary clarification.
- Roof areas, substrates, repair materials and prior coatings mapped; sampling and exceptions documented.
- Dry/sound assembly qualification completed separately from bond testing.
- Actual production cleaning, rinsing, repair and drying represented by each patch.
- Primer/base coat identity, batch, rate, drying and recoat sequence recorded.
- Fabric or dolly geometry and test build match the named method; actual width retained.
- Application conditions and cure timestamps meet the relevant system instructions.
- Wet/dry conditioning and any covered soak identified; different conditions separated.
- Instrument ID, units, range, resolution, stated accuracy, zero/check status, peak/average mode and underlying readings retained; uncertain boundary readings resolved.
- Every strip/patch result, both fracture faces, limiting plane and mixed failure observations recorded.
- Applicable threshold met without rounding up; failed conditions, invalid tests and retests retained separately.
- Matching preparation/system accepted only for represented roof conditions; changes trigger matching requalification.
- Test disturbances repaired with compatible materials and photographed.
- Repair completion, coating thickness, warranty application/registration and manufacturer review requirements completed separately from the adhesion pass.
- Accepted assembly and test locations filed with the roof asset record; owner receives test and repair documents.
File the accepted cleaner, primer, coating, test locations, low readings, retests and repair photos with the roof asset and capital-planning record. Give the owner any unresolved scope with that record.
Source use, review and update log
By Anvilfield. Reviewed November 5, 2026 against the 15 primary sources listed below. ASTM references use public publisher scopes; field acceptance values come from the named manufacturer documents.
Update log, 2026-11-05: rechecked all 15 primary sources; corrected Henry cure-condition wording; clarified silicone specimen instructions, wet-test reporting and retest decisions; added a peel-average example and a labor/schedule worksheet.
How to cite this page: Anvilfield, Henry vs GAF roof coating adhesion tests: which result passes?, 2026, https://anvilfield.com/field-guides/roofing/roof-coating-adhesion-test-comparison/.
FAQ
What passes a Henry roof coating adhesion test?
Henry's October 2026 Pro-Grade silicone system guide requires at least 2.0 PLI for each test; use the exact product's preparation and cure instructions.
What passes a GAF roof coating adhesion test?
GAF Version 2.2 p.31 calls for an average above 2 PLI, while quick specs use a 2.0 PLI minimum, so record the job's applicable acceptance wording.
Is roof coating peel testing the same as pull-off testing?
No; peel reports force per strip width, while pull-off reports perpendicular force per loaded area and its fracture location.
How do you prepare a silicone adhesion test substrate?
Match the actual roof's approved cleaning, drying, repairs and primer sequence before applying and curing the exact coating system.
Is manufacturer adhesion acceptance the same for every silicone?
No; named manufacturers specify different patch geometries, curing instructions, force wording and failure-mode criteria.
How long should Henry adhesion patches cure?
Henry's 988 single-ply guide specification requires at least 72 hours, while its 280 acrylic instruction requires at least seven days.
Does 4 pounds of pull mean 4 PLI?
Only for a 1-inch strip; four pounds on a 2-inch strip calculates to 2 PLI.
Does cohesive failure always pass?
Not automatically. Henry's numerical criteria still apply to a cured, valid specimen. Polyglass TB 2015-02 Rev 013025 separately permits a completely cohesive visual result when no gauge is available; that exception belongs to its procedure.
Can a passing adhesion test replace a roof moisture survey?
No; bond testing and verification of dry insulation answer separate questions in both Henry and GAF systems.
Should roof coating adhesion be tested wet or dry?
Follow the specified conditioning; GAF Version 2.2 p.31 recommends a covered soak after cure as best practice. Report wet and dry results separately.
Can I test adhesion after the roof has already been coated?
GAF Version 2.2 page 31 describes a butyl-strip fallback when the standard test was omitted. Identify that method and its failure plane, retain the result and repair the test site.
What should happen when roof coating fails the adhesion test?
Identify the failed layer, correct preparation or the approved primer/system, and repeat the applicable test after its required cure.
People also ask
Codes cited in this guide
This guide is written and reviewed against the published standards below. Always confirm the current adopted edition with the authority having jurisdiction.
Sources checked
- S1. Henry, Pro-Grade Silicone Roof Coating System Application Guide 142.3US.E059 (10/26), copyright 2026. Reopened for final editorial review 2026-11-05. Opened with curl.
- S2. Henry, Pro-Grade 988 Silicone White Roof Coating TDS Revision 09/27/2023. Reopened for final editorial review 2026-11-05. Opened with curl.
- S3. Henry, Pro-Grade 280 Acrylic Adhesion Test Instructions 142.3US.E047 (02/23). Reopened for final editorial review 2026-11-05. Opened with curl.
- S4. Henry, Pro-Grade 988 Single Ply Roofs Guide Specification Undated guide specification. Reopened for final editorial review 2026-11-05. Opened with curl.
- S5. Henry, Acrylic Restoration & Fluid Applied Membrane Systems Application Guide 142.2US.E419 (04/24). Reopened for final editorial review 2026-11-05. Opened with curl.
- S6. GAF, Liquid-Applied Roofing Manual COMCO180 Cover/page headers v2.2; footer ©2024, COMCO180-PDF-0124; introduction contains older v2.1 wording. Reopened for final editorial review 2026-11-05. Curl blocked; read through web tool.
- S7. ASTM publisher, D903 D903-98(2025), active, publisher updated Feb 21, 2025. Reopened for final editorial review 2026-11-05. Opened with curl.
- S8. ASTM publisher, D4541 D4541-22 active, updated Jul 29, 2022. Reopened for final editorial review 2026-11-05. Opened with curl.
- S9. ASTM publisher, D7234 D7234-22 active, updated Dec 22, 2022. Reopened for final editorial review 2026-11-05. Opened with curl.
- S10. ASTM publisher, D6083 D6083/D6083M-24 active, updated Jan 8, 2024. Reopened for final editorial review 2026-11-05. Opened with curl.
- S11. Polyglass, TB 2015-02, Guidelines for Field Adhesion Test for Elastomeric Roof Coatings Rev 013025, not the year in bulletin number or URL. Reopened for final editorial review 2026-11-05. Opened with curl.
- S12. ASTM publisher, D6694 D6694/D6694M-25 active, publisher updated Jan 9, 2026. Reopened for final editorial review 2026-11-05. Opened with curl.
- S13. RCMA, Coatings Tech Note 2, Use of Primers Prior to Roof Coating Revised/published March 2025, posted March 31. Reopened for final editorial review 2026-11-05. Curl returned 403; full primary text read through web tool.
- S14. Progressive Materials, What Is an Adhesion Test? Aug 23, 2023. Reopened for final editorial review 2026-11-05. Opened with curl.
- S15. EPDM Coatings, Field Adhesion Test Procedure Undated, no revision printed. Reopened for final editorial review 2026-11-05. Opened with curl.