SSPC-PA 2 dry film thickness inspection field guide

How does SSPC-PA 2 determine DFT conformance?

Under SSPC-PA 2-2015, editorial revision 2018, ordinary metal-coating inspection averages at least three gage readings at separate positions within a roughly 1.5-inch-diameter circle into a spot, then averages five randomly spaced spots for each sampled 100-square-foot area. Its default Level 3 allows spot averages from 80% of the specified minimum to 120% of the specified maximum, while each area average meets the specified range. Adjust the gage for the actual substrate first, then map thickness failures even when the overall average passes. [S10]

SSPC-PA 2 compares dry film thickness, or DFT, with the project's required range; it does not choose that range. Its scope covers nonmagnetic coatings over ferrous metal and nonmagnetic, nonconductive coatings over nonferrous metal. Type 1 magnetic pull-off gages apply to ferrous substrates; Type 2 electronic equipment needs the correct probe for the substrate. [S1] [S9]

Thickness conformance answers one acceptance question. Cure, adhesion, continuity, surface preparation and service readiness require their own evidence. Use the [industrial protective coatings and abrasive blasting guide](/field-guides/concrete/industrial-protective-coatings-blasting/) for the preparation and application sequence; use the records below for reproducible thickness decisions.

Which documents control the job?

ANSI lists SSPC-PA 2-2022 as the most recent edition, and ASTM lists D7091-22 as active. The numerical PA 2 procedures identified here are attributed to the publicly available 2015 text with its November 2018 editorial revision; record and use the edition invoked by the contract, including its selected alternatives. ASTM D7091 addresses gage measurement, while PA 2 supplies the area-based conformance framework. [S1] [S9] [S10]

Before setting instrument limits, record substrate, product, coat stage, layer and total thickness ranges, restriction level and any geometry-specific procedure. Resolve conflicting specification and product requirements in writing before selecting the instrument mode.

What DFT range and restriction level should you enter?

Enter the specified minimum and maximum for the coat or system being inspected. ASTM D7091-22 assigns those limits to the governing specification; a product's broad application envelope does not automatically become the project's acceptance window. [S9]

Under SSPC-PA 2-2015, editorial revision 2018, default Level 3 criteria, a 4.0 to 6.0 mil coating specification permits spot averages from 3.2 to 7.2 mils, while each sampled area must still average 4.0 to 6.0 mils. Anvilfield calculates those spot limits by multiplying the specified minimum by 0.80 and maximum by 1.20. Level 1 permits no spot deviation from the specified range; Level 5 leaves the spot and area maxima unrestricted. Enter the actual restriction level rather than applying Level 3 to every contract. [S10] [S3]

Keep layer acceptance separate from system acceptance. A project specifying a 3.0 to 4.0 mil primer and a 4.0 to 6.0 mil finish has a calculated cumulative range of 7.0 to 10.0 mils. Preserve the primer inspection before covering it: a passing final total cannot identify which buried coat supplied the thickness. [S9] [S10]

The 2015/2018 PA 2 single-target provision uses a range of plus or minus 20% only when one thickness is specified and the manufacturer provides no recommended range. Establish that range before applying any selected spot restriction; the single-target provision and Level 3 spot allowance perform different jobs. [S10]

Why do manufacturer windows differ?

Tnemec N69's January 7, 2025 product data sheet specifies 2.0 to 10.0 mils per coat. Jotun's historical 2015 Penguard Express brochure lists 75 to 250 micrometers for Express and 100 to 250 micrometers for Express CF. These are different products and service designs; a present job uses its applicable local-market product data sheet. One mil equals 25.4 micrometers. [S12] [S13]

The worked atmospheric N69 inspection below uses a project-selected 4.0 to 6.0 mil range within that product envelope. Save the instrument settings with the report; a software default cannot override a specified maximum.

How do you verify and adjust a DFT gage on rough steel?

Calibration establishes instrument performance through controlled reference measurements; verification checks accuracy against an identified reference; adjustment compensates for the actual measurement conditions. DeFelsko's calibration terminology distinguishes a Type 1 base metal reading, which is subtracted from coated readings, from a correctly adjusted Type 2 result, which reports coating thickness directly. [S5]

DeFelsko's PosiTector 6000 Full Guide v7.7 calls for accuracy verification at the beginning and end of each work shift and when the instrument has been dropped or erroneous readings are suspected. A failed end check makes results since the last successful check suspect: identify and remeasure the affected locations with a verified instrument before acceptance. Set the periodic calibration interval from use and environment; one year is the manufacturer's typical starting point. [S6]

PA 2-2015/2018 section 5.3 also advises checks during acquisition, for example hourly when collecting many measurements. Choose and log those intervals before work; shorter intervals reduce the amount of data requiring remeasurement after a failed check. [S10]

Type 1: determine the base metal reading

SSPC-PA 2-2015/2018 requires at least 10 distributed bare-substrate readings for the Type 1 base metal reading, or BMR. Worked inputs of 0.6, 0.7, 0.8, 0.7, 0.6, 0.9, 0.8, 0.7, 0.6 and 0.6 mil total 7.0 mils and average 0.7 mil. Subtracting that BMR from a 5.7-mil coated indication gives 5.0 mil DFT. [S10]

Verify a Type 1 gage on certified coated metal plates with a thickness within its expected working range, as PA 2-2015/2018 section 6.1 requires. Use plastic shims only with explicit manufacturer approval; their movement can distort pull-off readings. Do not mechanically zero the gage to hide substrate bias. [S10] [S5]

Type 2: adjust on representative steel

A representative coupon must match steel response, thickness, curvature and blast preparation. DeFelsko v7.7 describes averaged one-point and two-point adjustments, typically using 3 to 10 readings per average; Elcometer 456 Issue 10 recommends Rough/2 Point with foils below and above the expected film thickness. [S6] [S8]

Verify the certified smooth standard in the appropriate reference setting, then restore the job adjustment and confirm it with a representative-substrate shim check. Applying the rough-steel setting to a smooth plate can shift its reading. [S6] [S17]

The following record uses worked inputs. Its 4.8 to 5.2 mil verification window assumes a documented sum-of-limits rule: a 5.0-mil standard with 0.1-mil tolerance plus an assumed 0.1-mil probe tolerance. That window checks the instrument/reference combination; the coating acceptance limits remain separate.

Roughness correction decision table

Never subtract peak-to-valley blast profile as though it were gage bias. A 3-mil profile does not turn a correctly adjusted 5-mil DFT result into 2 mils. Elcometer Issue 10's ISO 19840 zero-offset route uses 10, 25 and 40 micrometer corrections for fine, medium and coarse ISO 8503-1 categories; apply those only within that selected route, without adding them to an already adjusted Type 2 result. [S8] [S17]

How many readings, spots and areas do you need?

Using the ordinary US-unit sampling schedule in SSPC-PA 2-2015, editorial revision 2018, a 2,500-square-foot coated lot requires five sampled areas, 25 spot measurements and at least 75 gage readings before nonconformance expansion or special geometry provisions. Select three areas in the first 1,000 square feet, one in the second 1,000 and one in the remaining 500. Anvilfield estimates the initial inspection quantities below from that schedule. [S10]

For lots at or below 300 square feet, measure each 100-square-foot area or remainder; above 300 through 1,000 square feet, select three areas. Above 1,000, add one sampled area for every further 1,000 square feet or remainder. In words, subtract the first 1,000, round the remaining thousand-square-foot increments upward, then add three: areas = 3 + ceiling((coated square feet − 1,000) / 1,000). [S10]

Define the lot, coat stage, work shift and actual coated surface before choosing locations. A tank's footprint is not its shell coating area. Keep the standard's metric and US-unit schedules independent: it pairs approximately 10 square meters with 100 square feet, although exact conversion gives 9.29 square meters. Do not switch threshold systems halfway through an inspection. [S10]

Average readings into spots, then spots into areas

First average accepted readings within each spot; then average the five spot results with equal weight. Spot mean = sum of readings / reading count; area mean = sum of five spot means / 5. If one spot has six readings averaging 4.0 mils and four spots have three readings each averaging 6.0 mils, the correct area mean is 5.6 mils; pooling all 18 readings gives 5.33 mils and overweights the first spot. [S3] [S10]

Preassign area IDs and randomly distribute five spots across each selected area, as PA 2-2015/2018 section 8.2 requires. Move the probe within each spot for each reading; three repeated placements at one point do not sample its thickness variation. Separate targeted edge and defect checks from ordinary sampling. Repeat suspect readings locally; replace an excluded nonrepeatable reading to preserve the minimum accepted count and retain the exclusion reason. Repeatable thin film stays in the record. [S10]

Can thin spots fail when the overall average passes?

Yes. The selected Level 3 spot limits and area limits are separate tests. Each low spot, excessive spot and under-thickness area requires disposition. DeFelsko's PA2 instructions explicitly organize conformance around both spot and area results. [S3]

Individual gage readings contribute to spots; separately specified local requirements and visible defects need their own decisions. [S10]

Decide near-limit results at retained instrument precision under the agreed decision rule. A calculated 3.17-mil spot fails a 3.20-mil minimum even if a one-decimal display shows 3.2. Instrument accuracy is not an extra allowance to add to the coating limits.

Worked conformance example: 2,500 square feet of coated steel

Anvilfield estimates the inspection quantities and labor below; thickness decisions are calculated from the stated worked inputs. LOT-2500 is one atmospheric N69 coat-stage lot with a project range of 4.0 to 6.0 mils, Level 3 selection and US-unit sampling. The Type 2 gage has been adjusted over matching rough steel, so these displayed DFT readings receive no further subtraction.

Areas A through E are selected 100-square-foot areas. Each contains five spots with three readings per spot; all 75 inputs are retained below.

A fails on A1 despite its acceptable area mean. B fails its area minimum even though every spot passes; C fails on its excessive C5 spot. The passing 4.92-mil grand mean cannot reverse those decisions.

Map the failed areas to establish repair quantities; calculate the repair proportion from measured repair footprints and the full coated area.

What causes suspect DFT readings at edges, welds and old coatings?

Edges, curvature, thin steel and changing metal response can shift a gage result. Verify the probe on representative uncoated geometry before accepting measurements from those locations. DeFelsko's accuracy guidance requires model-specific geometry checks rather than a universal edge setback. [S17]

Use the appropriate probe and the invoked geometry procedure at welds and edges. Stripe coating controls application at difficult details; local DFT and any required continuity testing still need valid measurements. For overcoating, the magnetic or electronic total includes measurable old layers between the probe and metal. If added thickness matters, retain precoat baselines at relocatable positions and preserve the uncertainty of both measurements when subtracting. [S6] [S9] [S10]

Measure when the film is dry and firm enough for the probe without indentation. Dry-to-touch, recoat readiness and immersion cure are different endpoints. A dry-film gage does not supply wet-film thickness or prove that a lining has cured. [S6] [S12]

Concrete needs its applicable substrate measurement route. See [resinous floor installation](/field-guides/concrete/epoxy-resinous-floor-coating-install/), [concrete sealers and coatings](/field-guides/concrete/concrete-sealers-coatings-protection/) and [traffic-deck waterproofing](/field-guides/roofing/traffic-deck-waterproofing-coating/) for those systems. [Roof coating thickness and yield](/field-guides/roofing/roof-coating-mil-thickness-yield/) addresses application consumption, a separate calculation from metal-coating acceptance.

When do scanning and special geometry change the plan?

Identify the invoked beam, pipe, edge, panel or scanning procedure before collecting data. The public 2015/2018 PA 2 sample already includes scanning Appendix 10, and Elcometer's rapid-measurement evaluation distinguishes scan populations from ordinary spot sampling. Preserve scan grouping and software settings. Elcometer's evaluation also found that probe-tip wear can shift scanning results, especially on rough coatings; include tip inspection and accuracy checks in the scanning plan. [S10] [S11]

PA 2-2022 lists a separate Appendix 11 for intumescent/fireproofing and cryogenic spill protection. Record the invoked special-coating procedure and approved system limits before measuring those systems. [S19]

How do you map the extent of nonconforming thickness?

The public SSPC-PA 2-2015/2018 procedure delineates nonconforming thickness with spot checks at 5-foot intervals in eight directions, using the specified thickness range and continuing to two consecutive conforming spots in each available direction. Expansion starts from the failed 100-square-foot area and extends within the surface coated during the work shift. A boundary spot of 3.8 mils fails this example's 4.0 to 6.0 mil mapping range, despite clearing the ordinary 3.2-mil Level 3 spot floor. [S10]

Area A is a 10-by-10-foot square centered at zero, with east as positive x and north as positive y. Anvilfield's map starts cardinal rays at side midpoints and diagonal rays at corners, measuring outward along the surface. N5 is at (0,10), N10 at (0,15), and N15 at (0,20); NE5 is at (8.54,8.54), five feet along the diagonal from (5,5).

The worked coated-shift surface extends at least 20 feet beyond every ray origin. Keep the original failed area, failing-spot neighborhoods, repair footprints and work-shift boundary as distinct map layers. The map and table locate and retain all 51 boundary readings.

Under the 2015/2018 procedure, mark the area within five feet of a failing spot noncompliant; retain compliant locations preceding a later failure on the same ray as suspect for reinspection after correction. On repeating units smaller than ten feet in their largest dimension, use the specified repeating-unit route where five-foot spacing is impractical. Record physical and work-shift stopping limits; inaccessible coating retains an unresolved disposition. [S10]

Area A requires 17 added spots and at least 51 readings. Its north ray needs a third spot because N5 fails; the other seven rays stop after two conforming spots. B and C need separate delineation. Measure the actual repair footprint separately.

How do you repair and retest thin or thick coating?

Repair thin film with approved preparation and a compatible added coat within the product's permitted build and recoat conditions. Thick film needs product-specific disposition: another coat cannot remove excessive existing build. Preserve the original measurements, identify the repair envelope, then record new readings under linked retest IDs. [S12]

Tnemec N69's January 7, 2025 data sheet lists, at 5-mil DFT without accelerator, 9 hours to handle, 12 hours to recoat and 7 days to immersion at 70°F; at 50°F the values are 24 hours, 32 hours and 12 to 14 days. For N69 recoating after 60 days, immersion service requires scarification; atmospheric service requires scarification or an epoxy tie coat. The same sheet gives earlier preparation triggers for some other topcoats over exposed N69 primer. Select the actual topcoat and exposure route, and allow for surface temperature, air movement, humidity and film build when scheduling cure. [S12]

The worked repair plan adds 2.2 mils at matched A1 points and 2.0 mils at matched B points, meeting N69's per-coat minimum. C5 is removed, prepared and reinstated at 5.2 mils. The repair disposition retains the project stage's area and spot limits and defines cumulative acceptance; protect adjoining sound film from excessive added build.

A1 and C5 retests each use 5.0, 5.2 and 5.4 mils. A2 through A5 and C1 through C4 have new retest triplets equal to their original values. B1 through B5 use (5.2,5.4,5.6), (5.3,5.5,5.7), (5.4,5.6,5.8), (5.5,5.7,5.9) and (5.6,5.8,6.0). These inputs reproduce area means of 5.64, 5.60 and 5.04 mils for A, B and C.

N5 receives a 2.0-mil repair coat; its readings of 5.6, 5.8 and 6.0 average 5.8 mils; all other 16 boundary points receive new retest records, with worked values matching the boundary table. A and its boundary close when those checks pass. B and C remain open until their separate boundary results and required dispositions are attached. Use the [tank coating and interior lining guide](/field-guides/concrete/storage-tank-coating-lining-interior/) for cure, continuity and return-to-service release beyond DFT acceptance.

How much inspection time should you allow?

Anvilfield estimates 73 minutes for the accessible 2,500-square-foot initial sampling exercise. The estimate includes setup, measurements, relocation and reporting; expanded defect checks, access arrangements, repairs and return visits need separate time allowances.

Method: the November 5, 2026 estimate adds the four task allowances below. Reading time is the required reading count multiplied by the assumed seconds per placement, divided by 60. The five-area quantity comes from PA 2-2015/2018; task times are Anvilfield planning inputs.

Area A's 51 added boundary readings add 8.5 minutes of acquisition at the same assumed pace, plus travel and mapping. B and C need their own expansion, followed by repairs and retests. Split fixed setup time from repeated measurement work in the [storage tank coating bid scope](/field-guides/concrete/storage-tank-coating-bid-scope/), and schedule inspection while the required access is available.

Field checklist and auditable DFT report

PA 2-2015/2018 section 10 requires instrument and reference-standard identification, calibration date, adjustment-shim thickness, BMR where used, spot/area measurements, operator and measurement date. The record below adds raw readings and linked repairs so another inspector can reproduce each decision. [S10] Keep the original export read-only and append repair records; replacing failed rows with passing rows destroys the history that explains the repair. For garage work, attach exposed-metal coating records to the phased closeout map in the [parking structure restoration guide](/field-guides/concrete/parking-structure-restoration-repair/).

Test the export before production sampling. Elcometer 456 Issue 10 states that Counted Average saves its average without the individual readings, so use a verified separate capture route when retaining raw values. [S8]

1. Identify substrate, product, coat stage, lot, shift and actual coated surface area; assign the applicable measurement route.
2. Record the invoked edition, layer and cumulative ranges, restriction level and geometry provisions; resolve conflicting inputs.
3. Identify gage and probe models/serials, calibration date, certificate and reference-standard IDs/values; verify shift-start accuracy and retain actual results.
4. Document representative substrate and Type 2 adjustment or Type 1 BMR; save reference and work setups separately.
5. Label sampled areas and spot locations; confirm the export retains individual readings and setup IDs.
6. Acquire the required readings; retain repeatable lows and document excluded nonrepeatable readings and replacements.
7. Calculate spots first and areas second; record each decision under its applicable limits.
8. Map each failed area and expanded boundary, including stopping reasons, suspect locations and inaccessible coating.
9. Link approved preparation, repair product, added build, cure conditions, retests and photographs to the original defect.
10. Verify shift-end accuracy; recover affected measurements after a failed check, and deliver separate system-release records.

What columns reproduce the decision?

Use one row per reading with record_id, lot_id, shift_id, coat_stage, area_id, spot_id, local_x_ft, local_y_ft, reading_index, reading_mil, correction_method, correction_mil, corrected_mil, accepted and exclusion_reason. Add spot_mean_mil, area_mean_mil, spec_min_mil, spec_max_mil, restriction_level, decision_type, decision, defect_id, retest_of, setup_id and source_edition. Finish with ray_id, distance_from_area_ft, boundary_stop_reason, timestamp, probe_id and operator_id.

For this Type 2 dataset, record rough-steel shim adjustment as the method, zero additional subtraction, and corrected thickness equal to displayed thickness. Boundary rows use the decision type “delineation versus specified range” and leave area mean empty. Missing values stay distinct from zero; every retest receives a new record ID even when readings match the original.

What inspection scope should a tank owner buy?

Buy a defined measurement and closeout package: asset ID, shell/roof/interior zones, actual coated area, substrate, coat stage, lot/shift divisions, access, hold points, boundary mapping and linked retests. Specify inspection frequency by application stage and the required evidence at each hold point. Visible dimensions support planning; field access and owner documents determine the measurement scope.

Attach owner records, coating schedules, service exposure and repair maps to each vessel in the account register. Coating contractors building a tank account register use [PorchRocket Asset Search](/asset-search?for=tanks) for a reviewed list of visible storage tanks and the owner of record at each site.

Use the [aboveground storage tank glossary](/field-guides/glossary/aboveground-storage-tank/) and [water tower glossary](/field-guides/glossary/water-tower/) for consistent asset names, and the [water tower inspection and coating guide](/field-guides/concrete/water-tower-inspection-and-coating/) for the broader owner program.

Sources, review date and how to cite this guide

Byline: Anvilfield. Sources checked November 5, 2026; editions and revisions listed below. Catalog entries identify standards and scope; the numerical PA 2 procedures above cite the public 2015/2018 text.

Update log: 2026-11-05: Reviewed numerical rules and manufacturer instructions; clarified sampling, recoat conditions and retests; corrected map spacing and expanded verification records; removed prices.

• S1: ANSI: SSPC-PA 2-2022 catalog, current edition and metal-substrate scope
• S2: Industrial Physics: SSPC-PA2 manufacturer standard overview, undated
• S3: DeFelsko: PA2 instruction addendum, manual v7.0
• S4: DeFelsko: PosiTector 6000 PA2 mode video and transcript, undated
• S5: DeFelsko: Calibration Terms for Coating Thickness Gages, undated
• S6: DeFelsko: PosiTector 6000 Full Guide v7.7
• S8: Elcometer: 456 Operating Instructions, TMA-0489 Issue 10
• S9: ASTM International: D7091-22, active practice catalog and scope
• S10: SSPC: PA 2, January 26, 2015, editorial revision November 1, 2018; public publisher-authored sample hosted by Normsplash
• S11: Elcometer: Statistical and Technical Evaluation of Rapid Dry Film Thickness Measurement, undated
• S12: Tnemec: Series N69 product data sheet, January 7, 2025
• S13: Jotun: Penguard Express range brochure, historical 2015 product ranges
• S15: ISO: ISO 19840:2012 catalog, separate measurement procedure
• S16: Accuris: SSPC-PA 2-2022 catalog metadata
• S17: DeFelsko: How to Ensure Accurate Dry Film Thickness Readings, manufacturer guidance
• S18: AMPP: Using SSPC-PA 2 Effectively course, undated education resource
• S19: AMPP: Materials Performance, July 2022, New Standards Published; Appendix 11 scope

Cite individual tables using their linked fragment, table name and underlying authority/edition.

How to cite this page: Anvilfield, SSPC-PA 2 dry film thickness inspection field guide, 2026, https://anvilfield.com/field-guides/concrete/coating-dft-sspc-pa2-conformance/

Related tools

FAQ

People also ask

• Why is surface prep important for coatings?
• What is a tank lining?
• Why does parking garage concrete spall?
• How often should a water tower receive a comprehensive inspection?
• What should a storage tank painting bid include?