BRENNEN

Industries / Construction and property / Commercial and residential construction / Painting

Surface preparation

Surface preparation is the work done before paint goes on: washing, scraping, sanding, filling, caulking and priming, and checking that the surface is dry and sound.

4

Milestones achieved, from 1927 to 2001.

0

To watch for: announced, not yet confirmed outside.

2

Open deliverables. Nobody has shipped them.

The road

AchievedWatch forOpen

Achieved

  1. 1927

    Peeling traced to wet walls

    In a report first issued in May 1927, Frederick L. Browne, a chemist at the US Forest Products Laboratory (FPL) in Madison, Wisconsin, wrote that house paint failing in under two years almost always fails because water reaches the back of the siding, through leaky joints or as moisture from inside the heated house, and not because of the paint, the wood or the painter. He judged special primers and paints sold to fight it "of very doubtful merit". The laboratory reproduced the failures on small heated and humidified "miniature houses" set outdoors in winter, and Otto R. Hartwig of the West Coast Lumbermen's Association went on to study more than 1,500 affected houses, publishing in 1929.

    LessonStop the water before repainting, because a better paint does not cure a wet wall.

    inspectapedia.comresearch.fs.usda.gov

    Credit

    Frederick L. Browne, Forest Products Laboratory; Otto R. Hartwig

  2. 1990

    Grades for drywall before paint

    Gypsum board is the panel most interior walls are made of, and its joints are taped and covered with joint compound before painting. In 1990 four trade bodies, the Gypsum Association, the Association of the Wall and Ceiling Industry, the Ceilings & Interior Systems Construction Association and the Painting and Decorating Contractors of America, published "Recommended Levels of Gypsum Board Finish" (GA-214), which grades how far a wall has been finished, up to Level 5, which adds a skim coat of compound and is highly recommended "where severe lighting conditions occur". The first text was read as requiring extra coats at the lower levels and was rewritten in 1996, after four years of discussion.

    LessonA grading scale settles disputes between trades only when every trade reads its words the same way.

    wconline.com

    Credit

    Gypsum Association, AWCI, CISCA and PDCA; the Drywall Finishing Council joined in 2008

  3. 1999

    Vacuum sanders cut dust

    Joint compound is sanded smooth before painting, and the US National Institute for Occupational Safety and Health (NIOSH) had found drywall sanders breathing up to 10 times OSHA's limit of 15 mg/m³ for total dust. In a 1994 study at the painters' union apprenticeship center in Seattle, NIOSH engineers found that five commercially available vacuum sanding systems, three on poles and two held in the hand, cut dust exposure by 80% to 97% compared with sanding without them; NIOSH published the result for contractors in June 1999. It also found that the warnings on the compound makers' data sheets were seldom followed.

    LessonWritten warnings changed little on site; a tool that catches the dust where it is made did.

    govinfo.gov

    Credit

    Kenneth R. Mead and colleagues, NIOSH; International Brotherhood of Painters and Allied Trades

  4. 2001

    Bare wood spoils the paint

    In 1984 FPL researchers left smooth-planed western redcedar siding boards outdoors unpainted for zero to 16 weeks near Madison, then primed and painted them and watched them for 14 years. Paint on boards left bare for 16 weeks cracked within three years; boards left bare for as little as one week failed sooner than boards painted at once; and boards painted with no weathering were in almost perfect condition after 14 years.

    LessonPrime planed wood promptly or scuff-sand it first, because one or two weeks of weather are enough to shorten the life of the paint.

    paint.org

    Credit

    R. Sam Williams and William C. Feist, USDA Forest Products Laboratory

Open

  1. 01

    Bare-wood test, today's primers

    A repeat of FPL's weathering-before-paint test with the primers and siding woods sold today. Boards from one batch are left outdoors bare for 0, 1, 2, 4, 8 and 16 weeks, half of each set is scuff-sanded, and all are primed with three widely sold primers, then tested for adhesion and photographed outdoors each year after.

    The brief
    Credit · unclaimed
  2. 02

    Why the paint peeled

    An open set of house paint failures, each with photos, moisture readings and the cause an inspector found on site, sorted into causes such as water through joints and moisture from inside the house (Browne's groups), paint over weathered wood (Williams and Feist), chalky surfaces, and coats that do not bond to each other. Then a test of whether a vision model picks the recorded cause from the photos alone. Older paint may contain lead, so record by camera and moisture meter only and do not scrape, sand or cut any paint.

    The brief
    Credit · unclaimed

Open 01 · the brief

Bare-wood test, today's primers

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

A repeat of FPL's weathering-before-paint test with the primers and siding woods sold today. Boards from one batch are left outdoors bare for 0, 1, 2, 4, 8 and 16 weeks, half of each set is scuff-sanded, and all are primed with three widely sold primers, then tested for adhesion and photographed outdoors each year after.

Why it is still open

The long-term test of how little weathering harms paint used western redcedar weathered in 1984 and the paints of that time. A Crossref search in November 2026 found no later repeat with current products.

paint.org

Built with what exists

Pine or cedar siding boards; an outdoor rack facing south; three current acrylic primers; a portable pull-off adhesion tester used to ASTM D4541, in which a metal disc glued to the paint is pulled off and the force recorded; a camera.

  1. First step

    Cut and label 36 boards from one batch of siding and put the 16-week set outdoors this month.

  2. Done looks like

    Adhesion readings in MPa for at least six boards per primer, exposure period and sanding choice, published with the raw data within six months of the first priming, followed by yearly photos of the same boards.

Open 02 · the brief

Why the paint peeled

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

An open set of house paint failures, each with photos, moisture readings and the cause an inspector found on site, sorted into causes such as water through joints and moisture from inside the house (Browne's groups), paint over weathered wood (Williams and Feist), chalky surfaces, and coats that do not bond to each other. Then a test of whether a vision model picks the recorded cause from the photos alone. Older paint may contain lead, so record by camera and moisture meter only and do not scrape, sand or cut any paint.

Why it is still open

Public image sets label the defect, not its cause. The "Building defect on walls" set on Roboflow has 472 images in five classes, one of them peeling_paint, and a search in November 2026 found detection sets and papers but none that records why the paint failed.

universe.roboflow.com

Built with what exists

Phone photos; a pinless moisture meter, which reads moisture in wood without piercing it; a fixed case form; an open labeling tool such as Label Studio; a vision-language model asked fixed multiple-choice questions about each case (see the Note "A thousand questions, one pass"); Python for scoring.

  1. First step

    Write the case form, then record ten failures on buildings whose owners have agreed.

  2. Done looks like

    At least 300 cases from at least 100 buildings, published under an open license, with the model's rate of naming the recorded cause set beside the rate at which two inspectors agree with each other on the same cases.