Plaster

Veneering plaster

This formula should be pretty close to the same as USG Diamond Veneering Plaster.

Seems to be misusable to fill holes in drywall, but is much harder than drywall or compound and may crack at edges of broad gaps. One place in USG instructions seem to indicate that you can use it for taping, at least if what you’re putting over it is more veneer. USG says to seal any drywall compound before applying.

By weight:

• 8 parts plaster of Paris (tested with Certainteed)

• 2 part hydrated lime (tested with DAP)

• 5 or slightly more parts retarding gauging fluid (below).

  NB: USG says to mix diamond veneer with roughly half its weight in water. Original notes said 5:2 dry ingredients to water rather than 10:5, but this is in fact too dry on retrial. In fact, harried, possibly inaccurate, and now lost measurements say I used 323g water to 500g dry on second trial. Many subsequent trials seemed to work well with exactly the 8:2:5 ratio.

Coverage (untested, based on commercial product data)

EXPERIMENT SEEMS TO INDICATE THIS IS OPTIMISTIC; SEE PLASTER TEST OF 2025-05-02.

USG Diamond Veneer Plaster is rated at 5000 to 6000 ft^2/ton of dry plaster mix, applied at a nominal 1/8” thickness (they say at least 3/32”), for smooth coats, apparently depending on smoothness of underlying substrate. “Sand float finish sanded (1:2 sand to plaster)” 4000 to 4660 ft^2/ton of sand+dry plaster, “Heavy Texture finish sanded (1:1 sad to plaster)”: 3000 to 3600 ft^2/ton of sand+dry plaster. 3/16” coat for heated ceilings 2300 ft^2/ton sanded 1:1.

From this, we get a dry plaster/lime mix requirement of roughly 1/3 lb/ft^2 = 150g/ft^2 = 1.63 kg/m^2. Mixed plaster at 5:2 dry ingredients to gauging fluid would then weigh about 0.47 lb/ft^2 = 210 g/ft^2 = 2.28 kg/m^2, still assuming roughly 1/8” thickness and a smooth finish on an optimally smooth substrate.

Retarding Gauging Fluid

All molarities assume that my Amazon “Spicy World” sodium citrate is trisodium citrate dihydrate (molecular weight 294.1). All tests to validate this were done with in-house deionized water, taking its density to be one and measuring by weight.

Standard gauging fluid is now 0.01M (0.00375M having made my life hell).

If dry proportions change, the plaster probably controls.

It’s easier to mix if you make a 0.25M stock solution and then dilute that to working strength, but I have suspicions that stock solution kept around too long may go bad (because 0.00375M made from old stock was even worse than I might have expected it to be). The stock solution can be made with 73.5g dihydrate in 1l DI H2O.

Mixture experiments

These are meant to replace earlier experiments for which the notes were lost. I had settled on 0.00375M gauging solution, but tried to do a wall with it and utterly failed, in part because of too-fast setting. It seemed as though it was probably actually setting faster than earlier small-scale tests, but there’s some possibility that it seemed to be setting faster because it was more difficult to distribute setting plaster in a thin even layer over the large wall than over the smaller test area.

“Stock” here is 0.25M assuming the commercial “sodium citrate” I have is trisodium citrate dihydrate.

* Details for “Gauging” column

All mixes start with 4:1 Certainteed plaster of Paris and DAP hydrated lime. Column gives weight of gauging fluid as a fraction of weight of dry ingredients, thus 0.50 is 8:2:5 plaster:lime:gauging.

Earlier trials up through 2025-04-14 premixed plaster and lime, then added the mixture to gauging fluid (starting the timer at first contact between the dry mix and the fluid). Subsequent trials mixed lime into the fluid, then added plaster, and started the timer when the plaster went in. Results of both methods seem similar, but lime-first is easier especially for large batches.

Details for “Coupon” column

Plaster was applied to coupons at roughly “veneer plaster” thickness, vaguely between 1/8” and 3/16”.

• 1: 1/2” drywall, unsealed, 2-3” by 3-4”
• 2: 1/2” drywall, sealed with full strength Quikrete PVA concrete adhesive, 2-3” by 3-4”
• 3: About 3 square feet of drywall about 1/3 covered with previous plaster tests, sealed with 1/2 strength Quikrete PVA. Trying to approximate pantry wall I have to fix.
• 4: About 10 to 12 square feet of drywall mounted vertically. On first use, only small amount of upper part was covered with old plaster. First application covered 6 or 8 square feet with a somewhat uneven layer of veneer.

Old gauging fluid experiences

The only real remaining information from the earlier experiments is this statement, which refers to 8 parts plaster of Paris, 2 parts lime, 5 parts 0.00375M gauging fluid:

This [8 parts plaster of Paris, 2 parts lime, 5 parts 0.00375M gauging fluid trisodium citrate] is workable, or at least wet-trowelable, for about 45 minutes, although it got kind of iffy at around the 40 minute mark in testing. Should probably at least get it into the wall with 30 minutes.

Thus these notes said to use 0.00375M trisodium citrate in H2O. I apparently lost original paperwork. I had made up a bunch of 0.25M stock solution to dilute to working strength. Maybe a couple of months later, tried to use it at approx 0.00375M and ended up with plaster setting way faster than 45 minutes. Mixed wrong? Deteriorated? Who knows. Experiments (all with 4:1 plaster to lime mixture):

Formulas (mostly from US NBS Circular #151 1924)

Source gives these in volume. I’ve converted to weight using the densities measured for materials I had (see that section). Looks like proportions aren’t exactly critical.

• NB: these have been recalculated. Previous table used a sand density of 1.43 and seemed to be way, way too lean when used with the actual Sial #0 sand of that density. I mixed my 4:1 finishing plaster with sand at 0.30 by weight (per the old calculation), and the result was pretty frangible. Worse, when I tried to use it to fill a hole over lath, it wouldn’t stick and seemed like it would fall apart in the back instead of forming keys. Added an unknown but significant amount of additional plaster to actually patch the holes. This still may not be enough.

NBS circular says the finish coat should consist of lime and gypsum, mixed by the plasterer according to the desired setting time.

Another section, not the formulary, suggest that main reason to mix lime and gypsum is to retard the set. It also talks about mixing lime with Portland:

Lime and Portland cement may be mixed in any proportions. A little lime added to cement will improve its plasticity and sand-carrying capacity, and will have no deleterious effect except that the mixture will set more slowly. A little cement added to lime will increase the hardness and fire resistance, and will have no deleterious effect except on the color. About 15 per cent of one to 85 per cent of the other, by volume, is a good proportion to use.

Chemical retarders like citrate aren’t mentioned by name in the formula section, although commercial retarders “similar to glue” and effective up to 2 hours are mentioned. The formula section also alludes to the commercial availability of “plastic gypsum” (US patent 1392574), which I suspect is just PoP with chemical retarders added. Except that another part of the document also talks about “gypsum neat plaster” with retarder pre-added, and suggests that other premixes are available.

It also suggests that “quick setting lime” will “soon” be available. It suggests that either would be usable alone as a finish coat.

Other sources

Found by ChatGPT (but manually verified)

Canada’s Building Code for the North (1968, p. 89): “When 3 coat plaster is used [which it requires except on gypsum lath or masonry] the first or scratch coat shall consist of 1 part gypsum plaster to 2 parts sand by weight. The second or brown coat shall consist of 1 part gypsum plaster to 3 parts sand by weight. The finish coat shall consist of 1 part gypsum plaster to 3 parts lime by volume [that’s a lotta lime].

“Manual of Gypsum Lathing & Plastering”, Gypsum Association, 1972, suggests 1:2 or 1:3, probably by weight, for a base coat on lath.

1949 Virginia Fire Safety Regulations supposedly mention it, but I couldn’t find it. OCR was pretty bad.

… yet O3 summarizes this as 1:2 gypsum to sand by weight (“tight range 1: 1.8-2.3”). Or 1:2.5-3 for lime (which it sources at https://www.southerntrowel.com/services/diamond , https://nrc-publications.canada.ca/eng/view/ft/?id=e167367b-995a-4f4c-9ab2-983685799a3f (which I think is the 1968 Building Code for the North again), and https://www.coursehero.com/file/249818357/09210pdf/), none of which I have verified).

For Portland

These sources seem to be talking about finishing a masonry wall.

concretecaptain.com says 1 part Portland to 4 sand (I’m guessing by volume), which is even leaner than NBS.

civilsir.com:

Raw material densities

Measured values are probably good to a percent or two for the materials I had. They apparently vary pretty widely.

Setup checklist

• Cover floor
• Disposal plan for leftover plaster (probably best diluted heavily and dumped into soil).
• Tool-cleaning bucket with water.
• Spray bottle for wetting trowel
• Tools:
  • Hawk
  • Trowel
  • Corner trowel
  • Drywall knives
  • Loading scoop if possible