Modeling with Plaster of Paris

Model ceiling mold ready for pouring.

How to do the ceilings in our tiny cottage has been the focus of much thinking and experimentation around here in the last few weeks. I would rather not do too much plastering over my head. The idea of it dripping down in my face isn’t pleasant. What I’ve come up with is a way to pour the ceiling plaster in a mold and then pour the concrete for the floor above directly on the plaster. This will create a unified solid with everything well bonded and structurally strong. That’s my theory…

To test this out I made a model. The mold consists of a curved sheet of 1/2″ thick pink foam which has a plastic sheet on it that acts as a release agent. I also put on some packing tape and some shrink film to experiment with those materials as release agents. The idea is to get a smooth finish with a minimum of effort.

Plaster of paris poured into the mold.

Here the plaster has been poured into the model mold and a piece of metal lath embedded in the plaster part way. The plaster also has PVA fibers from Nycon mixed in to give it added strength. The plaster sets up quickly so it is important to work rapidly. I vibrated it some by tapping the mold with the goal of moving bubbles up off the bottom surface which will be the visible ceiling from below. This worked fairly well. I have a concrete vibrator which I’ll hook up to the scaffolding on the real ceilings to get better vibrating.

Concrete floor poured over ceiling.

Once the plaster layer was in place I left the top surface a little rough and then quickly mixed up a small batch of high slump concrete with PVA fibers to be the structural floor for the attic. In my model it is only about 1/4″ thick in the middle of the arch but in the real loft and attic it will be about 1.5″ thick.

The demolded model after 48 hours.

I was impatient and removed the model from it’s mold after only 48 hours. It was strong enough and nothing broke, even the very thin part of the middle of the arch. Really I should have waited 72 hours for 80% strength. In the real ceiling I’ll probably wait at least seven days and maybe to the full 28 days if there is no pressing reason to remove the scaffolding. I’m sure I can find things to do, I just need to control my impatience.

Soon, after the seven days of initial cure are done, I’ll stand on the middle of this little model, provided that the progressive weights before me don’t break it. It will be interesting to see how much force this can take. I intend to do the experiment without buttressing the ends as they would really be in the house. The buttressed ceiling will be far stronger. When will it break? I intend to find out…

Corner detail with pins.

The real shelves and ceilings will lock into the walls of the tiny cottage – thus the screws that were molded into the model’s concrete layer. To achieve that I’ll drill holes in the walls and set pins, scar up the surface of the walls where the ceilings will bond, pour down the partition cores and use the rebar that we built right into the structure in preparation for the ceilings. I don’t want these massive ceilings coming down on us. Each ceiling will weigh about 2,600 pounds. These screws set into the model represent the pins of reinforcement. The lath and fiber will also help as they’ll keep the entire thing together.

As you can see on the side there are still many air bubbles in the concrete. I vibrated it a lot by hand but didn’t manage to get the bubbles out. The top did end up fairly smooth and later I steeled it smoother which would make a good floor surface. The electric vibrator on the real thing should do better. Fortunately, even with the little bubbles it is still much stronger than needed. For a small project like this an electric sander might have done the trick, probably better than my hand shaking.

Upside down ceiling cake.

A view of the ceiling from below. Looks great except where I wrinkled the shrink film. I would love to not have to paint the plaster. I’m still investigating that. One thing I’m looking at is mineral paints. I would like to stay away from chemical paints that produce fumes within our home.

Closeup of plaster.

From a normal viewing distance all of the areas of the ceiling look good. A close up shows the diamond and texture pattern of the foam board mold, the smoother area where the tape was and the very smooth area where the heat shrink film was. There are some small air bubbles showing that I didn’t adequately vibrate the plaster but I don’t see those when I step back to a normal ceiling viewing distance. An advantage of being short!

In addition to using this technique for our ceilings we’ll be doing similar things for making shelves, benches, planters and the like. The next test project after this model is to make a shelf in the bathroom under where the aquarium will go. It will be a good chance to do something about four times larger than this model that still is not deadly serious. I like gradually stepping up with small projects.

Speaking of fun concrete projects, go visit Karl’s root cellar project.

Outdoors: 37°F/23°F Overcast, Light Rain most of the day
Farm House: 56°F/50°F three logs
Tiny Cottage: 57°F/45°F

About Walter Jeffries

Tinker, Tailor...
This entry was posted in Uncategorized. Bookmark the permalink.

7 Responses to Modeling with Plaster of Paris

  1. Podchef says:

    Kudos Walter–I think your onto something with yer tinkering. Have you thought of using MDF plywood for the form instead of the foam? I think you can get it in 1/2 inch which should bend to your curve.

    I’ve used a brand which had a paper face on it and was very flat and smooth. When I was building houses I used to pour monolithic mantle pieces and counters and the like with the MDF to give a smooth, flat and shiny finished surface without any major effort. You might also boost the amount of cement in the mix 6-sack instead of 5 because the “cream” is so much richer and allows for a better finish–oh, that’s right, your using plaster. . . .Well, remember the concrete ideas for those counters and mantles ye might someday encounter. . . .
    (BTW–Love the irony of the Cattle Max software for tracking livestock being displayed as as add when I clicked the comment link. . . )

  2. pablo says:

    But how good is the bond between the plaster and the concrete? They look pretty discrete to me. Will the plaster begin to break away over time?

    Would luan plywood work for your curved form?


  3. Pablo, There is a piece of expanded metal lath at the interface between the concrete and the plaster plus the surface of the plaster is not smooth and the plaster was soft when the concrete was poured into it.. Between the these gives a good bond.

    On the Luan, we thought of that, and I have a sheet, but it does not demold as easily as the foam board plus the resulting surface finish is not as nice.

  4. Anonymous says:

    Now that you are doing some more cosmetically sensitive work and developing labor savings through composite molding, does it give you ideas about doing a molded roof design in order to incorporate the plaster in the original poor instead of coming back to it later? (how’s that for a run on sentance?) There were time pressures on the first arched roof, but maybe with later sections doing the foam mold, imbedded wire and integrated plaster could be done on the vault. Do you think the plaster would layup on the steeper arch?


  5. Aye, Charles, it does lead my mind to wander over such thoughts… The plaster will adhere to a vertical surface so I don’t think the steep parts of the arch would be a problem. It would be pretty incredible to be able to do the whole interior barrel vault as one unit instead of plastering later. Doing a much large surface area like that will be more challenging so I’m going to work my way up to it gradually…

  6. Pete says:

    Spraying form oil (used in concrete forming) onto the form will give you an even smoother surface with better release. Even diesel of fuel oil will work in a pinch, but you will get much better results with the real stuff.


  7. Pete, yes, I’ve used the form oil some times but not this time. It does help with the release.

Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.