3D printed PLA mold for concrete casting

Hi, a new project and I’ll be adding to the documentation coming weeks. I hope it gives you useful information or inspiration.

Usually for casting concrete a wooden form is made which is filled. For simple shapes that’s easy, but some shapes are hard or time consuming to make of wood. I wanted to try this already for some time, so I decided to make a 3D printed mold. Interesting study project but for such a simple shape way over the top 🙂

.

This is what I want to make: a sign for use at competition days at my athletic club

(Tion Herfstloop)

We thought of a simple wooden cross with a weight (three tiles or a bag filled with sand) placed on it, a bucket filled with sand (or concrete). It has to be transportable, not too heavy, and it will be decorated (painted).

Rather simple, and it could well be made quickly without any problem. But for learning I’ll use 3D printing. The added value might be: nice shape, own logo, durable.

Questions:

-How thick should the mold be

-Will the concrete release well? Release agent or grease necessary?

-How many times can it be used.

-Aluminium or Portland cement ? Alu is much faster and fine for details. Portland is cheap.

A block with a center hole for a post, 2 handles to pick it up easily and two holes to put a rope through.

These parts would form the sides of a wooden mold. However, I abandoned this idea of separate parts and adjusted the plan for 3D printing the mold.

The 3D print design. The walls are still 9 mm thick, remnant of designing for wood 9mm.

The resulting shape. Looks nice to me. Print it

27 hours printing.

Settings: 0,20mm Standard @BBL A1M (Bambuslicer):

0.2mm layer height, 15% fill, , gyroid.

Nozzle: 0.4mm.

PLA: Basis PLA (Bambu), printed at 100% speed.

The next print ( the shape to form the opening for the hand) I printed at 160% (ludicrous). It turned out well.

Next: sanding before appling a coating.

K120 sanding paper, primer spray paint, 2 coats applied

I bought this material:

Gietsteen (casting stone). You can also use aluminiumcement mortar with sand. Alu cement is not sold everywhere like portland cement. I found it at sculptors supply store Beeldhouwwinkel. Others that have it: Keramikos and Hazelaar. 20 kg costs about 40 euro, transport 12,50 euro. That is much more than 25 kg portland ( around 8 euro / 25 kg mixed with sand (3x 25 kg, less than 10 euro). It better looks awfully nice for such a price. Bought it anyway. If this works well there will be follow up projects.

https://www.beeldhouwwinkel.nl/nl/caltra-gietsteen-standaard-snelwerkend-801

For the through holes I use a PE tube around a metal rod, so I can remove it easily: just pull it out, it is flexible and does not stick to concrete. Metal tubes or wooden rods might get stuck. Then you have to do a lot of work to get them out and keep the mold intact.

For the big hole I cut a 2-piece foam core with the hotwire cutter (smoother than with the belt saw), made a 2-piece foam core (might be easier to remove than a 1-piece core). Taped around it to keep it together. I can, I suppose, remove this core quite easily with a piece of metal, a screwdriver or other sharp tool. We’ll see 🙂

The mold parts. Now: applying release agent ( bekistingsolie (NL), Schalol (GE) ). Oil or vaseline can also be used as well but is not recommended as it may stain the concrete. I found a jerrycan with still 3L in it at Marktplaats for a very reasonable price. That will do for many molds to come.

A thin coat of oil applied, 1 liter is good for 20 -30 square meter, that means a layer three times as thin as a paint layer, so apply sparingly. I applied it with a dry-ish brush and removed excess with a paper towe

Cement and water. Ready to go

Mixed. Now tapping against the bucket with a stick, to get the air out

I waited 5 minutes before casting, following the instructions. But it set very fast, I could hardly poor 80% into the mold and it got to stiff already. Probably too hot today (over 25°C). So I quickly prepared another batch, mixed, did not wait 5 minutes this time and still it set very quickly. But now I succeded to fill the mold to the rim and hit it gently with the plastic hammer to get the air out and level the surface. Even filled a couple of small test molds. Covered it with wet cloth and a plastic bag to prevent that the concrete dries out. Tomorrow we’ll see how it ended up. Will it release properly, will there be a visible separation between the first and the second cast? I must go home now and not have a look (and rip the fresh concrete apart)

Ok, some positive and some negative points:

1 cast came out easily,

2 tubes and center hole worked out well

3 overall shape and weight are convenient

but:

4 shrink cracks,

5 characters were damaged

6 casting in two times gave an ubly joint

next cast:

Working at lower temperature or using a slower setting cement so I have enough time to fill the mold in one time

adding sand against fracture

Piz Roseg model

foto: google earth

3D printed model of the mountain on the border of Switzerland and Italy. Usually a cad program that uses heightfields (greyscale to elevation) is limited to 256 steps. That would result in stepped slopes, not nice to look at. I used Rhino 3D and the grasshopper plugin Bison to import the original geotiff files with a higher resolution (16 bit = 65K levels). Much better, smooth slopes.

However, it seems that Bison does not work for the Dutch DEM files (from the Algemene hoogtekaart Nederland) ahn.nl which is annoying.

New 3D printer

After the maker residency where I worked successfully with a Prusa Mk 3 and Mk4 printer, printing TPU for stamps, I decided to buy a new, modern printer. The one I already had dates from around 10 years ago and is gathering dust on the shelve now since around 2 years. I left 3D printing for what it was, focused on the laser and the router and if i had to print at all, i went to Fablab Enschede with it’s farm of Ultimakers. Even though it was not a bad printer, I never managed to get the multimaterial working properly (dual head, always leaking material from the head not in use and it had to be aligned very carefully) . Printing took just too long, and it had to print slowly to get a decent quality. Besides, the slicing software and the display were not user friendly to use. I mastered the levelling of the printing plate by hand (lots of trial and error), but it took too much time. I also wanted to get rid of the Peek connections in the printing head, they caused to much leakage and jams. Too much trouble to keep tinkering with it….

But now i enjoy printing again. The new machine, a Bambu A1 combo with multimaterial device, prints pretty good, straight from the box. And it prints fast. There was a flaw however, the Z axis was not exactly perpendicular to the building platform but with a small script in Grasshopper / Rhino3D I corrected that quite effectively and it prints at a 90 degree angle now. Which is very useful for architectonical maquettes (models).

Sure, i could have purchased a Mk4 as well, but i have there is such a huge difference in price… and the multimaterial option is also important. So i gave in and choose for the closed source Bambu with doubts about client helpdesks and so on. Still, i do not regret it.

TPU (rubber) printing works fine and the material is surprisingly strong, it just does not tear apart. I use it for stamps.

And last but not least I can print landscapes with it, which is useful for my work. Next test will be with multicolor, such as blue for below sea level, white for the 0-meter line and other colors for the dry land. It will be a nice addition to the lasercut, stacked models i made up to now.