Another Lost PLA Casting Method

Home Model Engine Machinist Forum

Help Support Home Model Engine Machinist Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.


Well-Known Member
Staff member
HMEM Supporting Member
Global Moderator
Jul 2, 2021
Reaction score
MidSouth, USA
I ran across this video the other day showing a lost PLA process using sodium silicate sand.

Pretty interesting.
The method needs some refinement, but the results are impressive.

This method is quite a bit easier (I think) than the investment casting process, and uses materials that have a long shelf life (investment casting slurry has a stated shelf life, which is somewhat short; perhaps it will last longer than advertised).

I will have more comments when I get a minute.

1. Rule No.2 of aluminum casting: Never stir the molten aluminum.
(Rule No.1 of aluminum castings is don't overheat the melt. Pour around 1,350 F if possible).

2. Using excessive amounts of sodium silicate makes the cores extremely hard and difficult to remove, as you can see in these videos.
If you use the recommended ratio of sodium silcate to sand, the cores will break out easily, especially in water.

That was very interesting! I couldn't tell if the pla was burnt out in his oven or by the pour, or a bit of both. He also used an induction coil to melt the metal, I would like to know more about that. All in all, very interesting, and the castings were flash free.
I have seen small import induction melters, and they work, but would seem prone to problems, since the components are water cooled.

An induction melter is the ultimate in my opinion.

I did talk with one individual who purchased one, and ended up burning up his electric service electronic meter, by feeding excessive harmonics back into the power company system. He had to add a line reactor.

If you could get replacement parts, an induction melter may be ok for small melts.

For larger melts, the electric service required is generally 3-phase, with 480 volt preferred, and you still have to be careful not to feed harmonics back into the power system.

And some power companies will include a demand fee, and bill you 12 months for a demand you set in 5 minutes.

I call my oil-fired furnace a "poor-man's induction furnace".

At a burn rate of 3 gal/hr, I can produce about 415,500 Btu/hr (121.77 kW) with my oil burner.
I could never power that from my home electrical system.

I have used my oil burner in experiments up to 10 gal/hr = 1,385,000 Btu/hr (405.92 kW), but normally you don't need that sort of burn rate for a home foundry.

Below is the induction unit I saw demonstrated.
It requires a chiller and pump to keep the cooling water flowing to all the components and to the coil.
Worked well, but a rather complex setup, and not really inexpensive compared to an oil burner.

A used electric kiln can be found for under $500.00, cost to run around $2 per melt takes around 2 hours which is around the time to mull sand, ramm up and get ready to go. Can be used in the shop with doors closed and degauses perfectly enough as there is no combustion of fuel and time to naturally diffuse.
Rather important rule in aluminium casting is, if using bits of scrap/offcut aluminium, do a "heat test" on a small piece out in the backyard first.
Any aluminium containing magnesium will start a fire you cannot put out.

I would consider that as rule No.1.:)
Rule No1 for me would be get a decent surface on your patterns before you even think about casting from them ;)

He got a good result but it reproduced every layer in those printed patterns
The print lines are a bit of a problem with the lost PLA method, since if you fill the 3D printed pattern, or paint it, it may not burn out cleanly.

It should be noted that many/most patterns can be the split type, and a core can often be made from the interior of a 3D print.

Some folks use the lost PLA method when a split pattern would also work as well or better.
I am of the opinion that a split pattern with a core is the easiest way to go for most castings.

A split pattern with core eliminates the entire burnout process.
The lost PLA really only needs to be used when the casting is very intricate.

One benefit of the lost PLA process is that you can completely ignore draft angles, radius, etc.

I don't recommend ignoring radius, since this creates stress points in the casting.

And I like the look and feel of a casting that was made from a pattern that has draft angles.
Castings made from patterns without draft angle and radius start to have a Lego-block look to them, with all the flat surfaces.

I have seen some spectacular iron castings made with lost PLA and multiple layers of slurry (SuspendaSlurry).

I am told that sodium silicate molds will work with iron, but I have not tried that.

Another benefit of using a split pattern, and a two-part bound sand mold is that a ceramic mold coat can be sprayed onto the inside of the mold.
I am not sure if several coats of ceramic mold coat would fill all the lines, but it may.

I will stick with my filled 3D printed pattern halves, that can be repeatedly used, but I can see that the lost PLA method is here to stay.

To me the lost PLA method makes much more sense than the lost foam system.
With the lost foam system, the pattern has to be hand-made every single time, which seems rather counterproductive, when you could just 3D print another pattern.

In this video, the same individual makes some pretty impressive cast iron sleeves for one of his engines, however he snatches defeat from the jaws of victory with a bad arrangement for the sprue, runners and gates.

And he is does not need a high percentage of sodium silicate for the simple cores he uses in this video; 3% is generally the most you would use for a core.
Some folks mix a little sugar and water into their sodium silicate cores to make them break up easily, but I have found that if you use a 3% ratio with sodium silicate, it breaks up easily when water is applied.

A sodium silicate core with more than 3% will be hard as a rock, as you can see in his video.
He is adding a lot of extra and unnecessary work when he uses excessive sodium silicate in his cores.

The sprue he uses deadends into a flat bottom (see 16:18), and so when the molten iron hits the bottom of the sprue, it is creating a mushroom of metal that is rolling back onto itself, and entraining air and slag in the process.

To avoid his sleeve defect, he should use a sprue that transitions into a horizontal V-shaped runner, with a curved transition at the bottom of the sprue into the runner. The runner should run past the two sleeves and terminate in a spin trap that extends out the top of the cope.
Off of the runner he should use a knife gate at the top of the runner, into each sleeve, with the gate being almost as wide as the sleeve, and a gate height to match the thickness of the sleeve.

The cores should have a hole through their center (between 1/8" and 1/4" hole, using a wood dowel rod when making the core), and those holes should be vented out the top of the cope.

The highpoint of each mold should have a 1/16" vent hole out the top of the cope, to prevent trapping air in the top of the mold.
Some say vents are not necessary, but bound sand molds seem to trap a lot of air, and I have found that the high point of each mold cavity must be vented out the cope with bound sand.

The secret to making good castings using any metal is to control the velocity, avoid sudden transitions, avoid any path that will cause the metal to splash around. High velocity molten meltal will almost always create castings full of defects.
You want a low velocity laminar flow of the molten metal into the mold cavity.

You want to let any entrained air, loose sand, and slag to be swept down the runner and into the spin gate, before the metal reaches the gates and begins filling the mold cavity.
And it is best to fill upwards into the mold cavity, and not have the metal waterfalling down into a mold cavity (he got this correct).

The lip of the crucible should be as close to the top of the sprue as possible, to avoid the waterfall/high velocity effect from pouring down the sprue. The pour should be continuous without interruption once the pour starts, and the sprue should be completely full during the entire pour, to prevent aspirating air down the sprue.

There are some very subtle things one has to pay attention to if one wants to avoid casting defects.

I am not on ytube. Perhaps someone can post this comment on his channel.

Edit: I sent him an email with this information at the email address he has listed on his channel.
Perhaps it will reach him.
Very clever and creative fellow for sure.

Last edited:
I would normally spend a lot of time preparing a pla print before using it to cast something. In a way, his method is easier. Pla is a pig to sand/scrape, fill, prime etc. All he has to do is hit the finished casting with a variety of sanding disks and files (and filler?) and he will have a surface that will take paint and look good. It won't look like a cast surface, but does that matter?

For what I do, my home made small electric foundry will suffice. If I ever felt the need for more heat (mine will hit 1000C), I would build or buy an oil burner.

He is a talented man though.

Latest posts