Edge gating technique

[CI]

Here is a video of someone casting some parts for a 1/4 scale V8 engine.
This individual has done a lot of work making lost foam castings using a CNC router and pink hardware store foam.

The lost foam method is well established in the industrial world, and used for engine block castings, intricate pump castings, etc.
Typically the industrial method uses beads that are expanded inside of a heated metal mold, to make the foam pattern.
The molten metal vaporizes the foam as it flows through the mold cavity.
The key is to get an exact pour rate, pour temperature, and an investment coating that has enough porosity to allow the vaporized gasses to pass out of the mold cavity in front of the molten metal wavefront.

While the work I have seen with hobby lost foam castings is impressive, I don't plan on using this method because the surface finish is not as good as I want, there are potential cold joints where molten metal meets from two sides (the mold filling process is very slow), and there are sporatic defects visible in some of the castings.
For many, these castings would be perfect for their build, and the quality is most impressive.
The fact that these thin castings were even made, by any method, is a significant achievement.

The fins and lettering on top of the valve covers is really challenging, and it is difficult to pull patterns like that from the sand without damaging the mold.
CNC'ed foam patterns can be made with little or no draft angle on things like the lettering and fins on the valve cover, and that is not a problem with the lost foam method since you coat whatever shape your pattern is with investment coating, and things like draft angle don't matter.
From a structural standpoint, draft angle on a pattern makes a stronger casting I think, and the radius at corners eliminates stress concentrations at adjoining surfaces.

One drawback to the lost foam method is that a new pattern has to be CNC'ed out every time you make a casting, since the pattern is melted in the process.
Not really a big deal if you have a good CNC machine.

The fact that there is so much visible gassing out the sprue during the pour is concerning, since the gasses are flowing past the clean metal that is flowing into the mold cavity, and that is going to entrain slag, sand, air bubbles, etc. into the pour.

What is impressive about this video though is how thin the castings are.

The other takeway from this video is the edge-gating system that is used.
I have used a lot of knife gates, and they work very well, and keep metal velocity and turbulence to a minimum.
The gate used in this video is almost the entire edge of the casting, and there is a large runner behind the gate.
I tend to use oversized runners and let the gates regulate flow, and this is pretty much what I think I am seeing here.

Quite an interesting video from many aspects; hobby/foundry/model engine/3D modeling (SketchUp was used).

<a href="https://www.youtube.com/watch?v=Mmze9FxQ1XQ" target="_blank">http://www.youtube.com/watch?v=Mmze9FxQ1XQ</a>

.

[CI]

I saw a thread yesterday that I guess is several years old, but had forgotten about, which was a Merlin engine built from a set of those fantastic investment castings that someone in California made.

I think potential problems can multiply when one gets smaller than perhaps a 1/3 scale with a complex engine like a V8 or V12.
The Merlin thread mentions some tedious looking casting straightening, which is really risky since the method used was heating the casting and bending it.
My experience with heating aluminum casting and trying to bend them is that they break.

A 1/3 scale V8 would be much more forgiving in my opinion.

Everyone has their favorite scale.
I admire those who can build working engines on a tiny scale.

If the previous post's foam castings are not exactly square, then they will have to be either machined flat if possible, or straightened using some method.

I also wonder whether machining allowances were included on the above patterns, or whether the patterns are a CNC of a finished engine.
I suspect it is CNC'ed patterns of the finished engine, so that is not going to work out too well when you start machining.

If close attention is not paid to the shrinkage factor, then the castings will be too far out of spec, and will not be machinable to a size that matches the drawings.

Many factors to consider for sure.

The evolution of the hobby into 3D modeling, and then the offshoots into various casting methods like lost foam and lost PLA is quite interesting though, and I think these techniques will see further refinement over time.

For a while, I thought that they would perfect the wax filament for 3D printers, for use with investment castings, but I don't recall any viable uses of that in the hobby.

.

[vtsteam]

I've done lost foam casting in zinc, aluminum, and one failure attempt in iron. It's not my favorite method of casting, but there are times when, for me at least, it's useful. Mainly for intricate thin walled castings. It's not an enjoyable process, and for what I build I generally I don't need a thin walled intricate casting. I do enjoy conventional greensand casting in all of its more traditional aspects, but if I need something in whatever method is required, I'll use it.

[CI]

That is some impressive lost foam casting work, especially that extrusion bracket thing !
I can see that it seems to work pretty well.
You must have the coating and pour process figured out.

I wonder if the iron failure was due to the higher iron temperature creating gas faster than the coating could dissipate.
I know they use the expanded bead lost foam method with cast iron engine blocks successfully, but hadware store foam is not expanded beads.

Edit:
I have seen some discussions concerning hardware store foam density, and apparently it can vary.
Hardware store pink and blue floam seem pretty lightweight, but the white expanded bead foam seems to be lighter.
I would guess the lower the foam density, the faster the mold fill, and probably a higher success rate.

There are also discussions about machinability of hardware store foam, and apparently the more dense foam is more easily machined (CNC'ed).


I wonder if 3D printed sand mold machines will ever be cheap enough to hit the hobby market.
That would allow some very complex molds to be made, with internal passages, etc.

.

.

[Jasonb]

Question

If you are coating the foam with layers of investment and forming a shell why does the foam have to remain and be burnt out by the metal.

Could the foam not be melted/burnt out once the investment shell had been dried avoiding the need for the slow pour and gasses that Pat mentions? In effect you are using the foam like you would a wax.

[Jasonb]

If the previous post's foam castings are not exactly square, then they will have to be either machined flat if possible, or straightened using some method.

I also wonder whether machining allowances were included on the above patterns, or whether the patterns are a CNC of a finished engine.
I suspect it is CNC'ed patterns of the finished engine, so that is not going to work out too well when you start machining.

Pat, is this a problem you are likely to come up against as you say you intend to cast "Net" size parts. OK if they come out flat and true but what do you do if they don't? I can't see that lost foam would be any more likely to give problems than other methods.

[CI]

Pat, is this a problem you are likely to come up against as you say you intend to cast "Net" size parts. OK if they come out flat and true but what do you do if they don't? I can't see that lost foam would be any more likely to give problems than other methods.

Apologies in advance for this book; just thinking outloud.


There is a water quench involved in getting aluminum 356 to a T-6 temper, and the books recommend a hot water quench, to minimize warping.
I have tried hot and cold water quenching, and have not really noticed a difference in warping, or any noticeable warping for that matter with either method.
I think as the scale factor goes down, the tolerance for warpage becomes much more critical.
For a casting that is 12 inches long, a1/16 inch deflection is not that big of a deal, and generally can get machined out.
For a casting that is 4 inches long, a 1/16 inch deflection may be a big deal.

Those Merlin castings that came out of California were investment cast, and there did not seem to be much machining allowance on them, which is probably why that one individual was straightening some of his castings.

I was getting heavy into T-6 tempered 356 aluminum castings because I could not get the iron process to work consistently, and I thought that would be the best I could do.
Then I figured out the iron casting process, and so I don't deal with aluminum castings much anymore, except perhaps the BHM permenant patterns, which are not tempered.

I have not had to temper iron parts (so far), and I have not noticed any warpage on my iron castings, but I tend to make castings intentionally on the large side as far as model engine building goes, just to avoid having to work to very tight tolerances on castings.

I use to use quite a bit of machining allowance in the 3D model, and over time, I have been reducing that more and more.
I think I am using about 1/8" machining allowance, but I suspect that with resin sand, I could go to 1/16".
I don't see a big advantage of going to 1/16" machining allowance though; it just takes a minute to skim off 1/8", and that gives a bit of insurance against any warpage or misalignment.
The benefit of approaching net size on castings is that it speeds up machining time.

The major advantage I like about bound sand is that the patterns are not rapped side-to-side before they are pulled from the sand, as is done with traditional green sand molds.
The rapping side-to-side makes for oval parts that are designed as round.
I can chuck up a bound sand flywheel casting, and there is no wobble either on the inside or outside of the rim or hub, which is really nice.

The net size thing I think is just a natural progression as one's molding techniques improve.
There is no reason to make a casting that is less accurate than what can easily be made.
Sort of like machining; no sense in maintaining a 4 thou machining tolerance when making an engine if you can maintain 1/2 a thou tolerance easily.

I think the machining allowance should accommodate any warpage, but that assumes the warpage is minimal.
I think if one is getting warpage in iron parts, then something is wrong with the method or mold.
I have heard of some folks aging their iron castings so that any movement happens before the machining process.
I have not built enough engines to know whether iron aging is really a problem in model engines.
It is important to let iron castings cool overnight in the mold, to avoid hard spots, and this may also help to avoid warping.

I think there is a place for hardware store foam castings, as VT illustrates.

I am wary of the fumes generated by burning foam.
There are stories on the casting forums about people burning the insulation off of electrical wiring, to make scrap copper.
They make the mistake of inhaling the smoke, and as one person mentions he had one brief whiff of burning plastic smoke, and had permanent lung damage that has disabled him for life.
Buring plastic fumes is said to be able to knock a bird out of the sky dead from 30 feet up in the air.
At any rate, I take extreme measures to avoid exposure to toxins in the foundry, and if I did lost foam castings, the same would apply.

The biggest problem I see with lost foam castings is the slow fill rate.
For things like model engine blocks and parts, I would worry about cold joints, there two molten metal fronts meet, and both have a bifilm coating on them.
For a structural part, chances are a cold joint will fail.
If the molten metal fronts are hot enough, then the joint will merge, and there will not be a problem.
I can see the cold joints in many examples of lost foam castings.
I have never seen anyone try to break the castings at the cold joint to find out if there is really a problem.

If you can fill the mold cavity without to wave fronts meeting, then there would not be a cold joint problem.

With regular casting methods, what I have always heard from the "experts" (experts being those in the commercial foundry business with a long history of mearureable and verifiable success) is that the mold cavity should be filled as fast as possible, but not so fast that turbulence and splashing occur inside the mold cavity.
Laminar flow of the metal, and metal that is as close to pour temperature as possible is what you want to fill the mold cavity.
The pour temperature used with lost foam is much higher than sand casting pour temperatures, and this may affect surface finish.
I think vacuum assist is mentioned in the above video, to speed up the mold filling process.

I like resin-bound sand molds because the results span many of the good qualities of many other types of castings, such as investment, etc., and with a ceramic mold coat, you can also get excellent surface finish.

I feel like if I can get near investment casting quality by using resin-bound molds, then there is no reason for using the much more complex investment process.

In the end, I think the casting method has to be selected that best fits the application.
Like cutting bits:  do I use carbide, tool steel, etc ?
Depends on the application, but generally I use carbide for almost everything I machine, and I get acceptable results for my needs.

.

[Jasonb]

I don't see a big advantage of going to 1/16" machining allowance though; it just takes a minute to skim off 1/8", and that gives a bit of insurance against any warpage or misalignment.
The benefit of approaching net size on castings is that it speeds up machining time.

That seems a bit of a contradiction, in the Monitor and other threads you have said Net will save machining time yet here you say it only takes a minute to take an 1/8" off.

Surely if it is going to be that quick to machine something you may as well leave the 1/8" which would guard against warpage and that minute to machine would likely be less than the time spent with mylar or other plastic films when trying to avoid draft angles as they can be part of the machining allowance - typically sloping a surface from 1/16" to 1/8" give sboth draft and machining allowance.

You must have been lucky with the two flywheels you cast for the green twin ( don't recall you machining any others), having machined 50 or more I find there is almost always some oval, pringle or mould shift that means you simply can't get a casting to be totally true

The blue/pink foam is used as it gives a far better machined surface than white which would only really work for basic shapes that need all surfaces machined. Different matter where it is expanded into a smooth mould.

[CI]

That seems a bit of a contradiction, in the Monitor and other threads you have said Net will save machining time yet here you say it only takes a minute to take an 1/8" off.
If I mentioned "net" in the past, I think I actually said "near net".
They do make commercial castings that are net, and they do not require any machining.
So it becomes a matter of how close one wants to get to net.
As I mentioned, I am gradually reducing my machining allowances, and as I am able to measure more castings, if possible I will reduce the machining allowance, but not more than about 1/16".

The iron castings I have made seem to be very flat.
The casting will be as flat as the resin bound mold (see below).

I really don't like the draft angle on the exterior of flywheel rims, since it makes the flywheel very difficult to grip in the lathe chuck.
Sometimes the draft angle on 12" flywheels begins to make the flywheel too large for me to machine too, so a good reason to eliminate that.

I played around with resin-bound molds a lot when making the green twin, and casting other miscellaneous parts.
The green twin flywheels were indeed very round, but that is no surprise because the flywheel pattern was made on a commerical high-end 3D printer (farmed out that print), and the resin-bound sand does not move at all; it just sets exactly to the pattern.
Resin bound sand will pick up texture as fine as a hair, or less.
I suspect most casting kit castings are not made with resin-bound sand, and if they are made with resin-bound sand, the molder may not use as much care in making and aligning the mold halves as a hobby model engine caster can or would.
If a molder has to cast 50 or 100 castings, chances are he is not going to use the same standard of care as could be used in a home setting, making only a few molds/castings.

Two problems I had with resin bound sand, both of which were solved were:

1. Resin-bound sand sets after X minutes (x varies depending upon the amount of catalyst), but there is a period of time after set has been reached where the mold half must be place on a perfectly flat surface, such as glass, or a very flat manufactured coated wood shelf.
I discovered that if the set mold is placed on a surface that is just 1 thou out of flatness, then the mold will warp, since it is not fully cured.
Two mold haves each with 1 thou out of flatness equals a lot of flashing that runs out between the mold halves, and also makes for a casting that is not true and round.
The solution is to immediately place the set mold half on a very flat surface for perhaps 30 minutes, before combining the mold halves.

2. The second problem I had was trying to use commercial inserts, such as Adolfs, to create locking parts on the mold.
It is extremely difficult to get these inserts exact, and even deviations as small as say 1 thou created problems with closing the mold completely, and thus again excessive flashing and not a perfectly round part.
I stopped using the commercial inserts, and instead started drilling 1/4" holes through the cope mold into the drag mold, after the cope was rammed on top the drag.
Then 1/4" stainless rods were inserted in the drilled holes after separating and mold halves (to remove the pattern), and then reassembling them.
This gave a very accurate mold alignment, little if any flashing; and I still use this method today.

I have not tried the mylar film yet.
The assumption is that I will be able to ram the mold, let it set, and then be able to withdraw the mylar without damaging the mold.
I will probably mold up a test piece, just to verify this theory, before I get too far along with the BHM patterns.
It may be possible that the mylar breaks the sand-pattern bond enough that the pattern can be removed from the mold in the usual way (using a small slide hammer).
The slide hammer pulls the pattern directly perpendicular out of the mold cavity, which is why flywheels are very accurately round using this method.
An initial impact is made in one or more places on the pattern with the slide hammer, with just enough force to break the bond.
Once the bond across the entire pattern is broken, the pattern comes out of the mold straight up, with no side-to-side motion.
It is a very accurate process, and more accurate than I had hoped it would be when I started using resin-bound sand, especially after resolving the two problems above.
Not quite investment casting accuracy, but close.

.

[Jasonb]

Best practice is to hold a flywheel by the inside of the rim if using a chuck as that allows you to set the cast surface that won't be machined true while at the same time allowing you full access across the face of the rim. It also means you don't need such a large lathe. To grip a 12" flywheel by the OD you would want at least 14" swing so the jaws clear the bed and even more if the carriage legs get in the way.

Other oprion is a faceplate which again gives access but a bit longer to get it running true. It is often needed with 5 spoke flywheels as the spokes tend to clash with the jaws.

[vtsteam]

Sorry to be so late with answers to a couple of questions above.

Jason, people have tried removing the foam pre-casting via solvents and via burnout. It just adds an extra step and time. For aluminum, at least, it is possible for me to get good results without doing that.

I have only one (failed) trial with iron so I can't say with that. From what I believe happened, the problem for me was the coating broke up before solidification -- but I was trying Satanite as an investment, rather than my usual gypsum based wallboard compound coating, or Plaster of Paris (for aluminum or zinc), both of which I thought might calcine at iron heat. No proof of that either since I did not try them, and I try to avoid definite statements about things I haven't actually experienced!

There are professional investments for iron, but I typically avoid most professional products in favor of home brew and readily available inexpensive or free ingredients, where possible. Just a personal quirk. If I can't do something homebrew style, I move on to something else that does work that way. For the record I certainly did not exhaust all the possibilities on the one failed iron pour. But since I get good results with greensand with iron, I didn't have the interest level to explore DIY lost foam iron. Maybe some day I will revisit.

re. blue/pink extruded polystyrene foam, and white expanded bead polystyrene foam. I've used both extensively in building model airplanes for years, and so I'm very familiar with them. As you say CI, expanded bead (white) foam is used extensively in iron lost foam casting. But that is because it can be molded itself -- generally using steam to expand the beads in the mold. Surface finish can actually be controlled quite well -- as can density. Bead size temperature, pressure and many other factors mean it isn't necessarily going to have a rough or beaded surface like a cheap ice cooler, packaging, or a coffee cup.

When a DIYer uses it -- it will generally be cut to shape, and will come from a rougher or, bigger bead, or lighter density source, and depending on how it is cut (knife, saw, hot wire, needle cutter) will vary in surface finish by all of these factors. But it would be an over-generalization to say white foam will have a worse surface finish than extruded foam, for casting purposes.

I agree, CI, that the fumes from lost foam casting are objectionable, period. It also impregnates the sand (which does not have clay or moisture in it, as greensand does). I hate that smell, so my use is only occasional, outdoors, and for small parts, if and only if I can't do what I want using greensand molding.

Or by fabrication. Even though I like casting, I also enjoy fabrication. Particularly brazing for small steel parts -- which not a lot of people seem to do these days. It works quite well for cast iron, too.

As far as machining allowances go, that's different from shrinkage allowance. Shrinkage allowance in applies to all of a casting. Machining allowance applies only to surfaces that will be machined. If you like castings in old engines, most of the surfaces of many parts will be left in as-cast condition. So you only need to apply shrinkage allowance to those surfaces. Where you will machine, you add the machining allowance. Of course you guys know this, but sometimes it's possible to forget in a general discussion about pros and cons.

In my experience, lost foam casting with extruded foam leaves a finer as-cast surface finish on aluminum, than greensand casting. This can also mean that a machined surface may need less of an allowance to get down to clean shiny metal. Really it's a judgment call to determine machining allowances for a pattern. That will be based on personal experience with what your own castings are coming out like in a particular material, and with a particular casting method.

As a dyed-in-the-wool simple DIY hobby caster, I use only scrap aluminum and scrap iron, and I do not heat treat aluminum. I also wait for aluminum molds to cool for at least an hour and allow at least 5 hours or more for iron before breaking out. I don't douse aluminum parts with water to cool them, as I see so many do on YouTube. I have never experience aluminum or iron part warpage due to rapid cooling.

I have occasionally experienced chill in cast iron, and in that case have heat treated in the furnace and slow cooled it to anneal it, and this has generally worked out well.