This will be geared towards green sand, since that's what I have. A lot of it should be applicable to other types as well, but don't hold me to that.
Pictured are my tools. From left to right on the board: 6x6" flasks, cast aluminum rammer, 8x8 flasks, heat sink fin and wood pattern, 12x12 flasks with 6x6 and 10x10 chassis patterns laying inside; buckets of molding sand in front row.
The most versatile process, and probably the oldest, is sand molding. Here's how I do it.
The sand can't be plain, dry sand because it won't hold it's shape - it needs some sort of binder. The classical one is bentonite clay, though it seems to me any clay will do (I've heard of fire clay working, and mine here uses KT1-4 ball clay). The way it works is the sand grains are coated with the clay, so when it is compressed it's like kneading the clay together, and it holds its shape.
Well why not just pure clay? It would trap all sorts of gas and the metal would never be able to get in. Why not dry it out? Clay shinks a lot, so the result would be far off from the pattern. It would also have to be fired, needing extra heat and time out of the process, and even if it made it to this point, it would probably crack from the stress of the metal rushing in, as it heats the mold on contact.
So why the sand? It adds a refractory quality which allows it to take more heat. This doesn't matter much with aluminum, but it starts to matter with bronze, and definetly does with iron and steel. It also allows gaps between grains, making it porous, allowing the steam and gas trapped between the mold and metal to be released.
(Check out the molding section for some info on sand, which I don't feel like retyping here. :^)
Pictured is my 120 lbs. heap of sand. No, I don't leave it on the floor...it would dry out completely in a single day! I keep it in buckets as seen at top, covered of course.
You wouldn't think there would be much to just a heap of sand. It sounds kind of boring. Well it's actually a very finely-tuned heap of boring sand! :-p
Pictured above is the sand layed out some time after using it all up in molds (it almost exactly fills the 12x12 and 8x8 flasks, for two molds), ready for mulling and remoisturization. Mulling consists of stomping it against the ground until it's nice and solid, like when it's rammed in a mold. Then it's broken up, fluffed and mixed nicely to spread around the moisture and such. Once the consistency is right, I load it back in the buckets.
Ok, so what's all this about fine tuning and consistency? Well if you have too much clay in the mix, it will stick to itself better, but everything else too. It also seals the pores - remember the mention of solid clay, above? Too little and it won't stay together. Moisture has kind of a duality: more and it gets stickier (I guess it makes the clay work better), but it also again seals pores, and makes more steam in the process. Too little and it'll fall apart, but you'll get less steam.
One of the most famous tests for sand is the hand squeeze test. You grab a handful of sand, squeeze it in your fist and destuctively test it for tensile (pull on it lengthwise) and compressive strength (crush one of the halves) as well as moisture (if it wets your hand it's no good) and porosity, if you're so inclined (seal your lips to the flat side of one of these cores and blow. Don't come crying to me about grit in your mouth after that one, 'cause it happens to me too! ;-)
The first step in sand molding is deciding on how you're going to do it. This pattern is pretty simple. In fact, being a casting itself (a die casting to be specific), it already has a parting line on it, showing where the dies seperated. It looks good to me so I decide on using that for reference. You also have to decide on the positioning and gating (gates are the horizontal paths, that aren't part of the casting, which the metal flows through). Here I placed it off a little to the corner and diagonally, my theory being the heaviest section of the casting, the ground part of the scenery in this case, will spread the metal out the best, filling the mold quickly; and that on cooling (when the metal starts shrinking), this area will cool last, thus it should be closest to a store of molten metal. In a previous casting similar in theme, I got some shrinkage on the opposite side of the casting, in the outer rim. I decided on adding a riser (vertical coulmn) to supply molten metal on that side.
Dad always said "you got the thinking part done, now all that's left is the doing." Well we've decided on how the casting is going to be arranged, let's put some sand on it.
Here I've riddled (sifted) some sand on top of the pattern, about 2" worth. I could use straight sand, but this removes clumps, improving the surface finish. I then ram (compact) the sand on top of the pattern using the blunt end of my rammer, and ramming in the edges around the flask with the flat end.
Once the first layer is rammed, I fill the flask to the top with sand (clumps don't matter here anymore) and ram it in, and continue until the sand is built up over the top of the flask. Now we strike off the excess sand, scraping it level with the top of the flask.
The mold is turned over, exposing the pattern and what sand was in contact with the floor, er, molding board. (This is supposed to be over a special molding bench, but I don't care! :)
Now, this pattern has a very simple parting line - it runs in a plane across the whole thing. If the pattern were flat altogether, I could leave the mold as-is and ram the other flask right on top... but it isn't, there are those three little feet which stick out, leaving sand sticking under it. So now what? I carve out the sand until I get to the parting line. This is known as coping out (even though I'm doing it in the drag - the bottom flask - not the cope (top flask), go figure).
Now we can proceed with the cope. Apply parting dust (fine powder to keep the sand from sticking to itself; I use baby powder) to the mold face and start ramming the cope on top. At this point, you can include cones or cylinders to pre-form the sprue, and riser(s) if any; if not you'll have to carve them out later. Anyway, ram the first layer (which should also be riddled sand) tightly against the drag, pattern and edges as previously with the drag, to ensure good finish and strength. When done, strike off, then seperate the mold. With any luck, it should seperate clean and easy.
The two mold halves. I've carved the gates in the drag (left), but have yet to carve the sprue and riser in the cope (right). Tip: use a few inches of EMT conduit tubing for this. Line it up and just push it into the sand. It won't disturb the surrounding sand and comes out with a sand core inside when you remove it, leaving a smooth hole.
Here she is, one complete sand mold, all ready to be poured.