I'm interested in the idea of building rigid high strength frames, for example for a lathe or milling machine frame, by burning out shapes in plate steel on a CNC plasma cutter, then assembling those plates with 3d printed support jigs holding them together and bolting through them with long bolts. Then pour the interior with concrete and after the concrete sets up, tension the bolts to prestress the concrete...
It's a bit limited in shape, you can't really include the type of features you see in cast iron, but machine bases typically don't need them that much.
When you're attaching a component to the base which needs precise orientation, you attach it to a jig which holds it true, then you bring the bed up to it and you epoxy the mating surfaces so the epoxy acts as a shim. After the epoxy hardens you can tension the hold-down bolts.
Caleb James DeLisle
in reply to Caleb James DeLisle • • •I realize that there's a general purpose way to make a high precision surface on something:
1. Lay a piece of surface ground steel, maybe 3-5mm thick, down on a surface plate. This steel must have a smooth finish but it does NOT need to be precise in any way.
2. Tape around the perimeter of the steel with vacuum tape and attach a vacuum hose on one corner.
3. Start the vacuum pump, sucking out all of the air between the steel and the surface plate. At this point, the vacuum will flex the steel, making it conform to the surface plate.
4. Take any old rigid piece of steel, stone, or concrete, paint it with epoxy, and stick it to the exposed surface of the steel.
5. After the epoxy hardens, release the vacuum and cut the tape free.
At this point, you have a high precision surface that has been twinned to the surface plate. The epoxy has hardened in the form of a "perfect shim" between the rigid body of the part, and the strip or plate of steel.
Caleb James DeLisle
in reply to Caleb James DeLisle • • •Problem with the above is you don't know how uneven the part is that you're sticking to, too uneven and there's not enough epoxy to fill the gap and you get a void.
A related concern is if too much epoxy is painted in places to fill up the voids, the thick areas will be subject to more shrinkage than the thin areas. Epoxy glues shrink by 0.1 - 3% depending on the type.
Two thoughts come to mind:
1. This process can be done using tinfoil, and then you can hit it with a wire brush to scrape out all of the voids since unbacked tinfoil will tear out instantly. Then the process can be repeated again with a little bit of extra epoxy painted in the void areas. When the entire surface comes up with no voids, then you do the final mating with sheet metal.
2. A related idea is to pour concrete down against the sheet metal on the surface plate. This will have more uniform shrinkage so voids should be less of an issue.
CC: @souldessin
Soul Dessin
in reply to Caleb James DeLisle • • •In the flooring business, they use a very liquid cement that is self-leveling, they dam it up and then pour it in, and then it will naturally level to perfectly flat and even.
That could be replicated into something else by just using a very watery epoxy. Something with a very low surface tension.
Caleb James DeLisle
in reply to Soul Dessin • • •That'll work for making something like a table, but when we're talking about precision, things get "weird"...
The thickness of a human hair is about 2 thousandths of an inch. Paper is 4 thousandths, a laboratory grade surface plate is guaranteed flat and true to 0.045 thousandths.
When you get to these precision levels, you can put a 1/2 inch piece of steel under a measuring instrument and watch it flex as you push it lightly with your finger. You can hold a part in your hand and see it expand from your body heat in real time. Pouring even the thinnest epoxy and letting it cure in air might get you down to 1 thousandth, but you can get that by purchasing a piece of bar stock and making a simple surface grinder.
Surface plates are made on very high precision surface grinders, those grinders are themselves made using master surface plates. The masters are made in groups of 3, they coat them with a die called Prussian Blue, this die creates an extremely thin and uniform coating of about 0.01 thousandths. They lay one plate down on top of the other and this transfers blue marks from one to the other, then they scrape it using a tool which can take off around 0.001 thousandth per pass. They use 3 because when 1 and 2 match, and 2 and 3 match, and 1 and 3 match, then they know the surfaces are actually flat...
So I would say if you want to build precision machines, the easiest thing to do is buy your surface plate and then make everything else by cloning it.
Off topic: A surface grinder is actually quite easy to make, you clone a surface plate twice, once for the bed and once for the carriage. The carriage slides on a film of oil over the bed and you just use rudimentary chains, tracks and screws to control the right/left and forward/back because these are not precise. Also the angle of the head is not precise, everything can be eyeballed.
Caleb James DeLisle
in reply to Caleb James DeLisle • • •A surface grinder is the most rudimentary machine, it requires only 1 flat surface.
A lathe requires 4 flat surfaces:
1. Vertical ways (up-down truing)
2. Horizontal ways (back-forth truing)
3. Vertical cross-slide - parallel to horizontal ways
4. Horizontal cross-slide - square to vertical ways
Then the headstock needs to be parallel to horizontal and vertical ways.
The milling machine is the worst, because it has all of the surfaces of the lathe, but it also has a Z surface which needs to be trued on two axes.
Fortunately the lathe and mill are not generally expected to produce work better than +- 1 thousandth, but the cylindrical grinder and the EDM have all of the same problems as the lathe and mill, but are held to a much higher standard...
vandys
in reply to Caleb James DeLisle • • •Caleb James DeLisle
in reply to vandys • • •I've been thinking about this, I have two possible solutions:
1. Air bearings which are obviously self-cleaning
2. Oil bearings with an oil pump which forces a continuous stream of fresh (filtered) oil into the bearings and sweeps away dust
Air bearings have really good precision, but they're expensive to buy commercially and not that easy to make, so I tend to lean toward #2, but I don't know how much oil flow you can push without it beginning to lift the carriage up off of the ways...
I imagine somebody has studied this because it seems weird not to...
Caleb James DeLisle
in reply to Caleb James DeLisle • • •