Robust milling machine for aluminium wheels?

You won't find a DIY milling machine that can handle the requirements for making a product like a full size car wheel.

 

It's almost certainly gonna be a 5-axis mill to true up the outside and finish up the spokes in any kind of turbine style. A 3- or 4-axis you might be able to get away with if you only want pure 2.5D prismatic designs and design your mounting flanges accordingly.

You're probably not going to get the surface finish that you'd really like from hobbyist machine-building components, but you could try some prototypes with your OX and see. I suspect that a good-size heavy mill is really what you need- the design of which requires months or years of learning (check out CNC Zone to get started) and a fair bit of money (5 figures, realistically) to then build. I'm as done as it's ever possible to be with the learning aspect, now figuring out when the optimal time to start the building. Hopefully next year sometime.

If you're lucky, there may be a middle ground, where you can frame up with aluminum, mount linear rails instead of using V-wheels, use ballscrews instead of threaded rod, and then backfill as much as you can with some kind of cement product. That could potentially get you the surface finish you need since any chatter induced by the aluminum's resonance should be mostly absorbed by the concrete backfill. Relatively low stepover and the slowest feed/highest speed you can reasonably get away with might take you the rest of the way there.

In any case, plan for expensive axis drive motors. NEMA 34 steppers or hybrid steppers, at the least, I'd say. You can pick up a hybrid stepper and drive for $250-300. Maybe DC servos, since there's not an enormous difference in price at that point and you'd be able to move faster and get smoother finishes. You might even get lucky and find a job lot of used AC servos and drives, which would be ideal.

 

I originally had a huge screed along very similar lines- lathe/UMC, multi-spindle/single point tooling, etc, casting methods and issues, pressure testing, etc- written up, until I looked up exactly what a 3-piece wheel is. Turns out, probably less relevant than you'd imagine when hearing "wheel". The center section is just a little pancake of material with the design in it, which shouldn't require as much of an intensive machining process. A high speed spindle and a moderate speed 5th axis with good quality cutting tools in a vertical package should do the job in terms of getting the mounting flange and center journal round and concentric.

A picture would have been handy prior to writing that out, admittedly.

 

Right, but look at the scale of his machine from another video:



Huge weldment structure, similar to This Old Tony's build but more commercial-styled. This is extremely similar to my original "big build" design, but I've moved to more of an interest in horizontal designs lately, and become increasingly concerned about the dimensional stability over time of weldments (ie. they move due to stresses, even after cooling). You can get it heat-treated, but obviously that requires a massive oven, so I've been thinking of ways around that. He should fill those tubes, too, not sure why he's left them hollow.

Anyway, point being; steel is a better construction material for CNC machines than aluminum. Alu is just cheaper to ship, easier to build with, and very rigid for its mass- ideal for laser cutters and 3D printers, but a bit of a hack for milling. Obviously many people get workable results from it in materials that are fairly finish-tolerant like woods, plastics and composites, but as you start moving into metals- which will clearly show an imperfection only 0.0005" big -you're getting a little closer to the edge. Steel is more damping and you can generally fill it with sand or concrete to increase those properties. Cast iron is a better material again, but the downside is... It's cast. You need a foundry to cast it and the ability to post heat treat it. Lots of DIYers have been looking into mineral epoxy (literally sand/rock aggregate and encapsulating epoxy) because it's also very dimensionally stable, high compressive strength, and vibration-damping like cast iron. Some commercial outfits use it too. Basically, rigidity isn't really all that good by itself. What you're really looking for is mass and a crystalline damping structure- you get the rigidity by default through sheer scale.

Then the massive steppers, ballscrews, linear rails (could be a HGR25/SFU1605 like my intent), 2kW+ spindle... And yet, he's working on parts only a couple of inches across. There's a direct correlation between spindle power and feeds and speeds, and you need specific feeds and speeds to prevent tool rubbing, chip thinning, etc, which are themselves dependent on tool material and geometry, which is dependent on part size, material, geometry... Then you need axis drives which can push the machine weight at the required speed whilst overcoming the spindle power pushing in the opposite direction so you need specific speeds and torque (therefore power) on the axis motors and screws that can take the strain. CNC is one long chain of interconnecting calculations, and it's hard to find a ballpark figure anywhere for anything!

So your OX may be able to make the same parts, much more slowly (use HSS tooling in that instance, you'll burn up and chip carbide in friction), with a slightly lower quality surface finish (due to the spindle and steppers sending vibrations through- and possibly flat out bending- the aluminum and especially the wheels. Worth a try, all you have to lose is the odd blank (assuming you don't mount it to a vehicle!). But if you're seriously looking for a machine that can do that scale of work, it's go big or go home. Extrusion machines are primarily for cutting wood and acrylic, realistically. Like I said above, you might be able to rig a "hybrid" style machine that's reasonably capable in aluminum, but probably not great if you try to go to steels. There could be a middle ground, but it's still gonna be fairly expensive.

Obviously, if you don't need any turbine capability, then you can just do a 3-axis machine, and only machine the front side, or do both sides with fixturing and two operations. Slow, but workable. Definitely cheaper and easier to build. You'd be able to do stuff that looks like the part in the video. Once you want to start branching out and doing more organic forms, the price (and complexity of CAD/CAM) goes up enormously.

 

Tormachs are basically mini-mills with desktop PCs running LinuxCNC built in (there's no point buying a Tormach and then not getting PathPilot with it). If you don't need the support or the ecosystem, you'd be better off getting a decent benchtop mill for maybe $2500, build a $450 (including the parallel FPGA stuff) Linux box to go with it, and convert it for maybe another $1500 in motors and ballscrews. You end up substantially below the price of a 440 but get the capability of a 770 or 1100. (Unless you convert the spindle to a high-speed one as well, then it might be closer to the same price- but it's still a physically larger machine, and travel is king).

You could probably get away with a cheaper machine for this kind of work- maybe $1800- since you can utilise a 4th axis to circumvent the y-axis travel limitations. But if you ever conceive of potentially going to 5-axis, make sure you get ALL the z-axis travel. Too much isn't a word there. Otherwise (probably like me, I'm suspecting) you're gonna be looking at building a column standoff to raise the entire z-axis, and losing the ability to do small work directly on the bed in the process- vice-held minipallets will be the order of the day there.

The point itself is valid, though; if you're seriously looking at larger metalwork, converting a pre-built Grizzly or Precision Matthews import machine is gonna be FAR easier, faster and more precise than DIYing any kind of frame. I'm gonna convert my Grizzly G0758 here sometime in the next few months instead of building a compact heavy-duty machine- partially the heat-treatment of weldments thing, partially pure cost. It just makes more sense, even though I'd rather maintain access to a manual machine at all times- the solution there, of course, is to have jog wheels and run it manually fly-by-wire through the CNC controller. So it's not too big a deal.