This is an extremely expensive, low volume product. I would not be surprised if manufacturing uses the same process as prototyping.

I bought one in ~2014 - when I picked it up at the Apple store, the representative told me it was the first time he'd ever seen someone buy one (apparently he'd worked there for several years).

I exclusively use Linux distros on it now - hopefully they make it easy to do that on the new machines.

While I don’t doubt that the MacPro was a low volume machine, I doubt that most people buying one were buying it from a physical Apple Store where customization was limited.

It's impossible to install Linux to the preinstalled SSD on any recent Mac, though on this one you can at least add your own additional storage and it probably won't be too hard to get Linux booting off that. Aside from the SSD the hardware seems to be pretty normal stuff that already has Linux drivers.

Though I don't see why anyone would want to pay the rather large price premium of this machine if they're not running MacOS on it.

Looks more like a fun weekend project[1] than an impossible task. At least for my mid 2017 MBP. I’ve always had issues with WiFi cards doing a hackintoshes or Linux on macs.

[1]https://gist.github.com/roadrunner2/1289542a748d9a104e7baec6...

The T2 chip that makes installing Linux to the internal SSD impossible was introduced in the 2018 models.

May I ask why do you buy Mac of you're working exclusively on Linux?

It was for work at the time, they had iPhone apps in their portfolio

You could do this a variety of ways. Ball mill would work fine.

If you were doing this in "real" volume, you would likely likely just cast aluminum into this shape.

However, companies like apple/amazon/etc have a tendency to just find a supplier who can sell them semi-specialized machines and buy a lot of them.

Apple has bought out (either the machines or capacity) for things like laser drilling.

Amazon echo/echo show's were made by buying a ton of semi-custom machines from these folks: https://mdaprecision.com/

(They talk about it in various places)

You'd want a ball mill. Keeping it in good condition and sharp might be an issue, though. (Although not nearly as big a problem if the case is aluminium.

But the holes are spherical, not cylindrical. How could you do that with an end mill the size of the hole?

With a ball end mill, theoretically.

Not that it's a good idea - you need to be able to actually evacuate material from the surface you're cutting, and milling is different than drilling in this respect. But I imagine you could come up with a very fast two step program using a cylindrical end mill for plunging/roughing and a relatively large ball end mill for finishing.

My understanding is that using a ball end mill to "drill out" a hemisphere while perpendicular to your surface will create very high forces that are bad for the tool. I'd imagine you could offset the tool axis and then achieve the same effect, but using a smaller tool might end up being easier.

Please note that I am far from even a beginner machinist.

Right, that's the idea I was trying to express. Don't expect anything good to happen if you try to plunge a ball end mill straight down into the material.

The thought was only that a large-radius ball would be able to approximate the sphere to within tolerance with larger stepover than a smaller tool. After bulk material removal with a more suitable cylindrical tool of course.

With a little bit of googling I discovered a different class of end mill called a circle segment tool, which could potentially make short work of the problem in a 5-axis machine:

https://www.moldmakingtechnology.com/videos/circle-segment-t...

Would love to see how Apple does it, but it's never happening.

I'm not really a machinist either, but let's not allow that to stop us from incorrecting each other on HN!

The stock material is aluminium alloy, it’s not that hard on tooling.

"milling is different than drilling in this respect"

Is it?

With drilling you need to evacuate material also.

Or am I misunderstanding something, I am far from an expert in such things.

With drilling, the fluted drill bit evacuates material for you. When milling, you don't have the same luxury.

Could you perhaps have an end mill with catch-holes on the insides between the teeth(?), leading to a central hollow in the end mill, on which a vacuum is pulled to pull the material through and out, allowing the end mill to just bore continuously? (Sort of a very tiny solid-state version of a tunnel-boring machine, if you can picture that.)

That’s not feasible for a few reasons. The size difference between a cutting tool and the chip it produces is much smaller than the difference between a digging machine and rocks/gravel. The relative strength is also an issue, strings of steel or aluminum aren’t going to be sucked into a hole as readily as loose material. Lastly the geometry of a cutting head would be tricky if you need the chip to go towards the center instead of being thrown out. I guess it could work if you were milling something that turns into a fine material, like graphite. (Now THAT was a messy day in the shop)

I can see the problem with wide chip, yeah.

Specifically re: strings, though, I’ve always wondered whether—as a separate problem—it’d be possible to have an end mill or drill that dices long stringy chip into shards. Seems like it’d make operation/maintenance easier in some respects. Maybe a bit with two parts, where the tip (either end mill or drill) has counter rotation against teeth running down the rest of the bit, such that the chip gets sheared as it encounters the interface between the two parts?

Also not an expert here, but doesn't an end mill have flutes too?

It does, but the ability to evacuate material straight out of a hole is very limited. Some are designed for side milling only, others can do plunge cuts but it's mainly done to expose a new "shelf" so that the bulk of the material removal can be done using the side of the tool.

Right ok, thanks :)

Ball end mill.