Buying a milling machines

Jewelry Discussion
21

Hi Peter,

As I said in my direct reply, .005" is so imprecise that it would
cause a machine shop to junk any CNC milling center that couldn’t do
better. Normal machine centers hold .0005" or better. .0002" is
pretty typical. For monumental sculpture, nobody’d care if you were.
005" out, but that’s the thickness of a piece of paper. For small
stuff decorative stuff, it might not matter. For any kinds of parts
where two of them would interact with each other somehow, it’d pretty
much guarantee scrap parts.

The issue isn’t exactly the size of the machine, so your idea about
"hiring big people to carve big statues" doesn’t exactly relate. The
issue is how rigid the machine is, and how much it deflects under
load. The “arm” style bots you’ve been looking at have long
cantilevered spans that articulate on each other, so that a little
slop in the shoulder joint adds to the slop in the elbow, and both of
them add to the slop in the wrist. By the time we get out to the
cutting head, all that slop means the machine doesn’t exactly know
where the cutter really is. It knows how many degrees of turn it set
all its joints to, but it doesn’t have any about what
they actually did, or how much those giant arms are bending under
the cutting load. It also doesn’t know anything about how much the
cutter’s bending under the load, or how much force the cutter’s
exerting, in which direction. (which can either pull the arm forward
into the cut, or keep it from advancing as fast as the machine wants
to, depending on circumstances.) For monumental work (or welding a
car together, which is what these things were originally built for.).
005" is plenty close enough. For serious precision work, it’s not
even in the game.

Milling machines and lathes for serious work are built on a very
different plan. They’re built so that each axis of movement is
isolated from the others, so that errors in one axis don’t foul up
movements in any of the others. They’re also built much more
solidly. A CNC mill that can handle a workpiece 30x16x15 will weigh
24,000 pounds, probably two or three times what those little
arm-bots weigh. The problem with them for monumental sculpture is
that the machine needs to be big enough to swallow whatever it’s
working on. You can’t pull it up to a marble block the way those
arm-bots can.

I’ve never been entirely clear on just what you’re trying to make.
Sometimes you’re talking about carving a jade boulder, and sometimes
you’re talking about figurines. You do realize that it’s going to
take two entirely different machines to handle both jobs, right?

Rather than cruising around looking at machines, (which is neat)
you’d be better served to figure out (or explain) what kind of part
you’re trying to make, and then proceeding from there to figure out
what kind of machine you need.

Regards,
Brian.

22 
My only comment is that we do not hire big people to carve big
statues and small people to carve figurines.

I’m not a CNC expert, but I am, or was, a stone carving expert in my
way.

I gave it up because there’s no money in it, mostly Hardstone,
anyway. And Brian (Alberic) discussed some of the issues related to
robotics and stone today. AND I scanned through Peter’s video links,
too. Interesting detour this thread has taken.

There are two big issues at work in cutting stone with CNC. First,
as Brian almost said, is the number of axes. It’s easy to cut a
kitchen sink on a mill (people do it all the time, nowadays), and
some of the vids show sculpture that’s only vagely sculpture - more
like relief carving. It’s quite another thing to carve “The Pieta”
on a mill. Or even to rough it out. Brian pretty much addressed that
this morning. Rigidity VS machining freedom/multiple axes. Those
robots can be purchased used in various places, it’s writing the
drivers for your work that gets you.

A much bigger problem is, “what do you mean by stone?” All of
Peter’s videos show cutting in plaster and marble with carbide
tooling. You can do that, though chipping is a bigger issue the
smaller the cuts get. Jade or even granite is a whole different
animal. Jade must be ground, it cannot be cut with toothed tooling,
even diamond. Then you would need to program the entire machine so
it knows just how fast the wheel is cutting and when to move on to
the next section - part of “toolpath”. You can’t use a toolpath
designed for steel and carbide tooling on jade with diamond grinding
wheels, IOW. That’s not even mentioning the design of the tooling
itself, once you get into finer details. 1/8" shank diamond wheels
just aren’t going to cut it.

One solution that industry has come up with is ultrasonics.
Ultrasonic drilling has long been used to cut cameos and the like,
but there’s also ultrasonic milling, which is mostly used in
aerospace and such. The leader in that these days is SonicMill:

In my research this morning I also found some vids of ultrasonic
milling:

It hasn’t evolved to the point of carving “The Pieta” just yet, I’m
afraid.

23

Hi Brian -

The rest of your posting is excellent, helping us to bridge the
communication gap between those of us who speak a language of the
machine like C and those who speak only SEE (Standard Everyday
English). I took a BCIT course in C so I speak it a little.

Let me guess that >99% of the 11,500 Orchidians would have to
communicate in SEE with “Cyber Sculptor” as the Minister called it
for the press release of a 617K grant/loan to New Age Robotics and
Controls in Ontario. Perhaps they can help us to answer our
questions on Orchid. BTW, C is an intermediate level language which
has been the accident-free controller of our rapid transit since I
first rode it across Vancouver Expo grounds in 1986. Lives for
millions of trips have been entrusted to it. If only we had robocop
at transit stations to stop the muggings. SEE is highest level and
some robots today speak SEE albeit in limited scripts.

I am recommending 100K as a starting budget to get a Cyber Sculptor
system operating for our project and we want the best of the best
worldwide.

The best test of trustworthiness is how willing Cyber Sculptor
suppliers are to answer questions here on Orchid where thousands of
gemologists are the judges.

See Robotic Solutions, Inc. for a video on carving a bust
of Beethoven.

I've never been entirely clear on just what you're trying to make.

Read futurejade web site http://www.ganoksin.com/gnkurl/18i

What we want to make is the most amazing park and revenue-generating
tourism draw in Canada out of a block of jade which is tens of
thousands of tons in size.

IOW the park will be one huge sculpting.

It will surpass Queen E Park, Butchart Gardens etc as a tourism
attraction.

Sometimes you're talking about carving a jade boulder, and
sometimes you're talking about figurines.

Yes, the 100 ton boulder can be carved as a “Fisherman’s Throne” at
this world famous salmon and trout fishing location on the Canada
Trail (which crosses all of Canada) and the entire rock face
(thousands of tons) can be carved into the jade escarpment as a
stepped garden.

The proposed theme is First Nations, especially Sto:lo Nation, art
and culture in the carvings, first done by hand in clay and then 3D
scanned for the carving machines.

Figurines can be an earnings engine for the project until the in
situ carving is advanced enough for opening day and gate-revenues.
The film rights of the step by step work can also generate revenue. A
tele-robotic rock drilling machine operated from City Hall a few
miles away for the placement of dexpan (we will try to avoid
explosives) would be interesting Robotic driller + dexpan would be a
third robotic carving machine. First Nations people here want to get
into closed containment fish farming and a pond for local fish
species at the foot of the gardens could be an excellent experimental
station for them.

You do realize that it's going to take two entirely different
machines to handle both jobs, right?

Thank you for that opinion. Do others agree? Or is there an
intermediate sized machine which can do both figurines and in situ
reliefs and bust-sized sculptings? Will it be available in a few
years?

Rather than cruising around looking at machines, (which is neat)
you'd be better served to figure out (or explain) what kind of
part you're trying to make, and then proceeding from there to
figure out what kind of machine you need.

Parts:

(1) figurines, amulets etc. carved from suitable local stones (jade,
argillite etc) in First Nation themes for immediate sale.

(2) in situ rock face carvings like a Fisherman’s throne,
bas-reliefs on rock faces and bust-sized figures or larger carved
into outcroppings. Tests with a laser-burning machine also prove it
could be used. Burning onto argillite worked well but not onto jade
although a more powerful laser could do the job. We could probably
burn the entire Sto:lo Atlas by McHalsie et al into the rock face.
Again - all in First Nations themes because Sto:lo will inherit this
park. Thus the park also becomes a First Nations museum.

PS - The Hope area Sto:lo modestly are reluctant to use the title of
“Chief” so I will cc to him as Dr McHalsie rather than Chief
McHalsie. IMO present photos do not do justice to the Transformer
Stone carving for which he is custodian. If it were 3D scanned and
re-carved in jade or argillite like the Sphinx of Naxos at Delphi we
could all see what an excellent work of art it is.

That would in effect be a job interview to give us a demonstration
of the proficiency of the Cyber Sculptor I want to hire - the best of
the best.

24 
I am recommending 100K as a starting budget to get a Cyber
Sculptor system operating for our project and we want the best of
the best worldwide.

For what you are proposing you will need a lot more than that, 1
million would probably be a better guess begin to develop a new
machine with the requisite capabilities.

You do realize that it’s going to take two entirely different
machines to handle both jobs, right?

Thank you for that opinion. Do others agree? Or is there an
intermediate sized machine which can do both figurines and in situ
reliefs and bust-sized sculptings? Will it be available in a few
years?

As Brian mentioned rigidity is one of and possibly the prime
attribute of a precision milling machine. To cut the material you are
talking about you are going to need a big rigid machine, possibly you
will be able to do smaller work on the big machine but I don’t see
some kind of intermediate scale machine doing both and certainly the
robots you mentioned before will not work for either application they
are just nowhere near ridged enough.

James Binnion
James Binnion Metal Arts

25 
Yes, the 100 ton boulder can be carved as a "Fisherman's Throne"
at this world famous salmon and trout fishing location on the
Canada Trail (which crosses all of Canada) and the entire rock face
(thousands of tons) can be carved into the jade escarpment as a
stepped garden.

That’s quite a vision, Peter. As with all visions, you need to deal
with= reality to make it a reality. I’ll leave the CNC part to Brian
and Jim,= though I’m not ignorant either.

Jade. Not marble, not plaster, not plastic, not alabaster, but jade.
Hard stone carving, and it’s tough stuff. You can’t laser it because
it can explode. You can’t blast it because that only works with more
friable materials, like limestone. IOW, you can blast it but it will
blow up, cleave or fracture indiscriminately. Mount Rushmore is
granite, which has a granular texture to it. Blasting jade is more
like banging on solid glass. Well, it’s considerably harder than
glass, but that’s the idea. I guess you could drill-and-blast in
lieu of sawing, like in construction, but that will take months for
each hole - see below.

You can’t cut hardstone of any kind with edged tooling. That’s not
the stone, it’s the tooling. When the edges hit the material, they
break. It’s an inconvenient reality, but it’s true. So, the ONLY way
to cut such things is by sawing and grinding. Drilling in such
material is a grinding operation - core drills, edged with grit.
Rock saws are circles of metal, edged with grit, usually diamond.
That opens up a whole problem with CNC using peripheral tools
(grinding wheels) instead of end mills or shell mills, which is the
preferred tooling for the usual processes, but that’s not
insurmountable.

More importantly, grits, whether grinding or sanding, only cut as
fast as they cut. If you get a rock on a diamond wheel and push it
into the wheel, it will cut at 2mm/minute (say). If you push harder,
it will cut at 2mm/minute. If you push really, really hard, it will
cut at 2mm/minute.

Another inconvenient reality.

The top level of this pen holder: http://www.ganoksin.com/gnkurl/18w
is nephrite, about 3" x 4". It took a couple of weeks to cut and
finish, just grinding away.

So - you have a hundred ton boulder, if not more. First you need to
planit and draw it on the stone. Then you need to saw the rough.
That will take years at 3 feet a day, which isn’t unreasonable. Then
the grinding will start, which will take decades and that’s not even
polishing. It would take hundreds of years to actually sand and
polish such a thing.

This is all based on the fact that abrasives take as long as they
take, and there is NO getting around that. Sure, you can open your
park anytime you like, and have the work be in progress. I suspect
that you know your plan is ambitious. I suspect even more that you
don’t realize just how ambitious it reallly is.

26

Hi Peter,

I honestly think you’d be better served to sit down with somebody
local to you who knows both CNC mills, and arm-bot systems, and can
advise you seriously on what you need. Some of the things you’ve
said make me worry about whether or not you know the systems well
enough not to end up spending a lot of money on a system that isn’t
quite what you need. When put in the perspective of a purchase in the
hundreds of thousands of dollars, a bit of time and money spent
educating yourself is a good investment.

For example, you do know that none of the CNC systems (of any
commercial flavor that I know of) are running on C, right? Most of
the controllers run on some variant of G-Code, which is… Well, it’s
a programming language, but not a computer programming language.
It’s designed for programming machining centers, so it’s a little
different. (I keep trying to liken it to C, or basic, or assembly,
but it isn’t really like any of them. It is its own thing.) (Some of
the hacker community homebuilt controllers probably do run on C, or
some variant of it, because that’s what the engineering geeks know.)

You’ve been trying to use the peanut gallery on Orchid as a
consulting service, which is fine, we all volunteered our answers,
and (you hope) our knowledge. But you’re talking a very large
investment here, and there comes a time when you have to sit down
with real people, in front of you, and evaluate things in a
businesslike fashion. You’re getting to that point.

This isn’t a “buzz off”. You’re just getting to the point where it’s
more responsible to do research with real people who normally
consult on these things, rather than a bunch of random voices on the
internet who could be Irish Setters, and perhaps a Chihuahua or two
for all you know. (Personally, I was raised by my cat.)

Regards,
Brian Meek.

PS–> John’s entirely correct about the difficulties of machining
jade. You can’t. You can grind it (very slowly), and you can build
CNC systems that can actually sense pressure resistance, but you’re
talking years of time, and vast sums of money. Before you put the
tribe’s money into something, the responsible course is to make sure
you really know for sure that your plan will work, and to do the
homework to doublecheck. At the moment, I’m not optimistic.

27 
Most of the controllers run on some variant of G-Code, which is...
Well, it's a programming language, but not a *computer* programming
language. It's designed for programming machining centers, so it's
a little different. (I keep trying to liken it to C, or basic, or
assembly, but it isn't really like any of them. It is its own
thing.) (Some of the hacker community homebuilt controllers
probably *do* run on C, or some variant of it, because that's what
the engineering geeks know.)

Your comments like the one above are very helpful. When I saw the
videos of those machines carving exact replicas from figurine size to
small statue size in stones of hardness 3 (marble) and greater, I was
amazed. And given the acceleration of technological progress what do
we expect in a few years? Nobody I know predicted the small, powerful
desktop pc when I took punched card decks into our university’s big
mainframe. We are probably guessing on robotic carvers as well. If
small and affordable carving machines parallel the pc development
what happens to the jewelry industry? Everyone here is as capable of
forecasting as a technical expert. Didn’t Bill Gates say 64k should
be enough for anybody? IMO the “middle men” between those who create
the original artistic piece (which may be Computer Assisted Design
or a hand-crafted clay figurine or a sketch) and the marketable
product will be phased out by the automation. So I put this forward
for general discussion:

Jewellers’ activities are then limited to that first artistic stage
and the final marketing stage.

Whatever this “variant of G code” is, the feature which all 11,500
people here can relate to is that it is INSTRUCTION TO THE MACHINE.
Does everyone here know what he or she wants the machine to do? Yes.
They know it in a language of instruction to the machine which is not
binary or C or Fortran or APL or… but SEE (Standard Everyday
English). Here is how SEE code is generated: I want this machine to

A technical expert may say You can’t have that now but maybe in 50
years _____ will happen. Fair enough. That is a SEE answer to a SEE
question. But as for what a 3D scanner or CAD or CAM machine does
now, it is also fair that right out of the box the robotic jeweller
will respond to the everyday language of its owner-buyer.

28 
As with all visions, you need to deal with reality to make it a
reality...... Jade. Not marble, not plaster, not plastic, not
alabaster, but jade.

Well said. It is the purpose of this thread to discover the reality
of robotic stone carving.

Caveat emptor. What do we see in all those impressive videos?
Gypsum-alabaster perhaps and nothing >H3 of marble. Last winter I was
amazed when I shattered a brand new hickory handle on a jade stone.
Didn’t know my own strength… ha, ha. But I can cut the same stone
into blocks and slabs easily with a diamond saw as it is only H
6-6/2. Thus I would predict that the engineers can fine tune the
machines to use H10 diamond end-effectors on H7 granite etc. by
slowing the speed of the machine. Are there any robotic engineers on
Orchid?

So there’s one reality test among others. I do not spend one dollar
on a milling machine until the seller can take these slabs or blocks
and work them into figurines of bas-reliefs etc.

And thanks for the tip on jade exploding under laser. What is your
source of We tried laser on jade and argillite because
one thing we can do is burn some First Nations history right into the
rock faces. In effect laser etching is very shallow bas-relief
carving. Laser burns into argillite nicely but hardly touched the
jade. I was thinking of a more powerful laser but who wants to cause
a dangerous explosion? Another issue of reality.

29

I saw that my reply went straight to Peter - I’m putting it on
Orchid, too. I imagine there’s a general interest to this topic…

thanks for the tip on jade exploding under laser. What is your
source of

Years ago I made some stone burnishers and considered going public
with them - making production. I called a laser machinist/driller
with a big-time CO2 machine. Industrial stuff, and we had a nice
conversation. His advice was to not use laser on stone because the
stone can shatter (explode is a dramatic word) because they don’t
take heat very well in general and also the possiblility of latent
water. That doesn’t mean that it WILL, just that the risk was too
great and there would be a high percentage of failures. It’s not a
dangerous explosion except for the danger of fracturing your work.

You can’t used edged cutters of any sort on hardstone of any kind.
Meaning things like indexables, carbide or diamond. The inserts will
simply break pretty much immediately because of the shock. You can
use 20 grit or something to gain speed. And generally diamond is a
high speed abrasive. Not slower, faster.

30 
PS--> John's entirely correct about the difficulties of machining
jade. You can't. You can grind it (very slowly), and you can build
CNC systems that can actually sense pressure resistance, but
you're talking years of time, and vast sums of money.

If the black jade amulet already presented in this thread is bona
fide, then the difficulty problem has been solved although the time
problem as you note is another issue (the fact that robots can work
24/7 may offset it).

My guess is that the machine which carved the jade does not have a
pressure sensor built in but rather a manual adjustment is made to
slow it down in allowance for the hardness of the stone compared to
marble or alabaster… Jade is not as hard as granite and the QD
machine is sold as a granite carver. What are QD end-effectors made
from? I would also think that diamond end-effectors are not required
for jade with H only 6 - 6 1/2.

But again, caveat emptor. One would be foolish to spend a lot of
money on a carving-milling machine without first having a demo of its
effectiveness by the seller.

PS - I prefer the more specific “carving-milling machine” for this
thread. Gold mines like the ones I used to work in have mill
buildings with many milling machines. I worked on the crushing
machine, ball mill, initial filtering machine post-ball mill and even
the final filter/press which yielded a black mud about 12 carat just
before smelting and I had almost the entire mining hierarchy watching
every move I made. BTW, are there now robotic solutions for small
stone faceting? What does one call these machines?