(Also asked on rec.nodels.engineering)

I'm thinking about how to make some ornate guitar roses, similar to this
one: http://www.maxwellplace.demon.co.uk/...onium/rose.jpg.
The original is cast brass but I'm wondering whether a CNC mill could be
used to mill/carve the basic shape in a dense hard wood so I could then
hand-finish the carving and gild it with gold leaf. Has anyone used a
CNC mill for anything similar, or are there any CNC experts that could
tell me how easy or hard a task this might be?

(Hidden agenda: this would also provide the excuse I've been looking for
to buy a small CNC mill ;-) )

Alternatively, is there a sensible way to produce a brass casting to
replicate the original?

Dave

On Jun 19, 8:43*pm, NoSpam wrote:

This would be a reasonable piece of CNC work. Take it to the usual web
forums for more on home-built CNC.

CNC is not a good approach for perfect replication of cast pieces.
Casting and patternmaking are additive, CNC is mostly subtractive. A
fine intaglio detail in the piece can be simply impossible with really
good CNC, because it's smaller than the tool diameter. For a heavily
decorative piece like this you might need to go back to a fairly
simple version of it and then re-add detail in a way that's more
amenable to milling.

You could also machine it additively, with something like a RepRap (as
if!), a Makerbot, or a credit card and a trip to Shapeways. If you
want it in metal, Shapeways and the like are a good option.

Alternatively, is there a sensible way to produce a brass casting to
replicate the original?

Brass sucks, because zinc is awkward (and unpleasant) to deal with.
You're better using a bronze or bell metal alloy, made from copper,
allowing maybe a smidgen of brass & lead solder, and 10-15% tin (old
pewter mugs). This can be much more fluid to work with. Leaded bronzes
can be handy if you want a nice clean engraving or CNC metal
afterwards.

Most of my bronze casting is done as lost foam patterns in Petrobond
sand (try MUTR for small quantities). Sometimes I use wooden patterns,
if it's a batch run. Wood or foam will both CNC quite easily for
patternmaking - usually MDF is best. This would work nicely as an MDF
flat-back pattern. I'd then consider hand engraving it to add detail.

3D scanning is getting easier, but for this I wouldn't worry too much.
A flat lattice, then engrave it, would do for starters.

On 20/06/2011 02:05, Andy Dingley wrote:
On Jun 19, 8:43 pm, wrote:

This would be a reasonable piece of CNC work. Take it to the usual web
forums for more on home-built CNC.

CNC is not a good approach for perfect replication of cast pieces.
Casting and patternmaking are additive, CNC is mostly subtractive. A
fine intaglio detail in the piece can be simply impossible with really
good CNC, because it's smaller than the tool diameter. For a heavily
decorative piece like this you might need to go back to a fairly
simple version of it and then re-add detail in a way that's more
amenable to milling.

You could also machine it additively, with something like a RepRap (as
if!), a Makerbot, or a credit card and a trip to Shapeways. If you
want it in metal, Shapeways and the like are a good option.

Alternatively, is there a sensible way to produce a brass casting to
replicate the original?

Brass sucks, because zinc is awkward (and unpleasant) to deal with.
You're better using a bronze or bell metal alloy, made from copper,
allowing maybe a smidgen of brass& lead solder, and 10-15% tin (old
pewter mugs). This can be much more fluid to work with. Leaded bronzes
can be handy if you want a nice clean engraving or CNC metal
afterwards.

Most of my bronze casting is done as lost foam patterns in Petrobond
sand (try MUTR for small quantities). Sometimes I use wooden patterns,
if it's a batch run. Wood or foam will both CNC quite easily for
patternmaking - usually MDF is best. This would work nicely as an MDF
flat-back pattern. I'd then consider hand engraving it to add detail.

3D scanning is getting easier, but for this I wouldn't worry too much.
A flat lattice, then engrave it, would do for starters.

Andy, this is a great reply with some very useful advice - but it led to
me spending the past two hours reading about 3D printing and now I want
to make a 3D printer. What have you done! :-)

I don't have any experience of casting and I think I'd get someone else
to do it for me if I took that route. It sounds like you're suggesting
that it wouldn't be possible to cast the level of detail shown in that
picture, so how did they do it in the 18th century?

Dave

On Jun 20, 1:32 pm, NoSpam wrote:

Andy, this is a great reply with some very useful advice - but it led to
me spending the past two hours reading about 3D printing and now I want
to make a 3D printer.

Don't bother - they're slow, they're crap, the finished results are
lumpy and half the machines out there are vapourware.

3D printing is great - but you do it by uploading files to Shapeways,
who then do it for you properly, out of better materials.


I don't have any experience of casting and I think I'd get someone else
to do it for me if I took that route. It sounds like you're suggesting
that it wouldn't be possible to cast the level of detail shown in that
picture, so how did they do it in the 18th century?

Practice.

I can't do lost wax casting. Keep trying, never works right. The worst
part is that I know some barefooted bloke in Memphis was knocking this
stuff out thousands of years ago, and all he had was a lump of copper
and some bees. I just can't get a good enough burn out to remove all
the wax, and I get blow-holes. I've a new 1000C kiln now, maybe that
will help.

Bronze and pewter casting is great fun - other sorts of casting
aren't, because they're either too difficult, too hazardous or (for
silver) usually too expensive.

You need a foundry to do it in. You really do need a foundry - a
reasonably large fireproof shed where you can do serious foundry stuff
without causing trouble. It needs space, it needs space to spill
molten metal and still leave room to jump. It needs to be fireproof
against gas burners and hot metal. It mustn't have a lawnmower in
there too, either petrol or a set of flammable plastic garden chairs.
You also need good ventilation. You also need a lean-to shed alongside
your foundry shed, to keep the propane cylinders in. It mustn't also
be used for woodworking, so that every corner is full of flammable
shavings. So it's a serious investment in space and clear space, used
only for dirty hot metal and nothing else

If you have such a space, go to it. Lots of amateur books around on
how to do it, lots of jewellery-scale stuff, then get some 1950s
engineering books on how to do it right (runner and riser design needs
you to read the books first).

You need a hot crucible. This is a real crucible (a tenner), a bunch
of firebricks in a welded steel cage to make a heating furnace, or
else a small gas cylinder filleted and lined with kiln insulation.
Your burner is a big gas burner, which you buy from a charming
evangelical Christian guy in the USA. They're made cheaply from
standard plumbing parts (someone's web site design) which everyone
uses, but if you can't just go down to Walmart and buy a #2 nozzle,
it's easier to buy a ready-made from the Christian chap - he puts
little tracts in with the shipping package. Bless.

Your casting is done in a bed of Petrobond sand, which you keep in a
closed box or sandpit when not in use (too expensive to do a whole
floor). Mine's in a big Quality Street tin, which is plenty big
enough. Don't mess with greensand, it's too difficult. Lost-foam
casting is an interesting process (yellow insulation foam, cut with
scalpel & sandpaper) but it needs ventilation. With the Petrobond, I
can also use re-usable patterns, often CNC-ed from MDF or multiple MDF
pieces. MDF patterns need to be shellaced, to seal their edges against
moisture and to smooth them.

Safety kit is important, but fairly cheap & simple (in comparison to
the heating rig). Every part of you needs to be covered with either
thick leather, Kevlar or polycarbonate. You also need to be as
waterproof as a roof, i.e. any splashes will fall off, not in. A
welder's apron and a welder's jacket are cheap. Rigger boots are
fairly splash proof, so long as they're leather (not vinyl) and
they're inside your trousers, not outside.

You also need some tools to handle a hot crucible. Usually a one or
two person ring carrier for carriage and pouring, and a set of long
top-grab tongs for lifting it from the furnace to the carrier. Also a
stand for the carrier, when you're loading it. This is easy stuff for
a bit of welding and some basic smithing.

Andy Dingley writes:

Andy, this is a great reply with some very useful advice - but it led to
me spending the past two hours reading about 3D printing and now I want
to make a 3D printer.

Don't bother - they're slow, they're crap, the finished results are
lumpy and half the machines out there are vapourware.

RevK's blog currently has some images of a mis-shapen pink Dalek and
burnt out stepper motor controller.

http://revk.www.me.uk/
http://revk.www.me.uk/2011/06/extrudiate.html

--
Alan J. Wylie http://www.wylie.me.uk/

On Jun 20, 11:51*am, NoSpam wrote:

Maybe the answer is to
look for a secondhand CNC mill.

You can't afford it. Unless you're phenomenally lucky (hang around
tech colleges) there's not much useful to be had S/H. Industrial kit
is too big, educational kit is too wimpy and also ridiculously priced.
The only S/H stuff worth having is that very narrow range of
apprentice training, with machinery bought specifically for
apprentices.

CNC mills are not a practical home option. With a couple of £k, three
phase, air supply, forklift access and a concrete floor, then you can
have a a nice S/H machining centre with a tool changer (now you're
talking!) Add another couple of £k for tooling.

What most home "CNC mills" are are actually high-speed routers. These
have manual tool change in a collet chuck, so there's no end
registration. Repeatable depth setting after a tool change is thus a
bit hit & miss. Still useful though. They trade low leadscrew forces
(and thus allowing cheap steppers turning leadscrews, rather than
needing expensive ballscrews and servomotors) for needing a high-speed
small-diameter cutter that will work plastics, brass and aluminium.
You also start to care about precise free-cutting alloys and distrust
chilled castings. Still fun and useful though.

For home-build, the popular option at present is a US-built Taig mill
and doing your own CNC conversion. I'm not a fan of these - they're
cheap, but you throw too much away. The electrics are the usual
hideous American crap and the 240V conversion from taig is just plain
unsafe (no earth continuity through the 110V transformer). It's also
usual to throw away the Taig spindle (either one, they're both too
slow) and to replace it with a clamp bracket and something like a
Kress router at 20,000 rpm. Stepper mounts are bootstrapped or bought
from someone on the same web forum who already has their mill up &
running. Steppers, coupling, stepper drives and a copy of Mach 3
(interprets G code and turns it into stepper movements) all cost
money, but not much these days.

For CAD, take a look at CamBam. This is mostly there to import .dxf
from AutoCAD (i.e. designs) and make G code (i.e. machining
instructions) from it, but it's also usable as a CAD tool directly.
Just don't ask me to _like_ the UI. Good web community.


IMHO, home CNC means about three machines as a reasonable level of
comfort. Something small for doing jewellery, engraving and PCBs
(maybe a Proxxon), something Taig-sized in the middle for making
"model engineering" and something with two and a half axes and a 4'x4'
flatbed for routing out MDF & plywood. If you have a fast spindle on
a small machine, you can afford to have a slower spindle on the mid-
range mill and do something resembling milling in the softer steels.
Not easy to make the leadscrew work mind you, and that typical model
engineer's vertical mill that everyone has just isn't precise enough
to convert. An MDF cutter is great fun, easy to make your own frame
for, and you can put it into a box with its own dust extract.

Toymaking in MDF gets fun with CNC, as you can make repeats of
complicated parts.
http://quercus.livejournal.com/300504.html
http://jarkman.co.uk/catalog/cnc/batwheels.htm

OTOH, these were done on a Taig machine, took an hour for each big
wheel (1/2 hour if I'd optimised the tooling and toolpaths) and made a
filthy amount of MDF dust. They'd have been much cleaner on the router
machine since.