BENCHTOP CNC SYSTEM

Here is a posting I made to a couple of newsgroups a week ago. Stephen
Barmash responded (thanks Stephen) and suggested I check out Onelist and
post my "proposal". Believe it or not, I hadn't even heard of Onelist and
this list before. I spent some time going through the archives (way to
long, ask my wife) and this seems like a much more appropriate forum than
what I found in the newsgroups. It's a fairly long posting, I apologize,
but I just got to writing...

To all benchtop CNC mill enthusiasts/dreamers:
I've just finishing designing my 3-axis CNC benchtop milling machine, and I
figured before I start making chips, I'd do a sanity check and try to get
some feedback.
I'm a mechanical engineer working as a product development consultant, and I
have long had a need for a means to produce prototype parts and molds for
various projects I work on. I always work in 3D solids (Solid Edge, mostly),
so I have 3D CAD data available for export to a CAM package. I've been
looking at the 3D CAM packages out there, and with my computer and machining
experience, I'm sure I'll be able to produce the G-code needed for
simultaneous 3-axis machining of complex parts.
I was hoping to buy a system, but I couldn't find anyone who made what I was
looking for. I've looked at Sherline, Taig, Unimat (haha), converted
mill-drills, Minitech, Light Machines, Denford, MHO Millright, Emco,
Defiance, converted Bridgeports, Techno-isel, MAXNC, and whoever else I
could find, and I could not find one system that met my requirements for a
small CNC mill.
Here is some of what I wanted:
* zero-backlash
* high rigidity (for limited steel-cutting, and accuracy)
* decent feedrates (120 ipm)
* generous travels (16" x-axis, 8" y- & z- axis)
* low friction slides and screws (for speed and power requirements)
* PC-based (Windows, pref.)
* affordable (under $5k minus PC)
* compact (forklift and crane not required...)
Here's why I ruled out what I found:
* The price limit ruled out Light Machines, Millright, and Defiance,
although they are definitely sweet machines and meet most of my other
requirements. Except for Millright, they are also shorter in travels than
what I hoped for.
* The Minitech machine is getting closer to what I wanted, but the price is
still on the steep side, and the travels are too short.
* Sherline, Taig, and both the MAXNC 10 and 15 are just too small and slow,
and I have serious questions about wear and accuracy after extended
machining. As hobbyist-oriented machines they seem wonderful, but I need a
fast, rigid, reliable machine for extended use.
* Denford and Emco have nice systems, but, again, they seem to be targeting
deeper pockets than I have, and their travels are just too short.
* Techno-isel's are nice, just not rigid enough, and the price is a little
too high anyway.
* A converted Bridgeport is too big and the price adds up (bigger motors
needed, high friction, leadscrew backlash, etc.)
* A converted mill-drill was about the closest to what I wanted, but it
would need to be retrofitted with ballscrews, and the quill feed would need
extensive work to maintain repeatability, let alone get 8" out of it. Five
inches on the quill could make do, however, since the headstock can be
adjusted up and down.
To buy or not to buy?:
Through all my browsing, searching, and asking over the past 1-2 years, I
couldn't find what I was looking for. I would have to wait, or do it myself.
So, I decided to apply my design expertise and design my own machine in
order to get what wanted. The penalty, of course, is that time is money, and
in the end the money and hours that go into this one machine will probably
exceed what it would cost to buy a more expensive machine. Even if I was
willing to spend the extra money, I didn't find anyone that produced a rigid
enough machine with the travels and compactness that I needed.
Now, many hours later, I have finished designing my machine and have a small
stack of drawings ready for oily finger prints and pencil notes. In
designing and building my own machine, as a product designer, I could not
control the compulsion to design for manufacture and assembly, and to design
around cost-effective components. It's just a habit of mine, which finds its
way into anything I design. This is usually a good quality, but the extra
design time required is not always economical for one-time projects. So, in
the back of my mind, I hope I can impress enough people with the performance
and potential cost of my machine, and that maybe others will be interested
in one of their own.
Here's what I ended up designing:
Mechanical:
* 3 independent slideway systems (linear stages) mounted in x-y-z fashion
(c-frame) like a full-size CNC mill
* each linear stage is built up independently and precisely mounted to a
low-precision cast iron or welded steel (or...) machine base via a precision
alignment fixture and polymer chocking material
* slider/saddle of X and Y stages are mounted face to face, and rail of X
stage is used as table
* variety of spindles, routers, dremel tools, engravers, etc. mountable to
z-axis plate (starting off with Sherline setup until I design a beefier
setup)
* dovetail slides with adjustable gibbs (tapered design)
* slide locks that clamp the gibbs against the dovetails, 2 per slider
* low-friction, low wear polymer bearing surfaces cast (injected) between
slider/saddle block (the shorter part) and rail/table (the longer part)
* rail/table is anodized (Type II or Type III, depending on wear test
results)
* rails and sliders are designed as potential aluminum extrusions and have
integral limit/home switch mounting channels
* each rail/slider system has a mated cross-sectional envelope of 8" X 3"
* sliders/saddles are 8" X 8"
* .631" X .200" pitch ballscrews and 2 ballnuts (preloaded against each
other) on each axis
* motor mounting plate bolted to bearing end blocks; couples motor to screw
via toothed belt. Pulleys are changeable from 2:1 ratio to 1:2 ratio
* bearing endblocks, ballscrew nut mounting block, and screw support block
are based on common design (extrusion-ready)
* each linear stage is identical with the exception of rail length and
corresponding ballscrew length
* X-axis: 16" travel x 24" table
* y-axis & z-axis: 8" travel x 16" rail
Motors/drives system:
* 315 oz-in steppers, 200 full steps/rev, geared down 2:1
* .200"/rev screw
* microstepping driver: 40V, 5A/phase, set to half-stepping
* 110 ipm at 354 lb thrust (without losses)
CNC control system (currently available):
* FlashCut CNC (max 7300 steps/sec)
* serial cable to signal generator (MPU) box
* Windows-based
* lookahead capability
* reads G-code and DXF
CNC control system (if/when it is available):
* Windows-based CNC control interface to IndexerLPT device driver
* parallel port step/direction output
* 90,000 steps/sec (allows more steps/rev for microstepping)
* unlimited lookahead
What do you think?
I'm packing alot of punch with the system I'm composing, and I believe I've
designed it as inexpensively as possible for what I want it to do. Is there
anything I've overlooked or that you would want to see in your own system of
this caliber?
How much will it cost?:
Well, I was willing to spend $5K for a complete system, minus PC, but I
think I will be spending more cash than that to build my own. The economy of
scale is working against me on a one-off. Fortunately, I have access to a
small machine shop. The mechanicals will require some tooling fixtures for
assembly, most of which I can jerry-rig with a granite surface plate and
some parallels, and there are other things that are hard to amortize over a
production run of one unit.
In any case, the breakdown of costs for purchased items (new), and estimated
costs for machined items (in medium production quantities), not including
tooling, assembly costs, etc., is:
* MACHINE - $1500 - 3 Linear stages (complete, inc. ballscrews, bearings,
etc.), machine base (welded 8" x 4" x .250" wall steel tube), and basic
headstock
* DRIVE SYSTEM - $1800 - Motors, drivers, power supplies, cables & switches
* CNC CONTROL - $1100 - software and signal generator box
* TOTAL: $4400
These costs are just about rock-bottom for a system of this type. I've been
wondering what a reasonable price would be if this system were to see
production, taking into account tooling, assembly, profit, R&D, advertising,
and all the other expenses that companies incur to produce, promote, and
sell a product. I admire Sherline's approach to offering excellent price for
the quality, while still making enough profit to stay in business with a bit
of dignity. If I were to actually sell this machine, I would definitely try
to follow their example.
Please, if anyone has already found what I'm looking for, let me know so I
don't have to "reinvent the wheel". Otherwise, if you have any comments,
suggestions, warnings, etc., they would be greatly appreciated. I'm
designing this for myself, but it would surely be nice to generate enough
interest to make it profitable to build it for others (if it hasn't already
been done) at a really good price.
If you have made it this far, thank you for your interest. If you would like
to respond personally, please email me at the address below.

Sincerely,
Carlos Guillermo
VERVE Engineering and Design
email: carlos@...
or cnk@...