It begins

Recently I have become aware of the existence of open source 3D printing technologies, most notably www.reprap.org.  A bunch of reading of that site, and other associated blogs, and I am determined to build my own printer.  It's not my intent to make a direct copy of the Reprap design, or of several commercial designs out there, but rather to build the best printer I can make for a reasonable budget.  Self replication is not one of my prime movers.  The RepRap project has helped me an enormous amount, so I do hope that by documenting my build, some folks working on that project might gain some insight or ideas.

I do have a lot of tools, both hardware and software to use to do this.  Over the past year and a half or so, I've put together a pretty cool 4 Axis CNC machine.

The major structural components are made from 3"x1.25" 6061 bar stock aluminum.  At my last job there was about half a pallet of cut off ends of 20' sticks, maybe 22" long.  The fact that this material was free to me was a major factor in the construction of this machine.  Most of the work done building this machine was done with a HAAS TM1, which is a small toolroom/prototyping CNC mill.  Each axis has two frame rails, which are connected by cross bars.  These cross bars are doweled into the rails, and bolted.  The sections were then machined flat to accept the recirculating ball linear guides.  These are a relatively cheap variety, but almost certainly overkill given the envelope, overall speed of the machine, and spindle horsepower.  They do provide a very low friction, as well as slop free motion.

To get things moving ball screws were chosen.  Two nuts on each screw with Bellville springs between them take up any backlash between the screw and nuts, and dual angular contact bearings at the driven end eliminates it from that area as well.  An old cast iron machine table was "recovered" from a surplus place, and I've reused the head from my old Sherline mill.  It's not an ideal solution, one of the current projects involves replacing it with a Sherline "industrial" cartridge spindle, and reworking the head of the mill a bit.

Driving the system are a set of NEMA 23 sized steppers driven by Gecko drives.  I really like these drives, as they have a good feature set at a relatively reasonable price.  Good service and support as well.  Control of the system is a bit of a compromise.  Due the the vagaries of legacy hardware support,  coupled with newer tech such as power management systems, modern parallel port implementations can frequently be lacking.  Hours of testing revealed problems at higher step rates that I needed to get the precision/speed I was looking for.  There are relatively niche motion control boards out there, but they are pretty expensive.  After testing several options including EMC2, I found the most cost effective way to use the computing hardware I had at hand was to pick up a Smoothstepper USB motion control board.  It does force me to use the Windows program Mach3 for control, along with several other minor limitations.  Although annoying, living with them is a small price to pay for a machine that cuts parts accurately and without missed steps.  A friend had a trunk full of surplus 8020 so I built a stand for the machine.

After the big pieces were squared away I converted the Sherline rotary table to CNC by building a table mount for it and it's stepper.  4 Axis programing is "interesting", either with or without CAM.  It has enabled me to make a bunch of cool parts that otherwise would have been impossible.

I'm sure anyone who is familiar with the RepRap and RepStrap is thinking, "Why doesn't he just make a thermoplastic extruder, and slap it on the side?"  The short answer is that these mechanics were optimized to move a heavy table and vise, and cut metal and plastics at moderate feeds.  At higher feed rates, the increased mass needed to handle the stresses of machining becomes a liability as you have to exert more energy to move it.

The plan is to build a thermoplastic extruder first, testing it with these mechanics.  The cartesian bot is relatively trivial to build, especially with the experience of having built the mill.  It will be optimized to move lighter loads faster, while maintaining good accuracy.

Extruder design is based heavily on nophead's work.  Here is a model of the current design, nothing cut yet, as I still have a few things to button up.

The heater block will be made of 7075 Aluminum, with a split clamping design to grab the resistor.   7075 is chosen over other grades because it machines very nicely, is very strong, and I happen to have an appropriately sized piece of stock handy.   Nozzle is threaded male and made of brass.  The intent with both the nozzle and the thermal break is to have the flat bottom of the part bottom out on the machined flat bottom of the hole.  Threads on the nozzle and thermal break are relived from the end to eliminate tapping to the bottom of the holes.  The thermal break will be made of stainless steel.  I'm using Alloy 303, as it is considered the most machinable of commonly available grades.  It will thread, drill and ream easier which is a great help when dealing with stainless.  The thermal break will be turned down for a section to lower the ability to transfer heat upwards, and the heat sink at the top will hopefully cool the top of the thermal break enough.  The black portion of the diagram is the mount/duct for a 40mm fan.  If this arrangement does not suit, I plan to investigate water cooling.  The only PTFE used is to insulate the heater block, and for wire insulation, none under mechanical stress.  There is also no adhesive used in this assembly, making rebuilding, adjusting, and modifying easier.