Syndicate This Blog |
Vitamins and mineralsThursday, March 05. 2009In which your narrator takes a bit of a different direction vis a vis the
question of replication...
This isn't one of those stories that you that has a beginning, a middle and
and an ending. I didn't really start out to do what I'm blogging here. Actually,
I wanted to design a printable linear stepper motor. Tommilise 2.0 uses
excellent, but expensive (~US$75/axis for stepper and lead screw). They are rock
solid and accurate. They do have a bit of an Achilles' heel, though. Their lead
screw is a bit thin and flexible for serious milling.
I set out, therefore, to somehow put together a linear stepper with a
heavier lead screw. That's where I started. It quickly became apparent that this
was a pretty big job and so far I've come at the problem from several directions
which I will not recount here. That said, I've been worried for some time about
the direction that the mainstream Reprap project has taken in the past year. I
commented on that a bit earlier.
When you scrape away all the nonsense, however, my worry has been that
Darwin just isn't replicating. It's getting made with laser cutters in short
production runs. Not to put too fine a point on it, it has become a disassembled
industrial product, not unlike an Ikea sofa. Now that's fine, as far as it goes.
It gets a lot of Darwins out in the field and gives a lot more people a chance
to participate in the designing of a next generation Reprap machine that
actually CAN replicate easily.
Recently, eD at Bath University has been designing something that I think
falls somewhere between Darwin and Mendel. More recently still, Vik has been
knocking together mockups of eD's design to work the bugs out of it. Vik and I
began our respective systems using lead screws made from readily available
studding {threaded rods}. Darwin took the direction of using belts. I won't get
into the relative merits of the two approaches. That ground has been gone over
any number of times in both the blogs and the forums.
What I will discuss is that both Vik and I have been rather... perturbed is
probably far too strong a word, but lets leave it there for now, about the
dependence of mainstream Repraps on NEMA 17 and NEMA 23 stepper motors. The
bottom line is that while they are wonderful pieces of engineering, they're
rather expensive and, more importantly to both of us, they're bloody difficult
to lay hands on if you don't live in one of the G-7 {I'm not sure about Russia}
countries or China.
Now Zach, at the RRRF, has moved Heaven and Earth to get the cost of the
NEMAs on Darwin down to something reasonable. US$25 for a NEMA 23 is a brilliant
deal ... if you're in the US and the RRRF has stock. If you're not, you find
yourself twisting in the breeze. If you are serious distances from the US the
shipping cost of NEMA 23's gets to be a real problem. Vik tells me that you can
lay hands on a NEMA 23 in New Zealand for about $100.
On the other hand, tin can steppers are pretty easy to lay hands on pretty
much anywhere, either new or salvaged. Vik can buy a pretty nice tin
can stepper new in Auckland for about US$14. I can do the same thing for about
$12.50. These are typically unipolar puppies with either a 15 degree or 7.5
degree step. With a lead screw approach that kind of stepper can get you
considerably better than 0.1 mm positioning accuracy. They are a touch slow,
however.
One of the approaches that I took to designing a linear stepper was to take
a 15 degree tin can stepper motor, tin can is a pejorative term for permanent
magnet stepper, and slap on a gear pair onto it that makes sure that the
positioning accuracy is about 0.1 mm instead of a lot less than that. I decided
to use 3/8-16 studding which is extremely common here for the lead screw. The
gear pair to get a 15 degree stepper motor to yield 0.1 mm positioning accuracy
is about 1.6:1.
Vik and I began to informally pursue a tin can stepper/lead screw solution
a few weeks ago. I snooped around on the web and scored half a dozen old Airpax
steppers on the web at a true Scots-Irish price of $2.50/stepper.
These Airpax steppers are heavy duty ones weighing a half pound each. They
are so many of them around used because they were a part of the old DEC
minicomputers twenty-odd years ago. You can get this stepper from that era new
in the box for about $5. I've found listings for them in Spain and South America
among other places. The same stepper is still made by Portescap {part of
Thompson and Danaher Motion} for about $23.50. You can get knock-offs from China
for less than half that.
This is not to make a case of this is the stepper that should be used.
There are a bunch of different ones that could do the job easily. It just
happens to be the one I chose.
Anyhow, I was busily trying to shoehorn one of these Airpaxes into a
gearbox for a linear stepper drive when it suddenly occurred to me yesterday
that I was focussing on entirely the wrong problem. What I should be focussing
on was making something very conventional that was easy to make. Vik favoured
using the tin can steppers to turn lead screws rather than turn a nut on a lead
screw. A few calculations led me to the same conclusion. The Airpax weighs about
half a pound. So does a foot and a few inches of 2/8-16 studding. It really
doesn't matter whether the studding moves or the Airpax or the studding rotates.
Probably rotating the studding takes the least energy.
With that in mind I took the gear pair that I'd milled and knocked together
a lead screw mockup not all that unlike what Vik had done with eD's
design.
 Here you can see the whole ensemble... 
Here's where things get interesting. Traditionally, while Adrian has
referred to "vitamins" as the bits that a Reprap can't print, viz, things like
steppers, IC chips and the like, we've had an informal and unspoken definition
of vitamins as being the bits you can't acquire locally. When I first joined the
Reprap project at the beginning of 2006, Adriaan kindly forwarded me a set of
printed parts for a Mark II Extruder plus a gearmotor to run it. The rest of the
Mark II I was able to acquire locally. Eventually, I wound up with Tommelise 1.0
and 2.0 from that little box of vitamins.
Now, look at the first picture of my hacked lead screw drive. The vitamins
are that gear pair.
Here you can see me milling the gear pair. It took me about 40 minutes to
do the job and would have gone a bit faster if I hadn't been fooling around.
The gears have big teeth, which means that they could be printed on a Darwin by
somebody MUCH less able than Nophead.
I'm tooling up to design a gear train that employs a cheap GM-17 gearmotor
to run a pinch extruder using the latest Bowyer configuration. The point is
that three pairs of gears for axes and a gear train set for a pinch extruder is
all that anybody would need to make a next-generation Reprap machine anywhere
assuming that they could get salvage or new tin can steppers wherever they
happened to be.
The point is that the controller board kit that I use plus these simple set
of parts which I could knock out in an afternoon can be put in a padded envelope
and shipped anywhere in the world for maybe $10. The other bits could be
acquired locally.
Now THAT's replication that will get Reprap everywhere.
Trackbacks
Trackback specific URI for this entry
No Trackbacks
Comments
Display comments as
(Linear | Threaded)
Brilliant!
Now you're talking. So $10 + 3 x 2.49 + 20 for mdf + 6 for rod + 5 for nuts/bolts ~= 50 bucks + the controller board which could be done on a breadboard for 10 more plus a programmed chip is going to be well under $100! ---
Plus a print head or dremel. ---
Finish one and lets see it print!
I'm not sure about the numbers and of course they're going to vary by where you are, but that's the general idea.
there's nothing new about a CNC mill made with stepper motors. you should return to the original goal of increasing the ratio of simple raw materials to bought'n parts.
It is easy enough to increase the ratio of simple raw materials to bought parts. There are a lot of ways to do that. The problem that Vik and I are working on right now, though, is how to diffuse the technology cost-effectively into most of the world instead of just the G-7 plus China. The laser-cut systems aren't doing that.
Anonymous above was me. No matter how many parts you can reprap if the cost is too high how many can afford to build it? Cutting costs by buying just hardware parts or scrounging for boards and bolts will go a long way to allowing the proverbial 12 year old to get one up and going. If one 12 yo has one the whole school will eventually have one or be able to access one. Every 12 yo that makes one will reduce his and his family from reliance on the rest of the world and its economy. Rant over!
"If one 12 yo has one the whole school will eventually have one or be able to access one."
That's rather what I have in mind.
Making RepRap "locovores" is a great goal, as is reducing vitamin costs. RepRaps can be thought of as a particularly useful (if tricky to keep fed and happy) domesticated animal.
Taking a page from Guns, Germs and Steel, we can increase the impact of RepRap on society by broadening its "habitable range," which this kind of innovation does really handily.
Once again Forrest, you prove yourself a powerful force for RepRap's market penetration.
Barrier of entry also applies here in the G7, though more along the line of tools required and time invested to getting an entry level solution in place.
I think the current dawin design is too fiddly.
If a person could wander down to his local hardware store and say I want a 3x3 foot piece of plywood cut out for a 24 inch sink as a staring point for mounting the X & Y moving stages. Oh and yes I have a use for the cut out (Z platform) and I need four small table legs, etc.
The rest could almost be done with a saber saw, the starter kit goal should be a weekend or two project.
Most defiantly something the 12 year old could hack together as a first generation in wood shop.
Let's give a better alternative to a McWire?
"Most defiantly something the 12 year old could hack together as a first generation in wood shop." That's always been my plan.
The gears are nice, but...
It seem to me you could reduce the vitamins needed (the gears +) buy just direct driving the threaded rod.
Better step accuracy,
Less backlash,
More torque,
Simpler and
Cheaper.
The rest can be handled in software.
Lewis
Software won't give you speed and software won't deal with axial thrust on your stepper motor.
If the threaded rod is captured at one end, there will be no axial thrust.
Speed is important but so are simplicity and cost.
Lewis
I think that wood/plywood/MDF is an excellent choice of material for structural parts. Glue, instead of screws and bolts, would further reduce the quantity of vitamins. And glue can be made with flour and water if need be. It is a material that most people can work with minimum tools and it is accessible all over the world at decent prices.
The only problem with glued joints is that you need jigs and clamps to hold the pieces in place while the glue hardens. Flour and water makes paste, not glue. You can't make paste do duty as a proper wood glue to my knowledge.
|
