Re: 3D printing, powder spreading

G, B & F -
I have a particular model of the 3DP unit in question, and although it
is really simple technology, I will honor the manufacturer's request
that their trade secrets remain such. Given that, I can share a lot of
information still that one would find from open sources, such as trade
shows, discussion, and that which I have learned during my time with it.

Although MIT's patent lists compatibility with any loose powder, and
any binder, I have typically used a plaster dust for hi-res parts, and
a corn starch for lo-res or wax-cast parts.

Our binder tank is not heated (I have the monochrome version), One
important note is that the ink in the HP10 printhead is purged out
when first inserted. The first few prints on a new head are a little
darker due to "black" residues, but after that you have to look hard
at the printing to see the mostly-clear binder being deposited. It's
also important to know the technologies that Epson and HP use in their
printheads - Epson uses piezo transducers to form the droplet, while
HP uses a heating pinpoint element. The thermal characteristic is an
important one.

The manufacturer has spent considerable time fine tuning their
binder/powder relationship, which is why they will charge $1200 for 80
lbs of plaster, and $500 for a gallon of binder. Needless to say, we
won't be copying their formulae anytime soon for hobby purposes. But
we don't need to.

After spending considerable $$$ on materials, we looked at alternative
options. First, we tested their plaster. It really is good quality
molding plaster. We have replaced it on occasion (when we ran out)
with good quality artistic molding plaster and had equal results. We
were also thinking dental alginate might be a good substitute, too.
For the binder, while I by law had to post the MSDS sheets on my wall,
they weren't of too much help (and apparently trade secrets again) but
my chem background told me to use my nose first. There is nothing
super-duper reactive in the binder, else it couldn't be handled in an
office environment, and the bullet points they use to sell the
technology with would be inaccurate. The "nose" says think back to
your grade school days and remember Elmer's brown liquid glue. Yup,
natural horse glue, really thinned down. I'd expect that the
manufacturer uses a synthetic version for shelf stability, but recall
where and how this technology was originally developed - (as the
legend goes) a couple of MIT students with little $$$ and some dead
ink-jet printers, probably printing whatever fluids they had at hand
in the kitchen.

Ok, so you have your plaster, you have your horse glue, your HP
printhead, 2 bins lifted and retracted by stepper motors on single
center screws (really boring assembly and a b!tc# to perform
maintenance on), and of course your rotating snowplow.

After printing at a rate of about a half-inch per hour on a full build
area part (I typically print small things and often don't use the
entire build area, so it's quicker than that), the videos stop giving
you information. This is where the major players in RP start to have
their "performance wars". FDM says that 3DP is messy and weak. 3DP
says that FDM is hazardous (the older Dimension dissolvers, at least -
that's changed now too). LOM is an expensive waste, and SLA is just
expensive. The new PolyReacts aren't proven. Thus the 3DP camp,
doesn't advertise what you do next, too often....

When the print is done, you leave it inside the printer for about an
hour. Newer versions of the cabinet are now slightly heated (about 80
degrees) - and typically before, during and after printing. (Explains
the comment about the heated tanks). You can take a fresh (or "Green"
) part out right away. You can also wreck it easily, too. Thin walls
(1.5mm or less) are very fragile at this point, and would likely chip
off. We usually wait the hour - there is a world of difference.

After that time has passed, you "de-powder" - in two ways - first you
brush off (like an excavation dig) all the loose unused powder. It's
recycled back into the powder bin. This is where it gets messy. You
can choose to vacuum it all off, and have a difficult time reclaiming
it, or you can brush it all over the place and spend more time
cleaning your printer. (Note: powder with binder in it is NOT reusable
- thus if you have a bad or broken part, it goes into the trash.) The
second depowdering is inside the expensive "spray/vacuum booth" they
don't tell you about in the sales demo - overglorified vacuum cleaner
and airbrush. You get the majority of the powder off and leave it in
the printer. Then you move the part to the booth (cabinet, really) and
dust the rest off with the airbrush, or smaller bristle brushes. Once
again, powder without binder may be reclaimed.

Although this part could be measured, colored, and looked at, you
wouldn't want to pass it around the table. Still too fragile. Thus, we
fill the air inside with something. CA, Epoxy, wax are typical. Yup,
we even put CA into a spray gun. It's a commercially available kit
from the manufacturer. Crazy. (I don't often use CA, as it forms a
skin on the part and does not penetrate the part very much - about
1-2mm - but is good if you need it NOW.) I prefer a 2 part medium
epoxy, such as MAS (though the manufacturer sells their own version as
well), but after trying lots of finish epoxies, the MAS gave me the
best performance at a very good price. If you want to do investment
casting, you'd infiltrate with wax (and use the cornstarch power
instead. It works, you need to make a lot of patterns to get it right
the first time, though.)

In retrospect, it appears that heat is a common theme for best
performance in this technology - once the part comes out of the
printer (hour wait), I put it into a 90 degree C convection oven for
about 20 minutes. Don't let it burn, just get really hot. After
removal, I paint the epoxy on - the plaster wicks the epoxy in quite
deeply - about 12mm on a good fresh pull. I keep adding epoxy until a
shiny skin just starts to form on top - the part is "full". After it's
24hr cure, you can sand, drill tap, mill, paint, bondo, throw, drop,
mail your part now. I even used plaster models in a skateboard part we
were making. Once you get this post-processing done correctly (it is
an acquired skill), the parts are extremely rigid. Note - if you
"skin" the part with epoxy too much, you have to sand down epoxy,
which is a challenge - a slightly rough texture during infiltration is
what you aim for, and a bit of finegrain sandpaper makes for a
beautifully smooth finish afterwards.

The best part of the technology, I think isn't that it makes cheap
parts quickly - it's that it makes cheap PATTERNS quickly - yes, we
can and do use the part right out of the printer as a prototype.
Better than that, we can use the part as a pattern for making other
parts, or different parts that an FDM can't do - such as flexibles -
just like you can get rigid casting agents from companies like
Smooth-On, there is a 2-part polymer called Por-A-Mold which will
infiltrate the plaster or cornstarch part. A well designed model
(you're designing the 3d model for printing this time instead of
dimensional accuracy) with the Por-A-Mold polymer will flex - very
much like a CV boot or rubber ball - this time you aren't making a
hard part, your're providing a surface for the polymer to bind to; the
powder/binder bonds disintegrate after the cure when you squish the
finished part...

Ted.

--- In CAD_CAM_EDM_DRO@yahoogroups.com, "ballendo" <ballendo@...> wrote:
>
> Fernando,
>
> Excelent link. Worked for me. (Yahoo didn't break it up!)
>
> Thank you,
>
> Ballendo
>
> --- In CAD_CAM_EDM_DRO@yahoogroups.com, "Fernando" <fer_mayrl@>
> wrote:
> >
> > Graham,
> > maybe if you get rid of the leading zero on the number. Here is the
> > complete link, it will most likely get truncated my yahoo, but you
> > know the drill of copy pasting it:
> > http://patft.uspto.gov/netacgi/nph-Parser?
> Sect1=PTO2&Sect2=HITOFF&p=1&u=%2Fnetahtml%2FPTO%2Fsearch-
> adv.htm&r=10&f=G&l=50&d=PALL&S1=6610429&OS=6610429&RS=6610429
> >
> > --- In CAD_CAM_EDM_DRO@yahoogroups.com, "Graham Stabler" <grezmos@>
> > wrote:
> > >
> > > --- In CAD_CAM_EDM_DRO@yahoogroups.com, "Fernando" <fer_mayrl@>
> wrote:
> > > >
> > > > What kind of binder are they using on the plaster that hardens
> so
> > > quickly?
> > > > Regards
> > > > Fernando
> > > >
> > >
> > > Web searches suggest that sugar water mixes are used with
> plaster, I
> > > think the tanks are slightly heated and also read that the model
> > > should ideally be left in situ for a while before disturbing if
> > possible.
> > >
> > > I couldn't find that patent, are you sure that was the full
> number?
> > >
> > > Graham
> > >
> >
>