Makerbot 107

Preface: I'm a pretty hard-core builder and hacker, and I'm very comfortable
with electro-mechanical devices, even when the instructions don't match what
I have in my hands. I suspect that most people on this list are similar,
but this was a build that required a lot of care. I took it slow, and
made sure at each step that the mechanics appeared to be very smooth and
functional. On the Heathkit scale, the makerbot is a color TV.

My makerbot (number 106) pretty much just worked.

Prebuild:

After sanding them with 180 grit sandpaper, I painted all the wood
parts a nice gloss white. This was done with two coats of auto body
primer, a light sanding with 220 grit sandpaper, and two coats of
gloss white. All paint was Rustoleum spray cans - 3 cans of each
color. Partially due to weather delays, the painting took longer
than the rest of the build.

During the build I had to use a hobby knife to shave the paint out of the
mating tabs and any holes where things were assembled, including bearing
holes, the holes for the stage sliding guides, and the holes for the X
and Y stage rods. The Rustoleum paint stays soft for a long time, so this
wasn't too tough, and probably easier than trying to mask the insides
of the holes.

The threaded rods I received had some machining chips still stuck to the
threads, so I cleansed them off with a spray solvent (Carb and choke cleaner)
and then gave them a coat of WD-40.

During the build:

I found that the secret to getting the X and Y slides very smooth is to open
the holes in the stages that hold the slides until they can wiggle and align
themselves. Fastening them with hot melt worked great. This is well
documented in the instructions.

My experience with other sliding mechanisms is that using oil (3-in-one is
recommended for the makerbot) eventually attracts grime and dust, gumming
things up. This can be cleaned, but I decided to start out using some
food-grade spray silicone lube - the can says "Sprayon(tm) lubricants
S00210 Food GRade Silicone Lube". A warning - keep this stuff away from
any wood you ever intend to finish, it will prevent you from getting a
good finish. So far, this is working great, with about 6 hours of total
print time on the first spraying. I did not sand or steel wool the rods.

I leveled the Z platform by using a piece of rod I happened to have around,
adjusting the nuts until the rod just fit between the platform and the
top of the makerbot case in each of the four corners.

I was short a few long screws to build the extruder, but the makerbot
folks sent some out by first class mail, and it didn't hold up my build
at all.

I overtightened the screws holding the motor to the extruder, and there
are some star-shaped cracks in the acrylic at the screw holes. It doesn't
appear to affect functionality.

I used Loctite removable threadlocker 242 on almost every screw. I put
this stuff right up there with Kapton tape, it should be within reach
on the work bench. It only takes a tiny drop to do the job, so a bottle
lasts forever. I hope this will keep things from vibrating loose.

Postbuild Checkout:

Some stages were reversed, and I fixed that by reversing the motor
connector. I overlooked the fact that you also have to reverse the
endstop cables, but figured that out pretty quick. I installed the
endstops and got popsicle sticks installed for X and Y so that they
work. I'm planning to print some Z end stop clips soon.

First Bringup (Saturday):

Being a bleeding edge kind of person, I found ReplicatorG version
007, and tried it. This was Saturday, a day before Adam released it,
and I had bad results - It wanted me to install later firmware, and
I couldn't get the firmware to build. Knowing about the problems
with 006, I dropped back to 005, and it worked fairly well for me.

I had a lot of problems with print pauses and blobbing, but about
75% of the time I could get through a small print. The default
settings for skeinforge did a pretty nice job - here's a blog post
I made about that first print: http://www.antitronics.com/?p=66

Yesterday, I was pretty much useless due to a cold, which made it
a perfect day for slow, iterative troubleshooting. I suspected
that perhaps my host computer was causing latencies in replicatorG.
I'm running the latest daily development release of Ubuntu Karmic
(I'm an Ubuntu developer). I took a careful look using some
instrumentation (latencytop, ptrace, etc), and it appeared that
there were no problems at all from the O.S. side of things -
the delays appeared to be entirely internal to replicatorG.

My builds were successful enough to continue to tweak skeinforge
settings to get a better build.

The prints I started with were not filled well - I'm not sure
what the terminology is, but they were sieve-like - there were
spaces between the parallel lines of extrusion that were intended
to create a smooth surface. Also, there was sagging between the
support structure for infilled layers.

By the way, the makerbot coin on thingaverse is really a great
test pattern - it has the features that helped me tune the bot,
and is quick to print.

So, the first thing I did in skeinforge was change
Fill->Infill Solidity(ratio) from 0.2 to 0.5. I think I'll
end up changing this back, or to an intermediate value, but
having it at 0.5 sure creates a very solid print - although
I think this may have contributed to a curling problem I had
last night with one large print.

The second thing I changed was:
Carve->Extrusion Width over Thickness (ratio)
from 1.9 to 1.7, then 1.5, and ultimately to 1.4. This results
in a very well filled print - beautiful results.

I recommend changing these setting a little at a time. My
methods carry over from years of darkroom work (back when
photos required chemical processes). I make a print, label
it with a number, and make notes about the settings and print
number. Change one variable, and repeat. It's really helpful
later when you can examine several points along a process
variable, and see the outcomes.

At his point, my prints were great except for occasional
pauses with blobbing. Then I read last light on this list
that I shouldn't have the control panel open (with
temperature monitoring) while printing. I closed the
control panel and - problem solved.

I got brave and tried printing the Screwable Box:
http://www.thingiverse.com/thing:413

It came out just about perfect, except for a minor problem
that I understand (double-stick tape failure led to the
build platform not being level). I did still have one minor
blob during a 40 minute print, but I easily fixed it with
a craft knife.

Oh - I've seen discussions about build platforms - My kit
came with the new laser-etched acrylic platform from
makerbot, and it's been working well for me.

No problems so far with the extruder - I've been pulling
the plastic out at the end of each build session, I don't
know whether that makes a difference or not.

I'm looking forward to trying relicatorg 007, and especially
to designing some of my own objects - there are things in
my head, waiting for physical manifestation.

Steve Conklin
http://www.antitronics.com

Unless otherwise stated, the content of this page is licensed under GNU Free Documentation License.