All-plastic 'Snappy' printer

User 'revarbat' has created something wonderful:

Snappy, an all-printable-PLA (well, except for the hot end, electronics, one (1) 686 bearing, and 2 threaded rods) snap-together printer!

As soon as I get my Delta calibrated, I know what I'm building from its first spool!

Very cool.

I wonder how big a build volume you need to be able to print the parts? Looks like maybe 150x150?

Edited 1 time(s). Last edit at 09/11/2015 05:47AM by frankvdh.

This is indeed wonderful. Every time I am about to give up on RepRap, something cool like this shows up and gives me new hope.

It is impressive that the printed plastic-on-plastic linear rails work so well (youtube video link here for convenience). I wonder if the printed parts show any steps or artifacts from where the rails are snapped together? Do the rails need sanding or filing to make them smooth after assembly? What is the quality of the printed parts? So many questions. I hope the developer will visit us on the forums...

Quote
Ezrec
and 2 threaded rods

I'm actually experimenting with a direct drive Z axis that will eliminate the threaded rods.

Quote
frankvdh
I wonder how big a build volume you need to be able to print the parts? Looks like maybe 150x150?

All parts should be printable in a 160x150x120 build volume. My main driver for choosing the given part sizes is that it should be printable by the printer itself, which has about a 190x190x180mm build volume. It should be 200x200x180mm, but I still need to work on limit switch positioning.

Quote
MattMoses
I wonder if the printed parts show any steps or artifacts from where the rails are snapped together? Do the rails need sanding or filing to make them smooth after assembly? What is the quality of the printed parts?

There IS artifacting on the prints from the rails, but it has mostly been due to the looseness of the rails. I've been able to improve print quality a LOT by tightening up the slider spacing. I've also modified the slider designs so that the edges hit each other at an angle, which smooths things out a lot.

I'll post photos when I finish printing my latest revision of the printer and print parts on it.

Would these artifacts increase or become reasonably stable generation after generation, if a Snappy printer had to print a second Snappy printer, and this one printed another one, and so on?

For reference, here's photos of some of the first items I printed on the Snappy, before I fixed a lot of the issues.

• You can see a lot of Z banding due to the vertical lifter-rod misalignment in the v0.93 printer, and from the Z axis motors being held too loosely. Direct drive Z shouldn't be prone to this, I think. Also, tightening up the motor mount and slider spacing should reduce this a lot.
• The three major horizontal defects on the support legs print were due to a poor filament feed, that has been fixed by adding the filament spool holder.
• The banding on the X and Y axes is mostly due to the extra slop in the sliders. This has been tightened up a lot since these parts were made.
• Another cause of the X and Y banding is the sliders crossing over the rail part joints. Ends of rails have been rounded to reduce this effect, and the ends of the sliders themselves have been angled to reduce this effect even more.
• Backlash issues can also be seen in the round parts, which was due to improper spacing between the pinion drive gears and the racks on the sliders. This spacing has been fixed, and the herringbone design should make backlash minimal.

I haven't yet tried sanding the rails. There may be quality to be gained from that. I mostly just lubricate the rails with mineral oil, and manually run the slider over the rails a number of times to smooth out the few flaws I usually see.

I'm still printing out parts for my latest improvements, so I'll post photos of resulting prints when I get that printer together.

Edited 1 time(s). Last edit at 09/11/2015 08:24PM by RevarBat.

Quote
cristian
Would these artifacts increase or become reasonably stable generation after generation, if a Snappy printer had to print a second Snappy printer, and this one printed another one, and so on?

Unfortunately, I would expect generational degradation.

Very neat! That's an appealing design - congrats.

I've released Snappy v1.1b, which re-instates the lifter rods, as the direct Z drive just did not work out.

You can get the latest files from the repo at https://github.com/revarbat/snappy-reprap, under the release name "release_1.1b".

This is a known working release.

Quote
ReverBat
the direct Z drive just did not work out

Can you elaborate? What were the problems?

For Z motion, do you think a printed lead screw like this one or this one, or a printed Sarrus linkage like this one or this one would work?

The direct drive Z failed in that it just didn't have enough torque to lift the extruder platform against gravity and slider friction. Also its layer resolution was sub-par.
It COULD be geared down to work, but that introduces more slop into the system.

Sarrus linkages are fascinating, but very complicated and hard to make pin-less without being very wobbly.

I do find the printed lead screws interesting. Would have to do some tricks to be able to print it in a way that can be assembled from shorter sections, though. I fear the layer ridges would add a lot of friction.

Edited 1 time(s). Last edit at 09/28/2015 04:39PM by RevarBat.

As promised, here's some photos of the parts from the 1.1c revision of the Snappy. It's a lot better, but there's still some artifacts in the printing. They are so extremely consistent, though, that I have to wonder if it's viable to make software compensate for it in the G-code.

Looking much better than some Chinese Deltas I've used!

You can now find the assembly instructions for building a Snappy RepRap at:

[github.com]

Those instructions are beautiful. Did you generate all the pictures manually in OpenSCAD?

Wow, yes, great work on those instructions. That must have taken a lot of work!

Yes, all the assembly instructions images are generated by OpenSCAD.
Yeah, it took me a while to do that, but the benefit is, I can make changes to parts and the assembly instructions are updated with a simple 'make instructions'. (And a half hour wait.)

Quote
RevarBat
the benefit is, I can make changes to parts and the assembly instructions are updated with a simple 'make instructions'.

Now that is a really neat idea.

I like this design, would be fun to try to build one!

Would it be a good idea to make a small "calibration test", that is just a couple of sample assembly attachments? It'd be nice to have a fairly small print to test printer/plastic calibration, to ensure that segments fit tightly together before starting huge prints. This could also be useful to make sure parts will fit well when changing filament or to make sure parts from several printers fit together. Or is this what the "slop_calibrator_parts.stl" is for (which I just noticed while poking around)? If so an even smaller test-part might be nice.

Just a note, I tried to build this on Ubuntu 14.04, and the bundled openscad is too old. I added the openscad PPA and upgraded openscad like this, and then "make" worked fine.

sudo add-apt-repository ppapenscad/releases
sudo apt-get update
sudo apt-get install openscad

The slop calibrator part is indeed what you are looking for. I could make it smaller, but I want the calibration text on the side of the part to stay readable.

I'm adding a note about updating the Ubuntu packages to the Make.md and README.md files.

I know that you have developed this with 3D printing in mind, but I wonder...

How possible would it be to adapt this to injection molding?

If you can do an injection molded variant, I think this would be a fantastic kickstarter project.

I don't actually know what all constraints injection molding adds. I did a quick google on it, though, and as best as I can tell, this design would need to have thinner walls, and the extruder platform may be problematic.

Have you considered the CoreXZ mechanism? That would allow you to get rid of both threaded rods and one motor. A substantial savings.

[www.youtube.com]
[forums.reprap.org]

I am in the process of converting my prusa mendel to coreXZ (via [www.thingiverse.com]).

Moving either the X or Y motion up onto the gantry would be problematic for introducing a lot of wobble into the printing process, as plastic isn't as rigid as metal rods.
I have been looking at ways to use string to remove one of the Z motors, though.

Although the motion goes on the gantry, with CoreXZ both motors are grounded at the bottom and do not move So not a lot of mass will be added.

And printed pullyes and spools are much easier to make perfect than gears in terms of artifacts they introduce, you just attach it to your stepper or drill at full speed and hold sandpaper to the side and it will become perfectly round pretty quickly. So any imperfections are less likely to make it to the next generation.

And some plastic may not be as rigid at metal, but I bet you can make it a heck of a lot more rigid than the 8mm threaded rod that was good enough for a few generations of mendels.
You can print plastic parts to be much stiffer for the same mass (and print time) by making them hollow with minimal infill. You only sacrifice volume. I think the only weak point when it comes to rigidity will be the snap joints, but that can be solved i think
if you print in abs, throw some acetone in there before snapping together and you effectively have a single piece of rigid plastic.
Otherwise, you just need some sort of tensioning mechanism, I am confident something can be come up with, at worst, a single screw isn't the end of the world. (I have not actually printed your snap joints yet so don't know how much play there is)

Spectra is also very slippery, i don't even think you even need pullys as a first pass, sliding over PLA may just work. Pullys can then be printed at ones leisure.

Edited 1 time(s). Last edit at 11/05/2015 12:08AM by John Meacham.

Quote
RevarBat
Moving either the X or Y motion up onto the gantry would be problematic for introducing a lot of wobble into the printing process, as plastic isn't as rigid as metal rods.
I have been looking at ways to use string to remove one of the Z motors, though.

On the other hand, having the bed move with CoreXY kinematics would be highly useful, as the XY rigidity is built right in to the kinematics.

Basically take the very first CoreXY, the Bot of the Cloth and put a lightweight print bed on ball rollers onto it in place of the pen holder. Then put your gantry over that.

I've got a Printrbot Play with a 101.6mm x 203.2mm x 127mm build volume, and I'm quite taken with the Snappy. I know the requirements from the github BOM are 150mm x 135mm x 110mm, but I'm wondering if there's anyway to scale this down a bit? That, or "cut" some of the parts in half so they snap together?

I have no idea about these printers when i listen plastic snappy can anyone tell me what is this

Quote
Golkar1
I have no idea about these printers when i listen plastic snappy can anyone tell me what is this

Snappy is a printer that composed of 73% printed parts. That's the most printable printer I've heard of yet.

Meaning, by some measure (volume? weight? part count?) there are only 27% of the printer that is NOT 3D printed.
This makes the printer

1. Easier to source
2. Cheaper
3. Easier to assemble

Since it's designed to be printed in PLA, it's accessible to even those enthusiasts who don't have a heated bed.

have you tried to incraese the buid size yet