By now you should have noticed that Makerbot has announced their new 3D printer, th Replicator, and over the past few months has started shipping them. Further of note, the Thingomatic(TOM) is now End of Life(EOL) though Makerbot is committed to supporting existing owners.
The major announcements on the Replicator design were dual MK7 nozzles, integrated interface kit, and a removal of the build platform. Sound familiar?
The big difference nobody saw coming seems to be a move towards the z axis being on the platform now and the nozzle being on the x/y, much like the Ultimaker. In the short term this will mean a slight speed increase as the platform won’t wobble the print all around during prints. In the long run though, theres such legacy weight in the design that I don’t see capability for massive speed increase without further improvements.
We made some fun of the new design as simple a TOM turned upside down and got some laughs on the forums, but in all honesty these are all great improvements. There are shiny new linear bearings ,injection molded parts, laser cut parts, and custom machined parts–all indicating a more mature product and, in addition to the extras, is a big reason why they’ve had to increase the price to $1749 with one extruder, $1999 with dual extrusion.
I got a chance to see one while in China at the Maker Carnival Beijing Seeed Studios booth and it was printing smooth as you’d expect. The print quality wasn’t particularly superior to a TOM with MK7 nozzle, which isn’t really surprising. I think we won’t know how much better the machine is until we assemble, or rather disassemble one as they come assembled now. I think my hope, and Makerbot’s, is that these machines will need less maintenance, and when they do it won’t mean massive disassembly.
Makerbot announced on their twitter account this past tuesday that there was a ‘big announcement’ coming at CES next week. It floated the hash tag #makerbotpredictions to spur discussion of what that might be during the leadup. Some on Twitter have postulated a whole new printer while others the announcement of a new 3d scanner. For my part I don’t think Makerbot is going to diversify into something like 3d scanning at the moment. The VC money they received means they will be focusing intently on scaling production and getting more Makerbots in more people’s hands. I also don’t think they can announce a whole new printer yet. Makerbot has probably a couple thousand existing orders that they can’t ship through January or so at least. Those customers could easily cancel orders if some really big news was announced. My prediction is another rev to the existing Makerbot design. I believe they see their main printer line as hardware in beta, constantly being upgraded. The obvious suspects will be an integrated interface kit, dual extruders, and or maybe a change to the underperforming automated build platform. Hold me to my words though, as we’ve only got a few days to wait!
This is the first in a series on a series of guest posts from Prescott Ogden (@prescottogden) discussing print defects, why they occur and what to do about them.
In the course of raftlessly printing dozens of mrule’s Polyhedron Vertices, I’ve encountered a print defect I’ve termed “flashing” after a similar defect in injection molded parts. This is a slight flaring at the base of a part, where it meets the print bed, as well as a finite reduction in the finished part, e.g. 12mm parts are 11.5mm, 50mm parts are 49.5mm.
This is normally not a big deal, but the polyhedron vertices snap together near the base, so the flashing interferes with mating between parts. It can also cause problem if you need parts with a very specific height.
As far as I can tell, this is because our Z axis offset is set too low. For good print cohesion, we make sure smear down the first few layers of plastic as hard as possible. This causes flexing in the axis, and that layer is shorter and therefore wider. On the next layer, the same thing happens, but there is less flexing, so flare is slightly less, until the flexing is constant. This gives rise to both the height reduction and the trianglar flaring at the base of the part.
The simplest solution is to adjust the starting point up by the amount of the height reduction, however, we then lose the bed adhesion benefits of smearing the first layer, and raftless printing becomes impossible. In order to get correct, or at least workable geometry while maintaining the print bed adhesion, I’ve developed the following OpenSCAD script where example.stl is the stl you’re altering:
file = "example.stl";
height_reduction = .8;
chamfer_radius = 0.6;
chamfer_height = 2.1;
epsilon = 0.001;
envelope_size = 1000;
//Chamfer
difference() {
//Height Boost
union(){
translate([0, 0, height_reduction])
import(file);
linear_extrude( height =height_reduction, center = false) projection(cut = true) import(file);
}
minkowski() {
difference() {
translate([0, 0, epsilon/2])
cube(size=[envelope_size, envelope_size, epsilon], center = true);
linear_extrude( height = 2*epsilon, center = true) projection(cut = true) import(file);
}
cylinder(r1=chamfer_radius, h =chamfer_height, r2 = 0, center = false, $fn=10);
}
}
This code simply pads the bottom of the part, so the overall height will be correct, then chamfers it using a minkowski sum, so that when flashing occurs, it brings the part out to the desired perimeter, rather than beyond it.
This hack is not ideal, it still distorts any features close to the bottom of the part. A better solution might be to modify or configure skeinforge to adjust the z-height and width over thickness on a per-layer basis, or more simply to put and extra height_reduction between the first and second layers. Despite it’s shortcomings, the chamfer approach solved my mating problems with the polyhedron vertices, and it’s a good first step.
We’ve been printing for 2 years now at HeatSync and I thought I’d put together some of my top tips for new printer owners. Some of them are hard won, others are from other bloggers in the Rep Rap/Makerbot community over the years so as such I’ll take no credit for any of these innovations.
We’re doing a ton of 3d printing at the lab right now with Moheeb’s Thingomatic and our Fakerbot. Moreso, Prescott and I are talking about a custom MK7 extruder for the Fakerbot, Moheeb and I are talking about reviving the Mendel with new Prusa parts, and Raymond and Corey are silently slaving away on their custom Ultimaker builds. Stop in!
[video]
I’ve personally moved away from the idea of the Rep Rap for now. I'ts a beautiful goal, but I’d rather spend the next 5 years actually 3d printing rather than figuring out how to replace all the precision parts and processes with cheap and ubiquitous not-terrible ones. Precision in precision out…
A sexy underdog right now is the Ultimaker which has x and y movement on the print head and the z on the platform which, combined with some good firmware, are allowing them to do much larger and faster prints. It’s got some of the community of Rep Rap and Makerbot, but still the precision of laser cut and machined parts.
Or we could move towards some of the more commercial machines like the impressive, but delicate, Z-corp that made our Theo Jansen Walker. Those retail around 20k and the binder and powder are costly as well. You can’t deny the output though…
Finally there’s the obvious answer, just stick with the Makerbot guys. My philosophy has always been to stay where the public’s attention is, even if there are shortcomings. They won’t be behind for long given the amount of engineers they have working on the problem, both within and without. And, the only thing keeping the Fakerbot from being a thing-o-matic is the mk7 extruder and the larger build size.
What say you guys? Where do we put our money next?
@VirgilVox’s new Thing-o-Matic that arrived this weekend! It was a pretty easy build for him, except for the usual firmware programming issues (for the motherboard I just had a better timing touch, for the extruder we re-burned the bootloader and still couldn’t upload until we finally we disconnected the molex and not-ethernet and it took… shrug)
The real annoyance was the mk7 build. Its their most recent creation, and the build instructions belay that. Theres 4 different positions you can assemble the bar mount and its incredibly difficult to understand which they actually want you to use. I built it one way and had to redo it later to make it fit the mounting bracket. Then @virgilvox reported having to rebuild it again after I left. Then the filament wouldn’t go through and we had to rebuild it again. Finally @VirgilVox wanted to argue about which side the filament pusher was supposed to be on so he rebuilt it. Not a great experience.
After that we just had to tighten down the x and y belts and it was printing gorgeous ‘out of the box’. I have to say it was a definitively advanced experience compared to the Cupcake days.
The real star of the show is certainly the new Mk7,though. Check out the resolution and z transition seam of the Mk7 and 1.75mm filament on the left and our custom Mk5 with 3mm filament on the right.
When we ordered our Gen4 electronics from Makerbot they came with 3 endstops, though the cupcake wasn’t terribly suited for them out of the box. I developed brackets to install 2 of the endstops but couldn’t find our third endstop. Since this IS a hackerspace I went ahead and etched a board utilizing our laser cutter and laser transfer technique for the third and installed that as well.
Fakerbot now acts almost entirely like a thing-o-matic by homing it’s x,y and z. No more heating plastic or using paper to align the z!
@intel_jim brought over a new set of 3d prints for our little ‘showcase’ of 3d printing technologies. These are from the local print house PADT and show off several prints mixing various rubbers and plastics in various configurations. My understanding is he got these by touring the facility. Is PADT our next HeatSync field trip?
They join our Theo Jansen Walker from the Z-Corp inkjet binder process, the hand from the Solido LOM process, and our Rep Rap parts made on the Dimension Stratasys.
