Best 3d printer, Best Price

So I want to build a reprap, not an ultimaker or makerbot because while they are helping to raise awareness, the reprap better embodies the spirit of 3d printing, a platform designed for everybody and to make it in everybody's hands by one method: crowdsourcing. No one profits, but everyone profits.

I'm under the impression that the prusa is the best newest, most user friendly, cheapest, and best printing plan for the reprap out there yet. But I have also been observing the ultimaker's speed and definition.

How does the prusa compare, and have ther been any hacks yet to apply the ultimaker speed and definition to the reprap?

That's the spirit! Welcome!

In my personal opinion, a self sourced Prusa with lm8uu bearings is the best quality/price option for any 3D printer as of right now. It doesn't score as high on the beginner-user-friendly charts, but it more than makes up for it in bleeding-edge-stuff and versatility. And when you learn to use it, which other printer can beat "free" spare parts? The biggest problem - which is also the biggest strength - is nothing is standard. Once they have built one printer, most people start tweaking parts of their own, and over time very few - if any - RepRaps share the exact same parts.

The ultimaker does awesome prints, and one could get into the whole bowden/speed/nozzle size/oozing discussion, but I wont. I'd say this though: on any diy 3D printer, the quality is depending on the operator. I don't want to mock anybody, and my first prints on my sells mendel were utter terrible, but here's an example on how prints on most (if not all) printers start out: [ultimaker.soup.io] Sure, a "locked" design is easier to troubleshoot, and theoretically you will achieve top quality faster on such a design if you're just starting out. But that is often due to several other people having the exact same setup. See where I'm going with this regarding RepRap? You need to be willing to learn, often through trial and error and a LOT of googling.

On that note, after only a couple of days, and with very minimal calibration, my non-standard-sized Prusa printed this: [www.thingiverse.com] By no means a stunning print at all, but it prints better than my year old sells! Some is due to the experience I've gained, sure, but I attest it to mechanical superiority!

---

--
-Nudel
Blog with RepRap Comic

Awww I see. Good thing I love to learn. However it does seem that an ultimaker print I saw didn't even have visible layers, it was of a yoda head on their website I think. I'm sure it probably has a lot to do with the camera angle, distance, and lighting, but it still makes me wonder.

But as to the speed, I love that hack because it takes an organic approach to building; thus it is most efficient with your resources. From what I saw in the video, the interior of prints were not solid, but consisted of a pattern that kept the prints integrity while using less plastic, kind of like how our bones look in the interior (a mesh of calcium organized in such a way to resist forces in all directions).

If anyone has figured out this hack for the reprap, I would love for you to post it.
[blog.ultimaker.com]

That Yoda head is the result of some members of the Ultimaker community pushing to find the limits of what can be done. They are using very thin layers. When I first saw those, I did some experimenting with that myself on my Prusa. I didn't get nearly as far, but it came down to patience, of which I had little. I wanted to get back to another project, which is more structural than cosmetic at this point. There's nothing inherently more precise about the Ultimaker. Those same techniques could be used on a Prusa. I may go back to super-fine printing at a later date, just to see what I can learn to coax out of my printer.

The big thing is that the Ultimaker folks who are pushing the limits of print quality are using Netfabb to generate geode rather than Skeinforge, and it's much more advanced. I've always had trouble getting Skeinforge to behave properly with layers smaller than 0.15mm. I see no reason why Netfabb couldn't generate geode for a Prusa that would result in prints as good or nearly as good. In fact, I've just purchased Netfabb and I intend to find this out for myself.

kyleeamonahern Wrote:
-------------------------------------------------------
> If anyone has figured out this hack for the
> reprap, I would love for you to post it.
> [blog.ultimaker.com]


This is a standard Skeinforge feature; just reduce the infill ratio to below 1. Most of us run with Line infill at about 0.2 to 0.35 solidity ratio.

All of the printers out there support infill where the interior isn't solid. If you're using skeinforge to generate your gcode, you can have your infill be crosshatched lines, boxes or hexagonal. Hexagonal infill makes amazingly strong lightweight parts.

The infill method is nothing special, RepRap (and most 3D printers?) have been doing that pretty much from the beginning. I personally prefer 15%-30% hexagon-infill and three shells on my prints. The bowden cable has also been use for a good while, with various - mostly not very good - results.

I hear the ultimaker nozzle is 0.4mm, and that's not anything spectacular. [www.flickr.com] The high quality is obtained by stretching the filament a lot and running a bunch of very very thin layers. From what I understand this is a trick done with netfabb. Now, more layers = more time, a lot more time in fact, so they make up for it by going insanely fast and very hot (and at the same time they bypass some of the issues with the bowden cable).

Theoretically you could use any printer to do that, but since most printers have a much larger mass to move around, they'd have to do it slower, and it'd take ages to finish even a small print.

This method is awesome and creates stunning prints, though it does have its drawbacks. The lack of control over ooze with a bowden, means it'll string more when traveling over spaces. For all I know you might be able to brush the strings off, and might not be an issue at all. But since so many people seem to make their pride in "straight-off-printer"-shots, and a lot of people react badly to easy-to-remove plastic string (I've had that discussion a fair amount of times) it's only natural people print what shows off the printer the best. I think this is why most ultimaker test prints are either single-walled awesome structures, and/or objects with few to no bridges in them. I haven't looked very much into this though, so please add pictures, links and more information.

Point is, if you count the weight of the plastic laid down, a (well tuned and built, I must say) prusa with 0.5mm nozzle can be faster than an ultimachine running at super-low layer height. For mechanical parts, that kind of speed and a good enough quality is what matters, now how fast the head travels along in itself. The limiting factor is often the hot end not being able to melt the plastic fast enough, and not the mechanical construction. You can overcome this by running the hot end very, very hot, but you will have a lot of ooze...

So can we get quality and speed at the same time? Sure, one way of doing it is running dual head extruders with one tiny nozzle (0.25mm?) for detailed perimeters, and a bigger nozzle (0.6-0.8mm?) for the infill. The infill nozzle could be on a bowden cable to keep the weight down, as oozing inside the model wouldn't matter much.

I would also love to see a bowden-type extruder with a lightweight motor close to the nozzle to support the retraction of the main feeding stepper. It could add that extra little fine control that's needed.

This discussion, as many others, boils down to, "what are you going to use your printer for?". If you want to print the type of objects ultimaker excels at, and don't want to tinker and build or care much for the RepRap idea, an ultimaker might very well be your best choice. Personally I find myself more interested in designing, testing and building printers, than printing (more or less useful) stuff off thingiverse. Which quite frankly stuns me, as it is not at all why I started building a RepRap in the first place.

Edit: Oh man, that did become a wall of text. And it took so long to write most things were already covered by others. Oh well..

Edited 1 time(s). Last edit at 11/03/2011 02:08PM by Nudel.

---

--
-Nudel
Blog with RepRap Comic

Quote

I've always had trouble getting Skeinforge to behave properly with layers smaller than 0.15mm.

What does it get wrong?

---

[www.hydraraptor.blogspot.com]

I'd have to agree with what everyone has been saying so far. I have had what seems to be a very similar experience with Nudel. My first RepRap was a Prusa Mendel and my first prints were, at best, globs of plastic with a vague resemblance of what they should actually be. That time, very young in my infancy of RepRap, was only a short 5 months ago. It was a standard Prusa, PLA bushings and such, and with some patience I was able to get decent prints. As I learned the ins and outs of my Prusa, mostly by breaking and upgrading it, I really got my feet wet in the idea of "hacking" and changing my Prusa for better prints. My most recent upgrades, LMU88 bearing on my x and z (y is next), I saw a major improvement. Also if you get a smaller nozzle, I'm using the Makergear hotend with a .35 mm nozzle, the layer resolutions seems to be a little bit easier to handle. I easily print at .26 mm height and my prints look pretty great... they just take forever.

So kyleeamonahern, I applaud your RepRap spirit, if you go the route of the Prusa and stumble along to great prints, the first few months you are going to be frustrated as hell, but when you get that first print that looks good... ooo baby. So its totally doable, I mean a Prusa that prints really good quality objects, just might take a little while for you to find that sweet spot of your printer.

I hope this helps. Good Luck out there

bdcannon

nophead Wrote:
-------------------------------------------------------
> I've always had trouble getting Skeinforge to
> behave properly with layers smaller than 0.15mm.
>
> What does it get wrong?


Everything becomes much more fiddly at that low layer height. In particular, getting the extruded filament wide enough at a speed I find acceptable has been a challenge. It always ends up stringy and insubstantial, even after raising the number of extra shells and increasing the WoT values to account for the lower layer height.