What are the limits of the standard Prusa i3?

Hi All,

I'm half way though my first build (a Prusa i3 box frame), but my mind already turns to changes I may want to try for my next build (does that mean I'm already an addict?)

I see lots of posts on these forums from people wanting to use exotic designs to build the biggest, best printer, but is there any consensus on what the practical limits are on the standard Prusa i3, and its iterations (without radically changing the design), or is the i3 already optimised for print volume, speed, etc?

It would be interesting to hear from people who have successfully (or unsuccessfully) increased X,Y or Z sizes, just by scaling up the current design (increasing rod lengths).

Any other limits people have investigated?

Edited 2 time(s). Last edit at 11/07/2014 05:46AM by Mr Potato Head.

I find the real challenge in scaling a printer design is maintaining stiffness and precision. The amount of deflection of most of the structural elements (essentially beams) under load increases as the square of the length. Therefore a 50% increase in printer size will result in 2.25x the amount of structural deflection. So you then make the structural members bigger, but if any of those members move, then your also increase the weight and dynamic loads, and hence, again more deflection.

Related to this is the precision of the components. For example, the build platform should be flat (ie, free of cupping or twists, etc) to within about 0.1mm across the entire surface. Meeting that spec with a 300mm build platform is much more difficult than with a 200mm build platform. Especially when you throw in a headed bed where any non-uniform heating can cause cupping or warping. You have similar issues in maintaining the straightness of your linear rails or rods.

The other limitation is in speed. From what I can tell, a typical 0.4mm nozzle tops out extruding about 10mm^3 of filament per second, independent of how hard you can push. With a 0.2mm layer thickness and 0.5mm line width, this limits your printing to about 100mm/sec. You can print faster with thinner layers, of course, but with more layers, your part will still take longer. So with a fixed volumetric print rate, actually printing something that requires a really big printer requires many hours to days of non-stop printing. Being able to print that long without some printing issue coming up (clogs, feed issues, etc, etc) is in itself a big challenge.

So keeping within the limits of the standard design, how far can the print envelope be pushed before these factors become a problem?

Lots of people must have tried this, is there any consensus on how much each axis can be extended before you need to start modifying the design?

+1 with LoboCNC

Jo Prusa increases just his i3 in Z amplitude fron 8" to 10". You've got your answer. But keep in mind i3 is a generic principle, maybe each i3 is different.

Do you realy need a bigger print volume ? If 80% of the parts you want to print exceed the standard i3 print volume, so you're right. If it's only 20% you should stay with your i3 and find alternative ways to get that parts. By split/glue for example.

With a computer you can scale a pic, a 3d model easily. Through, if it's not a vector file, definition stays fixed. In the real world it's different. A machine's engineering is not scalable or parametic. When you design a machine, you start by the result you want to get. For the i3 it was "I want a 200x200x200mm print volume with at last a 0.1mm resolution, simple and lowcost build". So, everything of the machine is calculated to reach this goal and nothing else. Increasing the Z capacity of 25% does not seems to be a big factor as far as the weight of the piece is not too high.The i3Rework version changes the 8mm rods for 10mm ones with improoved print quality, so you may appreciate that the original i3 design already reach it's limits with no print volume changes.

IMHO, if you want a bigger print volume forget the i3 and take a look to other Reprap designs. Larger Repraps goes on a cube frame or a delta frame, wich is wise. Keep in mind you will need more accurate structural and mecanical parts to get the same print resolution as size goes by. Bigger and accurate means expensive.

+1 to what's been said, especially the $$ comment. If you want big, you might consider concepts beyond a stretched i3.

If you are the type to tackle it, use 80/20 extrusions, in the larger sections, like 4040 and 8080. You can quickly construct a strong framework.
Ebay has lots of overstock sections. You can buy someones used collection of screws/nuts/corner brackets for cheap if you keep your eye out.
You might also look at recycle yards and industrial salvage yards for 80/20 frames and scrapped assemblies. Usually sold at aluminum prices.
Factories are always scrapping their 80/20 in assemblies.

The axis can be moved easily, just like the smaller machines, using long belts for X/Y and screws for Z. Many ideas for frame and motion.
I'd suggest the bed only move in Z, and move the head in X/Y. you can buy Parker Daedal linear or THK or similar stages that simplify Z greatly.
Or one stepper driving a belt around 2 or 3 screws for Z. I find coordinating two steppers for Z to be a pain.

Dimensional setup is always going to take a lot of time, precision measuring (dial indicators, large calipers) etc.

I used 20mm rods to create a moving beam (think Ultimaker/Tantillus style) and even those flexed at 30 inch lengths.
When the printer is infilling and doing rapids, the weights and bearing slop become an issue. framework
I'd recommend strong cross-sectional members, with a high moment of inertia.

I eventually changed to traditional Cartesian movements and a 4040 y axis and linear bearings for X and Y and found the stiffness and accuracy I needed.

It can be done. 2nd gen

i also recommend a parametric CAD design starting point, as that's your first and cheapest prototype.
Then constructing it is just assembly of your well designed parts.

---

Dave

Praxis3D, Kickstarter campaign X-truder,

Thanks guys, interesting answers, but at the moment I have no ambitions to build a much larger printer, my question was rather "how much wiggle room does the standard design have" (ie, by just extending rod lengths)? My gut says the answer would be on the order of 10 or 20%, probably no more, but I'm certain people must have already tried this, so what are their real world results?