MakerBot Print is our newest print-prepration software, which supports native CAD files and STL assemblies,
allows you to interact with all your printers via the Cloud, and many other exciting new features.

Download Now

Hey! This thing is still a Work in Progress. Files, instructions, and other stuff might change!

Square Delta

by tempo502 Mar 8, 2015
Download All Files

Thing Apps Enabled

Please Login to Comment

What is the brand filament and what is the color.
I find color very beautiful but can not find it anywhere.

Its been more than a year since the last question about this four-armed delta printer. Is it still working nicely? Has the extra arm proved beneficial enough to justify the trouble. Does that extra arm and rectangular arrangement allow for printing on a larger square surface rather than the traditional delta type circular surface? I've been thinking of ways to build 3d printers with minimum vitamin content and delta printers have potential, but I really want to see a box shaped build space.

To be honest, I haven't touched it in a long time. Nothing wrong with it... I just have too many printers.

I do like the square build plate for procurement and mount design convenience. The downside is, with three arms on a square frame like this, the "undriven" corner region of the build plate has fairly poor stability and is susceptible to calibration errors. This is very similar to the reachable space outside the frame triangle on a regular triangular Delta. When one of the arms goes close to horizontal, the non-linearities in the Delta kinematics start to get nasty. It's an issue with all Deltas but is somewhat worse in this design because the undriven corner is just so far from any of the towers. Adding the fourth arm (which I am too lazy/busy to do) would largely resolve that issue and make the printer much more consistent across the build area.

No Delta technically has a rectangular OR circular build space... the true reachable region is too complex for most people to think about, so we just inscribe a simpler shape inside it and limit printing to that smaller volume. Resembles a https://en.wikipedia.org/wiki/Reuleaux_triangle but that isn't exactly right. You can put a box build area inside a regular Delta just as easily as a circle... it just gives up a little more usable edge region. Printing near the edges of the Delta plate is a bad idea in any case, because it's harder to calibrate.

Thanks for replying. All this, for me, is in the context of a personal goal of building a 3d printer that I can get through a 24" doorway and still have a a full 16"x16" non-moving build space using a minimum of "vitamins" It seems to me that a three armed unit would have a harder time doing that than a four armed system with lifting at each corner. I'd be strongly inclined to use a kind of closed-loop drive system which might be able to avoid binding if the motion-math is right. I can definately see the need to slow down near the corners, but I don't see how it would be problematic. Anyway, thinks for your time. :-)

could you have all four towers connected to the effector if two of the towers were connected to the same axis?

No, those two arms would be fighting each other.

Just a thought: what if instead of implementing a fourth arm, you attached the extruder to the sliding block on the fourth column, and had it move to the current Z-height? Then you could shorten your bowen tube, reduce hysteresis, etc. I've seen "flying" bowen extruders before, but the way they're implemented probably doesn't reduce the effect of their weight on the other columns. But it would if it were being moved by its own motor, on its own column.

That is totally possible. It would require custom firmware though.

Awesome ... what are you using for the heated bad?

200mm square silicone heater, around 200w, RTV'd to the aluminum heat spreader. Then 0.5mm silicone heat transfer pads (non-adhesive) and a removable boro glass sheet, with 1mm PEI taped on with 3M 468-MP tape.

what you have done here is pretty rad. I wonder if you would get a benefit by adding a 5th arm.

You'd probably get diminishing returns on stability, and the frame would be a lot more complex. If you go for the full six independent arms, you've got a six-axis hexapod, and those are pretty awesome. Not much reason to go for five.

Congratulations ! I made a kossel mondrian delta (reverse) with auto-calibration and i think the square basis can be excellent to build some big columns :-) One day we will have the firmware for 4 arms and then these bots will be a lot.
Thank you for sharing. So this firmware version is for 3 arms at 90 degrees of angles ? Did you print ?

Yes, you just need to put the attached config.txt file on your Smoothieboard using a recent "edge" build of Smoothieware and it will do kinematics for the towers positioned 90 degrees apart. Specifically, the config file tells Smoothie to apply an angle offset from the standard tower positions: 15,-15,45 degrees. This moves the normal "front left" tower right a bit, "front right" tower left a bit, and shifts the "rear middle" tower to the "rear left" position.

Print pics are included as the second photo in this Thing if you look through the list.

I bought a Smoothie but don't have it in a printer.. I've been printing Kossel Mini parts.. But maybe I should switch to what you have here.
Where did you get your T-slot bars from? I need non-printable hardware..

There's a full BOM in the Thing files :-)

Any 20x20mm extrusion should work. I use Misumi HFS5 2020.

Very impressive! Congrats on your original design.