meet Lander, the multi-configuration 3d printer

[realthor]

I wanted so much to have a Wallly until i learned Nic is not working on him anymore and that the math was too complicated for the cheap electronics used in Cartesian 3d printers.

GUS looked great and simple but lacked the dimensional accuracy to print itself rep-rap style.

Then I found Morgan, the single arm SCARA bot that looked so alien while being so friendly... but it used linear rails and a lot of metal inside its center pole.

I needed something else. I could have waited for the next versions of the mentioned printers or some forks but why not try to learn something until such a time and design one that suits my needs. Understanding 3d printers will be important for engineers in the future.

This is how after many posts in the "Ideas for the next Wally" section I ended up creating a new thread for this project.

I introduce you Lander, the bot that can land on your desk, on the floor or on the wall or can print in a continuous fashion on even ground. It's a Lander, it must cope with on-grid, off-grid and other adverse conditions and must have everything on board.

This is the nomad offspring of Wally and Morgan.

The goals for this project are:

1) rep-rap to the bones
2) no specialized or hard to find parts (linear rails, linear bearings, aluminum extrusions, etc)
3) low vitamin count, preferably more of the same kind than different ones (for emergency replacements/etc)
4) no timing belts, until we can print them, fishing line or other type of strong thread will be used instead
5) large printing area
6) multiple configurations and enhanced mobility: on the wall / off the wall, folded for transportation
7) as little metal as possible because of its energy intensive extraction and processing
8) precision over speed (complicated printed parts are allowed if precision permits; this is to ensure self replicability)
9) can be easily scaled up by modifying a small number of components (grow the double parallelogram arms and lengthen the SCARA arms )

This is an initial sketch that I will use as a reference while refining it:

screenshot.200.png (85.38 KiB) Viewed 80771 times

As the solution I have chosen:

1) a single arm SCARA that functions just like Morgan, being constrained on the left or on the right of the possible printing area, so that I can reuse Morgan's firmware,

2) and a Z movement via a winch and a double parallelogram system.
- The double parallelogram is built with a Double Lamina Compliant Joint to force a controlled symmetrical movement
- The winch will be designed in such a way as to prevent the triangulation errors associated with the way the cable is spooled. In commercial winches this is not a priority while for a 3d printer this is a must. As a consequence, there will be a limited length of wire that will be possible to spool but this works well within the scope of a 3d printer (20-30cm Z).

3) there is no solution chosen yet for the other functional elements of the bot like extruders, electronics, power source, etc. These will come as I go but the intention is that this bot can eventually print with different materials in the same print, possibly through the same nozzle.


Note: bear in mind that i am learning as I go, I am not an everyday maker, my engineering skills are limited and I have to learn pretty much everything as I go. Moreover English is not my first language and a lot of this engineering lingo is making knots in my brain. I will try to learn as much as I can from the people that give me constructive feedback. Many thanks to Nic for his professional approach.

[Nicholas Seward]

I am a little curious about the horizontal continuous configuration. If you use the arm, then it isn't continuous is is? Can you clarify?

[realthor]

This is something I realized is possible with this bot after I have gone through several design iterations and it's a long shot. This is this bot's nirvana . The idea for the continuous printing is a little fuzzy even for myself. I will try to explain:

I can take the bot off the wall and use the floor as a very long Z. The double parallelogram arms could be locked in place acting like a carriage's chassis. Depending on the availability, rollers or dolly tracks (pipes, rails) can be used to move the bot backwards as it prints. I don't exactly know how long an object can be horizontally printed until it's too heavy and will fall off but solutions can be found (base structures that are clamped after they are printed, etc)

The winch we have discussed about can be scaled in such a way to be able to roll quite a length of line. Let's say 100cm. But for long items the cross section is often similarly large so the bot will stay quite a long time printing one slice. On the last slice there will be enough time to have a procedure to unwind the bolt completely and anchor it again 100cm back.

I am thinking of objects maybe 50cm-1m long but using some other materials (new ones seem to pop up every week) larger structures can be achieved, pipes can be printed in remote locations, kayaks can be printed in one piece maybe, blades for household wind turbines, etc. I haven't thought that far but I have found this on the web while searching for similar concepts to exemplify:

http://diy3dprinting.blogspot.ro/2014/12/freebot-mobile-3d-printer-concept-with.html


Please note that I've put this idea at the bottom of my sketch. It is not something I will focus on from the beginning and I don't know if it can be achieved at all with acceptable results. That is a goal set far away so achieving closer goals will seem easier

I say let's discuss this further when i have the first two configurations taken care of

Counter intuitively maybe my first configuration to achieve is the desktop mode. I have chosen that one because it is more prone to errors and I will learn a great deal about calibration and what not while trying to fix it. Bolting the SCARA arm to a wall and having the bed mobile will be easier. By the way, I understand that moving the printer will require each time a calibration. I will have at some point an automated bed leveling.

[realthor]

You could even drive a set of rollers with the wormgear to avoid the string altogether for the "on the floor" configuration. I have quickly added them on my WIP sketch for exemplification:

screenshot.201.png (65.17 KiB) Viewed 80740 times

Ok, I'm done with the dreamworld stuff, you asked

[Nicholas Seward]

Image.jpg (25.17 KiB) Viewed 80759 times

I wanted to model the whole thing but it is a bit tricky so I will leave you with this until I get more time.

Specifics on this model: Each part has an approximate outer diameter of 25mm. The worm has to be oriented with an approximate 3.5 degree angle from the perfect 90 degree orientation to meet the straight teeth at the correct angle. Every turn of the screw rolls up about 64mm of string. 1 turn of the stepper pulls in 4mm of string. That is 20um of string pulled in per full step. The pitch of the string wrap is about 10mm. For a 300mm travel you only need 5 wraps. That will make the winch fairly compact. It will be 50mm longer than the width of the static nut assembly. I would make that no wider than 50mm. So 50 mm wide with a 50mm buffer region on each side of the winch assembly for the screw to poke into.


It is most definitely printable. As drawn the string spiral aligns with the tilt of the worm so you could possibly see artifacts. It is an easy fix though, I will switch one of the spirals to a left handed spiral and then we should get smoother tooth engagement.

I can also make the spur gear section and the spool section two distinct sections. It would actually be a spur gear with a screw coming off both sides. We will have to experiment to see what works the best.

I do have some concerns about the ultimate accuracy. I will have to build a test rig to measure this.

[realthor]

Wow you are awesome! That looks nice... and printable. I will not exceed 50mm in-between nuts, in my sketch the stepper motor goes inside the case holding the nuts and that is 30-40ish mm wide.

As drawn the string spiral aligns with the tilt of the worm so you could possibly see artifacts. It is an easy fix though, I will switch one of the spirals to a left handed spiral and then we should get smoother tooth engagement.

you lost me here.
1) Why see artifacts if the string rolls parallel with the worm (matches its tilt)? how would a left handed spiral generate a smoother tooth engagement?

I can also make the spur gear section and the spool section two distinct sections. It would actually be a spur gear with a screw coming off both sides. We will have to experiment to see what works the best.

Hmm, yes, that was my first design but as long as we can print this I'd rather prefer the whole thing to be like in my last screenshot (and your render too). This would eventually allow for a longer Z at some point.

I do have some concerns about the ultimate accuracy. I will have to build a test rig to measure this.

2) Can you explain a bit what could affect accuracy? I just need to understand, I am not challenging your experience

---------------
I am curious about how the nuts will look. Based on how wide the teeth tops look, if each nut would accommodate let's say 3-5 spirals, they would get pretty wide themselves. I guess that if you model the case with nuts and the spurred bolt it could be eventually printed and tested by many. I will print and test one myself on a friend's printer.

3) Regarding the string, in my sketch I have placed the wormgear on the diametrical opposing side of where the string is fed on the spool via a small pulley. Are the 3.5deg the wormgear is tilted away from the perpendicular position enough so that at the top of the worm it gets away from the centered feeding line?

[realthor]

DCLJ - Double Lamina Compliant Joint

I have skimmed through the "New arm design for Simpson style printer." on the reprap forums as I am trying to put together a design for the DLCJs in my double parallelogram linkage and I've read that a solution with Nylon strapping would be better.

Will a Nylon strapping-based DLCJ get rid of a method for adjusting the tension, like the guitar tuning peg for the braided line?

I have sketched a double parallelogram system using the DLCJ with braided line and no tensioner at this point. I want to start a discussion about this so I know which direction to take. It looks fairly complex to me and the number of bearing this would need is staggering

DLCJ.png (550.75 KiB) Viewed 80711 times

The spooling drums that would slot into the arm will have to have some 3d-printed grooves that would guide the line in an figure 8 pattern. Not sure this is an real issue though.

Is this overkill regarding the number of bearings? I hope it is and can be simplified but I don't yet see how.
@Nic: How could this be translated to what you refer to as "boltless and bearingless solution"? Mybe boltless not in this case but bearingless? Z movement is fairly slow and I don't think bearing are justified but a smooth movement without choppy jumps seems to be something a bearing can offer.

Sorry for not being patient enough until i get my previous post answered.

Thanks.

[realthor]

Spur-Bolt Update: All movements on the bot can be actuated by this spur-bolt system and it would somehow greatly simplify, even if only by repeating parts:

screenshot.204.png (113.43 KiB) Viewed 80708 times

Regarding the SCARA Arms, the M8s that are necessary with the current design (using extending couplers) so that a proper MA of around 15 can be achieved aren't now the only option. There is reduction from the wormgear to the spur-bolt and then there is another reduction from the M8 (or whatever ends up in its place) to the SCARA drums/pulleys.

I am worried that this would not work due to the backlash issues associated with wormgears. For Z this is not an issue because the movement goes only in one direction which keeps the teeth of the worm and spur in permanent contact while for the SCARA Arms it will be back and forth. But I've seen wormgears used for Simpson in reprap forums. I don't quite understand yet how the "preloaded passive joints" can be applied to my design but I can think of the Anti Backlash Double-Roller Enveloping Worm Gear. At least some of the design features of this worm can be borrowed so they can work on a wide spur gear.

I found http://www.allytech.eu/index_fichiers/Backlashfreewormandwormwheel.php to be a good read for anti-backlash wormgears. The spur gear seems to be fairly large compared to the worm which would work against my idea but experienced engineers can tell better. Could this be used in a 3D-printed solution that would work with the above design?

[realthor]

threaded spur gear stock

I started modeling the thing and it's not easy at all ... The thinking behind the dimensioning may be easier with experience but I have many questions. It seems to me that almost any worm design I would think of only touches the spurred bolt on one tooth at a time, so having beefy enough teeth for the "bolt" part is a must. As they are plastic, "beefy" gets new meaning when considering that I will carry the whole weight of the bot with this winch.

In the design below there are 3 components that need separate dimensioning:

1) The bolt has a 25mm root diameter, accommodating 78.5mm of string per revolution. The height of the threading is ~7.5mm, making for an outer diameter of 40mm. The pitch of the helix is 10mm.

2) The spur gear that is cut through the bolt's threads creates teeth that are about 8mmx5mm at the base with an ~7.5mm height. Seems beefy enough for me but I can't do a stress analysis yet. There are 12 teeth per revolution, with a ~8.5mm circular pitch.

3) The worm is a smaller diameter than the bolt's root, it is 14.5mm, with an identical thread profile. This throws me out of my mathematical comfort zone so I can't tell exactly how this would translate into mm/stepper motor step. I tried to tilt the worm so that a better contact can be done between the two gears and I got to a 10deg tilt that it seems to me would insure an almost continuous contact as the worm rotates, but this is eye-balling, no math for backing this up.

screenshot.214.png (91.59 KiB) Viewed 80596 times

I could start re-designing it with new values if they are suggested to me but with my current knowledge this is all I have. My next update will be a similar (if not identical) design with the motor support and the necessary nuts through which the spurred bolt moves. I want to get to a printable design.

PS: is there any open source or easy way to calculate a globoid (or Double Enveloping or Hour Glass) worm? I have played with one nice online calculator at me-bac.com and the result looks promising. I now need to learn how to design the globoid worm in CAD and appreciate the tilt for the best contact line.

[Cozmicray]

Might the winch line get caught up in the gear teeth and not spool nicely?


Here is a stab at a KISS design?

winch possibility

winch_jpg.jpeg (20.47 KiB) Viewed 80562 times