Originally shared by Nicholas Seward
One of my students just made this sweet 1-arm SCARA prototype. I have been meaning to try this geometry for a very long time. It checks out and will easily support 3D printing.
We are accepting suggestions for the Z. I really like the idea of a backpack printer that can print huge 300mm+ objects. That said…how can we add a Z and still keep it cute and foldable as it is now? I would be okay with a smaller Z such as 75mm or 100mm.
I’d say, lift the part with the motors. Keep the bed in place, probably ad a back wood for stability, yeah, with two z rods that should be doable 
Foldable? How about a scissor lift/arm under the XY motor mount? Use a single horizontal Z motor with left hand and right hand threaded rod coming out opposite sides. The extension rate would not be linear but it could be adjusted and compensated for by firmware.
You may make the second hinge (one that is closer to the effector) tall and place a threaded rod there. So the effector elbow sits on the nut on that rod.
The idea is to move only the second elbow along the Z axis.
Does it maky any sense? I can draw a scetch if you like.
But personally I think you won’t get an acceptable resolution with that approach. Too much of a moving mass on a long lever. And the second issues is too many bearings. Their play will be multiplied by the long arms.
@Anatoly_Makarevich as is there is virtually no play. The resolution is good enough for low quality prints but we plan to up the resolution. This is just a mockup after all. If you need more proof then check out the RepRap Morgan. This is very similar to it on a kinematic level.
use a pistol or hydro system and you need a pump instead of motor
@Nicholas_Seward I know of Morgan style printers. The idea is quite attractive indeed. But as far as I know they are not widely used. I think there is a good reason why.
@Anatoly_Makarevich Yes, it is a bit of a design challenge to make to work. These are not common on a hobby level because most hobby builders don’t have the engineering chops. It is easy to see in industry that they are much more common.
Let me run through the standard concerns.
Weight: the arms look massive but they have infill and are quite light. Rigidity is proportional to bh^3 where b is the width of the arm and h is the height. Moral of the story, go huge with low infill. Additionally the steppers of static mounted. If I had to guess the arms in total weigh less than 250g. Eventually a hot end will be added but those are pretty light also.
Bearing play: this is easy to handle. You just shim radially and preload axially. Trust me that there is zero play is this system. Flex from the arms and stretch of the string is an order of magnitude more than bearing play.
Resolution: I forget exactly what the worst case resolution of this prototype is but I think it is 300 micrometers. That varies linearly down to 0 as you move in towards the central pivot. If those numbers are accurate, I would get decent print in the middle. That said…we are aware that it should be better and all we have to is increase the mechanical advantage.
Leveling: This one can be tricky. It is possible for the effector to not sweep a plane. You can solve this by building micro adjustment into one of the proximal arms. That said, the prototype was close enough to planar that we won’t do that. Besides, leveling probes are all the rage.
Let me know if I missed anything. All of that was to say that it can be done. That said…there are tons of advantages. I will just leave you with one…this little bot with 150mm arms could print objects 500mm+.
@Nicholas_Seward That’ll certainly work no doubts at all. All I’m saying you trade speed/accuracy for lightweight design and small printer dimenstions.
Say 200 gramm effector (with all that fans and sensors). It’s like 2000 gramm of mass on the spool. (radius of the spool is ~10 times less then the length of the arm). That is not counting the mass of the arm itself.
You won’t get good acceleration at the end of the arm.
By the way I have a delta printer using fishing lines. They actually stretch. You may have hard time catching resonance on rapid movements.
If you know all of that and that doesn’t defeat your goals then go for it
By the way no matter how hard your arm is it will bend under it’s own weight. And the more straight the arm is the more it bends. You’ll have to compensate that somehow.
@Anatoly_Makarevich Yep…all good points. Resonance is the number one problem here. Ultimately, I will reduce the natural frequency by switching the fishing lines for belts and adding a hot end (strangely enough).
Acceleration is a none issue for me. I am not interested in a speed demon. I just want a good large portable printer. I will just limit the acceleration to half of the acceleration when resonance artifacts become visible. It will probably add up to 50% in print times but I am okay with that.
Bending is not really an issue. The arms are massively over-engineered for this. A human can apply enough force to deflect it 3mm or so but under sane loading I have no worries. Also, gravitational loading is constant so can be adjusted out if required as I mentioned with micro adjustment on the arms.
I don’t think we misunderstand each other here. I am just writing all of this because there is a lot of misinformation about robots like this. There are a ton of examples of this being done poorly on the hobby level which gets translated into false conventional wisdom. Typically there is no math or analysis being done to meet a specification so they are usually doomed.