Delta design considerations/theory

Hi everyone,

I'm looking into designing my own small delta Reprap from scratch, mainly for the fun of designing machines. However, I'm a bit at a loss when it comes to details on delta Reprap theory; I can not find anything useful in the Reprap Wiki (I know it's probably hidden somewhere, but where?). I'm interested in knowing some fundamental design principles and challenges and how people solved them in order to learn from them and adapt them for my own design.
Some examples of what I'm interested in include (but the list is not exhaustive):
- How do I calculate the optimal arm length?
- What bearing options do people use for the arms and what are their tradeoffs? (I've seen ball bearings and magnets, but I haven't seen any discussion on WHY some of these designs offer desirable properties and what they are)
- How important is the mass of the print head?
- How accurate should positioning of the three axes be?

Unfortunately, the Reprap Wiki does not seem to be very conclusive (or very well-structured) when it comes to these questions. Where can or should I look for answers? Given that everyone and their dog builds a delta Reprap these days, they surely must exist.

Felix

- How do I calculate the optimal arm length?
I know there is a nice math paper floating around on this topic but someone else will have to track it down. Optimal can mean lots of things. Shorter arms get you more height. Longer arms get you a better build area to printer area ratio. Basically, there is no optimal length. You need your target specs to zero in on a target arm length.

- What bearing options do people use for the arms and what are their tradeoffs? (I've seen ball bearings and magnets, but I haven't seen any discussion on WHY some of these designs offer desirable properties and what they are)
Lot's of ways to skin a cat here. Traxxas rod ends are cheap and pretty good. They do have some play so bearings or magnets can help you step up your game. Depending on the implementation bearings or magnets could be better than the other.

- How important is the mass of the print head?
It depends on how important high accelerations are to you.

- How accurate should positioning of the three axes be?
It depends on your delta radius, arm length, and your target effector positioning accuracy.

I wish this was more clean cut. However, if there was a BEST way to do this then we would only have one machine design. We all have different ideas about what is important in a machine. Lots of factors to balance. Cost, speed, accuracy, acceleration, size, looks, etc

---

ConceptFORGE
Wally, GUS Simpson, LISA Simpson, THOR Simpson, Sextupteron, CoreXZ

here are some Delta robot kinematics articles:


http://forums.trossenrobotics.com/tutorials/introduction-129/delta-robot-kinematics-3276/
http://cats-fs.rpi.edu/~wenj/ECSE641S07/delta_calibration.pdf
http://www.eej.ulst.ac.uk/~ian/modules/EEE516/Appendix-%20Riona%20Corrigan%20B0439354-%20Rostock%20Mini%203D%20Printer/Digital%20Appendix/Written%20Report/

---

All about delta 3d printers in one place [www.deltarap.org]

arm length discussion: [forums.reprap.org]
math based calibration discussion: [groups.google.com]
dynamic positioning error discussion: [geekhack.org]
Edit: and delta segmentation error discussion: [forums.reprap.org]

Edited 1 time(s). Last edit at 08/27/2014 05:35AM by hercek.

- What bearing options do people use for the arms and what are their tradeoffs? (I've seen ball bearings and magnets, but I haven't seen any discussion on WHY some of these designs offer desirable properties and what they are)

Traxxas are much simpler and cheaper to implement.
Yes, bearings and magnets are precise, they also weigh a lot, cost quite a bit and are a challenge to get right.

Despite what many think, the magnets aren't even that much better IF you do things right and think about what you are doing.
First off, magnets cannot handle high speeds. Before any of you complain, I've had them, at some of the speeds I run at, they will simply not hold on. It just won't happen. Some of this comes down to what you expect speed-wise. I had them chatter on small infill at 150mm/s, which is about the absolute slowest speeds I run my printers at these days for infill.

Second, if you want them for precision, you are probably running slow speeds anyhow. If you slow down any printer, slop and tolerances become less of an issue. Granted excessive slop is a problem, but if things are decent, your quality should be decent regardless of which rod end system you use. However, if you need a little extra precision, rubber bands on the rods will tighten it up when necessary (I use hair ties).

My biggest issue with them though is safety, I've had mag joints separate and the effector come off in my hands when I wasn't expecting it, luckily it was cold at the time. I don't like the idea of flinging around something that hot, untethered. It won't fly across your room, but it could land on something and start a fire or burn someone/something pretty bad.

- How important is the mass of the print head?
Depends on your priorities, if you want precision, it probably won't matter as your speeds will be low anyhow.
If you go for speed, then it does matter. Higher weight means slower acceleration and problems bringing it to a halt as well. This is a problem with mag joints.

- How accurate should positioning of the three axes be?
Using a probe and some software can tell you exactly how far off things are without a lot of effort. Probes make deltas much easier to use.

I've tried all three options, Traxas, magnets, two versions, magnetic cup and steel ball, steel cup and magnetic ball and my current preferred option which is steel ball in printed cup held together with a spring and Spectra line. See page 4 of my Cherry Pi thread for details. I gave up on Traxas early on as I couldn't seem to get rid of the backlash issues. I know lots of people use them successfully, they just weren't for me. Magnets were great up to a point, as sheepdog says above, they don't hold up well at high speeds. The best I managed on Cherry Pi was 150mm/s infill. After that it was down to luck if they held up for an entire print. The spring and string solution I now use is great. There is absolutely NO movement on the nozzle if you try and move it with a finger. I opted for 8mm drilled ball bearings as they are way lighter than the 10mm magnets I was using and as the cups they sit in are printed into the plastic parts I also saved the weight of the milled bolts I used as cups. My current effector is extremely light. Non travel moves on my machine are now at a reliable 250mm/s with acceleration set to 9000 in Marlin and jerk at 20. I can't yet print higher than 100mm/s as my current extruder can't keep up with a 0.4mm nozzle. Precision at these speeds seems equivalent to 50mm/s speeds with other iterations of delta rod connections I have tried. I use Cura and my printed parts are, in my opinion, excellent. Unless something new hits the scene I will stick with this arrangement.

Andy

The ball and string method is one I want to experiment with, but using Delrin or ceramic balls to keep the weight down.

Thanks a lot for the answers so far, that was quite helpful and insightful. Can someone explain the backlash issues associated with "Traxxas" bearings? I'm considering using more precise plastic bearings, but I'm not sure of the design constraints yet.
For my current printer (pretty standard cartesian XZ;Y, geared extruder) I'm using 3mm filament. I see that usually, delta printers use 1.75mm filament. From my experience, 3mm filament can be pretty stiff, but is this such a huge problem with delta geometries that it will absolutely necessitate 1.75mm filament? I'd rather like to avoid buying two different filaments for each printer. Also, my hotend design works quite well with 3mm filament.
ABS does not seem to be the most popular material, but I like it for its mechanical properties, thus I'd like to use mostly ABS and maybe some PA. Does this require further design considerations (apart from using a heated print surface) specific to delta printers that I might have missed so far?

Quote
fr
Can someone explain the backlash issues associated with "Traxxas" bearings? I'm considering using more precise plastic bearings, but I'm not sure of the design constraints yet.

[forums.reprap.org]
[forums.reprap.org]

Quote
fr
What bearing options do people use for the arms and what are their tradeoffs?

There is some discussion of tradeoffs on this page: Joints for Delta Printer.