Here I'll try to explain the beginins of the project, a bit of background of where I'm coming from, the idea, and where I'm trying to go with this project.
As a student, building a Reprap is a cost and time intensive undertaking. I find the concept of a self replicating very interesting. Not having much money to spare led to my interest in highly printable Repraps as a means of reducing cost by printing many parts of the motion system. I was very tempted to build a GUS Simpson and was also looking at delta printers of which a large cost is the linear bearings. I didn't like how the GUS Simpson uses two special bearings. What about building a traditional delta style printer?
Searching on the Internet I found that there have been attempts at creating a traditional delta printer (e.g, RappiDelta, PolyBot), but they were either not well documented, or had low resolution, or were otherwise incomplete (I couldn't find functional firmware), and so I started my last project, the Icepick Delta.
The Icepick Delta has been a success to some extent. It did not require a linear motion system, and therefor was cheaper to build. Some of the issues are as outlined below:
- Calibration is very difficult. Measuring an arm, and entering it in the firmware incorrectly by 0.5 mm will be a quiet a problem.
- The arms and end effector sockets have to be spaced exactly 120 degrees from one another.
- The end-stop distances are critical, there are three to calibrate. If one of these is calibrated incorrectly, parts are made dimensionally inaccurate, e.g. warped, wider as it gets taller, or other weird nonlinear transformations.
- Low repeatability, (banding in prints)
- It cannot self replicate: After days of calibration and fiddling it was printing dimensionally correctly apart from everything being 4% too short in the z-axis.
- The Arduino Mega with a custom version of Marlin firmware could only calculate 40 delta segments per second.
All in all it makes for a very cool looking, low cost minimalistic printer in terms of hardware usage, but it is very difficult to build and not suitable for a beginner. Lessons learned here will help with the Reprap Neumann project for sure.
I came to the conclusion that the GUS Simpson is a very impressive piece of engineering and is simply much better implementation of a highly printable delta printer. The GUS Simpson is practically an inverted delta which removes the need for large frame like the IcePick Delta uses.
Still the problem remains: How can a Reprap be made, so that it can create its own motion system, and still maintain dimensional accuracy when replicating?
I spent a lot of time looking at other designs, trying to come up with an original solution. Some are rather beautiful and elegant pieces of engineering. In my opinion notably: the GUS Simpson, Reprap Morgan and Wally. I eventually came to the conclusion that there are no such Reprap's as of now that can print its own motion system while remaining dimensionally accurate or easy to calibrate as it replicates from generation to generation.
An example would be I had an Icepick Delta, a Reprap Morgan or a Reprap Wally printer and the pulley radius has a slight error in the pulley radius, the next generation that is printed will now have more than the original error in the radius of the new pulley. As the generations are printed this error will increase exponentially. It is the same with the GUS Simpson's arm joints, the Wally's pulleys and the Morgan's pulleys.
Reprap Neumann aims to tackle this core problem by using a very simple design. The problem on the afore mentioned printers that when there is an error, it is magnified. As explained before, the radius of the printer arms are larger than the radius of that pulley that was being used. To achieve a demagnification of error size, the radius of the point of rotation must be larger than the radius of rotation of the print bed. Although I do not think this is practical with the afore mentioned printers, I think I have found a way to achieve this.
The idea is to build printer that uses linear motion with an Y axis as a base on the table, a Z axis that consists of two vertical rails and an X axis that runs in between the two vertical Z axis rails. The rails and guides would be printed. This way by placing the rolling points further from the build platform, any movements will be demagnified, because the radius from the center of the bed to the guide is greater than the radius of the center of the bed to the edges of the printbed.
- Easy to calibrate; by using standard pulleys and GT2 belts the only thing that will need to be calibrated is the extruder steps per millimeter. No dimensional print accuracy critical endstops.
- Highly printable
- Low on vitamins
- Easy to assemble (hopefully). Using nuts and screws to assemble should be straightforward using a diagram and instructions.
- Can self replicate (in theory, will need to do testing to confirm)
- Highly printable: needs lots of printed parts
- Takes a long time to print
- Assembly time may be slightly longer due to more parts, maybe offset by ease of calibration
To be continued..
If you have any questions or if what I'm trying to explain is unclear go ahead and ask and I'll do my best to explain :)