Working Notes, please log in and edit!
These should be autogenerated by the 'Development' catagory or template. Or integrated in some other manner.
As we mention when discussing the Combinatorics Problem, a new design can be "better" from some other design in many different ways.
Erik de Bruijn lists these broad areas of potential improvements: 
- Added functionality:
- Use of ceramics and pastes instead of thermoplastics
- The ability to function in subtractive as well as additive operating modes (e.g., Hydra-MMM).
- The ability to mix multiple materials.
- Embedding wire and conductive materials
- extended auxiliary tools (Erik de Bruijn, "On the viability ..."; "3.1.3 Technological innovations")
- Improved existing functionality
- Faster, more efﬁcient, more detailed and/or stronger output
- Increased ease of assembly and use
- Lower cost
- More suitable components (e.g. easier-to-acquire)
- Specialization towards a certain application
- Digital pottery system
- Interoperability with other systems
- Compatibility with G-Code common in industrial CNC installations
- platform independent software
- Improved design architecture (e.g. modularization, part consolidation)
- Refining operating techniques
- Improved sharing infrastructure
Better electronics is discussed at Alternative Electronics. Better host software is discussed at Host software Variations. Many ideas for making things better are listed at ideas to place and FuturePlans.
Ideas for a better website, enabling easier collaboration, are discussed at Library/Notes.
discussions of the pros and cons of various frame materials go in frame material.
discussions of the various materials that can be shaped by a RepRap are currently at MaterialsScience.
discussions of the various techniques used to shape various materials are currently at Materials/Appropriate Machines.
By SLS, are you referring to selective laser sintering SLS Printer ?
Increased Build Area
Mendel is great, but I can currently
- Carry it through a standard doorway
- Lift it by myself
- Fit it in the back seat or trunk of a car
- Use it on a corner of my desk (as opposed to a dedicated room in a house, or a shed, garage, or barn)
We at RepRap agree scaling is a real problem. That's why MegaMendel is here to help!
Gert Joergensen has built the MegaMendel, a mighty machine with a build area of 766mm x 453mm x 497mm.
Some "Need bigger surface!" ideas for scaling up to 200 mm x 400 mm heated bed, might apply to even larger scales.
There's been some discussion "RepRap / RepStrap for making Aerodynamic Models" circa 500mm x 500mm x 500mm.
There's been some discussion of a "1 x 1 m print area reprap".
Perhaps you can "Help me design a large Rostock" 4ft diameter (1.2 m diameter).
If that is not big enough, MegaRap is a (currently hypothetical) RepRap. It may be a PourStrap or WeldStrap and is definitely a CNC Router which can process 4 ft x 8 ft sheets of plywood, foam, etc. See CNC router#BigRap, RouterStrap.
Many people would like a CNC machine big enough to hold a "full-sized sheet" of 1.2 m × 2.4 m ( 4 feet × 8 feet, aka "four by eight" or "48 x 96") sheet of Wikipedia: plywood or MDF or other Wikipedia: engineered wood or foam, etc.; and then cut it into FlatPack parts.
The B&TRap -- using 3 cords -- can, in principle, be easily expanded to any size.
Better electronics =
- main article: Alternative Electronics#Goals
See Available Files for parts we already have in our library.
How can we tweak the design of a RepRap to make it more self-reproducing?
(Is there a page that talks about "tweaking the design to make it more self-reproducing"? Please move the following ideas be moved to that page.)
Adrian Bowyer's work changed the lives of some people such as architectural students who print models of buildings and entire cities, engineering students who print models of gears and get a much more intuitive understanding of how they fit together, etc. Before Adrian Bowyers work, those students couldn't even afford to rent time on a prototyping machine, much less own their own machine.
But some researchers feel that the current crop of low-cost fully-assembled prototyping machines based on Adrian's work, while they arguably work even better at printing models of buildings and gears, misses out on some world-changing ideas:
- A few people at a single location at a single company mass-producing identical machines is inevitably going to come up with ideas for improvement at a slower rate than a diverse collection of people all over the planet, each one building on the work of people that came before and sharing improvements to the people who come after.
- Rather than shipping a complete machine -- or even all the parts of a complete machine -- to someone that wants one, material costs and shipping costs and shipping time and time-to-build can potentially be reduced even further by taking advantage of local materials and local manufacturing capacity. (Ideally *everything* can be sourced locally, completely eliminating shipping time).
- Given a primitive early version of a RepRap, it is in principle possible to directly print out and assemble a much later, vastly improved RepRap, and then (indirectly) print out and assemble the very latest RepRap with all the latest capabilities.
- If everyone who builds a RepRap prints out 2 complete sets of parts and sends them to 2 other people who then build a RepRap (who continue the chain for 33 levels), everyone on the planet can have one -- see doubling time.
- Darwinian Marxism, as described in Wealth Without Money: when each machine is individually built and owned by the person who is going to use it, and each person contributes some improvement, no matter how small, then the improvements accumulate.
Many people find the idea of a Wikipedia: self-replicating machine, in particular machines that have a doubling time, fascinating to think about. Adrian Bowyer's Wealth Without Money essay mentions a "a rapid-prototyping machine that can make all its components other than" a short list of critical components, and hints at "the desirable aim of shortening or eliminating [that list] altogether." RepRap researchers often metaphorically refer to parts on that short list as vitamins. People discussing a RepLab or category: RepRap machines often use phrases such as the "reprappiness factor" or "replicability factor" or "RepRap Factor" or "more replicable" or "% RepRap-able" or "largely self reproducing" or "mostly self-replicating" or "largely self replicating" or "the percentage of the device that is printable".
Several researchers are developing ways to reduce the vitamins required to build a functional 3D printer:
- printable actuators -- eliminating non-printable stepper motors -- Actuator Fabrication
- printable frame -- eliminating threaded rod used in the frame -- Tantillus, GolemB, etc.
- printable motion control -- eliminating linear and rotary bearings, or the smooth rods and threaded rod used as screw drive, or both -- Bamboo Printer, PLA bushings, Category:Printable bearing, Ben Bot[], etc.
- printable electronics -- Automated Circuitry Making
- printable fasteners -- eliminating nuts and bolts and other non-printable fasteners -- MultiRep, MTM Snap, etc.
However, other researchers are keeping the same or increasing the vitamins, in order to reach #Other design goals.
objectively measuring more or less self-replicating
Alas, there seems to be a lot of confusion about whether a machine is "self-replicating", and about how to measure how self-replicating it is. (TODO: summarize comments on this topic from ConvertingARepStrapToAFull-blownRepRap, Talk:Orca#Re: whether it.27.3Bs a "true reprap" or not:, "Convergence to self replicating", "Dreaming of a Static Motor Arrangement and Less Vitamins", "95% reprappable repraps", etc.). Everyone agrees that a (common) machine built of parts, a machine that cannot print out *any* of those parts, is not self-replicating. At best it is a RepStrap. Everyone agrees that a (hypothetical) machine that one can drop on an asteroid, in total isolation, and the machine somehow harvests iron and rock and sunlight and transforms it into all the parts needed to build 2 machines identical to the original machine, is "100% self replicating", a fully printable reprap. But in-between cases are not so clear.
A variety of ways to objectively measure how "self-replicating" one machine is (a "metric") have been proposed. So far, all of these metrics have some flaw or another.
- minimize vitamin/total weight ratio. Flaw: adding a bunch of non-functional plastic spikes to each printed part increases the total weight of the machine, artificially improving this measure, but most people would say that is not a real improvement.
- maximize the "actual volume of printed parts" Flaw: adding a bunch of non-functional plastic spikes to each printed part increases the total weight of the machine, artificially improving this measure, but most people would say that is not a real improvement.
- minimize total cost of the vitamins. Flaw: because of currency fluctuations and shipping cost variations, this is arguably not an objective measure. Also, replacing two euros worth of bolts that I can get tonight, with a hundred euros worth of plastic parts that would take me days to print out, seems counter-productive.
- minimize total cost. through various Cost Reduction ideas. Flaw: one way to do this is to freeze the design and use mass-production high-volume manufacturing techniques to improve the economy of scale, but that misses out on the "continuous improvement" possible when each RepRap built can be different and possibly better than the next. Also, because of currency fluctuations and shipping cost variations, this is arguably not an objective measure.
- cost of vitamins (including shipping)
- cost of raw plastic feedstock
- cost of printing (electricity, especially for heated bed)
- labor cost * assembly time (see doubling time)
- minimize time to assemble. Many researchers have pointed out that time-to-assemble is far more important than initially realized -- RepRapBreeding, RepRap Breeder, Bonsai RepStrap, Walkabout, Category: Loaner Program, Print It Forward, etc. Flaw: If I buy some (allegedly) ready-to-go machine, this time is (allegedly) zero. But most people would say this isn't really self-replicating.
- minimize doubling time -- time to assemble from parts, plus time to print out a new set of parts and wait for the remaining vitamins to be shipped in. Flaw: If I buy some nearly-ready-to-go machine and insert a single part into that machine, and then the machine can print out only that one single part (the rest of the machine is "vitamins"), this time can be very short (overnight shipping plus time-to-insert), but most people would say this isn't very self-replicating.
- minimize the number of unique vitamins. Flaw: one obvious "improvement" that improves this measure is to reduce the unique parts from "threaded rod, nuts, bolts" to "threaded rod, nuts" -- by replace each bolt with a piece of studding cut to the same length as the original bolt plus a nut or two to act as the head. Is this really an improvement?
- minimize the number of exotic, single-source parts, replacing them with "jellybean" parts available from multiple sources
- minimize the cost and weight of the tools required to make the parts of the machine: Replace parts that can only be made on a quarter-million-dollar machine with parts that can be turned on a big $20,000 manual lathe. Replace parts that can only be made on a big lathe with parts that can be made with a small $2,000 desktop CNC machine mill. Replace parts that can only be made with a lathe or CNC with parts that can be cut with a $200 circular saw. Replace parts that require at least a circular saw with parts that can be cut with a $20 hand saw. For example, the ScrewRap with its "Minimal tools" goal. Flaw: it is unclear whether designs using lots of T-Slot should be considered highly replicable by this criteria -- considering the T-slot as raw material that can relatively easily be cut by low-cost tools -- or whether it is not very replicable by this criteria -- since it requires highly specialized equipment to make the T-slot from aluminum stock.
- Certain RepRap subassemblies have been designed to be "easy to make", "do-it-yourselfable" -- in particular, many Gen7 Stories show hand-built electronics using off-the-shelf prototyping board that can be easily customized. Designs using category: through-hole electronics and prototyping board are sometimes said to be more self-replicating than most that require a custom mass-produced PCB with higher up-front NRE costs and tiny parts that seem to be impossible to hand-solder.
- minimize the number of parts that have to be shipped in from distant places, replacing them with locally-sourced parts or printed parts. "build a printer from whatever you can find in a local hardware store."
- A set of machines, each of which can't make *any* of its own parts, but which collectively can make all of the parts of every machine in the set (what MattMoses calls "Cyclic Fabrication Systems"), are sometimes said to be self-replicating. See RepLab.
- Machines that can build all the parts for solar cell manufacturing factories that can produce solar cells that can power the original machines are sometimes said to be more self-replicating that machines that rely on the electrical power grid, but this idea isn't captured by any of the above proposed measurements.
Is there some objective measure of "percent self-replicating" that avoids these flaws?
Other design goals
Some researchers have yet other design goals or "design philosophy".
Some researchers deliberately tweak a design in ways that make it less self-replicating -- i.e., a "Vitamin-Rich RepRap" (see Kludgebot) -- in attempts to satisfy other design goals:
- fewer unique kinds of parts -- it's much quicker to find a certain kind of part in a pile of 100 parts of 2 kinds than to find a certain kind of part in a pile of 50 parts, all of them unique. Also, bulk discounts often make it cost much less to buy 100 bolts in 2 different sizes than 50 bolts, each one unique.
- fewer total number of parts -- reducing the assembly time. (LaserCut Mendel, MakiBox, R 360, etc. mention this as a goal).
- small and rugged to make transportation easier.
- Using a 3D printer helps develop certain skills. What features of a RepRap or RepStrap help people develop those skills?
- Building a 3D printer helps develop certain other skills. What features of a RepRap or RepStrap help people develop those skills? As a negative example, if a design is so difficult to assemble that many people give up on the project -- I'm looking at you, McWire (Death March: Do not build!!!) -- then those people can be tricked into thinking they aren't smart enough to assemble any 3D printer -- "learned helplessness".
A few other design goals are mentioned at ideas to place.
- these Development Pathway notes may slightly replicate and encapsulate the Gada Prize stuff.
- "Design a perfect 3D printer" asks: "If you can design a 'perfect' 3D printer, what would you do?"
- Erik de Bruijn. "On the viability of the open source development model for the design of physical objects. Lessons learned from the RepRap project". 2010. In particular, "3.1.3 Technological innovations" and "Appendix C".