Head swapping mechanism?

I think Adrian and Ed have thought about a head swapping mechanism. I haven't seen their thoughts written down anywhere, so I thought I'd write down my thoughts and let y'all have at 'em. I have an International tractor with a "Fast Hitch". It uses implements that have two bars projecting from their front. The tractor has two hitch receivers into which the bars fit, and are held. I'd reverse this mechanism, so that the cartesian robot has two bars sticking out of it. It would fit those two bars into horizontal holes in the printhead. When it lifts up the printhead, a lever drops down which locks the printhead onto the bars. To reverse the process, the robot moves the printhead over its home, and lowers it onto some vertial rods which hold it in place. This raises the locking lever, allowing it to move its two bars out horizontally.

Fast Hitch patent (expired, of course):
[www.freepatentsonline.com]

Love the method of connection. But one question? Electrical connections fit where?

---

Bob Teeter
"What Box?"

just mount a plug that is pretty easy to plug and unplug on the two seperate parts. Maybe a simple power delivery like USB. Then each head needs its own card, but I figure that some head might not have cards, just a motor that is always on (routing).

---

Jay

The proposed head swapping mechanism for Mendle was shown here [blog.reprap.org]

---

[www.hydraraptor.blogspot.com]

hey! instead of changing the head how about using shape memory alloy to create a single head that can change in size? its something im going to try when i get to it

It's not for a change in size, although it could be used for that, it is primarily for different types of head, like the paste extruder for support material.

---

[www.hydraraptor.blogspot.com]

you could also have only one load/unload home for heads, but also have a belt with different heads on it, so the belt presents the reprap with either a new head, or an empty space on the belt on which to unload the current head, this woud mean that it would be relatively simple to have 'n' heads, rather than changing the reprap design too much every time you want to increase the number of heads.

How many heads do you expect to need?

I get the impression that if the answer is something like 'lots!' or 'as many as I feel like!' then you're probably trying to make something that does *everything*, and that is rapidly going to become far more trouble than it is worth.

Consider a fairly simple tool palette... say a polymer extruder, a support material syringe and perhaps one more such as a lightweight milling head. Maybe a second material deposition head of either type for good measure.

This should allow you to make a pretty wide range of products, but the collection of heads could potentially be compact or light enough to mount on a single cartesian head, and almost certainly mountable on a moving-bed type system.

Not enough heads for you? Right, assemble yourself a second machine, and give that the extra heads you want.

Trying to make one machine do *everything* seems like a bad idea. Instead, make one machine *good enough* to help you build other machines, and avoid these sorts of pitfalls.

well, a non exhaustive list might be...

low temp polymer extruder
high temp polymer extruder
support extruder
metal extruder
ceramic extruder
epoxy extruder
foam extruder
silicone rubber extruder
food extruder (icing and chocolate )
miller
scanner
CO2 laser
colour printer
welder

some of these might be combined into one head, and others might be permanently on the head, such as a scanning laser... however I do think that the RepRap gets it's full use from the ability to eventually print multiple materials in full colour.

for version 1.0 this is of course ridiculous, but it is something that should be looked into for future versions

Edited 1 time(s). Last edit at 02/20/2008 12:02PM by deadgenome.

I'll stand by my multiple-fabricators argument

If nothing else, some of those are going to have power or weight-supporting requirements that massively dwarf the needs of other toolheads. Metal extrusion (I'm curious, how practical is that anyway?), cutting lasers, heavy-duty milling, and so on.

Food extrusion is best done in a somewhat cleaner environment than one you might use for more general pupose fabrication. Perhaps other processes need to be better cooled, or ventilated.

It has just occurred to me that if you made a fabricator with a toolhead that moved in all 3 dimensions so the bed could remain static, you could conceivably have an assembly line type arrangement with a conveyor moving the object between different fabricators with different tasks.

I just think that designing a system that can cope with 'n' heads instead of a fixed set shouldn't necessarily be more difficult and lends the system a level of futureproofing. All that is needed is to have a head delivery mechanism as a seperate component, the simplest probably being a turntable with three or four docks on it that can rotate the docks past the repRap's load/unload point. With this arrangement, people who want a nice simple toolset can use the basic 4-dock system, while those who want complex insanity can be happy as well.

The metal extrusion thing is by no means my idea, but is based on using metals with very low melting points so that 3d circuits can be constructed within a plastic object and from what I understand the core team are having some success with this concept already.

Cutting lasers don't necessarily have to be heavy, I have seen some fairly light CO2 cutters made from handblown glass. Also, if the laser is operating in the visible spectrum, fiber-optic might be able to be used to get the light to the target. A tool head with a rotating mirror is another option, with a rotating laser off to one side.

Heavy duty milling is of couse not an option for a lightweight rep-rap, however lightweght milling on a dremmel sort of scale might be and is worth looking at. working the object before it is fully cooled may also help on this (or might be a really bad idea).

For some jobs having multiple fabricators on an assembly line is the best solution, however an assembly line usually has the problems of specialisation of purpose that the repRap gets away from as if you are not making lots of identical objects, you get huge problems of timing (eg, the miller is still cutting something down while the object from the fdm machine is already finished and waiting to be milled, blocking things up for the next item to be made).

Since a core goal of the project is to make a machine that can reproduce its' own parts it should theoretically be simple to make a new printer for each printhead that is desired. In practice however there are at least 2 problems. First and fundamentally some heads such as thermoplastic and support material have to work together; even if alignment and other logistical challenges could be overcome it would be wasteful to move an item from machine to machine 2 or 5 times to deposit a single build layer. Second, while some parts of a reprap are "free"(plastic bits) or cheap and easy(rods and screws), others(motors and electronics) are unlikely to be so any time soon. Thus it makes sense to make one machine which uses many toolheads as long as the cost of making the toolhead work is less than that of new motors and electronics. Most deposition heads will be cheap to make. Some subtractive heads will be able to be adapted especially if the Cartesian structure is made slightly more robust than a Darwin. Some special cases might involve heavy milling, large (furniture size) fab volumes and medical applications. Also machines using alternate fabrication methods such as EDM and fused powder deposition would probably benefit from separate machines.

As a note to Ru's question about the feasability of metal deposition...

[gram.eng.uci.edu]

We aren't original--he says with a shrug--but we sure can read and copy!

Demented

oooo.... that is virtually identical to what I suggested after reading The Guy's electrospinning printer idea... [forums.reprap.org] & I swear I hadn't heard of this before, it just seemed obvious after reading about continuous inkjet printers