Comparison of various cartesian frames

There has been a bit of interest lately in different ways to do repstrap Cartesian systems, so I thought I'd open a general discussion of the various pros and cons of the different forms.

Given 3 coordinates: x, y, z; x and y being horizontal, and z being vertical, I can imagine three different systems:

1) The reprap method. The head moves in x and y, the table in z. This has the advantage of the most usable work space for the least actual volume and footprint. This is because the center of the head can get closer to the corner of the footprint than the center of the table, because the head is smaller. I hope this makes sense.

Unfortunately, this method requires a lot of stabilization and support, to hold two axis in the air. Thus, the reprap looks like the cube that it does.

2) The standard CNC method. The head moves in z, and the table moves in x and y. The advantage of this system is that because you are only holding the vertical axis over the table, it needs very little support. Often just a single arm overhanging the table. Also, because the x and y are at the base of the machine, their workings can be hidden under the table for a nice clean effect.

This method, however, has the least amount of workspace for the amount of footprint. In fact, it results in less than one quarter of the footprint. This is because the table needs to be moved so that the head can reach all of the corners. Again this is about as confusing as what I said about the previous method. I hope you can understand it.

3) The table moves in the x direction, the head in y and z. X and y are really interchangeable here, because you can just rotate the table 90

Given the want to do "high speed" FDM, the current darwin design has many things going for it. Small table space means you can put it somewhere that most other designs won't fit.

I personally prefer #3. This is often called a gantry machine, and Tommelise uses it.

There's some simple but interesting tricks that a gantry design allows. Not all of them would be useful for RepRap, but the possibilities are there.

Due to the machining nature of these, most gantries are actually four-axis. The support columns provide a large range but slow Z motion and the heads have additional Z motion, smaller and fast. This allows for maximum stiffness in the structure and still be flexible for use on tall and short projects.

Standard high speed plastic routers usually have two heads on the gantry, on one shuttle. Each cutter has its own driver. One is usually a parting saw, the other has two more axis for contour cutting in 3D. These also usually have the workspace divided into two tables that can be electronically locked together. This allows for unloading one part while another is being machined. A much smaller side workspace could be used as an independently moving head holder for multi-material projects.

If you replace your table axis with a rotary base, you have a simple swap-in polar design. Attach a cutting head instead of an extruder and you have an industry standard vertical turning lathe.

If you put a rotary spindle below the head with a vertical plane of rotation, you can lay down and build up an asyimetric lathe design given a thicker base support to start with (inverted from normal lathe work where you work down).

Now, what happens if you allow the table to completely leave the workspace? Easy unloading to start with. I've considered building a vertical table storage machine to put behind a gantry system. Need FDM today? Pick out the friendly surface for your particular plastic. Light plastic machining of an existing object? Perhaps a vacuum table. Vat based light reactive polymer? Roll the vat in and exchange the head for the UV laser. Need to do perfect circles? Pull in the extra axis rotary table instead.

What about a moving X, Y & Z like a docks container crane this gives full table usage the support is mainly low down and the Z is very simple. I have a small 300mm x 400mm hobby CNC router of this design and it works very well accurate to 0.01mm or better. I have been starting to look at changing the design to fit into the RepRap ethos of using FDM created parts and simple rods and i think it should be quite straight forward.

I was also trying to make it from of the shelf parts that could then be replicated to help with the RepRap chicken and egg problem if it could also house a small router head all the better especially in the early stages of development while the extruder is being perfected.

Possibly not for this thread and also possibly mentioned before (I'm very new to this project) why does the RepRap not use G code for positional control, as most routers do, it's very well suited to controlling X,Y & Z axis and simple parts could be produced without the need of an STL file

A big advantage of the reprap model for FDM machines I think is the inertia of the object you are creating itself. if it only has to move in the Z direction while printing, then it only needs to handle upward and downward forces, which it was already constrained to do because of gravity. if the whole table moved then the object will have to withstand lateral forces due to acceleration of the whole object in the x and y plane as well, which would put a limit on the x and y speed or/and the structure of objects you can build.

This isn't an issue with subtractive systems since the working media has to be held in place very well anyway.

I've built a (3) with Tommelise 1.0 and am going for a classic (2) with Tommelise 2.0. While in theory (3) has half the footprint of (2) in practice the advantage isn't that wonderful.

I'm currently puzzling out the X,Y, and I'm fixing the steppers to a cross-slide vice. For $40, I get 6 inches of screw-driven X and Y travel, all made of steel. And it has a vice, which means I can fix parts down in repeatable fashion.

Also, if you want to lessen the footprint of moving the table, you can make the table a turntable. That tool head (extruder or the like) can move in the z axis. Either the table or the tool head can move laterally.

Given the current realities of extrusion speed and duty cycle, I don't see much of a point in having a working area much larger than 15-20cm in each axis. A "conventional" machine as the OP's #2 is a cinch to build without fancy parts or tools, and can be made rigid enough for light milling or PCB routing, both of which could prove useful.

Whether acceleration is a problem I can't say. My sense would be that with the plastics now being used and the room-temperature extrusion method, the acceleration would have to be fairly high to affect things. With different deposition methods, though, this might change.

IMHO, all the action and energy belongs in the deposition systems. This is where the revolutionary part of the system lies--the rest is fairly low-level evolution.

I agree with Colin. The extruder is where it's at. We are just re-inventing the wheel with the cartesian system though it is necessary too for a complete system.

Demented