Introducing Eiffel

Hey folks.

Here's the link to my idea for [[Eiffel]].
[objects.reprap.org]

I think it will be fun to make, and I look forward to building one of my own when I've got the chance.

It's a working wiki pages, so you are extremely welcome to login and click "edit".

I'd suggest we keep attachments, images, and notes on the [[page(s)]], comments in the thread in order to minimize work for ourselves.

Cheers,
Sebastien

Sebastian (and anybody else reading),

I like the idea of a hybrid machine -- one that can do both additive machining (extrude plastic) as well as subtractive (route and drill PC boards.) What you've sketched out is similar to what I'm already ( *very* slowly) building. The main difference is that I bought an XY table for the horizontal motions. The links to spindle options are great. I was going to use a rotozip, but I like the larger collets (1/4 and 1/2 inch diameter) that several of those have

IMHO, a rectangular frame is easier to build (accurately) for most people, vs. the triangular frame of Mendel. I think that adding either corner braces (or full/partial face panels) would stiffen it up significantly. Or achieve enough stiffness to do PCB routing with a less expensive frame. Both the makerbot and Forrest Higg's machine(s) lend credence to the utility of a panel-based frame, and the variation to panel-reinforced seems reasonable to me.

The working envelope seems a bit bigger than strictly neccessary for PCboard routing, but will probably come in handy. Perhaps the Z dimension could be decreased a bit. However, I think good corner-joint design(s) are more critical than the exact dimensions you/I/others built to.

-- Larry

The working envelope seems a bit bigger than strictly neccessary for PCboard routing.

Especially once you try to scale up to 2"x2" or 4"x4" square steel tube, yes.

Also, face panels will help immensely help a lot. I'll try to pad out the background of the Matrix furniture technology stuff, rebadged 'gridbeam' when I have a chance.

Edited 1 time(s). Last edit at 01/26/2010 04:10PM by SebastienBailard.

One of the advantages of Mendel over Darwin was to get from
4 vertical rods down to 2.
You are going backwards in evolution here.

---

-------------------------------------------
* homeprototype free 3d design repository
* Blog
* Google+

MarcusWolschon Wrote:
-------------------------------------------------------
> One of the advantages of Mendel over Darwin was to
> get from
> 4 vertical rods down to 2.
> You are going backwards in evolution here.

Looking at the drawing, it looks to be proposing a single threaded rod to drive the Z axis, along the same lines as some DIY MDF CNC machines.

How much Z does one actually need to start making useful parts (Start creating a second generation). It seems one might be able to get away with something quite shallow if the machine is mostly flatpack / panel based, with a few reinforcing off the shelf corner/angle brackets, along the lines of the eiffel design -- 50mm or less? It would also seem able to manufacture more of it's own parts from simpler cheaper parts (MDF, copper clad board, plastic filament, and.. hmm.. anything else?) to produce a second generation machine.

One of the advantages of Mendel over Darwin was to get from
4 vertical rods down to 2.
You are going backwards in evolution here.

We still need to do up the gantry type design "generic cnc router". Want to have a go?

Greetings all,

Earlier, I wrote:
> The working envelope seems a bit bigger than strictly neccessary for PCboard
> routing, but will probably come in handy.

If folks are mainly interested in fabbing that uses only additive or subtractive machining on any given part, then I'm not so sure having a large horizontal workspace is critical. So, for circuit boards (which probably don't need any extrusion on them), manual switchover of the tool would be OK.

However, if we want a machine that can do both extrusion and routing, and *automatically* switch between the two tools, then IMHO, having at least one horizontal travel that is big (over a 12 inches) will come in handy, because it'd be much easier to offset the two tool heads along the long-travel than to make a mechanism that moves them (alternately) into/off of a single working axis. I think we'd still need a means to move them (independently) in the vertical with respect to each other, so the in-active head is above the active one, and thus doesn't hit the part during fabrication.

Why would one want to alternate additive and subtractive fabbing, automatically?
Speed with accuracy: With only an extruder, making fine-featured parts needs extrudsion of a fine filament of melted plastic, and that goes slowly. However, if one could extrude a fatter filament (e.g. for an entire layer) and then switch tools (automatically) and mill off the fraction of the layer that was "outside the lines," one might be able to make larger objects with fine features, much faster than otherwise.

Thoughts?

PS
Why Eiffel? (Not that it's a bad name, but it doesn't seem very tower like to me.)

A second benefit of adding sides, top, and bottom, even of acrylic, is that it would help maintain an even temperature of the piece being built if a hot plate is used under the part. And for future enhancement, if metal extrusion becomes possible, it might work better in a low-oxygen environment. This will be much easier to achieve with a fully enclosed reprap machine.

To make the machine more able to make its own parts, including outside walls, they can be cut (using the milling head) into sections that can then be glued together. With stiffeners added, it would be almost as strong as full size pieces. And since they are fairly simple rectangular pieces with mounting hopes, they would also be easy to make by hand if the builder desires.

As for multiple heads automatically switching, I would suggest that the Z-stage contain a gripper with servo control, and a 'rack' on one side at the top where multiple heads can be 'parked'. The machine would then move the current head back to its storage spot, open the gripper, move to the new extruder/mill head, grab it, and return to work with the X-Y coordinates as the last head. I would also suggest adding a fiducial point so that the head can be more accurately re-calibrated each time a new one is picked up. Three laser pointers aimed at three collimated detectors would be able to precisely re-locate the head no matter how far off the gripper is.

Mike

> frame of Mendel. I think that adding either
> corner braces (or full/partial face panels) would
> stiffen it up significantly. Or achieve enough
> stiffness to do PCB routing with a less expensive
> frame. Both the makerbot and Forrest Higg's
> machine(s) lend credence to the utility of a
> panel-based frame, and the variation to
> panel-reinforced seems reasonable to me.
>

How is Eiffel doing?

The wiki page [reprap.org] sais it works, which is great, but it has only photos of a roughly non-finished state.

Could you simply post some photo(s) of the finished machine, pleeeease?

I just ordered 5 6' steel beams off mcmaster... trying to reconstruct some CAD files from the photos on the wiki, figuring out what is going on/is necessary. Hoping to build the proposed build platform: 24" x 48" x 8".

Although I have no experience building a mendel/etc... I've been making my own pc's forever (custom watercooling loops/etc) and have some experience with lasercutting/routing/3dprinting/wood+metalshop work as an architecture student. Slowly wrapping my head around DC motors/steppers/servos/timing belts etc but would be happy to have a collaborator to help figure out what I'm doing and maybe work on documentation.

So far I have a gen6 mobo w/optos and soon a buncha steel

Edited 1 time(s). Last edit at 10/15/2010 06:40PM by Watanabe.

So after staring long and hard at the photos on the wiki and spending some time in the IRC channel I finally figured out what was going on with the design (overall extruded L shape, bed sliding in X axis under YZ)

My question is about the use of drawer slides in that mockup. Looking around it seems that to get the proposed 48" of travel would require a pretty expensive drawer slide:
[shopping.netsuite.com]

or maybe I'm totally overbuilding it and something like these are more inline?
[www.amazon.com]
[www.amazon.com]

Also, I'm wondering why these were used at all, instead of something like a smooth/oiled steel rod as a guide/structural support going through the perforated holes in the beam? Honest question, I'm on the design side as opposed to engineering but seems like this could be jigged up using a 3d printed part to ride on a rail and slide back and forth, being driven by a thread/rack and pinion/etc system... Does that make sense?

Have some fittings (x100 60mm m8 bolts, heavy-duty(8mm tall) nuts and washers) coming so I start playing with my big erector set.

edit: also interested in your motor selection. Googling the part # in the photos it seems like you've got a 150-160 oz-in holding torque... is that a good starting point? Looking on lin engineering site I'm finding the 5718M-04P/05S as being what I should look at(trying to keep under the 2A limit of my gen6 mobo)? When configured with a double shaft these are running $88 a pop from Lin... is there someplace else I can find an equivalent part cheaper? Figure I need 3 of these for the xyz axes, and a much cheaper nema17 used on a mendel to feed the extruder?

Edited 1 time(s). Last edit at 10/20/2010 09:37PM by Watanabe.

Discussion seems to have moved to reprap-dev.

http://reprap.org/pipermail/reprap-dev/2010-October/001641.html

Which makes excellent sense, since it's development.

Welcome aboard, Alexander, honored to have you.

---

-Sebastien, RepRap.org library gnome.

---

Remember, you're all RepRap developers (once you've joined the super-secret developer mailing list), and the wiki, RepRap.org, [reprap.org] is for everyone and everything!