Software Changes for FDM + Milling Technique for Higher Printing Accuracy

I am currently a senior in Mechanical Engineering and will be doing my senior project in the area of entry-level rapid prototyping. One of the techniques I want to investigate is the use for milling to improve the accuracy of FDM printing. Here is an example of how it would work.

The FDM process would proceed as normal, however, the software would add something like 0.2mm to the outline of the part. Then once the layer has completed printing, a very fine milling head will mill the outline of the part for superior accuracy. Once it is done, the excess material will have been removed and the part dimensions will be as they were originally intended.

My question for any software gurus out there who are more familiar with the Reprap setup is how would I go about implementing this procedure into the software/firmware? I think I can handle designing and building the machine to have 2 toolheads (one for FDM and one for milling), but a simple way to add this into the software has not occured to me yet? Any ideas?

I'm doing the same thing, but without an other mill, I'm just spreading the plastic thinner (down to .01mm) the head from the tip of the nozzle has enough thermal energy to melt and spread the layer thinner and to do so accurately.



what i ended up doing, and will be releasing possibly this weekend to everyone is this. in skeinforge the code at the thinnest thickness i want it to be. meaning .01, .02 or whatever the final thickness is to be. then i programmed it to look thru the g code, and rewrite the gcode with 3 passes 3x thick, 2times thick and 1times thick. each pass i calculate z a bit lower the third pass it is the final thickness. this is done so the stress on the extruder is minimized.

I am more than willing to provide this code for you or show how you would do it yourself. basically parse the gcode output. look at what z is, and say awe that is the final thickness, let me add .2mm to it, run the extruder, then go back to the same line thru gcode and rewrite it with extruder off, and set x+offset y+offset of your mill bit and m instructions to turn mill on and then set z to processed height, then machine plastic. i should warn you that milling plastic that is already thermally stressed will cause it to heat to melting point. so be careful. the plastic will be already 150 deg c or higher and have no way of dissipating the heat from the drill bit.


James

It's called Shape Deposition Manufacturing. (At least by some people.)

Here's a Stanford Lab that's done a lot of work on it.

[www-cdr.stanford.edu]
[www-cdr.stanford.edu]

[bdml.stanford.edu]

---

It's a very cool subject which we'd be very happy to have under the RepRap umbrella. It would mean lots of tiny gears, or gears which mesh nicely etc.

Here is a google scholar search on "shape deposition manufacturing"
[scholar.google.ca]

Let me know if you guys want me to give SDM its own forum.

Here's a place where we can keep notes on the wiki. (Just log in, click "edit")

[objects.reprap.org]

---

What do you think you're going to do hardware-wise?

Edited 1 time(s). Last edit at 01/15/2010 09:40PM by SebastienBailard.

Sebastien: For the hardware it will probably be our own design for some type of CNC/rapid prototyping hybrid. It is going to need to have a lot more mechanical advantage for some of the other tasks we have in mind for it. As far as how we will handle the two different heads, we have looked at a basic toolchanger setup as well as options to simply give each of the heads their own Z axis so the ones not in use can be moved out of the way. We are still a little early in the design phase to make a decision on that yet, but I will be sure to update.

James: I was a little confused by your post. Do you actually have multiple extruder heads each with a different nozzle size? If so, yes, I would be interested to see how you handled that with the software code as I am sure I could easily adapt that to our concept.

James,

Think about making a 'generic cnc router' with a gantry, but mount the extruder on the far side of the gantry, so they can't interfere with each other.

Toolchangers are cool, but hard. Do a design that lets you fit one in if you finish the other stuff early.

Hell, you could buy a taig, bolt an extruder next the spindel with a 6 inch horizontal spacer, and go crazy, focusing on 'bleeding edge' rather than 'another cnc router'

---

Also, I am salivating at the thought of a good RepStrap in the wiki, so I urge you to upload everything you like. Because we need 'another cnc router' badly.

Let me know if you want your own forum like these guys.
[dev.forums.reprap.org]

Sebastien: I'll talk to the team and see if they think a forum would be a good medium to document some information on our build. I used my blog at cpwebste.blogspot.com last time and that worked pretty well... Either way, we will find some way to get some info into the public domain.

Also we have debated between using a toolchanger or just having multiple heads on the same X axis carriage. If you do go that route, you would need to have independent Z axes (assuming the extruder moving up and down is the Z axis), right? You might be able to get away with having the 2 at the exact same height but then if you tried to mill a layer I could see the extruder running into things. But again, the challenge I think is going to be how to handle 2 extruders or 1 extruder and one spindle in the software/firmware. You will have to take into account that the 2 heads will likely have different XY positioning as well as a different Z height. That's where the majority of the new territory is going to be.

If anyone has any info about how to deal with 2 extruders or a single extruder and one milling spindle, please post it here or PM me as it is always helpful to learn from past experiences.

Just a thought: If you're trying to avoid the milling head bumping into the parts during the extrusion cycle, you don't need a continuous Z-axis, right? You just need to be able to flip the mill up when it's not needed, and down when it's in place (not set it to variable heights). So in that case, it might not be difficult to simply use a solenoid, or similar setup.

Raise the bed of the workpiece a bit. Then you can have the extruder bolted to the spindle, with a ~6 inch horizontal separation. When the extruder is moving around extruding on the workpiece, the spindle.is cutting air. When the spindle is cutting plastic, the extruder is cutting air.

"cutting air" = not whacking into deposition bed or workpiece.

I think that's right, anyway.

Sebastien: I'll talk to the team and see if they think a forum would be a good medium to document some information on our build. I used my blog at cpwebste.blogspot.com last time and that worked pretty well... Either way, we will find some way to get some info into the public domain.

I'll be ecstatic with every bit and byte that's added to the commons. [[Shape_Deposition_Manufacturing]] is a good place for notes, links, video feed, a full new formal machine, etc.
[objects.reprap.org]

Even a engineering report in MSWord is great, although I'm more in terms of the ' engineering report on RepRap.org' style. Also, I can lock down pages of the wiki so that only you have the key. That makes it easy for your Mentor to see that "yup, James did all that original work there."

cpwebste, how is the project coming along?

Just updated my blog with a little info about the project. It have been codenamed Hydra due to the fact that we hope to be able to use multiple heads simultaneously. Hopefully lots more to come over the next few weeks!

Edited 1 time(s). Last edit at 02/22/2010 04:38PM by cpwebste.

---

[cpwebste.blogspot.com]