Combination of FDM and milling

Hi all,

with my old CNC-mill i want to combine the additive fabbing of a FDM-extruder with the subtractive finishing of a CNC-mill.

In a german CNC-forum i found a simple methode for building very accurate housings and fittings - here i have some images ...

1 - build (reprap?) the inner core and finish with a mill:



2 - apply the material for the housing (maybe with the reprap too?):



3 - after curing finish the surface with the CNC-mill:




4 - separate the object from the core:




Viktor

Edited 1 time(s). Last edit at 02/09/2008 06:49PM by Viktor Dirks.

Nice.

I have been thinking about the advantages of each layer of FDM being followed by a milling pass as an object is built up. You get the additional accuracy of milling combined with the ability to make shapes impossible through milling alone. The milling only removes a small amount of material so far less wastage than normal subtractive milling and much faster. Also the milling bit only needs to be as long as an FDM layer depth, so it can be very fine.

---

[www.hydraraptor.blogspot.com]

Hi nophead,

... it's the same for the LOM-approach ( [forums.reprap.org] ), where i too have to mill only one sheet per step ...

So i'm actually accumulating different ideas, toolheads and software-concepts for my CNC.

I hope to show some more interesting tasks during the next months

Viktor

Yes but with LOM you need to assemble it afterwards and your z-resolution is equal to the laminate thickness as FDM's is limited to the layer thickness. With FDM+M the Z resolution can be as good as the mill.

---

[www.hydraraptor.blogspot.com]

Hi nophead,

... correct, but for LOM-fabbing i'm planning a "fast-LOM"-setup, where i would in a continuous mode roll a thin sheet of material (paper, plastic, 'green' ceramic, metall with 0,05 to 0,1 mm thickness) onto a cylinder, press/heat/mold the sheet onto and cut the separation- and contourlines with a scanning laser (or slower with a needle or mill).

Imagine a roll of paper, transferred on a second roll with continuous glueing+cutting, and when ready, you can break off the fabbed cylinder and extract the embedded objects - they should look like carved from wood, but should be as strong as wood too ...

Your accuracy is then equal to the thickness of the sheet and the cutting accuracy of your machine.

Viktor

Victor,
I don't understand your scheme. Suppose you want to make a cube: this would be embedded in the wall of a cylinder? How do you make the top and the bottom of the cube flat? I.e. how is anything made to separate in the radial axis? The laser cuts along the length and tangential to the cylinder?

Chris

---

[www.hydraraptor.blogspot.com]

Hi Chris,

here a sketch of some embedded objects in the fabbed cylinder:


The software have to calculate the object-volume fitting the actual sheet as in the plane slice'n'dice-approach, but in a cylindrical plane instead of a horizontal one.

The surface-accuracy of the objects is dependant of the sheet-thickness, so 0,1mm (normal paper) or thinner would be fine.

All the sheet-area not covered by the object-volume should be crisscrossed, so there would be separated square columns through the complete cylinder which can be outbreaked (look in the LOM-thread, it's like the hatch iside the mug).

For horizontal separation you can cut a much finer hatch-grid under the ontop overlaying areas, so you can separate them from the underlying areas too ...

Viktor

So presumably you cannot make some hollow objects this way because you have to be able to access the inside to remove the unwanted bits?

I think FDM+M with a soluble support material that can also be milled can make a wider range of shapes and has the advantage of being fully automated. The only post processing is dissolving the support.

---

[www.hydraraptor.blogspot.com]

Hi Chris,

... it's a bit tricky with hollow LOM-objects - when crisscrossing the inner volume you can crash and extract the inner bits.

Another way is to separate hollow objects and glueing them together after cleaning.

The main advantage of the LOM-fabbing is the possibility of building overhangs and hollow structures without a second support-material, as the sheets are completelly self-supporting over the building-process.

But i'm experimenting with some more ideas of mixing different RP-methodes, so two-material-fabbing is worth testing too, when i'll have some extruders or dispensers running.

Actually i'm on activating and optimising my CNC-mill, next is the diode-laser (with safety-box), then i have to finish my tripod-demonstrator and after this i can try with lathe-LOMming and paste-extruding ...

Viktor

OK

Milling and FDM in one machine at the same time.

Just been trying to figure out how to put some of this up in the builders blog.

Ianus is born (well the prototype is)

Ianus is a dual head Cartesian with Independent Z axis on a common X and Common Y platform. Ohhhh sounds clever but not really.

Just a gantry style unit with two heads planned one facing forwards and one facing backwards.

The tubes and fittings I wanted have arrived and I have assembled them sort of trial style in a space in the loft.

Piccy 1 initial assembly, Note working lunch on table.



Piccy 2 assembled and screwed down to base board gantry not tightened up yet as it need positioning etc.



Piccy 3 now onto the slides etc here is the assembled bearing thingys ready to mount on trolleys. The idea came from the Build it Yourself CNC guy that Sebastien mentioned what stars they both are.



Build is halted now until I put some more lighting in the loft (not to mention boards, how did that hole get there) and repair my knackered drill press. Gonna need it for the carriage trucks as I will be making these from standard steel sections mostly square tube and angle . V Cheep and strong.

I think I want to make the heads removable too so they can be used in whatever pairings work best for the job, Support + FDM, Support + UV Cure Resin, Mill + either deposition methods, Syringe & Mill. etc etc etc

Hope this is interesting, all comments welcome.

cheers

aka47

---

Necessity hopefully becomes the absentee parent of successfully invented children.

Hi Andy,

interesting setup

It seems to be really big, so the tubes can bend over the long range - maybe adding more rigidity with diagonal steel-wires would be good.

Good luck with Ianus

Viktor

Hi Andy,
I also was thinking on a smaller scale but Ianus made me realise that if the milling provides the accuracy then the FDM no longer needs to be a fine filament, allowing larger objects to made in a reasonable time scale.

Chris

---

[www.hydraraptor.blogspot.com]

Yup agreed

Some of the dimensions are a touch out as yet. This is what you get for over designing in your head without parts to play with.

The goalposts look wrong. Too tall perhaps. The Pipe-Clamps (Alvin Key clamps are the nearest equivalent with 1" bore pipe) I am using and the steel tube are really for safety handrails etc so are excessively over engineered for FDM, maybe even for CNC but give mass/weight generally considered to be a good thing in machine tooling.

As they are all clamps though I can take them apart and tweak away for a while before settling on final sizing. So long as it is oversize i can always cut some off.

They do clamps etc for angle's other than 90 Deg that could be separately pressed into service as wire fixings so steadying the upright is something that is easily doable as a later addition. I want to see how much it wags about once the dual heads and carriages etc are designed/fitted.

Overall I am wanting something like a 500mm X 500mm X 500mm workspace give or take a bit. The bed crosspieces are 500mm and the two side pieces 1000mm, the goal posts are made using 750mm uprights. I will probably need to make the base plate run out over each end by about 100mm or so to make sure that the end workspace (that which can be equaly serviced by both heads) is roughly 500mm x 500mm.

Overall though if it is a bit over/under this is fine for the prototype. If I ever get around to doing anything finalized with it I will be able to draw on some "Lessons Learned" and measure the dimensions I ended up with.

I don't like having the Z axis as a long piece because to me it feels like it will be prone to multiply the backlash/tolerance errors at the end of the tool head on the longest extension (towards the bed). But after studying the fab@home with two syringe heads I noticed the end of one was fouling the work of the other. So some method of moving either head towards or away from the workpiece preferentially is inevitable.

Thanks for the feedback guys, much appreciated.

aka47

Edited 1 time(s). Last edit at 02/14/2008 03:33AM by Andy Kirby.

---

Necessity hopefully becomes the absentee parent of successfully invented children.

... here i have some samples of 3D-milling with my old CNC-system - the rooster in the images is 70x60x10mm in size, but the millhead can move until 500x500x90mm.

Back-side only roughed:


Front-side finished:


This was a test-output with Cut3D from vectric, but with the basic-software from Isel i'm capable of 3D-constructing and milling with 3 or 4 axes.

I think with some experimenting and good hints from the CNC-guys it should do any parts for direct use or cast-molding i'll need, so there should be some progress during the next months ...

Viktor

Yet more excellence.....

Demented seems to be getting along well too.

aka47

---

Necessity hopefully becomes the absentee parent of successfully invented children.

Just out of interest, I worked on a project about 10 years ago where we had a cnc machining centre with a mig welder attached, we also attached the mig to a robot, both systems worked well but were a bit slow, the machining centre milled each layer, I think that Krup has a patent in this area but not sure, one of the first 3d printers, the Sanders modeller used wax through a peizo nozzle and had a milling cutter a bit like a cylinder mower to reduce the layer thickness.

Shinko Sellbic has apparently done a hybrid FDM/milling machine (from 1997, apparently): [www.sellbic.com]

They have some other neat things such as a filament-fed mini extruder head: [www.sellbic.com]

---

[haveblue.org]

Yes, thanks to you I've become a big fan of Shinko Sellbic. I wonder if they'd accept job applicants from overseas.