Tuesday, May 20, 2008
Partial replication. First functional parts printed!
"If I can build a RepRap Darwin, I can build almost any machine."
Because of the experience I got from it, and because of the fabrication possibilities it creates. This was one reason for me to start this project. There are many more, but I wish to share results, not philosophy (at least not now). Ten days ago I had completed the build of the machine (with many time lapses of builds of sub-assemblies). Yesterday I printed a pretty decent mini-mug. This was proof-of-concept 'working', but the prints were of mediocre quality (nothing was configured or calibrated yet). Now I'm able to print functional parts for more machines! (And I even had other things to do!)
My first functional print (see pic on the right) may also offer an inspirational way to make the undocumented Z-EndStop (the Z-origin is documented)
The part required some cutting and a drill to make the holes (for the screw) that should have been there. Right now, where my RepRap attempts to print a hole (nothing should be there) it prints more than anywhere else: a complete fill with plastic. This is because the outline is really thick and because of the curvature takes longer to process than a straight line. I thought that AngleSpeedFactor might need to be lower, but this didn't change it much (yet).
I'm printing in HDPE, and through the time lapses I make you can clearly see the material shrink causing the warping. If this happens a lot, the object will detach, even with my sandpaper-hairspray combination to adjust the stickiness. Also in this video, you'll see that the X opto flag is perhaps too thin to print? It doesn't attempt to print a layer that is too thin? Any ideas? The part was not designed to be printed on its side, which it could be (as a work-around).If you like, take a look at the RepRap section of my Blog.
HDPE is the most difficult polymer to work with. I never got it to stick to anything satisfactoraly except PE or PP chopping boards. Even then large objects need a raft which the RepRap s/w currently does not support.
I think the host slices starting at half the layer height so I expect if you have that set to more than half the object height you will get nothing.
But I guess I need more than that to achieve results like you have, nophead :)
I just found out that it sticks reasonably well to ordinary 'duct tape'. I'm not sure how any of this compares to a raft because I have to design and print one to know... and the question is, how will I make the raft stick ;)
Yes I think somebody else got it to stick to duct tape. The problem is that it will not be able to resist the warping forces for anything other than a small object.
The raft works on a chopping board because you can weld HDPE to PP or PE. The problem is that if you weld it strong enough to resist warping you can't get it off again. That is where the raft comes in. It overlaps a bit so it spreads the load at the corners so it is able to hold the object down with a less strong weld.
I haven't tried my peelable raft technique with anything other than ABS yet but expect it to work with other plastics with just a change in bed material and temperatures.
For HDPE I would use a PE chopping board and start with 180C for the first 1mm layer and 240C for the two 0.5mm layers. Bottom object layer start at 180C and subsequent layers about 220C. Just guesses, would need a few hours of experiments to get it right.
For PCL and PLA I would use a balsa bed plus a raft.
My layers are currently a bit thick. The orifice is about .5 mm (drilled M6 nut from bitsfrombytes) but I get 1.2 mm extrudate out of it. I guess that will matter a lot for the output resolution, but it will also complicate the speed vs. durability due to pressure (e.g. breaking out the PTFE barrel, motor slippage, etc). I have a chopboard that I could barely reuse after a print, but a hacksaw may still do the job. I stopped using it because it was concave from the dishwasher.
I am building a Darwin at the moment. I expect it to give results at least as good as HydraRaptor otherwise there would be no point and I don't think it will take long to get there.
Vik and Adrian are getting good results with PLA and PCL. HDPE is never going to be as good.
The only advantages HydraRaptor has over a standard Darwin are that it has a shaft encoder on the extruder and it doesn't have any comms delays.
To get fine filament you don't need to make the hole smaller, you just move the head about 4 times faster.
My machine has a 0.5mm orifice. It gives about 1.1mm HDPE filament in free air but I stretch it back to 0.5mm. I use a layer height of 0.4mm and a filament pitch of 1.2mm.
Plug the same figures into Darwin and it should produce the same results.
BTW, you can get HDPE from New Image Plastics
They sell it in 5 lb rolls for about $8/lb. Donna answers the phone there and Jim Waring runs the company. Both of 'em are great people. They make a very high quality product and last time I checked will colour your filament for free.
Great work first of all and very fast progress!
The preference you are looking for is SpeedUpFactor and SpeedUpLength.
I ran into the same thing a while ago when I thought playing around with AngleSpeedFactor would solve my problem. It did not because it is another thing.
BTW, I assume you are using this version of host: 0.8.3. If you take a look at the jar file you will realize that it does not contain the SpeedUpFactor and length yet. So you need a new version and once you have installed it you need to add the two parameters manually to the parameter list which is this: .reprap folder in your home directory under Win XP. I am not sure how you do it under Linux as I use XP but it solved my problems and my quality is quite fine without any warpage with CAPA.
I am working currently on supplying the filaments BTW. As planned three filaments will be available: PCL, ABS and PLA.
The extruder tool is ready, we have a 15m long extruder machine which will produce the 3mm filament 19-20/min. I hope to run it soon and solve the material problem...
He's kindly proving our theory that the RepRap will mutate. Look, it has sprung a new part!
The build quality is a tricky thing to balance, but once you get it right you know you've hit the sweet spot! I try not to ask much of the first layer or two except that they anchor very firmly. Any errors are soon bulldozed flat by the next layer.
Links to this post: