Makerware vs Replicator G - which is better?
What I was after from my Replicator 2 was fast repeatable print with a good surface finish. Basically I plan to use it to create models of products I am designing so I can demonstrate design concepts to others.
Makerware worked first time I installed it without a hitch. It sliced my models and the medium and fine prints all worked well - never tried the course setting. However I was looking at the stretchy bracelet model included on the SD card and quickly realised that Makerware in its current form could not generate this type of print. So to satisfy my curiosity I downloaded ReplicatorG to see if I could figure out how it was done.
At first I could not get ReplicatorG to work at all. It would seem to slice and dice the model but there was no GCode at the end. After a lot of stuffing about thinking it was the profile settings that were wrong I updated my version of python, tcl and reinstalled ReplicatorG - from then on it started behaving.
Another newbie mistake I made was to select the PLA profile in the material settings - it seems that the PLA profile is bogus and only the ABS one works - even though we are printing in PLA - go figure.
ReplicatorG has help buttons but none of them seem to work on my Mac installation. So no luck accessing documentation form within the app itself. The documentation on the web for ReplicatorG and particularly Skeinforge (the slicing engine) seemed all over the place, one site would recommend doing something and another would contradict it - so I was starting to feel like it was too much work to figure this thing out. However, then I read an article on the MakerBot website that said that the best documentation for skeinforge was embedded in the python code itself, and after digging around a bit on my mac I found what appears to be the compiled documentation from the python code embedded within the application itself. You can find it here
It is well worth a read.
I have zipped up the current profile configuration I am using. In ReplicatorG under the GCode menu, select Edit Slicing Profiles and click on the Locate button. This should open a finder window to
~/.replicatorg/sf_50_profiles/
You can unzip my profile into here is you want to use it.
Firstly, here are the Print-O-Matic settings I am currently using.
Object fill (%) = 15
Layer height (mm) = 0.2
Number of shells = 2
Feedrate (mm/s) = 80
Travel Feedrate = 150
Print temprature = 230
Filament diameter = 1.69 (measure your own and use that)
Nozzle diameter (mm) = 0.4
You will notice that I am only using a 0.2mm layer height - this is because I am using the skin module so the outer surface is printing at 0.1 but the internals print faster at 0.2 you can even bump this up to 0.25 with little visible difference to speed things up even more.
Here is a list of the main things I have been mucking around with.
Fill
If you want a work piece that is strong and light try adding some internal filled layers with
Diaphragm thickness = 15 (set to zero if you don't need this)
The main variables to consider are
Extra shells on alternating layers = 1 (this should be different to the base layer to help stitch the infill together)
Extra shells on base layers = 2 (this is controlled by the Print-o-matic settings)
Extra shells on sparse layers >=2
Grid Rectangular (selected) seems the fastest strongest option infill style
Solid surface thickness >= 3 (if you want a solid base on a 0.1mm layer print 5 works better)
Start from choice = Nearest (I know they recommend lower left, but I find that leaves a seam)
Jitter
Jitter over permitter width = 2 (this helps remove seams)
Raft
Even if you are not using one - leave the module activated
Skin
I tried using 3 divisions but I think 2 gives more reliable results on curved surfaces. On flat walled objects, 3 divisions and a 0.3mm layer height would probably work quite well.
Horizontal Infill Divisions = 2
Horizontal Perimeter Divisions = 1
Vertical Divisions = 2
Attached is a photo of a Makerware medium 0.27mm layer height print (BLUE) and a ReplicatorG 0.2mm print skinned at 0.1mm using my attached profile.