@Richard_Horne I’m in for a go at this, but I do have reservations as to how feasible it is, particularly for smaller more detailed prints.
Retraction is obviously an issue. It would be cool to be able to linearly actuate the screw as well as rotation. This could give you a functional equivalent to retraction a bit like how injection molding machines hydraulically push melt into molds. But backwards.
I have a large cartesian machine as a project where size/mass would not be problematic.
In an ideal world it would be great to get smaller pellets which would mean smaller screws and details would be possible. However that rather eschews the purpose of being able to take normal IM resin pellets and print with them.
For mass pellet transfer in factories lean phase conveying is used and could be useful here.
I’d echo +Sanjay’s comments - problems to solve: small melted volume, feed, screw, weight, retraction. Might be best to go for a larger volume printer, larger nozzle, and work down.
The David videos don’t show parts being built that need retraction. Simple parts with no breaks, no travels over open gaps.
I’d rather have a hobby extruder be a separate machine unless you can get quality, fast and sharp retraction or have some means of controlling ooze. So that should be in your design goals. I think you want to design the machine around the extruder like David, because it’s going to be a heavy extruder, compared to moving the build platform.
With a large extruder and feed apparatus, it might seem logical to fix that and move the bed, but the main advantage, to me, of pellet extrusion is to do large volumes. You might be moving a larger mass than the extruder. I think filament is better for smaller printers, say, under 12x12x12. Should we think, say, 18x18x18 or larger for this, and try to get the extruder mass down and use a fixed bed?
I’ve seen automatic filament winders, that turn on a motor when there’s enough filament coming out of your filastruder. Why not reverse the idea? Filastruder -> slack line of filament -> regular reprap. When there’s too little slack, the Filastruder is called into action to produce more - filament on demand.
You could even build it such that the filastruder is cold most of the time, depending on how much slack you can put in the system.
@Paul_Chase You can’t do this because filament extrusion is a steady state process - all the process variables (melt pressure, melt temperature, etc) must be steady for the output diameter to hold constant. By starting and stopping, the Filastruder will ooze and filament will stretch, resulting in wild filament tolerances…
@Tim_Elmore1 Hmm, that means the pellet extruder might do the same - you’d need a pressure feedback loop as well as a temperature feedback loop, wouldn’t you? I have your Kickstarter Filastruder but I haven’t ever had time to put it together. Have you got a pressure feedback loop in it or does the Filastruder (as I assume from your comments) depend on a steady state flow?
A (printer mounted) pellet extruder would absolutely do the same, but we can tolerate that! Filament must be uniform in diameter (assuming there’s no inline measurement) for a printer to work, but we deal with ooze from a printer hotend by retracts (and corresponding primes) all the time. It can be handled in that case, because the output of a printer mounted hotend does not have to be constant in diameter.
The Filastruder definitely does not have a melt pressure loop. A melt pressure sensor would cost a few times what the Filastruder costs. It is not sensible to include at this price point.
Rich, I love this idea and would love to see these open source designs and build one myself. I also have a thought about a very light-weight extruder for your 3DR Nano. Contact me if you’re interested.
As @Tim_Elmore1 pointed out earlier, the filament would need to be extruded uniformly and that wouldn’t happen here with the output of the extruder going into the bowden and getting pushed and retracted all the time.
