Extruder Design and Build Speed

Extruder design looks awfully tricky to get right. Frankly, I have trouble seeing how a .5mm extruder, laying down material at a "fast" 60mm/S, could approach satisfactory build speed.

Unless I have miscalculated, filling the RepRap's 430x430x100 mm (that's 18,490,000 cubic mm) work space 50% with material takes 8.92 DAYS (the 24-hour type) at 60mm/S. This seems impractical by orders of magnitude. That assumes (over optimistically) that the extruder spends zero time moving without laying down material.

Most product layers are, of course, wider than .5mm. A layer of 50mm width takes 100 passes to create. How about an extruder with an exit slot (instead of an exit hole) of programmable width? For example, a minimum width of .5mm, and maximum width of 128mm, could be controlled with a single byte in .5mm increments. If product geometry permitted an average extrusion width of just 50mm, build rate could be increased 100 fold.

Edited 1 time(s). Last edit at 06/21/2009 11:12PM by jdoll.

Build speed is a bit of a stumbling block, yes. Build times for larger objects can reach into the 10s, 20s, and 30s of hours even on commercial Stratasys machines. 2 and 3 day builds have been done on commercial FDM machines - "rapid" prototyping can be something of a misnomer, it's really only rapid compared to the alternatives.

The 8.92 days figure, though, is unrealistic for a real build. Besides the fact that a 430x430x100 mm block of solid plastic is relatively useless (not to mention expensive), it's doesn't take advantage of the quirks of the additive manufacturing process. Even if you needed a shape of plastic like that, your actual extruded volume could be closer to a third of that, using a sparse infill strategy for the inside of the object. Realistically, objects tend to be much, much, smaller than 430x430x100 build volume, and are almost always made with sparse infill. Now, the idea of being able to change the size of the extruder output hole on the fly is awesome, so if you think you can design it, go for it. Everyone will love you for it forever, because you're right - it would dramatically decrease build times. Otherwise, the workarounds in place are sparse infill and the ability to stretch and expand extrusion width by varying movement speed; and constantly pushing the maximum extrusion speed limits (with the advent of firmware featuring acceleration, the maximum speeds for longer build segments could be conceivably be pushed up by an order of magnitude, vastly accelerating the infill process).

I suspect the reason objects tend to be much smaller than maximum is that nobody wants to wait around days on end for a result; it cannot be that larger objects are useless!

As an electronics/software guy, I am probably not the best candidate to do a complete design. However, I have attached a concept sketch with proposed dimensions for critique by mechanical experts.

The concept is based on the present extruder. Imagine the extruder barrel stretched sideways, so that the extrusion hole becomes a slot. Add to this a "shutter strip" 0.7 x 0.2 x N, which can be positioned to block any portion of the slot by an external motor. N is a bit longer than the slot length. Slot length is not critical. A longer slot means faster builds, but more ungainly extruder dimensions. My initial thought of 128 mm is probably too large. 64 mm might be better. If the shutter strip is positioned to block all but .5mm of the slot, the extruder should operate about the same as the present fixed extruder.

The shutter strip must enter through a slit in the side of the extruder head, with tolerances close enough that melt will not leak out. (I am uncertain if this is feasible without additional provisions.)

To accommodate retainer slots for the shutter strip close to the extrusion exit, I reduced the nozzle draft angle to about 30 degrees. This leaves about .2mm wall thickness at the weakest point. If this is not enough, the edges of the shutter strip could be beveled.

Edited 2 time(s). Last edit at 06/21/2009 11:05PM by jdoll.

Sorry I may be totally off base here (had few this evening, so totally random firings of responses, isn’t that a British thing?)!

I do not think the design has any problems with a half millimeter extrusion, what we potentially need is something between half a millimeter and three millimeter full flow plastic welding extrusion for building objects.

So maybe assume a simple six by six matrix that touches down in a Z axis (why would we need this)?

Or to simplify it further full flow verse single channel of half a millimeter for internal spaces?


Run with it everyone?