Welcome! Log In Create A New Profile

Advanced

Mosaic Manufacturing's multi-filament add-on for single extruder printers, Anyone else excited?

Posted by Feign 
So not long ago I stumbled across an article about a flashlight printed as one piece using conductive filament (which is already exciting as is) and a machine from Mosaic Manufacturing. I've been looking around and have found almost nothing else about this machine except that Mosaic calls it "SEEM Technology" and is otherwise keeping rather quiet about it for now. I'm rather excited about the idea of pairing a machine with say a cheap Delta or something and having the same capabilities as a multi-extruder machine.

Any ideas on how it might work mechanically? I get the general concept that it's precisely cutting and splicing lots of short segments of different filaments into one filament and is somehow calibrated to know just how far ahead of time it needs to do so to make the switch between materials down the line... How is it making those cuts and splices while the filament feeds through it?

And if we figure out what parts are and aren't patented, how feasible would it be to make a DIY version of the same kind of setup?
I bet they are just keeping the printer-end of the filament running at a constant rate thru a tube but then periodically cut it off, and then jam the pre-melted end of the next filament thru the same tube into the trailing end. Or something like that. It may actually be a pretty simple mechanism. (Patented or not is neither here nor there for DIY'ers.)

I'm more curious about the planning & software. If the color splicing pattern is completely pre-determined, it seems like any irregularity in the extruder feedrate would accumulate into misaligned colors. If the Mosaic box was acting as a bowden extruder, this wouldn't be a problem, but in their video, it looks like the printer has its own independent direct-drive extruder.
I'm thinking it gets the amount of filament per color per layer (by mass rather than volume) It looks like maybe it has an option of making a small cooling tower in between each layer to handle the transitions between one filament and the other, from the looks of a few parts. That might be able to handle slight changes in feedrate due to variations in the filament diameter, since those would average out over time.

What if there is a sensor in the machine that can tell how much filament the printer is pulling from the machine, it might be able to correct for errors then?

If the printer is feeding erratically or downright jamming, then there's probably bigger problems than color misalignment, right?

I didn't think of it keeping the idle filament hot while the active filament feeds, that probably would simplify the splicing...
I've had circumstances where the filament is not feeding properly (like from a misbehaving spool) but then corrects itself. These circumstances might be difficult to handle.

I'm guessing, though, that you are right and there is some way to monitor the actual filament feedrate. If the Mosaic box feeds the filament under force control (rather than position control) providing a slight positive bias force for the filament, but then monitors the actual feed rate, it should be able to keep the colors in sync with the X-Y motions. Of course, it would need to be monitoring the G-code execution. Maybe the G-code is first fed into the Mosaic box which then passes it onto the printer, perhaps letting the printer's G-code buffer empty out periodically at safe times (maybe over their cooling tower?) to create sync points?

Of course, this is a lot of wild speculation - I suppose I could just ask Mosaic how it works.

EDIT: Actually reading the Mosaic FAQ is useful: the Mosaic box does not connect directly to the printer. Also they do have a sensor which measures the filament movement directly. Still not clear exactly how they keep things sync'd

Edited 1 time(s). Last edit at 12/19/2014 01:33AM by LoboCNC.
I am willing to bet that the Mosiac has a Software pass through to allow it to monitor/modify the gcode on the fly.

Also, there is probably a "calibration object" that has to be printed on initial setup, and you tell the software the calibration pattern, and it can adjust for feed length and rate.

Then, the Mosiac software would know exactly how much it has offset the color splice command (in the "future" of the gcode command stream)

I would also expect that the generated gcode will print a "clear and verify object" before the main print, that ensures the mosaic is synced (ie the word "mosaic" on the bed, where each letter must be a different color) so the feed path is of a known color pattern before the first build move, and that the user can do an early abort if the check pattern is incorrect.
Sorry, only registered users may post in this forum.

Click here to login