What is this witchcraft

[jimc]

haha thats funny. you think they would have pulled the tape off there before doing photographs for the home page of their website.

[jsc]

I'm still interested in getting an easy filament splicer to do that sort of thing manually

Just print a jig and use it as a guide to superglue it. According to the internet, the melting point of set superglue is 187°C, lower than any print temperature. If you are lucky, it will flow right through the nozzle. If not, then it will be the perfect nozzle obstruction. I would try it now except I am not at home and don't know where my extra nozzles are.

It does seem plausible that might work. I've found various numbers for cured cyanoacrylate melting temperatures, but they're all below 200C. Anyone feeling brave/lucky?

[insta]

I can try it, I have an already jammed up V3b that's getting blowtorched here soon enough.

[jdacal]

Would it be possible that the slicer code (USB cable) is plugged into the filament box instead of the printer. The box would then generate a new gcode synchronized to its color changes and feed it to the printer via its own USB connection?

[jdacal]

Would it be possible that the slicer code (USB cable) is plugged into the filament box instead of the printer. The box would then generate a new gcode synchronized to its color changes and feed it to the printer via its own USB connection?

EDIT: Oops! Pandelume had mentioned something similar, missed that post.

[mitchfrommosaic]

I may be wrong but it looks in essence to be a intermediary box between the extruder and the filament. Multiple colors of filament go in and a single comes out. My guess is then that the print is actually fed into the new box which is in turn connected to the printer itself. As the print progresses the third party box switches constructs a filament that corresponds with the required color

For example 20mm Red -> 10mm Green -> 100mm Yellow

Interesting idea but it will very much depend on how their software integrates with the current slicers. This is an entirely new layer above the gcode itself.

EDIT: Looks like they say the box doesn't connect to the printer directly ... they talk about a sensor that tracks the the filament movement. Now I am as confused as you Insta. No idea how the hell they determine where the print is at.

-------

Hey Bratag, I'm Mitch, a cofounder at Mosaic. You're right, the box ( The Palette) sits between the printer's extruder and the filament. It takes in up to 4 filaments (with different colours or properties) and through an automated process, cuts, and splices them together to form a 1.75mm continuous monofilament. The printer can pull in this filament just like it would off a standard spool of filament. The Palette does not connect to Rambo directly - in fact there are no hardware, or firmware changes that need to be made to the printer (except for the addition of a piece of velcro).

The Palette has a periphery device called the scroll wheel that velcros onto the side of the printer. The printer pulls filament through this high precision encoder which delivers the print progress information back to the Palette. This is how the Palette is able to stay in sync with the printer to deliver the right filament to the hot end at precisely the right time in the print. To account for skipping or unaccounted for over/under extrusion the algorithm has been designed to be adaptive. What this means is that if the g-code tells the printers extruder to pull in 200mm of filament but instead it only pulls in 195mm, the scroll wheel will figure this out and actively compensate for this in future calculations. It does this by comparing theoretical g-code lengths to the real measured lengths form the scroll wheel. This active compensation completes a feedback loop that allows the Palette to stay on track.

Integration with slicers has been done by leveraging the already existent multi-extruder options in most slicers. In Cura for instance, you import your files as separate STL's and assign each to a different extruder. You then slice the part to create G-code that would be suitable for a 4 extruder printer. We have developed a plugin that takes this G-code and pulls out the filament change information like you described above (20mm Red -> 10mm Electrically Conductive -> 100mm Yellow). It then removes all of the extruder change commands from the G-code so that the printer does not try and switch to different extruders during the print (good thing as it only has one). All of this happens behind the scenes in the slicer so that all you need to do it hit slice and you end up with this multifilament single extruder G-code, as well as a .SEEM file that gets inserted into the Palette with all of the filament change information (via provided SD card). Although this process is quite universal and will work with any printer that runs on standard G-code or x3g, we will be releasing the firmware and software open source to allow anyone to tweak it to their needs.

I hope that answered most of your questions.

[mitchfrommosaic]

That piece of tape on the extruder that says RW...those are Rob's initials. That is an internal QC tag from a batch test and should have been removed prior to shipping!

Haha this is the best part. Is Rob selling parts on the side?

Haha that is funny Rick. That tape was on the original M2 Hotend and we just never questioned it. The M2 was our first of many printers and is still my personal favourite. The geared extruder is much more reliable then some other simple direct nema 17 extruders and makes for great performance with The Palette.

[SouthSideofdaSky]

It looks like these guys just started a Kickstarter today at noon and they are already almost to their $75,000 goal.

Whoa.

The new Kickstarter page answers a lot of questions I have but the biggest one is still...is it actually going to be reliable? Would be awesome if it was.

https://www.kickstarter.com/projects/13 ... ng-evolved

[Bratag]

850 bucks though ... I could put on a kraken for considerably less than that. Was really hoping for a better price point

[mitchfrommosaic]

It looks like these guys just started a Kickstarter today at noon and they are already almost to their $75,000 goal.

Whoa.

The new Kickstarter page answers a lot of questions I have but the biggest one is still...is it actually going to be reliable? Would be awesome if it was.

https://www.kickstarter.com/projects/13 ... ng-evolved

Hey SouthSideofdaSky

Reliability has been a focus from day one. Nothing more frustrating then returning to a print after a few hours and seeing a plastic mess or stalled extruder and I think we've all been there. We needed to ensure that The Palette would not be another source for failed prints and so reliability is always the top design constraint. This meant designing The Palette's filament drive systems so that filament passes through reliably with no chance of jamming as well as adding several sensors into The Palette to provide feedback (to name a few).

To put some numbers to the prints you can see on the campaign, the orange, red, and yellow pyramid was a 6 hour print. Other more complex prints had hundreds of colour changes meaning, The Palette reliably spliced filament hundreds of times live during the print. Of course reliability is hard to quantify until you have a production unit tested for a few months but I hope that this gives some perspective as to where we are with the Palette now.