TLDR; I Started a BoxTurtle MMU Build and am pretty impressed so far.
About a year ago I started a survey of opensource MMUs and found most to be overly complicated for what I was trying to accomplish - 2 material switching (infinite spool / two-color / primary and support material).
I wasn’t thrilled with what I found and started to go down a rabbit hole of cobbling one together.
I didn’t get too far before the 3DChameleon went open source and I printed the parts for that and built the drive / selector assembly mostly from parts on hand (bought a few bearings). I was - and still am - impressed by how well that mechanism works. I was still going to “dumb it down” to just two lanes, but it doesn’t really have a intrinsic spool management solution which was driving me to look around for options on that front.
At that point I ran across the ArmoredTurtle/BoxTurtle project. It has 4 lanes in the base configuration, but each lane is fully independent with its own drive (and no selector) and powered respoolers. It is one of the more expensive (4 lane) options for a few reasons, but I reasoned many would be mitigated by only populating 2-lanes and re-purposing a controller.
It turns out that the components for lanes 3&4 are inconsequential and I’ll probably end up with the recommended special-built controller anyway, so I am knee-deep into a 4-lane build.
Impressions so far:
I started skeptical, and printed some of the parts to make sure I was going to have adequate quality / tolerance for the build and was very impressed with the the part design. The files are per-orientated and have supports modeled in where necessary. They are well thought out and remove easily.
The Github repository for the 3D printed parts is a bit hard to navigate, but they have an amazing “configurator” tool that walks you through some questions and then creates a zip file with all the necessary .STL files and no extra “optional” stuff to confuse you.
Part names are clear, indicating the color (for a base/accent color build) and quantity of instances required.
I have already adapted a BIQU H2 Filament cutter to my MP3DP-V3. It appears to work okay, but the ArmoredTurtle project has a “FilamATrix” derivative of the Voron Stealthburner I may try to adapt.
Running down that rabbit-hole, I stumbled across the NightOwl which is as close to what I originally set out to build as I could imagine.
If I wasn’t so far down the path I would probably do the NightOwl, but now I’ll probably finish the BoxTurtle first and build that too for a comparison. My MP3DP-V2 is still working fine and had a Klipper conversion anyway…
I looked at the Box Turtle but the $300 in parts threw me off. I love the idea but cost is a big factor. I am doing the TradRack now. Just finished the V5 and made some mods to accommodate filament cutting.
Wish i had found that H2 cutter sooner. I do feel like it will sacrifice some X travel but maybe not…
I think I lost about 8-10mm on X with the cutter. It’s a little bigger than that but I had extra travel on one side that I leveraged by shifting the extruder over a little. The primary interference is the cutter arm blocking homing in the y direction. I think I could home x, pull-off x, then home y to regain nearly everything I gave up.
My initial goal was a very low cost 2 spool option. As you point out this is one of the most expensive ones.
I had just enough stuff on hand to make it a little cheaper to self source than get the kit, but the big advantages were
I could purchase in increments as I built confidence it was going pan out rather than go all in on parts that clutter my junk drawer if I quit on it.
Buying bits and pieces spread the coat out so I could stick my head in the sand and just not tally it all.
Self sourcing usually ends up with “extras”; ie order a set of 10 bearings when you need 8. This makes the per part used cost look less even if I actually spent more to stock the extras that will likely never be used.
I looked pretty hard at the TradRack and printed a few pieces. I wasn’t happy with the quality of fitment of my parts so I didn’t go too deep.
The issue is my print quality, not the design, but I wasn’t too excited about all the tuning I was going to have to do. The BoxTurtle parts seem to fit okay using my default settings.
It is similar to the TradRack in as much as there is one feed stepper and one selector stepper, but it uses one cam shaft with 4 cam lobes rather than a moving selector and one cam lobe.
Very similar design philosophy and I suspect it will be fairly easy to build. It is significantly cheaper than the BoxTurtle, due to the reduced motor count that allows for simplier and lower cost controller.
It is also cheaper because it does not have any spool management, but neither does the default TradRack.
I actually just finished building a Happy Turtle Lettuce Feeder. I have been going round and round with the tradrack and the Happy Hare installation. The pause when it changes filaments seems to come and go randomly. I wanted to try a new setup and see how it did.
Have the thing build and contemplating where I am going to install it. Thinking it may go on the V4 and I will install the AntHead on it to allow cutting. But them I kind of want to convert the V4 to a V5… but I dont want to rebuild the frame…
I did the dual drive and the BTT encoder mods. I have changed to the AntHead toolhead with the Leafcutter filament cutter and gotten Happy Hare sorted out. It is pretty reliable now. I did have to take the encoder apart ones to tighten it all up and add some thread lock. Note: I only use it for the filament switch and have not implemented the encoder part of it in my install.
The Aforementioned HTLF is still not actually hooked up on the V4 yet. I have the new AntHead built for it but it prints PLA so well that I have not wanted to change anything out quite yet. I plan to use AFC on it as that is what it was designed for.
Rewind for the TradRack is done with spring spools inside Polymaker Polydryer boxes. They work pretty well once you figure out that you have to preload them some. The HTLF has 4 Filamentalist rewinders built for it.
