Better, faster, and stronger,
It's go time, fellas.
Modding on my Rostock has been on hold for nearly a half-year now, and I've briefly detailed my wishlist for a new design before, but I haven't found the time (or motivation) to make real, purposeful strides until now. My thoughts have expanded a tad, and it's helpful for me to try and summarize some of the priorities here.
I'd say that the guys/gals at Seemecnc are still at the forefront of delta-style, hobby-level 3D-printers, especially in the States. The alternatives, like Deltaprintr or DeltaMaker don't quite have the traction or anywhere close to community support. In terms of build quality (at a comparable price level), the only competitor that has peaked my interest is Atom.
Stuff I think the Rostock MAX and its derivatives have really locked down:
- Cheap Cheapskates: Using the 2020 extrusions and a roller-based sub-assembly for linear travel is probably going to be the most cost-effective solution for scalable printers. I'm still a little wary of the wheels will hold up over time (I've already had to re-tune them from time to time), but they seem like an easier problem to deal with than trying to track down the right type of rods and bearings for each design scale.
- Electronics Integration: It's actually really important to me that the PSU and control electronics are enclosed within the same overall Delta frame. A lot of Kickstarter-grade Delta kits conveniently hide the electronics and power supply off to the side, so you never get a sense of what the whole system looks like. I really like the idea of a portable printer that I can lug around, even if I'll never end up doing that. I don't necessarily want something as small as the Bukito or the Portabee, but there's just something nice about only having a single unit to worry about. Maybe I just want a fashion accessory to print with at a coffee shop or something. I don't know.
- Expansion Options: I don't know if Seemecnc did this on purpose or not, but the overall modular design makes it really easy to make upgrades, tweak existing elements, or add attachments. It's definitely clear that the engineers put a great deal of thought on making the different sub-components accessible even after the machine is completely built. This may come back to bite them later, but the Rostock series is (imo) clearly a printer made by makers for makers, even as it's now transitioning to a more consumer-friendly, ready-to-go option.
Considering the development of printers by various successful upstarts (ie. Printrbot, Makerbot, Ultimaker, etc), I think it's clear that design is like a snowball rolling down a mountain. Once you get started, it's often very difficult to change direction. Sometimes it's best to just start over again. I think you see this theme often in all sorts of startups as well: smaller and more nimble groups can often be more successful by making the sort of sharp pivots that a larger one just can't. With 3D-printers, I often look to the newest upstarts over the established heavy-hitters. The Printrbot Metal Plus is the only example I can come up with where a company with an established line legitimately tried something new.
As such, the Rostock MAX has some shortcomings that I don't see they escaping from with only iterative modifications and design band-aids:
- Part Count: The Cheapskate sub-assemblies, the belt re-routing, and the exclusive use of melamine sheets in particular means that the kit part count is a lot higher than most alternatives, especially the minimalist Kossel Mini design. I wasn't always a fan of 3D-printed parts in printers (even though it's basically the core tenet of Repraps), but I'm now recognizing that they're more than adequate for the loading requirements of printers, and they tend minimize part count a great deal.
- Tensioning: Man, there's got to be a better way to tension these belts properly. Filament-based drives have suggested the use of guitar tuners (though anything with a non-backdriveable worm-gear or something would probably work), so that could be a nice alternative. I would definitely like to avoid the Kossel Mini's approach of shifting the entire top frame.
- Electronics Decoupling: Sure, sub-components are accessible, but the electronics routing is still a bit haphazard. Easy enough to reach and adjust, but not at all optimized. It's been tougher than I'd like in extending some of the extra options of the control boards.
- Zeroing: Despite all the times I've done this, I'm still not very good at it, and still not very confident that I've done it properly. A lot of it has to do with my magnetic tool coupling. The curse of a removable tool-head, I suppose.
Design Focuses (Updated 2015-05-23)
Not in any particular order:
- Frame Rigidity: I'm a big fan of a lot of the newer printers' use of more aluminum or metal framing in place of acrylic and/or wood pieces. I want to print at higher speeds! And have a flashier printer! But I think increased speed is the only functional benefit.
- Extruder Torque: I'm being convinced by the interwebs that higher extruder torque is necessary, either through a bigger stepper motor or using a planetary gear attachment. The gearing would also help with much smaller layer heights. We're beginning to push the resolution limits nowadays, and it seems like extruder torque is one of the limiting factors. There's also a couple of designs out there using two gears instead of a gear and just a bearing, presumably leading to less filament stripping. After looking at the options, I think I'd prefer to go with a geared Nema stepper instead of trying to print add-ons like the Wade Extruder variants.
- Proper Cooling/Heating: The need for this is pure speculation. I've had some layer shift issues, and all evidence leads to either the stepper drivers or the stepper motors themselves overheating. I'm curious if this is the actual issue, so I recently got some fans and heatsinks for both the steppers and drivers. In terms of the printing space, I'm no longer a fan of the IR bulb, which seems a little dangerous to have with the print heat parked right next to it. I'd much prefer the space-heater approach. If space becomes an issue, I think proper enclosure sealing may also get me a long way. Designing for proper enclosure integration should be a priority.
- Abandoning Magnets: I still think magnets are magical, but I think a rigid joint is likely the better way to go. I want to give the Traxxas universal joints a try to see if they're better in the long run. I'm a little concerned that my entire hotend sub-assembly is just slightly lifting when printing at very low layer heights due to the extra build-up of pressure. If I'm limited by the limited travel in Traxxas joints, I think the Cherry Pi's use of extension springs may be a good alternative
- Linear Rails/Rods: Proper linear rods and linear slides aren't really that expensive when you know exactly what lengths you need. If only for the reason that they're traditionally the "right" way to do things, I'd like to give them a shot. I've heard that noise might be an issue with standard metal bearings, so I might also give the Igus linear bearings a shot (assuming price is not a huge issue)
- Auto-Level: This NEEDS to happen, whether through FSR's (assuming the temperature issues are worked out) or a limit switch (which can be manually swapped in). I'd rather not deal with inductive sensors, given how temperature-dependent they tend to be. Without tool-changing, I think I would try to design for a probe that swings out instead of the spring-loaded Kossel/Rostock z-probes. For repeatability, I think the goal should be simplicity above all else.
Tool-changing nozzles, whether through a detachable coupling or just a pivoting base, is still on my mind, but it's probably not going to be a core focus for a while. I think I bit off a lot more than I could chew when I was first considering the problem. But we'll see. Anything can happen nowadays.