The PrismX is elegant, strong, fast, and affordable. It's an open-source reprap 3D printer that traces its lineage back to Prism AliceBlue, Prism, and Mendel before it. Like its predecessors it prints 3D objects, can replicate itself, can be tuned, hacked, repaired, upgraded, and more. It's designed to be outfitted with deluxe components throughout, but can also be rep-strapped from the ground-up. Its plastic parts can be manufactured for higher degrees of precision. Frame strength and alignment have been improved. As a result the machine's performance is outstanding.
The PrismX is an evolutionary open-source design. The PrismX keeps the spirit of its forefathers intact, being a true reprap that can replicate itself. However it is also designed to be manufactured with molded or machined parts for improved precision and broader manufacturing options, and proposes several deluxe but affordable upgrades that provide significant return on investment.
To engineer the PrismX I've taken the OpenSCAD-based fork of the Prism and completed a process of creating new parts, replacing and revising existing ones, refactoring the code to improve it, and so on. One of the primary milestones I reached in January, 2013 was to create a complete OpenSCAD model of the PrismX. This was inspired by Johan Rocholl's work on the Rostock Delta Robot, which is also 100% OpenSCAD and was published with a complete machine SCAD file. This took a long time to write and debug - but in the end the investment payed off immensely and we've never looked back.
The PrismX has been developed by focusing on each area of the printer and revising it. The revisions were tested on a prototype named AliceBlue. Both PrismX and the earlier AliceBlue have the same geometry as the original Prism. They both share some original parts from the original Prism. In the PrismX much has changed with respect to the original Prism. It still looks similar, but refined and evolved.
|440mm (W) x 440mm (L) x 410mm (H)
|~8 kg (gantry only, not including motors, PSU, wiring, etc.)
|Max Potential Build Envelope
|215mm (W) x 210mm (L) x 200mm (H)
|Cost to build
|$??? US (complete w/deluxe upgrades, electronics)
- Parts designed for molding or machining - can be printed also.
- 4-motor design. Designed in OpenSCAD.
- Belt-driven z-axis with GT2 belt and 36T pulleys. Motor has 36T pulley so entire z-drive is 1:1 ratio.
- The 00str00der belt-driven extruder.
- 8mm smooth rails throughout.
- Y-Axis completely redesigned. Universal y-axis plate, GT2 belts and pulleys, custom idler and motor mount. The y-axis belt is designed to loop around the bottom extrusion to maximize z-axis travel. With the universal plate this design is light, fast, and accurate.
- Z-axis completely redesigned. Designed for acetal bushings, thrust bearings, and leadscrew nut. Acetal bushings work best because they're designed for axial motion - and they're affordable, reasonably precise, and durable. I've tried the original Prism z-axis design and some variations on AliceBlue; this setup works best so far. The precision leadscrews are optional but you'll need to modify the SCADs for other leadscrews. I'd say a printed leadscrew nut works best because it can be printed slightly tighter than the threads but still be secured and adjusted. Delrin is next best, metal is OK and probably has the best aesthetic.
- Z-axis rods moved 4mm outside of frame. This not only makes assembly easier as the original top-support is not needed (though I revised it). The frame is even stronger now.
- X-axis completely redesigned. Sheliak x-ends and Sheliak X-carriage are designed to be printed for precise alignment.
- Top-rod "armor" - supports that secure the top rod to the uprights. With four (4) of these the frame is very strong. I've had no problem carrying the entire machine w/PSU by the top bar and have traveled with it many times. It needs its own seat in the car though...
- LM8UU bearing mounts -- Mounts are designed to be molded or printed. Instead of clamping we use steel bearing retaining rings. They're affordable, durable, and work much better than clamping mounts - especially the printed clamping ones that are prone to break. More importantly, the bearings are rock solid, align more precisely, and travel smoother.
- Bottom-feet completely redesigned (in 2 parts). Can be made from multiple materials such as soft rubber.
- Frame can be resized. I don't see anything wrong with 1000mm extrusion here - except you'll need to make your own heatbed - the T-6061 Universal Y-Axis plate could probably hold up, and in fact you could make a larger one with our published source code.
- GT2 belts and pulleys throughout.
- Mounts for power switch and power supply on the frame.
- Filament guidler
- Filament spool holder. There are two versions - one that sits next to the frame, and one that attaches to any location on the frame. The attached version is standard.
The PrismX is actually based on Prism AliceBlue. Thus AliceBlue can be considered an early PrismX prototype. It's somewhere halfway between the original Prism and the PrismX. I started revising the original Prism in early 2012. After a few detours in life Prism AliceBlue was commissioned in December, 2012. It has been running quite reliably ever since - even with modifications we've made for proving PrismX. I'd say the main reason for its reliability is the belt-driven z-axis. The rare times AB has failed, it did so in a predictable and consistent way and repair was notably easy.
- y-rods moved above frame ~3mm
- Universal Y-Axis Plate with customized idler and y-motor support - however at first it had TommyC's y-plate
- z-axis drive system upgraded to GT2 belt and pulleys (not printed), thrust bearings
- x-ends with customized Kuehling's x-ends - the motor moved behind the z-axis bearing mounts
- power supply mount, RAMPS mount, ATX adapter board mount
- other goodies like power switch, filament guidler prototypes.
Evolution Into PrismX
AliceBlue has become a proving ground for PrismX modifications; all sorts of mods, actually. It's very dependable. However, it has limited travel on the X and Z axes, so I always have to make "short" plates for it. The customized Kuehling's setup and original frame design are responsible. My first goal with AB was to maximize travel. First, we outfitted AB with a Universal Y-Plate. This created belt-mis-alignment and the long journey of revision began. Then the z-axis was put to the test as the machine printed production parts, and we occasionally had to perform maintenance on it. We revised the z-axis supports, made them taller and thicker, and now we use acetal bushings that are designed for axial motion. The Kuehling's x-ends and x-carriage proved that they couldn't take the heat after many hours of fast-driving on AB, and had to be re-printed while we were still designing their replacements. Along the way we also logged thousands of hours of printing time with our MendelMaxes, and we happened to build a Rostock frame or two. All of this assembly, maintenance, and observation of the various models provided a trove of data to inform revision and modification on the PrismX.
Why a PrismX Instead of Another Printer?
Stronger frame, better alignment = faster and better prints. Frame alignment is actually more important than frame strength, and the PrismX scores well in both areas. The closed-triangle design on the Prism is stronger than the open-triangle or T-frame designs. There is no need for a double-lower gantry like the Prusa and MendelMax - that actually weakens their frames and was introduced into the design for the y-axis.
Alignment on the PrismX allows it to achieve print speeds that rival delta robots, with quality that outperforms them. We're printing at 200mm/s perimeter, 400mm/s infill, 750mm/s travel, and the constraints we've observed at high speeds aren't related to frame strength. While the PrismX can match delta robot speeds, few delta robot designs achieve the same frame strength and print quality as PrismX.
The PrismX has a complete CAD model which allows for visualization of the entire model including movement. The PrismX can achieve remarkable alignment with printed parts, not to mention the tolerances that are feasible with molded or machined parts. Finally, the PrismX is based on a proven lineage that goes back to the original Prism Mendel. The design is based on proven revisions. There has never been a need to make radical design changes from one revision to the next - the design changes have been evolutionary. The original Prism Mendel design is based on sound principles.
The printed parts BOM is contained in the Github repo at https://github.com/Terawatt-Industries/prismx-scad/blob/master/dist/stl/plates/README.txt. The set of plates required for the complete BOM is in the same folder. The plates are also SCAD files, found in src/plates.
The extrusion profile is the same as the original Prism. The complete BOM for hardware and vitamins is in the Github repo at https://github.com/Terawatt-Industries/prismx-scad.
Terawatt will publish a complete full-color assembly guide for the PrismX when kits are made available to the public. In short, assembly is still similar to the original Prism - but the parts are revised, the smooth rails have shifted, and the z-axis supports are different; amongst other assembly variations such as split-parts.
Nearly all plastic parts can be pre-loaded with t-slot nuts which expedites assembly. Also, X/Y frame alignment almost "falls into place" with very little effort. For the z-axis, alignment at the top of the frame is well constrained, and at the z-axis bottom we measure the midpoint (400mm / 2 = 200mm) or use a 196mm jig (200mm - 8mm / 2).
The PrismX can be scaled to any size. Its OpenSCAD files are parametrized, as well as the entire-machine CAD. Thus the design can be visualized and virtually tested for different extrusion lengths, leadscrew and smooth rail lengths, and diameters. Want a PrismX with 1000mm build volume on each axis, one with 200mm that's portable, or maybe one with 12mm smooth rails? No problem.
The PrismX is designed to provide an upgrade path for MendelMax owners. MendelMax owners can easily use existing 420mm extrusion, add 3pcs. 420mm extrusion, and re-use several important components such as motors, electronics, wiring, and hotend. Actually, you'll save a motor. We're making a variant of the printed/molded parts that work with M5 hardware - you can keep the t-slot nuts but the Prism parts are 2-4mm thicker so unfortunately we can't re-use most of the screws.
Prism owners can upgrade to the PrismX. Prism owners already know how easy it is to maintain and upgrade this design. For a complete upgrade you need to replace the Y-axis, X+Z axes, and Z-axis. You will basically need a complete set of plastic parts, except the 60-inside's and 60's. You should be able to re-use most of the existing vitamins (hardware), but you'll need some longer/shorter screws and a couple/few specialized vitamins like button-cap screws.
For a partial Prism upgrade there are several options. You could keep the Y-axis, and just replace X+Z axes. Alternatively you could replace the Y axis, keep the x-ends, upgrade the z-drive system and use an 00str00der. There are many possibilities for Prism -> PrismX upgrades.
Mendel, Selles, and Prusa - other reprap owners - can upgrade and re-use electronics, motors, and some hardware, depending on which variant they own. You'll probably save a motor. Print quality and speed are likely to improve.
We have a roadmap planned for this machine that potentially extends beyond 2015. What types of enhancements or upgrades would you want?