CloverPlus v2 Delta 3D Printer

thingiverse

CloverPlus v2 First and foremost this is a remix of the Reprap Clover by xAuSolarisx (thank you! I had a fantastic time building this) and contains the Compact Bowden Extruder by schlotzz (also thank you! I've used your design on multiple printers now). The CloverPlus is a small, mostly-printable 3D printer with delta kinematics. Please see the next iteration of this design, the CloverPlus v3 3D Printer:http://www.thingiverse.com/thing:1781851 Printable Diameter: 100mm+ Printable Height: 105mm+ With 105mm arms it's probably possible to print slightly beyond the 100mm diameter. Printable height is about 105mm with an E3D hot end and 300mm rods, more with longer rods and / or a shorter hot end, like a j-head. Mechanically the CloverPlus v2 can easily print at 100mm/s and move at 150mm/s or so, though this is highly hot-end- and controller- dependent. Printing video: https://youtu.be/dCSyj_GUx3U OctoPrint Timelapse: https://youtu.be/2u4MGgkmk7k Calculated v2 delta variables rod center to plate center = 100.429 carriage to ball center = 18.7 effector ball center to nozzle center = 14"delta radius" = 58.729 arm length = (105mm arm length + 7.5mm * 2 to center of ball joint on each end)"diagonal rod length" = 120 for 105mm arms"diagonal rod length" = 115 for 100mm arms Recent Updates Updated 9/21/2016 CloverPlus v3 3D Printer has been posted! http://www.thingiverse.com/thing:1781851 Source OpenSCAD source is posted for all parts except the Compact Bowden Extruder; see the Remixed Source Files for that. Build Details See the attached CloverPlus_BOM.xlsx for part details. Let me know if something is missing or something doesn't make sense. I used 300mm and 330mm Misumi P/SFJ8 rods and JB Kwikweld epoxy to secure the magnets and steel balls (after roughing with 120 grit sandpaper). The bed glass in the photo is 120mm in diameter and about 2mm thick, though 110mm would also work. I cut it with a circle cutter out of a sheet of (I believe) 3/32" glass from Lowes. Print volume is about 105mm high (300mm rods and E3D hot end) and 100mm in diameter. I recommend using glue stick (Elmers purple or extreme) on PET or Kapton tape for printing PETG and PLA without a heated bed. You can apply it straight to the glass, but this will inevitably result in scratched glass. Open Issues / Build Tips I recommend 10mm N50 magnets as these are strong enough to be forgiving of a slightly longer magnet gap than usual. There are arms with many different magnet slots posted - measure your magnets with calipers and if there isn't a .stl to match, tweak the .scad with the measured size and generate a custom arm .stl. The 0.2mm gap I have in the .scad has tested good with multiple magnet combinations. Use 8mm x 15mm x 45mm LM8LUU bearings instead of the shorter LM8UU bearings. The rod holes in the frame pieces may need to be tweaked for a good fit depending on how accurate your printer is at printing holes. After calibrating my extrusion width I have the rod holes sized at 8.2mm with spacing (center to center) of 30.1mm (carriage spacing is 30mm) to allow for the screw clamps to push them inward and close the spacing to about 30mm. Change the rod_extra and rod_spacing variables in cloverplus_topbottomv2.scad to suit your printer. Rod sizing can be anything in the 300mm range (I've used 300 and 330). You can probably even get away with 400mm as the frame is pretty stiff, but I haven't gone that far. Bowden tube sizing needs to be generous to avoid placing too much stress on the effector which could pop the mag arms. Sand both the plastic and the steel balls / magnets with 120 - 300 grit sandpaper and then clean with alcohol before gluing. Make sure the magnets slide all the way back into the arm recesses before gluing via the slots. Test the spacing to ensure the ball does not contact the magnet. After gluing, trim / sand / file the arm to ball contact points as necessary to ensure wiggle-free operation. I use cork spacers for the motors to reduce vibration and provide some heat insulation. PETG's glass temp is about 85C, but PLA's is much lower at about 65C, so watch your motor temps if you use PLA. 0.4mm hot end nozzle size or smaller is recommended due to the build volume Firmware setting delta variables are at the top. I highly recommend this delta calibration tool: http://escher3d.com/pages/wizards/wizarddelta.php I recommend a 32-bit Smoothieware or Reprap Firmware based controller board. An 8-bit arduino-based board running Marlin will work, but it will have trouble running at 60mms or faster - prints will be rough and the steppers will stutter. Repetier firmware on 8-bit machines is reportedly better suited to delta printers than Marlin, but as I was not able to get my Osoyoo board to correctly stream g-code using Repetier without utterly failing, I can't personally recommend it. For extra pizzazz (higher resolution) with a 32-bit controller, use 0.9degree / step steppers. I used 1.8d/step motors on my CloverPlus builds, but 0.9d steppers don't cost much more and are a good fit for small, highly-detailed prints. Acceleration of 1500mm/s^2 and Junction Deviation of 0.05 (Smoothieware) or Jerk of 10 or lower (Marlin) is a good place to start. If you have any questions, please ask in the comments! :) Print Settings Rafts: Doesn't Matter Supports: Yes Resolution: 0.3mm Infill: 75% Notes: I printed mine using Hatchbox PETG for strength and minimal shrinkage (50mms max worked best for me; I used a 0.5mm nozzle and 0.6mm extrusion width) and I highly recommend it for printer parts. PLA can be used for the frame, but be very careful with motor temperatures. ABS is a nightmare for parts with a wide horizontal footprint without a heated chamber, but if you have one, go for it. :) Support is required for the carriage belt grip offset (to accommodate the 2-pulleywasher design). 0.2mm layer height is recommended for the carriages due to the belt grips. Changes from RepRap Clover I've made a number of changes to the Clover to improve on various aspects of the design: All parts have been re-designed in OpenSCAD to improve maintainability and enable easier re-use for future revisions and remixes. Stronger top and bottom frame pieces now use 2xM5 screws on each frame joint to increase frame strength. 8mm rods are now clamped in from each side with a M5 screw so that the design can be broken down, reassembled, and adjusted. Bed mounts with clamp tabs for 120mm x 2mm glass plate Top controller mounts - Azteeg X5 v1.1 & v2.0, Osoyoo MKS Base 1.4, and Ramps 1.4 included, others available upon request Compact Extruder design is modified with mount spacing to match with the top frame parts, MK8 hobbed gear spacing, and 4mm M5 push-fit adapter Top and bottom frame vertical screw holes and nut traps are now sized for M4 screws Belt paths are now straight and M8 top pulley screw slot has been changed to M5 (with spacer for the 608 bearing) for proper belt path alignment PulleyWasher parts now go on both sides of the bearing to reduce belt rubbing; carriage belt grips have been extended to match and moved down for PulleyWasher clearance (support is required for this when printing the carriages) Effector and carriage ball mounts have been modified for better glue / epoxy contact area and less potential arm interference Arms have been modified for additional thickness (stronger magnet recesses) and a slot on the bottom to glue magnets in from the side so they slide in all the way Hot end mount height as been reduced to accommodate E3D fan and magnet effector attachment has been replaced by m3 screws and m3 lock nuts for better stability. A j-head hot end mount is also provided that uses a M4 screw to secure the bowden tube. Endstop switch mount holes added (plunger style, 9.5mm with 2.5mm hole spacing) Holes added to top and bottom frame pieces for zip-tie wire organization Increased stepper motor mount width to accommodate 42mm x 42mm x 48mm stepper motors Added power supply zip-tie mount bracket Added fan mount bracket Future Changes This section details planned design changes. See CloverPlus v3! http://www.thingiverse.com/thing:1781851 Older Updates Updated 7/7/2016 cloverplus_carriage_v2.stl and .scad have been slightly modified to include a 4mm hole to enable attaching a flying extruder. Updated 5/2/2016 The Bill of Materials has been reorganized, tidied, and improved with more details. Updated 5/2/2016 The BOM has been updated with more information on stepper motors and electrical parts (wire, etc.). Updated 4/26/2016 Updated bed clamps have been uploaded, including a separate .stl for just the bedclamp tabs (this is the part I modified). I've added a bump to the front of the tabs that should face down so that vertical pressure can actually be applied to the bed plate to keep it in place (novel concept... ;). The .scad source has been updated as well. Updated 4/19/2016 The blue v2 is finally up and printing! Accuracy looks to be better than the v1 build and pretty pleased with the prints it's producing. I've been printing Hatchbox red PETG at 60mms with 1500mms2 acceleration and 0.05 junction deviation (smoothieware). This seems to be working well; 2500mms2 and 0.1 junction deviation was a bit too aggressive with the heavy effector (6 x 15mm steel balls make it a bit porky). I've been printing on PET tape with Elmer's glue stick (purple or extreme, either seem to work) and using a 3-layer raft (first layer adhesion was still less than optimal without it). My LM8LUUs were rattling quite a bit, so I wrapped a 4mm ziptie around the bottom end of each and pulled it pretty tight. This doesn't eliminate the rattle, but it does make it tolerable. I've posted one fairly poor photo of the very first v2 non-calibration print; I'll be posting more print photos soon. Updated 4/9/2016 xAuSolarisx's choice to use 45mm x 8mm LM8LUU bearings was spot on; I shouldn't have even messed with the shorter ones. I received my 45mm LM8LUU linear bearings today and the play problem is gone. :) I've updated the BOM and the rest of the description to match. These will work fine with the posted carriage design. Updated 4/5/2016 I've been having trouble with poor linear bearings - there's too much play in the cheap LM8UUs I have. I could attempt to correct this by printing new carriages with the bearing recesses closer together or farther apart to put tension on the bearings, but I like my latest carriage prints and I'm not terribly inclined to re-print them. ;) I've been experimenting with printed bushings which seem to be working well, but not quite well enough for printing yet (too tight for unrestricted movement). Bushings are far quieter, which is nice, and there's less play in the mechanism which results in less effector tilt. I'll post more details when have a successful print using bushings. Updated 4/3/2016 I've posted a mount for E3D hotends with a recess for an external M5 push-fit connector. I recommend threading this first with a M5 screw before attempting to screw in the push-fit. I've printed the new carriage and it grips the linear bearings much better. This should resolve the slop, but I have some work left before I can attempt a test print with the updated design. Updated 3/27/2016 I've narrowed the carriage linear bearing recesses so they're a tighter fit. I've also added a pair of holes on the back side of the carriage to secure the linear bearing via grub screw from that side as well. My test build had too much play in the carriages with 0.4mm extra spacing around the bearings; 0.2mm should work better. Again, proper extrusion calibration is key. Updated 3/26/2016 I've posted slightly updated carriage and effector designs that don't recess the ball as far. There should still be plenty of surface area for adhesion, but this will enable more reachable horizontal space for a given arm length at the expense of a few millimeters of vertical space. Updated 3/25/2016 The arms are now 105mm long and have a magnet gap of 0.2mm - I've found this to be optimal in my testing. The OpenSCAD source is posted so you can tweak the arms to your content. The updated v2 arms are the same width as the original CloverPlus arms, but have a more pronounced cutoff and an optimized magnet gap. 105mm length should keep the arms from running into the ball to carriage edges; this is untested, but I plan on doing that tomorrow. Updated 3/22/2016 I've posted a power supply mount - cloverplus_ps_mount.stl. This attaches to one of the outside bottom 5mm frame mounts and uses 2 4mm zipties to secure a 58mm wide x 132mm long (length is less important) 12V 6A power supply. The top and bottom frame pieces have been updated to move the wire guide 3mm zip-tie holes 3mm out to clear the power supply mount. The carriage bearing clamp holes have been narrowed slightly to better grip M4 x 6mm grub screws. The new arm .scad has been posed along with the 11.4mm arm stl I used for my build (2 supposed 6mm x 6mm cylinder magnets came to 11.4mm total length). The BOM has been updated. Updated 3/21/2016 An updated E3D hotend mount (screw mounts were too thin) is posted, as is a j-head mount (does not accommodate a fan). The j-head mount has a M4 nut recess just above the hot end to retain the bowden tube. It will most likely be necessary to shave or sand the outside of the bowden tube somewhat to prevent the nut from crushing the bowden tube too much. The J-head mount has been successfully test-fitted. Updated source for both is posted. Updated 3/20/2016 Ramps controller mount .stl added and the controller mount .scad source has been updated. This is untested as i don't have a spare ramps floating around, but the spacing measures right when assembled in Sketchup. I'm mostly sure the mount stud size is okay. New carriage and effector STLs and OpenSCAD source for both have been uploaded. These use a better ball recess design that is more stable for gluing. Updated 3/15/2016 I've redesigned the carriage in OpenSCAD. It now uses a pair of M4x8 screws or M4x6 grub screws to secure the linear bearings from the sides. It's more sturdy and shrinks the delta radius by about 4mm, but this should not significantly affect the print area. The carriages should be printed with 0.2mm layer height so the belt clamp turns out right. Support is still needed for the belt clamp. The BOM has been updated and source is posted. The top piece has been adjusted to move the endstop switch out a bit for easier screw access as well as triggering away from the ball recess and up to increase build height by about 2mm (wasted space). The carriages should be printed with 0.2mm layer height so the belt clamp turns out right. Updated 3/13/2016 The effector and hotend mount have been redesigned in OpenSCAD. This resolves some dimensional accuracy issues and improves the maintainability of the design. cloverplus_effector_v2.stl, cloverplus_hotend_mount_v2_e3d.stl, and cloverplus_effector_and_hotend_mount_v2.scad have been added. Updated 3/12/2016 Added cloverplus_top_v2.stl and cloverplus_bottom_v2.stl. These are generated from my OpenSCAD redesigned top and bottom frame piece source. I've printed and tested a pair of these and everything lines up as it should. The new parts require 12 additional M5 x 16mm screws to clamp the rods in from the sides (I have tested that part of the mechanism and it secures the rods very well). Other improvements include slightly thicker edges and base dimensions as well as more consistency in general. They should be compatible with all existing parts, but NOT the original top and bottom frame pieces; you can't mix and match them. The OpenSCAD source has been posted as well. See the note under Build Tips about rod hole sizing. STLs and scad are now set to 0.2mm extra diameter for the rod holes; this enables as snug fit for the rods when my printer is correctly calibrated. New controller mount STL files have been added for the v2 frame - Azteeg X5 v1.1 and v2.0 and Osoyoo mksbase 1.3. The controller mount openscad has also been updated. Updated 3/8/2016 Added a fan mount bracket sized for 100mm fans with 91mm spacing. The OpenSCAD source is included so it can be resized for different fans. I've added a new photo of the fan mounted to the design; it attaches to one of the bottom frame connections. The BOM has been updated.