|Projects Publications Resume Links Contact About Youtube Donate||
This project is a delta 3D printer design inspired by the Tevo Little Monster (TLM) and the Anycubic Predator (AP). This page is not affiliated with either company. Both sources of inspiration are larger delta printers that cannot be purchased in the US as of December 2020. After seeing the TLM and AP are both discontinued, I tried to purchase a Homers Hector/Little Monster, but no luck. After several months and zero email responses, I requested a refund. But I still wanted to upgrade from my excellent Prusa i3 MK3S + OctoPrint to a larger, faster, enclosed printer. That's what led to this project.
If my TLM purchase had been successful, I would have made several changes and upgrades. Lots of reviewers don't love the TLM out of the box, but many are very happy after making some modifications. This project is definitely more effort and more expensive than my original modified TLM/AP plan. But it gives me an opportunity to cut fewer corners, learn more, and use more top-shelf components.
I'm dubbing this project DERP 3550, the:
I'll start with squirting hot plastic, but I'd eventually like laser beams and spinning sharp things mounted. The 3550 denotes the build diameter of 35cm and the build height of 50cm.
DERP 3550 by Gregory Alan Hildstrom is licensed under CC BY-SA 4.0
I chose a TLM-like design, but with more overkill. This seemed simpler to construct than an AP-like design. The design has the advantage of low part count. But frame rigidity and accuracy are directly related to part quality and accuracy, with less room for adjustment. Squareness and rigidity of the frame are derived from a few critical areas:
For rigidity and flatness, I'm using thick standard plate for the top and bottom frame plates. Standard plate proved insufficient for the bed and build plates, so I'm using cast aluminum tooling plate for those. The as-delivered c-beams likely have a chopsaw or bandsaw cut edge. The c-beams I received were quite good, so I don't think the ends need to be milled or ground square.
Why not core-xy? First, I think deltas look cool AF. Second, my Prusa is cartesian and I want to try a delta to expand my horizons. Third, to extend a middle finger to Tevo, Homers, and Anycubic for making interesting products and then not continuing to produce or improve them.
I definitely got some extras as I'm designing, building, and revising. I'll trim this list down once I have things working and stabilized.
|Duet3D Duet 3 Mainboard 6HC||Filastruder||1||received|
|Duet3D PanelDue 7i touchscreen||Filastruder||1||received|
|Duet3D Delta Smart Effector||Filastruder||1||received|
|Magnetic Delta Arms 400mm||Filastruder||6||received|
|E3D V6 Ultimate Edition Hot End||Filastruder||1||received|
|Panasonic Solid State Relay||Filastruder||1||received|
|Duet3D Duet 3 Connector Kit||Filastruder||1||received, extras|
|Duet3D Smart Effector Connector Kit||Filastruder||1||received, extras|
|E3D Collet Clips||Filastruder||1||received|
|PTFE Tubing Cutter||Filastruder||1||received|
|Groovemount Bowden Coupling||Filastruder||4||received, for filament dehydrator and dry box|
|Bondtech LGX extruder||MatterHackers||1||received|
|Keenovo Round Silicone Bed Heater 300mm 500W 120V||Keenovo||1||received|
|Custom Frame Plates (0.5"x32.0"x22.0" for 2)||Online Metals, waterjet service||2||received, 6061-T6 plate|
|Custom Cast Bed Plate (0.375"x16"x16")||Online Metals, waterjet service||1||ordered, cast aluminum tooling plate|
|Custom Cast Build Plates (0.25"x13.8125" diameter)||Midwest Steel and Aluminum, circle cut||2||ordered, cast aluminum tooling plate|
|Polycarbonate Sheet (0.236"x24"x48"), cut to 40-7/8"||Online Metals||3||considering|
|OpenBuilds C-Beam Linear Rail 1000mm||OpenBuilds Part Store||3||received|
|OpenBuilds C-Beam Gantry Kit||OpenBuilds Part Store||3||received|
|OpenBuilds NEMA 17 Stepper Motor||OpenBuilds Part Store||3||received|
|OpenBuilds GT2-2M Timing Pulley 20 Tooth||OpenBuilds Part Store||6||received, 3 motors, 3 idlers|
|OpenBuilds Smooth Idler Pulley Kit||OpenBuilds Part Store||5||received, 3 idlers, 2 counterweight pulleys|
|OpenBuilds GT2-2M Timing Belt By the Foot||OpenBuilds Part Store||24||received|
|OpenBuilds 24V Meanwell Power Supply Bundle||OpenBuilds Part Store||1||received|
|OpenBuilds V-Slot 40x40 Linear Rail 500mm||OpenBuilds Part Store||6||received|
|OpenBuilds V-Slot 20x40 Linear Rail 250mm||OpenBuilds Part Store||2||received|
|OpenBuilds 8mm Spanner Wrench||OpenBuilds Part Store||1||received|
|OpenBuilds Rubber Feet Set||OpenBuilds Part Store||2||received|
|OpenBuilds 3 Hole Joining Plate||OpenBuilds Part Store||4||received|
|OpenBuilds Slot Cover 500mm||OpenBuilds Part Store||6||received|
|OpenBuilds Tee Nuts - M5 (10 Pack)||OpenBuilds Part Store||3||received|
|OpenBuilds XLarge C-Beam Gantry Kit||OpenBuilds Part Store||1||received|
|OpenBuilds M5 35mm screws (10 Pack)||OpenBuilds Part Store||1||received|
|OpenBuilds Ball Bearing 625 2RS||OpenBuilds Part Store||2||received, spool holder|
|Hilitchi Stainless M3-M5 Socket Cap Screw Assortment||Amazon||3||received|
|Hilitchi Brass M2-M5 Threaded Inserts||Amazon||1||received|
|Socell T-Slot Nut Assortment||Amazon||2||received|
|Drill America M5x0.8 Tap and 4.2mm Drill Kit||Amazon||2||received|
|KKmoon M3-M12 HSS Counterbore End Mills||Amazon||1||received|
|Yodaoke Stainless Couplers (M5x15mm), bed standoffs||Amazon||1||received|
|Uxcell 5mm 304 Stainless Steel Rod||Amazon||1||received, shear cut distorts ends, needed sanding|
|16 AWG Silicone Wire||Amazon||1||received|
|20 AWG Silicone Wire||Amazon||2||received|
|IWISS IWS-3220M pin crimping tool||Amazon||1||received, |
thanks Teaching Tech
for the crimp connector guide video
|IRWIN VISE-GRIP wire stripper||Amazon||1||received|
|Qibaok 560 PCS heat shrink wire connectors||Amazon||1||received|
|Qibaok crimping tool||Amazon||1||received|
|Ferrule crimping tool kit||Amazon||1||received|
|iExcell M3 30mm-50mm Stainless Socket Cap Screw Assortment||Amazon||1||received|
|WINSINN 24V 40mmx10mm fan and blower combo||Amazon||1||received|
|Cylewet Roller Micro Switch AC 5A SPDT (20 pack)||Amazon||1||received, for filament runout sensor, could also be used for homing sensors if sensorless homing isn't wanted|
|Nylon String by the foot||Local Hardware Store||10||received|
|1/8" Stretch/Shock Cord by the foot||Local Hardware Store||10||received|
|Large Binder Clips||Local Hardware Store||6||received|
The hexagonal plates on the TLM look cool, but they make it more difficult to enclose. I reviewed some of the open source TLM models (jeowin, CC BY 3.0) for inspiration, but I'm not using or modifying any of those files. I decided to use a similar C-Beam spacing, but with a more triangular shape that is less material and easier to enclose. The tradeoff of this design decision is related to bed/build plate diameter, but there are options. I decided to use the same design for both the bottom frame plate and the top frame plate for simplicity. I designed the plate in OpenSCAD, but I also like to model parts in Blender. The larger holes are 5mm for M5 through holes. The smaller holes are 4mm that can be drilled and tapped for M5 threads.
I got the frame plates waterjet cut out of 0.5" (12.7mm) 6061-T6 aluminum. The quoted material size is 32.0"x22.0" for both pieces. An alternative to that approach is to print the layout full scale, glue it to the uncut plate, then manually cut, center punch, and drill. That being said, the waterjet service was worth every penny. With some much-appreciated help from Al, we countersunk holes for socket cap screws using a mill/drill press.
My first iteration used standard plate for the bed plate and build plate. That's probably a facepalm moment for more experienced printer builders. Standard plate isn't flat enough for easy calibration or good first layers. And there are a limited number of mesh bed compensation points in firmware. Live and learn. So I'm using 0.375" cast aluminum tooling plate, again, waterjet cut. The part fits within a 16"x16" piece.
If you want the bed plate to fit within the frame plate footprint, adjust the circle to 320mm diameter. I opted for a 350mm circle, which extends beyond the frame plate, but still fits within the frame extrusion footprint.
The bed plate will mount on standoffs to threaded holes in the bottom frame plate. That same threaded hole location is used for part of the stepper motor mounts in the top frame plate.
Again, don't use standard aluminum plate for this, unless you want headaches and crappy first layers. I thought about using magnetic flexible build surfaces like Wham Bam or BuildTak. But I thought I wouldn't be able to use the Delta Smart Effect nozzle probe while hot. Once you apply that magnetic base to a fixed bed, you're pretty committed, which limits options. On my Prusa's powder coated plate, low z offset made parts extremely hard to remove and that permanently discolored the sheet. A little higher was good for some parts and poor adhesion for others. I end up using glue stick for some prints anyway, so I though an aluminum or glass build surface would be ok. So I'm using 0.25" cast aluminum tooling plate. Two 350mm diameter build plates will fit on a 30"x15" piece. If printing on that turns out to be a headache, I'll try a stick-on surface on the removable plate or I'll try a glass plate.
Having two allows for different surfaces, preparing one plate while printing on another, etc.
I found Midwest Steel and Aluminum while searching for cast aluminum tool plate online. They offer circular bandsaw cuts of cast aluminum tool plate for a fraction of what waterjet parts would cost. They aren't as precise, but the material flatness and consistency are more important to me than diameter accuracy for this part. And if I get creative with a jig, I may be able to clean it up to a more accurate final diameter.
I deburred holes in the frame plates and holes in the ends of the c-beams. I also went over the mating surfaces of the c-beams and the frame plates with a flat stone. I used blue/medium threadlocker on the bottom plate to c-beam fasteners. I used anti-seize on the top plate to c-beam fasteners. I drilled and tapped all of the 4mm holes in the frame plates.
After securing the bottom plate and 1/3 of the top plate, I noticed about 1/2 hole out of alignment on the remaining two c-beams. Between that and checking things with a square, I was pretty happy with the factory end squareness on the c-beams. For now, I'm opting to not mill or grind the ends of the c-beams. If test prints prove inaccurate, I may revisit that.
You can print these parts or have them printed for you. I still have my first 3D printer while I'm building this printer, so I'm printing them myself. There are definitely other ways to make these parts, but I'm using the tools available to me. I printed the prototype parts in PLA or PETG with thin walls and low infill. For the most part, I printed the functional parts in PETG with 100% infill. If there are any shortcomings due to flex, heat, etc., I may reprint in another material. But I don't have any experience with ABS, ASA, Nylon, or PC yet.
I designed most of the parts for this project, but I sourced a few part designs from elsewhere as noted in the Source column of the table. I designed these parts in Blender because I like using it and I'm a big fan of open source. Maker Tales has a bunch of helpful YouTube videos for this sort of thing. Two of my favorite playlists are Intro to Blender 2.8 and Precision Modeling and Learn Blender 2.9+ Through Precision Modeling.
|M5 tap guide||tap_guide.blend||tap_guide.stl||1||printed|
|stepper motor mount||stepper_mount.blend||stepper_mount.stl||3||printed|
|timing belt clamp||belt_clamp.blend||belt_clamp.stl||3||printed|
|idler plate front||idler_plate_front.blend||idler_plate_front.stl||3||printed|
|idler plate back||idler_plate_back.blend||idler_plate_back.stl||3||printed|
|gantry magball mount||gantry_magball_mount.blend||gantry_magball_mount.stl||3||printed,|
330mm build diameter w/ mount,
350mm build diameter w/
drilled/tapped gantry plate
|power supply bracket||ps_bracket.blend||ps_bracket.stl||2||printed|
|Duet 3 mainboard box||duet3_box.blend||duet3_box.stl||1||printed|
|PanelDue 7i case front||Thingiverse LumberjackEngineering CC||7i_Case_Front.stl||1||printed|
|PanelDue 7i case rear||Thingiverse LumberjackEngineering CC||7i_Case_Rear.stl||1||printed|
|PanelDue 7i mount display||Thingiverse LumberjackEngineering CC||7i_Mount_Display.stl||1||printed|
|PanelDue 7i mount arm||Thingiverse LumberjackEngineering CC||7i_Mount_Arm.stl||1||printed|
|PanelDue 7i mount frame||Thingiverse LumberjackEngineering CC||7i_Mount_Frame_2020.stl||1||printed|
|counterweight idler 1||counterweight_idler_1.blend||counterweight_idler_1.stl||1||printed|
|counterweight idler 2||counterweight_idler_2.blend||counterweight_idler_2.stl||1||printed|
|solid state relay clamp||ssr_clamp.blend||ssr_clamp.stl||2||printed|
|extruder plate||extruder_plate.blend||extruder_plate.stl||1||printed, sandwich plate thickness maintains 100% contact between gears|
|simple fan duct||Thingiverse dc42 CC||40mm_fan_duct_v6_v0.3.stl||1||printed, opening trimmed with a razor, definitely Berd-air curious at this point|
|effector fan bracket||Thingiverse dc42 CC||EffectorFanBracket.stl||1||printed, may want to brace the duct to the heat sinc if using high accelerations|
|build plate spacer||build_plate_spacer.blend||build_plate_spacer.stl||3||printed in HTPLA, then heat treated|
|spool tower ends||spool_tower_ends.blend||spool_tower_ends.stl||2||printed, print one mirrored|
|spool bearing shaft||spool_bearing_shaft.blend||spool_bearing_shaft.stl||1||printed|
I considered three different extruder options:
I considered a counterweight in all cases, especially the flying extruder.
Each approach has pros and cons. The flying extruder approach is quite proven and it does well with fast printing, but it can struggle with retraction speed and flexible filaments if the bowden tube is too long. The direct drive approach excels with quick, accurate retractions and it can handle flexible filaments, but that adds mass to the effector and it significantly reduces acceleration. The remote direct drive approach seems like the best of both worlds, but the Nimble Flex is currently in preorder and I'm not sure how well the flexible driveshaft angles will work in an enclosure.
So the approach I'm choosing initially is the flying extruder with the Bondtech LGX.
The 100% infill solid counterweight gantry roughly balances the extruder and extruder plate. Pulley, wheel, and bearing friction is enough to hold that assembly in place wherever it is stopped. In motion, things settle in the gantry direction a little, so it is slightly heavier than the extruder alone. With just that amount of counterweight, the bowden tube pushes the extruder up during positive Z moves. For fast Z moves, the weight of the extruder may fall to the flexible shock cord attached to the c-beam gantry plates, which are supposed to be for centering. In order to keep some tension on the extruder and bowden tube, and to offset some of the effector and carriage weight, I added 400g to the counterweight gantry.
After attaching wiring and the bowden tube, I disconnected the timing belts to do some weight measurements. I placed a scale on the build plate under the centerd nozzle. With no counterweight, I measured 1137g at the nozzle, which includes c-beam gantry plates, delta arms, wiring, extruder, and effector. With counterweight gantry + 400g, I measured 497g at the nozzle. Each tower stepper is still holding up some weight, but less. I've seen a couple other builds using this counterweight approach, but I haven't done a ton of research. And I definitely haven't seen a detailed discussion of the counterweight mass.
config-20210716.zip is the latest. Currently working:
I started by using the RepRapFirmware Configuration Tool, then slowly iterating and making customizations/tweaks. I must say the RRF documentation and tech articles are pretty good. I was able to get things up and running without scouring forums or asking for help. I really like being able to fully configure the firmware with config files; no recompilation or flashing necessary.