Another possible title would be "Make a 3D machine using a 2D machine."
This Instructable is about using a CNC router and 2D designs to create a Delta 3D Printer.

In my research about 3D printers for hobbyist I have found a thriving community and design philosophy called RepRap. The way I understood it is that RepRap is about creating self replicating machines.

But this leads to a "chicken or egg" scenario. In order to make a RepRap 3D printer, you need a 3D printer to make the parts. If I had a 3D printer to make parts, then why would I need to make myself a 3D printer?

You can probably tell where this is leading... I don't have a 3D printer, nor do I have access to one or know someone who does.

What I do have is access to a Probotix Comet CNC router that's accurate to 0.0001" and about a year of self-thought knowledge of using CAM software to make "2.5D" designs.

the pages that follow is my adventure on making a RepRap (or RepStrap) Delta 3D printer on a very small budget (about $400) based on the RAMPS 1.4 control board. I will provide the CAM files, DXF files, and g-code files for the pieces i created.

Please note that I do not have formal training in CNC, CAD design, or anything related to the professional computer aided fabrication industry. i am self-taught, so don't take anything I give you at face value. Please feel free to validate the information I provide through other sources and give me positive, constructive criticisms. I welcome it and we are all here to learn and grow together.

I will also be sharing failed designs just to show how my design process works.

I hope people find this Instructable useful.

Here is a video of the end result - the finished project:


*** Update: 12-29-2015: This instructable is going to be more like a build blog. As I progress I will add more pages and update existing pages as things change.

Step 1: Internet Resources

Here are some of the websites that I used to get information, get parts, and get inspiration from.

RepRap.org main page

RAMPS 1.4 wiki

Marlin Firmware wiki

Repetier Host and Firmware - This is the firmware and the G-code interpreter I decided to go with. I went this route because the user interface was a lot easier and the firmware updates (which is critical for printer calibration) is a lot easier to navigate.

Delta 3D Printer Google Groups - in this group search for user Haydn Huntley and find the topic "Is anyone interested in inexpensive parts for zero backlash magnetic ball and socket joints?" I purchased 12 socket joints from Haydn for this build.

List of Delta 3D Printers

Steve Grave's Delta Robot Kinematics - This is a deep dive into the math of Delta Robot kinematics (the math of movement). Most of this was way over my head, but if you can wrap your mind around the math, more power to you.

Please note that I am not affiliated to any online store I have listed below. In providing their URL, it simply means I purchased something from them, so they are a known source for materials.

OpenBuild Store - Mark at OpenBuild has given me permission to share the SketchUp files and dimensional blueprints of their OpenRail extruded aluminum parts.

SanSmart Store - RAMPS 1.4 kits.

Inventables Store - Aluminum extrusions, parts. They also sell a hobby grade CNC router, the ShapeOKO 2.

Probotix Fireball Comet CNC Router - Info on the CNC router I am using. Melissa from Probotix has given me permission to use image files from the Probotix website.

And of course Google and Youtube.

Step 2: Materials

  • RAMPS 1.4 kit with LCD display

  • Three(3) equal lengths of extruded aluminum. AKA, linear rails or T-slot rails. Size is based on your preference, availability and fabrication capabilities. My build is using 20mm x 20mm x 750mm. If given the chance to redesign and build another, I would choose 20x40mm for increased rigidity.

  • Nine(9) rollers -aka, wheels - that fit into the track of the aluminum extrusion. I went with nine wheels since I have a low budget.

  • Half inch thick Medium Density Fiberboard (MDF), approximately 4 square feet should be enough material.

  • ABS plastic sheets 3/16 inch thick. Approximately 2 square feet should be enough.

  • ABS plastic sheets 1/8 inch thick. Approximately 6 square inches.

  • Four(4) NEMA 17 stepper motors, I used stepper motors with 200 steps per revolution.

  • Three(3) Nema 17 motor brackets

  • One Hot-end assembly set up for Bowden feed

  • One extruder assembly set up for Bowden feed and can fit a NEMA 17 stepper motor

  • One meter of PTFE tubing with diameter necessary for filament.

  • 12V power supply. I purchased a 30 amp from Ebay.

  • Three(3) GT2 pulleys with 16 or 20 teeth (all three have to have same number of teeth).

  • Three(3) smooth idler pulleys

  • Five (5) meters of GT2 timing belt, 6mm wide. Cut into 3 equal lengths.

  • Three(3) limit switches or optical switches

  • One (1) 300-400mm carbon fiber tube with 6mm inside diameter(ID). The ID has to be 6mm. Cut into 12 equal parts.

  • Three(3) 500mm carbon fiber tubes with an outside diameter (OD) of 6mm. The OD has to be 6mm. Cut into six segments exactly 250mm long.

  • 12 cylindrical neodymium magnets 6mm diameter x 10mm long

  • 12 ball joints or steel ball bearing 10mm – 12mm diameter - Remember HAYDN in Google Groups. Get this from him.

  • 12 nylon or aluminum spacers 1/4" long and large enough for #6-32 screws to pass through.

  • 50 T-nuts for the aluminum extrusions (we will not need that many but extra always helps)

  • 15 right-angle corner brackets

  • Gorilla glue Wood glue PVC / ABS solvent glue

  • 1/2" x 1/2" u-channel aluminum extrusion - purchased from Lowes. Cut three(3) 4" segments

  • 1/2" x 1" L aluminum extrusion - purchased from lows. cut three(3) 1" segments

  • Borosilicate glass print bed surface

  • Three 1" x .5" x .5" aluminum or plastic blocks. Something that can be drilled and tapped.

  • short length (about 6") #8-32 threaded rod, cut in half.

  • Two #8-32 wing nuts.

  • Assorted screws, nuts and washers. We will need lots of metric in M3 and M5 size and standard size in #4-40 and #6-32.

  • Cables and wires

  • Zip ties and cable ties.

  • Handful of 4 pin and 3 pin Dupont connectors.

Step 3: Tools

  • CNC router
  • Mill and various milling bits.
  • Set of work hold-down clamps and jigs.
  • Soldering iron
  • Drill and drill press
  • Band saw
  • Hacksaw
  • Screwdriver set
  • Heat gun
  • Digital calipers, very accurate ruler and dial indicator
  • Rafter square
  • Wire stripper, wire cutter, assorted pliers, forceps and tweezers.
  • Bubble level
  • Eye protection
  • Ear protection
  • ¼ inch router bit
  • 1/8 inch router bit
  • 1/8 inch Onsrud plastic mill bit – part 63-763 - Necessary for routing ABS
  • clothes pins

Step 4: Software

This is the suggested list of software you will need to learn to use.

  • Arduino IDE - To update the RAMPS firmware to the latest Marlin version configured for Delta Printers.
  • CAD and CAM software - I used CamBam. It is not that expensive and they give a generous trial period.
  • G-code interpreter - I use EMC2 or LinuxCNC.
  • Sketchup
  • Slic3r
  • Pronterface

I am sure there are others and feel free to comment on the software you like to use.

Step 5: Design the Frame

Since I have not actually seen a delta printer in person, the best I could do was make some guesses from images and youtube videos.

I concluded that making a delta printer is simply a matter of drawing an equilateral triangle, and placing vertical post on each corner. For the post, I decided to use extruded aluminum with t-slots in the 20mm x 20mm size. Most manufacturers of these extrusions provide dimensional drawings of their product. The company called OpenBuilds even provided the actual dimensional files in Sketchup. Mark from OpenBuilds gave me permission to share those sketchup files here. It is attached on this page. I took these dimensional files and opened a hole on each corner of the equilateral triangle that the extrusions snugly slide through and force them to be perfectly vertical. For the 20mm x 20mm extrusion, I made the holes 20.4mm wide. I have not actually personally seen or held an aluminum extrusion so I am just hoping my design works.

My MDF board is not thick enough to hold the extrusions securely, so I designed a second layer (auxiliary layer) that glues onto the base. I made sure that the placement of the extrusion made them equal distance from each other and that they are all facing the center of the triangle. Screw holes were also drilled for the stepper motor mounting brackets.

At this point I ordered one 1500mm x 20mm x 20mm and a 1000mm x 20mm x 20mm aluminum extrusions some, t-nuts, M5 screws, and corner brackets from OpenBuild,

Attached to this page are the cam, dxf and g-code files for the base. There will be no STL files as those are more common to 3D printed parts.

Step 6: Make the Base

Secure a large piece of medium density fiberboard(MDF) (1.75 square feet) onto the spoil board of the CNC router. Make sure everything is level. Zero on the CENTER of your mdf board. For the drill holes use a 1/8" drill bit. After the holes are drilled out change to the 1/4" wood router bit. Make two of the base boards.

After the base and top are made, make three of the auxiliary supports. For these pieces you will need an MDF board 8" x 14". Zero on the LOWER LEFT corner of the board.

For mdf, I have the cutter feed rate at 30 inches per minute and the router running at 3/4 from full speed. Slower feed rates and/or faster spindle speed causes the bit to overheat and the wood to burn.

Cut the aluminum extrusions (rails) into 750mm lengths with a hacksaw. Tape them together and put on the mill and mill the ends so that they are all exactly the same length, flat and smooth.

Once everything is cut out, get six(6) 1.25" long #4-40 screws and six(6) nuts. Get out the wood glue. Glue a support to each corner of the base. Secure the parts together with the screws. Temporarily insert an aluminum rail on each corner to make sure the base and the auxiliary support are lined up and square. Use a rafter square to make sure the rails are perfectly vertical.

Attach the stepper motor mounting brackets underneath using twelve(12) #6-32 x .75" screws 24 washers and twelve(12) nuts. It is not necessary to tighten these screws yet. Later, the stepper motors will need to be positioned so the belt clears the hole through the base without rubbing.

Step 7: Finish the Frame.

Once the glue is completely dry, place a corner bracket on each side of the rail using t-nuts and m5 screws. Mark the holes on the base and drill out the holes using a 3/16" bit. Secure the vertical rails with m5 x 40mm screws and lock nuts. as you tighten, make sure to keep checking that the rails are perfectly vertical and perpendicular to the base.

Attach the top cover using the same method with the corner brackets.

There is a third corner bracket on each rail at the top. That third bracket is for the belt tension screw. i will explain how that works when we get to the page about the belts and pulleys.

Step 8: Design the End Effector

The end effector is is an equilateral triangle just like the base, just much smaller. The design has a hole in the center for the hot end nozzle. on the sides are six steel ball joints.

The end effector needs to be light, so I initially designed and made it out of ABS plastic. End effectors are normally printed parts since the part is not mechanically stressed. Again, since i do not have a 3D printer, this will be machined on the CNC.

The initial design was for a generic end effector. However, once the j-head hotend was purchased and I made my measurements, i determined that the initial design was not going to work.

Version two is made of MDF because I fear the ABS may melt from the heat. It is also slightly larger to make up the much larger hole I had to put in the middle.

Attached are the cam files, dxf, and gcode (.ngc extension) files.

Step 9: Make the End Effector

The picture has a piece of ABS plastic being milled out but I changed the design and the material after I purchased the "hot end", the nozzle that melts and oozes out the plastic.

Get a piece of MDF and mill just like the base using a quarter inch cutter.

Install a #43 drill bit for the drill holes. These holes will be tapped for #4-40 screws. once the holes are drilled out, switch to the 1/8 inch wood routing bit. Zero the Z axis on the bit and start milling.Use the same settings used for the base. Only change would be the depth increment is smaller since the cutting bit has a smaller diameter.

Once the part is successfully cut, tap all the holes with a #4-40 tap. Screw in FIVE(5) ball joints into the corners. Set one ball joint aside.

Step 10: Design the Carriage

The carriages travel vertically on the rails. The up and down movement of the carriage on each tower is translated to x, y and z movement through six arms.

The carriage will have wheels that follow the tracks of the vertical rails. The wheels need to be adjustable to make sure they ride smoothly and do not wobble. On my carriage design, the wheel tension is adjusted by having the third wheel on an arm. A 1.5" #4-40 screw is used to adjust the tension on the arm (please see picture in step 12).

I went with 3 wheels per carriage because of my restrictive budget. The wheels are expensive.

On the carriage will be four components; wheels, anchor points for the GT2 timing belt, a flag for the end stop, and two ball joints to connect it to the end effector. Two pieces need to be machined to make the carriage.

The dxf, gcode, and cam files are attached.

Step 11: Mill the Carriages

Glue together 2 sheets of 6" x 10" x 3/16" ABS to make a thicker sheet. You can get the glue from the plumbing section of a hardware store. Let the glue cure for at least 24 hours, preferably 48 hours. The glue is a solvent and tends to make the ABS gummy in the middle until it is fully cured. If you start milling before the glue cures, plastic builds up on the bit and you risk damaging the work, breaking the bit, or both. Let the glue cure completely. I'm gluing layers of ABS together because it cost a fraction of what the material would cost if I purchased it in the thickness I needed.

Three(3) of the carriage.ngc file willl be milled from this sheet. The CNC machines feed rate needs to be slow. Nothing faster than 12 inches per minute. The plunge feed rate is even slower at 8 inches per minute.

The zero point of these parts are on the LOWER LEFT corner. Zero each part to maximize the number of items you can make out of the 6" x 10" sheet. There should be plenty of room.

There are two size drill holes on the carriage plate, six(6) holes are #35 drill bit, and two(2) holes for the ball joint use #43 drill bit. The gcode i generated should pause for a bit change. The carriage is milled using a 1/8" Onsrud bit.

Milling ABS requires a slow and shallow approach. Feed into it too quickly or have the spindle spinning too fast and the plastic starts melting. The melted plastic builds up on the bit and the work gets ruined or the bit breaks. You cannot start cutting into ABS and leave the work unattended. You must watch it with a hand on the emergency stop in case the plastic starts melting. I am describing in inches, but the design is actually in millimeters. I have the spindle speed set to 1/3 speed. the spindle power setting is actually so low that it is on the threshold of not spinning up when first turned on. The feed rate is set to 8-10 inches per minute and plunge speed is 5 inches per minute. The Comet can jog at 200 inches per minute, so as you see this is going to go very slowly. Layers are removed at 1/16" increments. Blowing cool air on the plastic when it is being cut also helps.

You may ask why I set up the gcode to make each part individually. Why not just stick all three carriage plates into one gcode(ngc) file and cut everything on one go? My reasoning is that ABS is a tricky material to mill. If something goes wrong and I have to reset, I would rather reset on just one part versus a batch of parts. It does mean I have to reposition my zero for each part, but that also give me greater flexibility, especially when something goes wrong.

Secure the 6" x 6" x 1/8" ABS sheet to the cnc router. Mill three(3) carriage_belt_retension.ngc parts. The part uses #35 drill bit for the holes and 1/8" Onsrud bit for the milling.

Step 12: Assemble the Carriage

You may have noticed that in my gcode files I don't drill through the material. The reason for that is that even at the lowest setting, the router spindle speed is too fast and the drill bit tends to melt the plastic. I am basically just using the CNC to mark the drill points, making a divot to accurately drill with a drill press. Once all the parts are milled out, take the parts to a drill press and drill all the holes all the way through.

Tap the carriage body holes with a #6-32 tap and the ball joint spacer holes with a #4-40 tap.

Take the carriage plate and drill a 5/32" hole a 1/4" from the end of the arm where the third wheel is mounted. Drill only the arm. See picture.

Now take the #43 drill bit, pass it through the hole in the arm and drill into the carriage body (see pictures). Tap the #43 hole with a #4-40 tap.

Get a a 1.5" x #4-40 screw, place a washer on it and pass the screw through the 3/16 hole in the arm and screw it into the carriage body. Do not tighten the screw.

Install the two(2) ball joints.

Pass four(4) #6-32 x .5" screw through the belt retention plate. Then place a washer, .25" spacer, and a washer on each screw. Attach this assembly to the carriage body.

Attach wheels 1 and 2 (not wheel 3 that is on the arm) using a M5 x 40mm screw and the provided spacer that came with the wheels. Secure with a lock nut.

Repeat two more times except SET ONE ball joint aside so that only FIVE(5) ball joints are used.

Step 13: Make End Stop Flags

A u-channel is machined into a blade that will interrupt the beam of the optical end stop. The screw hole is slotted to allow blade height adjustment.

To make this I cut a 4 inch section of u channel with a hacksaw. Then I made slots on one end that is 3/4 inch long. Finally, I used a hack saw to cut away one side and the top. I cleaned up the hacksaw cut with the mill. The end result is that I have a 3 inch blade or flag. Make two(2) more just like it.

Attach this machined piece to the carriage using two(2) half inch #6-32 screws. Add a spacer between the carriage body and the flag. I used 1/8 inch plastic shims that were meant for my quadcopter props. Do not completely tighen screws. The height needs to be adjusted later.

Repeat two more times.

Step 14: Attach Carriage to Rails

Insert a M5 x 40mm screw through the arm of the carriage, then the spacer. Slip the carriage assembly onto the rail and slip on the third wheel.

Do not let the carriage drop to the bottom. It could damage the wheels. Use clothes pins clipped to the rails to hold the carriage near the mid-point of the rail. Wood clothes pins don't work. Use plastic clothes pins.

Secure the wheel with a lock nut. tighten the #4-40 screw on the arm until the carriage glides smoothly on the rail.

Keep the clothes pin on the rail.

Step 15: Belt and Belt Pulleys

Cut your 5 meters of GT2 x 6mm timing belt into three(3) equal parts.

Attach a 20 teeth pulley to the stepper motor shaft. Tighten the set screw so the pulley does not slip on the motor shaft. Attach the stepper motor onto the motor bracket using M3 screws. Adjust the motor position so that the pulley is visible and centered on the hole in the base.

Loop one end (about an inch) of the belt around one of the belt retention posts on the carriage. Zip tie the belt loop in place. the belt teeth will keep it from slipping off.

Thread the timing belt through the hole in the base and loop it around the pulley. Make sure the teeth of the belt and the teeth of the pulley engage. Thread the other end of the belt through the center of the carriage, between the posts of the belt retention plate.

The smooth idler pulley will need to be mounted on some sort of sliding block. This block can be metal, plastic or hardwood. The material is not critical. What is important is that there is a hole in the center that a M5 screw can fit through and a #4-40 threaded hole on the top center for a screw to go through. I used an aluminum block that was 1" high by 1" wide by 1/2" thick. I drilled a hole slightly larger than 5mm about 1/4 inch from the bottom edge.

Pass an M5 screw through the smooth idler and through the block. Insert a t-nut into the rail and screw the M5 screw into the t-nut. Tighten enough so the smooth idler pulley is secure but the block can still move up and down the rail with some resistance.

Pass a long (about 2.5") #4-40 screw through the third corner bracket at the top portion of the rail and screw it into the #4-40 hole on top of the block. Screw it in just enough where the screw grabs.

Loop the timing belt around the smooth idler pulley and loop the end around the top post of the belt retention of the carriage. Pull the slack from the belt and secure the belt with another zip tie. cut off excess so only about an inch is left.. Add another zip tie on each end about half an inch from the first zip tie.

Adjust the stepper motor brackets so that the timing belts do not rub against the mdf base. Tighten the nuts securing the motor bracket once the belt is properly centered.

Add tension to the belt by turning the screw in the third corner bracket at the top of the machine clockwise. This will cause the block (and the smooth idler pulley) to move up pulling the belt tight.

Step 16: Six Rods, Three Parallelogram

The up and down movement of the carriages is translated into x, y, and z movement of the effector through six identical rods that form three parallelograms. In my build each rod has neodymium magnets that attracts the metal ball joints. This magnetic connection creates a linkage that has zero backlash/play.

Take the carbon fiber tube(s) with 6mm inside diameter(ID) and cut into 12 equal parts 1" in length.

Take the three(3) 500mm carbon fiber tubes with an outside diameter (OD) of 6mm and cut into six segments exactly 249mm long (I am accounting for loss of material from the hacksaw).

Get the two(2) ball joints that were set aside previously.

The most important part of making the rods is that the length of all six has to be identical. To accomplish this, I made two jigs. The first jig was designed to set the magnet inside a tube and create a "cup" that the ball joint sits in. The jig maintained proper cup depth while the glue dried. I'll call this first jig the "ball joint cup jig" because I lack imagination in naming things.

To make the ball joint cup jig, get two strips of mdf or wood that is 1" wide, 3" long and 1/2" thick. Drill a #43 hole in the center of one of them - this will be the bottom. Stack the two pieces of wood and drill two 9/64" holes equal distance from each end. Take the top piece of wood and enlarge the holes with a 5/32" bit. Take two #8-32 threaded rods and thread them through the 9/64" holes of the bottom piece. Take one ball joint that was set aside and thread it through the center hole.

To assemble the magnetic ends with the ball joint cups stick a magnet on the ball joint. Next carefully apply a thin layer of Gorilla glue inside the tube making sure NOT to get any glue on the lip of the tube - apply the glue about 1/16" away from the edge. Be sure to follow the Gorilla glue instructions (add moisture to the parts). Slip the tube over the magnet and press against the ball joint. Attach the top portion of the jig and tighten with wing nut. Wait for the glue to dry for 1-2 hours. Repeat eleven more times. After your done making these remove the ball joint and set aside.

the second jig is a piece of mdf that is 310mm long. Drill a #43 hole on one end and another hole 285mm away. Make one of the holes slightly over-sized so that the ball joint screw will slip in snugly without needing to be screwed in. Do NOT make the hole so large that the ball joint can wobble. The ball joint has to slip in snug. Take one of the ball joints and screw it into the other hole.

Take one of the magnetic ends (ball joint cup) and stick it to the ball joint. Put a drop of Gorilla glue inside the tube opposite the ball joint. Insert the one end of the 249mm carbon fiber tube into the magnetic end. Take the other ball joint and stick another magnetic end on it. Add a drop of glue inside the tube opposite the ball joint. Insert the other end of the 249mm tube into it. Stick the loose ball joint into the over-sized hole. Adjust the whole assembly so that the rod is perfectly straight and roughly equal amounts of the thinner tube is inside both of the larger tube stuck to the ball joints. Let glue cure for 1-2 hours. Repeat five(5) more times.

Because I made each part one at a time using the same jig for each part, each rod, is exactly the same length as the others. Unfortunately, the whole process took three days.

At this point connect the end effector to the carriage using the newly created magnetic rods.

Step 17: End Stop Optics

The machine has to be able to know carriage position. This is accomplished by the G-code command G28. But for the G28 command to work, there needs to be sensors at the three home points. The sensors could be micro-switches, hall effect, or optical. Which kind is used does not really matter. My design has three optical sensors purchased from ebay.

First purchase a L-bar aluminum extrusion from the hardware store the L dimension should be 1" x 1/2". Cut three 1" sections. On the larger surface mark and drill holes to mount the end stop sensor using a 1/8" drill bit. Drill two more holes on the other surface of the aluminum about 1/4" from the edges. Attach the sensors to the L-bars using #4-40 screws and a nut. It may be necessary to use 1/8" nylon spacers so that the circuit board does not short out on the aluminum L-bar.

Wrap tape around the tip of the end stop flags on the carriage so that the optic sensors can be wedged onto the flag. If necessary, tape the sensor to the flag to keep it secure (see picture). Raise the carriage so that the aluminum L touches the top of the printer. Mark the holes and drill through using a 1/8 bit.

Lower the carriage and remove the optical sensor. Attach the sensor to the top of the printer using #4-40 x 1" screws and nuts. Repeat for the other two towers.

Remember to remove the tape on the flags.

Step 18: Route Wires From End Stop Sensors

For the end stop sensors, i cut the wire that came with it and soldered on red, black and green 22 gauge wires for easier identification. I routed the wires through the center of the rail so they are out of the way. Where the wires meet an aluminum edge I applied shrink tubing for added protection.

Step 19: Update the RAMPS 1.4 Firmware

Please visit the Delta Robot 3D Printer Google Group to ready about different opinios on different firmwares.

Delta Robot 3D Printer Google Group

The go to RepRap Wiki for explanation on the different firmwares.

RepRap Firmware Wiki

I personally used the Repetier Firmware and using Repetier Host as my computer software.

Please go Repetier Website for more info and to get the firmware.

Launch the Arduino IDE and upload the latest Delta firmware to your Arduino Mega 2560. If you are not familiar with uploading sketches into Arduino boards, please search Instructables for Arduino tutorials and visit the Arduino website.

Introduction to Arduino by Rondofo

Arduino Home

Step 20: Design a RAMPS 1.4 Holder and Display Stand (Optional)

This step is optional but we need something to house our electronics. It is possible to just put the RAMPS board in a shoe box. But why? We have access to a CNC router after all.

The design basically connects to the base of the 3D printer and puts the LCD display at a comfortable 45 degree angle for easy viewing.

Please see attached cam, dxf and g-code files.

Step 21: Make RAMPS 1.4 Holder and LCD Stand

Use the same mdf as the base board cut 7.5" high by 12" wide. Secure to the CNC router - I screwed it into the spoil board.

Cut two of the RAMPS holder design with the CNC. Use a 1/4" wood router bit. for the drill holes use a 1/8" drill bit.

Cut a rectangular section of mdf 4" x 1.75". Either use the CNC router or cut with a band saw then mill to make the rectangle's corners as 90 degrees as possible and the edges as straight as possible.

Wood glue the rectangle to the RAMPS holder as shown on the pictures. Make sure everything is square and straight.

Once the glue dries, the whole assembly should fit snugly onto the base board of the 3D printer. Make it a permanent installment with wood glue or a removable structure with screws. I went with screws so I can remove and relocate the holder if I have to.

Place the RAMPS 1.4 on the lower shelf. Mark screw holes and drill holes in mdf.

Place LCD display on the 45 degree slope. Mark screw holes and drill holes in MDF.

Secure the electronics with screws. use nylon washers where necessary.

If desired, cut a clear acrylic shield to go over the display.

Step 22: Wires, Wires Everywhere

The crazy first picture is not the reality. It is how my mind perceives the wiring of this project. In reality, there really is not that many wires involved. Eighteen wires total for the motors, nine for the sensors, two for the thermister, two for heating element, four for power. So, not really that much wiring.

Follow the RAMPS wiring instructions found in the RepRap wiki.

RAMPS 1.4 Wiki

Also reference the technical specs and white papers of your motors on how to wire up to a stepper motor controller. Many build blogs and youtube video tell you to short two wires or test continuity to figure out how to connect the stepper motor wires. Well I say the most sure fire way to do it is to get the technical specification paper of your motor and follow the wiring diagrammed in there. You cannot really go wrong when following the diagrams that came from the manufacturer, unless the manufacturer could care less about writing an accurate white paper for their products.

If you look at the connection pins for the stepper motors on the RAMPS you will see "2B 2A 1A 1B" and as seen in the stepper motor diagram they designated there wires B2 = Red, B1 = Yellow, etc. This was specific to my stepper motors, but it takes the guess work out of wiring and no need to "short out two wires" to find pairs. We are making a finely tuned machine here. If white papers are available, use them.

As you can see the RAMPS board has stepper motor connections for X, Y and Z axis. Since all our motors travel the same directions (up and down) see the seventh picture to determine which motor connects to which axis.

For the stepper motor wires I kept the existing wiring.

For the sensor wires, the RAMPS board has an "S" for the green signal wire, the middle pin is the black ground wire, and the last pin is plus volts red. I am only sensing for the top end stop limit which is the MAX movement. That means that I will be connecting end stops on on Max-X, Max-Y and Max-Z. So, looking at the pins left to right I am plugging in at 2, 4 and 6. That is just me. You may decide to put six sensors and fill up all the end stop pins.

After all the wires are connected clean up the wiring. Do some wire management to make sure none of the wires accidentally get bound up into moving parts.

Step 23: Test Movement

With everything wired up, it is time to test the movement of the delta 3D printer.

Test the homing function. With the power OFF, move all the carriages to the middle. Apply power to the printer and navigate the menu on the LCD to "Prepare" then to "Auto Home".

If any carriage starts moving down, press the "Stop" button on the LCD circuit board. Turn the power off. Flip the stepper motor around and turn the power on again. Test movement with the "Auto Home" again. Make sure to keep your hand on the stop button. Stop the machine if the carriage crashes through the end stops.

If the machine crashes, double check your configuration.h settings. Change the size parameters or perhaps reverse some logic like configuring something from "false" to "true". Or perhaps a line of code needs to be commented out (//) like a command command for a pull up resistor.

The configuration.h will need to be tweaked later with all the peculiars of the delta, but for now, just getting the direction of travel correct and getting the machine to stop on end stop sensors is enough. I'm still actually learning the ins and outs of the Marlin firmware, so please forgive me if this part of the instructable is lacking. I will continue updating this ible as a learn more.

Step 24: The Extruder and the Hot End

I purchased a "bowden" extruder and version 6 J-head hot end from Ebay.

Please stay tuned for updates to this instructable as I show how these new items are used.

Step 25: Install the Heated Print Bed

I decided the printer needed a heated print bed. I purchased a set that included a 200 mm Pyrex glass disk.

Please stay tuned for updates with pictures and the process I used to install to the printer.

Step 26: Finished!!!!

So here is the end product after many, many months of working on it. Sorry it took so long but building something so complex by buying parts $20 at a time once a month will really make a project drag on.

Pictured is a "treefrog" I printed on the DeltaBot. Model by Morena Protti downloaded from Thingiverse. It was the second thing I printed. The first being the "lattice pyramid" by MINIFACTURE. The treefrog was printed at 75% actual size. i lost a little bit of detail due to the reduction. Print time was 17 minutes and some seconds.

Here is the video of the DeltaBot 3D printer going a movement test. People observing it keep asking me, "why do you have it playing weird music while you print?" and "Where are the speakers?" I have to explain that the "music" is actually the natural sound of the motors running.

OK, so I was thinking about it and thought, why don't I demo how the printer's stepper motors can sound like music and downloaded this gcode from Thingyverse. Credit for the gcode goes to Thingyverse user "Alwinman." You may need to turn up your sound on your speaker to hear the stepper motors.

NOTE: I will update this instructables with the last few missing steps - "Mounting the Hotend and Extruder", "Installing the Bed", "Place an Emergency Stop", and "Calibration."

<p>Is it me , but I don't see any DXF for the base or the auxiliary layer</p>
<p>Does any have Full size drawings that I can get?</p><p>Thanks</p>
There are some DXF to PDF convertion software available on the web. Download one and do the convertion and that should get you the full size drawing you want. Technically, my CAM and DXF files are full size drawings. I have stuck a pen to my CNC machine and plotted my drawings. Find a friend or a business with a plotter. Plot the dxf file. Remember, the plans are in millimeters. <br><br>Here is a forum where they talk about what you are asking and people provided some interesting (and hopefully useful) answers.<br>http://www.rcgroups.com/forums/showthread.php?t=72493<br>
<p>I use you project for basis for my own.</p><p>Now thinking about material for carriage, i have plywood 4mm and acryl 5mm. both them are soft. crumple under load.</p><p>how do you plan to fix the heating table</p>
<p>Hi!</p><p>Why are you using smooth idler pulleys? They're deformed belt teeth. Smooth rollers put in reverse bending, smooth side skirts smooth roller, toothed belt side skirts </p><p>tooth roller.</p>
Hi, Good point on the smooth idler pulleys deforming the belt teeth. I had not thought of that. I really appreciate the feedback. I used them because they were convenient. I did not find idler pulleys with teeth that had matching bearings.<br><br>Also, it should be noted that the with how tall the Delta printer is, the part of the belt that makes contact with the smooth idler pulley will never make contact with the geared pulley on the stepper motor.<br><br>I can see how deforming the teeth will slowly introduce slack on the belt. But the idler pulley is adjustable so that I can adjust the tension on the belt. If I am noticing too much slack, i can just add more tension.<br><br>Again, thank you for making me think about this. It helps me a lot.
20x60 mm instead of the 20x20
<p>Nice. Great idea.This goes back to what I was saying in the instructable. That if I could start over again, i would not have gone with a square rail. I would have gone with a rectangular rail that was 20mmx40mm. </p><p>THANK YOU!!! You're doing what i wish i would have done. Awesome 3D model by the way. what software is that?</p>
That is Sketchup. I love that software. Very easy to learn and use. Thanks for the great idea. I'm planning to build it sometime. I'll keep you updated. How's your project coming along? About finished?
<p>almost. I took everything apart because I did not seal the wood. I purchased a can of Zinsser clear spray shellac and gave everything 3 coats with light sanding between coats. That should protect the MDF. I'm planning on adding a step here about that.</p><p>I also moved the control box. It's now mounted to one of the rails. Because of that move wiring for 2 motors and 2 end stop sensors have to be lengthened and rerouted. I still have to purchase the heat bed and glass surface. I hope to have it finished soon.</p>
<p>Hi!</p><p><a href="https://www.instructables.com/member/MrTinkerer" rel="nofollow">MrTinkerer</a> You write &quot;If given the chance to redesign and build another, I would choose 20x40mm&quot;, how would you have arranged a wide side? along a circle or from the edge to the center?</p>
<p>Hi,</p><p>Please see JustinC76's revision of my design in the comments. He went one step further and did a 20x60mm. It's exactly what I would have changed on my design except 20x40mm with the 40mm side facing towards the center.</p>
<p>I really like your design I just built me a cnc mill machine and now want a 3D printer as well I was thinking to use vslot 20x60mm extrusion That way the belt stays in the center I already made a sketch up model that I could upload later on I will have to change the carriage to fit a wider vslot Great work, you've definitely inspired me to build something fun Thanks</p>
<p>Is there full size drawing some where I can get?</p><p>Thanks</p>
<p>I tried making a PDF of the DXF files but the actual dimensions gets messed up. Everything has to be machined to very tight tolerances so I gave up on making actual size pdf files.</p>
<p>Bought an extruder and hotend. Working on finishing this project now. There is going to be some design revisions.</p>
<p>Congrats on getting an extruder! Configuring a delta can be a huge pain in the butt. Let me know if you have any questions!</p>
<p>Any updates?</p>
<p>Sorry for the lack of updates. I have redesigned the RAMPS and LCD holder to bolt on to one of the rails. I will update that page this weekend with the new design. The old design took up too much room on the build platform. </p><p>I really don't like talking about personal finances... but i have been lean on funds. I have been setting funds aside weekly so that I can finally buy the extruder and hot end. But for every step towards that goal, i take a half step back with life's little surprises. That's just how life goes... </p>
<p>Boy, I know what you mean!</p>
<p>Is there full size drawing some where I can get?</p><p>Thanks</p>
<p>I have the DXF files on page 5, 8 and 10. All files are scaled in millimeters. Do you have a large scale printer or a plotter by chance? I do not have the ability to convert to PDF which would be the ideal format for printing on paper. Perhaps someone in instructables can do it for us.</p>
<p>Actually you do. You need to get something called &quot;PDF Creator&quot;...it will allow you to &quot;print&quot; to a PDF as it functions like a virtual printer, but it instead saves it as a PDF. </p>
<p>thank you. I will work on this and get the PDF files attached to this instructable.</p>
<p>Any updates? Did you get the hot head?</p>
<p>not yet. waiting on funds :(</p>
<p>Thanks for the great Instructable. Can't wait to see the next steps.</p>
<p>Looks great! What accuracy can it achieve? and how fast can it print?</p>
<p>i don't know yet. I have the stepper resolution set to the highest setting which should at the very least give me 100 microns. As for speed, again, i don't know, but deltas tend to be very fast. I think what will limit my speed would be the quality of the print head which I still have not ordered yet. My budget for this project is very strict. And the biggest cost is the print head. I will still be under my $400 budget when I am done.</p>
<p>so far as I know this tripod printer design suffers at accuracy at high speeds but even if you get 100 microns at high speed, you beat most DIY low cost printers :)</p>
<p>I'm not really after speed for my printer. I really just went with this design because I really like the kinematics.</p>
<p>Cant argue with that :)</p>
<p>Nice build!</p><p>One suggestion for improvement and making it simpler : mount the optical end switches onto the end of the v-slot profile and redesign your carriage to have a 'finger' which interrupts the optical end switch.</p><p>Furthermore I think rigidity of your base and top towards your v-slots is crucial for horizontal accuracy. So I would make it in another material rather than MDF..</p><p>As soon as my OX CNC is ready, I will embark on this idea!</p>
<p>Thanks for the suggestions. A soon as the machine is printing accurately, i plan on printing new carriages (got to love the concept of a machine that can build better parts for itself). I will probably build the carriage as you suggested and put the endstops on the rails.</p><p>As for the MDF. 1) the idea was to make a 3D printer with materials on hand or easily accessible/affordable. 2) 1&quot; thick MDF on each end is plenty rigid. Especially since I have the corner brackets bolted on as well. 3) MDF machines very well with very tight tolerance and is fairly uniform in thickness compared to plywood or other wood boards. 4) other materials such as aluminum, delrin, ABS, PVC, garolite, fiberglass, carbon fiber, etc would have been cost prohibitive in the size required (see again, point number 1). Perhaps give a specific example of a material that is comparable to MDF in cost, availability, and machinability (spelling??) that I could have used and if I have the chance I can try it out.</p>
<p>you are right about MDF, its unbeatable in terms of cost. I assumed you would use 18mm but now I see it is actually 1&quot; thick.</p><p>Over time, MDF absorbs moisture and then loses some of its stiffness, so you may treat it with some coating.</p><p>Instead of MDF, I had Garolite in mind.</p>
<p>Yes, Garolite is an awesome material. I actually use a lot of it on my camera platform builds (mostly .060&quot;). I love working with that stuff. Unfortunately a 3/4&quot; thick 24&quot;x24&quot; sheet of garolite xx is $107 at McMaster Carr. That's the cheapest source I found that does not gouge you on shipping. For the MDF, good point about moisture. I was planning on treating it with wood sealer.</p>
<p>good</p><p>thank you</p>
<p>Hello, In the next week or so I will be switching and reconfiguring the firmware of this 3D printer from Marlin to Repetier. From what I have been reading, Repetier seems to be easier to configure and understand. If it is successful, i will be updating this ible with more detailed steps on how to configure this delta for optimum performance and accuracy. Stay Tuned. Thank you.</p>
<p>Great build, I've been thinking of making a 3D printer as well, and I have a DIY CNC at home to mill parts. I was wonder why exactly you choose to do a Delta style machine, as opposed to a more traditional gantry style (I've also considered making a conversion kit for my CNC so I can swap back and forth and reused the same carriage system).</p><p>Just curious, again great job and excellent documentation!</p>
<p>A friend asked me the same question. My main response is the simplicity of the mechanics. Where a Cartesian 3D printer needs to have a separate design for each axis, with a Delta approach, I designed 1 moving part and repeated it 3 times. If someone else had not already figured out the math, i would not make the delta.</p><p>The other reason is aesthetics. I see the motion of the delta to be quite elegant.</p>
<p>Cool, that makes a lot of sense. They definitely look awesome. Is there an appreciable difference in accuracy/resolution between types?</p>
<p>If done right, the print quality should be the same. Rigidity is important in both styles of 3D printer. But Delta's tall towers are more prone to wobble. This is why I doubled up the MDF where the rails go into the base. I would even consider doubling up the MDF on the ceiling but for now I am not perceiving any wobble. Or maybe using 20mm x 40mm rails. I am learning as I go. I am sharing my adventure with y'all.</p><p>Oh, i forgot to mention a third benefit of Delta - <strong>SPEED</strong>. Since the end effector has less mass due to the Bowden system, the print head can accelerate, decelerate, and take corners more quickly than Cartesian. Delta's tend to move really fast.</p>
<p>grandi complimenti lavoro</p><p><br></p>
<p>grazie per complimento. Grazie per la visualizzazione delle Instructables.</p>
<p>great job compliments</p>
wow! really great job! I'm looking forward to the completion of the project!
<p>me too ;)</p><p>Thank you for the compliment.</p>
<p>great construction!</p>

About This Instructable




Bio: I like to tinker and create things. When I have time, i make stuff. The stuff could be as simple as my patent pending spoon-on-a-stick ... More »
More by MrTinkerer:3D Printers Stuff Expanded Spool Holder, Filament Guide, & Spool Hub Fidget Spinner - 3D Printed 
Add instructable to: