Theory
3D printers
the good:
- Easy complexity in parts, means low part count
- Easy to operate
- Common practice. Thingiverse and reprap
the bad:
- Slow
- Unreliable
- Inaccurate – (everything warps and parts require heavy post processing: filing, drilling, edge breaking etc.)
- Weak materials (even fiber reinforced filaments suffer from z axis delamination)
- Only suitable for making small parts
Laser cutters
the good:
- Super easy to use
- Fast
- Reasonably accurate (hampered by concave laserbeam and mirror alignment
the bad:
- Only 2D cut and raster engrave with tricky Z depth
- Weak materials (Delrin/POM is the only lasercutable material you can't break with your fingers)
- Designing 2D to become 3D requires more skills and assembly
CNC mills
the good:
- Very fast (a ShopBot alpha cut at the speed of a 700w lasercutter)
- Very accurate
- Real hardcore materials (Aluminum, brass, bronze, steel and thick Delrin/POM)
- High resolution 3D surfaces
- Reasonable geometry complexity
- Can make tiny and precise parts
- Can make really big parts
the bad:
- Twosided milling or mulitpart assembly require more design and operator skill
- Most Fab Labs have shopbots in terrible condition (unreliable, or low performance)
- Parts needs fixturing, requires skill and planning
- Big machines are scary
- Current software toolchain sucks
- Tensioned wire - example: Thorbjørns hang printer
- Articulated joints - example: Thor DIY robot arm
- Unsupported linear rail - example: most diy 3D printers, Nadyas MTM stages
- Supported linear rail - example: most standad fab lab machines, the chamfer rail system
Our chamfer axis design fascilitates both supported and supported linear axis config. For making good CNC milling machines, we bolt the axis to the frame or gantry of the machine at even intervals.
- Simple wire
- Timing belt
- Lead screw
- Rack and pinion
- Linear motor
We have identified rack and pinion as the easiest system to fabricate from scratch using standard fab lab machines (ShopBot). In order to have as little backlash and as high resolution/torque as possible we have developed [our own](https://github.com/fellesverkstedet/fabricatable-machines/Module development/cnc-friendly-rack-and-pinion) geometry generator for CNC mill friendly rack and pinon.
Three levels of steppers price and performance
- Wantai
- Oriental open loop
- Oriental closed loop
Closed VS open loop. The simplicity of steppers
- Polulo - integrated in ramps
- Toshiba - danieles satsetp
- Oriental motors paired motor and drivers
Stand alone:
- Ramps 1.4
- Replicape
Tethered to a PC:
- LinuxCNC
- Gestalt nodes
- Stepoko board
- Smoothieboard
- TinyG
- Windows terror
- Linux terror
- Custom full screen skin dream (what if you could see only the buttons you need, yet have easy access to advanced settings?)
Browser based
- Fab modules (score:6)
- Mods (score:7)
- Makercam (score:3)
Desktop based
- V-Carve (score:5)
CAM inside CAD program
- Fusion360 (score:8)
- Freecad (score:6)
- Bark beetle (score:10) (disclosure: Jens Dyvik has written this guide, made Bark Beetle and determined the score point criteria bellow)
Different scorepoints:
- Open source
- Free to install
- Do not require other non open source software to run
- Has Profile, Pocket and 3D mill
- Makes precise 3D milling
- Suports ramped plunge (for milling metals)
- Supports drill (for hold down)
- Supports tabs (for parts without internal holes)
- High speed/trochoidal milling (for milling metals on weak machines)
- Can be integrated with parametric models (for smooth worflows and future apllications)
- Integrated machine controll (its the future)
- Has good documentation
- RepRap. Focus on FDM printing
- Alu extrusion based systems with belt drive. Focus on quick and easy. Openbuilds, Shapoko, X-Carve
- MTM Stages by Nadya. Focus on modular, reconfigurable and expandable machines
- Chamfer rail by Jens and Jon, in Fabricatable machines development. Focus on making a larger degree of your own machine for maxiumum freedom in form, function and scale.
- Co-ops – buy together for discounts and efficiency (if a fab lab stock necessary parts and materials, the threshold for people to start to building their own machines gets much lower)
- Buy from source - find factory webshops
- Buy locally if the vendors don't suck
- Consider registering your own company – gives purchases almost half price in many countries (no VAT and reduced tax)
- Find sharp/new milling bits
- Clean collets and chuck
- Make sure that the sacrafial layer (spoil board) is well secured and freshly surfaced
- Grease rack, clean rails
- Tighten motors on all axis
- Check v-wheels for slop
- Configure ramp values/ acceleration in control software (shopbot control software corrupts its own config file while when it crashes, keep a backup config file at hand)
- Make sure you have good screws and washers for hold down
- Test a small part first
- Test the entire design in a cheap material
- Do it yourself, tell your instructor to level up and let you operate the machine
- Remember to document during the process
- Remember to RTFM (read the f***** manual)
- Be patient like a Zen master. Great machine calibration give great rewards.
- Tweak the pressure of you glide blocks/ v-wheels before with the pinion disengaged from the rack
- Tweak the pinion pressure on the rack with the help of the motor plates once you are done with the glide bloacks / v-wheels
- If you are real fancy you use software to compensate for uneven motion
- Make spare parts
- Make another machine for a friend
- Improve it (maybe make a new actuator for a new function)
- Dont be shy about giving back by sharing improvements (we use github because it makes it easy and scalable to be multiple contributors to a project)
- Make tutorials
- Make example files
- Tell authors about what you have made and what it meant to you (positive feedback fuel further open development)
The rack and pinion geometry is a central part of the fabricatable machines motion system.
It consists of a small round pinion spur gear mounted on a motor. This pinion gear fits perfectly with the gear teeth of the rack. The motor rotates the pinion and it pushes on the rack teeth and thus creates linear motion.
AIMATION
The rack and pinion geometry in the fabricatable machines are so far based on a cycloid curve.
The reasons for this choice are primarily to allow easy 3-axis CNC milling of both rack and pinion and that it allows for a low backlash design.
A cycloid is the curve described by a point on a rolling circle. This curve describes the path of the center of a pinion tooth.
This curve is a .SVG file of a cycloid for a pinion with radius 1 (Made with http://fooplot.com) It can be downloaded and and scaled, here is a clean version without the axis and grid and an Inkscape friendly version
Most industrial rack and pinion systems use gear profiles based on the involute curve.
There are two main ways of generating involute rack and pinon geometry. Generating the pinion curve for a know rack or designing a rack tooth profile and "cutting" the pinion to match it.
Standard involute curve Math
An umodified rack and pinon based on the involute curve would have completely straight flat sides on the rack teeth and the pinion teeth sides would follow the curve of a spiral. The kind of spiral you get if you unwind a piece of string from an axis.
Involute spiral
(Made with http://fooplot.com)
Straight rack teeth
Made by GearHeadsCC_BY_SA-3.0 via Wikimedia Commons
However, the geometry is often modified slightly to increase performance, smooth running, reduce wear, increase strength or for manufacturing reasons.
Modifying the involute profile
Base configuration: Pressure angle 6.8 deg, pitch 5 mm, 8 teeth, module 1.5915 mm, base circle Ø12,64 mm. Modifications: Rack tooth tip rounded to reduce undercut.
ANIMATION
Vs cycloid
ANIMATION
Both use 40mm effective dia and 8 mm core.
Involute
Cycloid
Cycloid DXF of rack and pinion
The size of the effective circumference of the pinion translates into the maximum pushing force it can develop from the torque it is rotated with. A large pinion will give a small force and vice versa.
Pcirc = Effective circumference of the pinion
Tmax = Max motor torque
Ma = Pcirc / 2 / Pi = Effective pinion radius = Moment arm in meters
Fmax = Tmax / Ma = Max pushing force
Example:
- Tmax = 3Nm Stepper motor (Aliexpress link)
- Pcirc = 60 mm pinion effective cicumference (pic)
Pcirc = 60 mm = 60 / 1000 m = 0.06 m
Ma = 0.06 / 2 / Pi = 0.0095 m
Fmax = 3 / 0.0095 = 314 N
1 N of force can lift ca 0.1 kg (0.22 lbs) so an axis with the above specifications can in theory lift a 31 kg weight (68 lbs) or push sideways with equal force
Please note that these are theoretical max values, friction will lower the actual value
All fabricatable machines so far uses stepper motors. The reasons for this are that they are easy to control, cheap and common. Many stepper motors have a built in position resolution of 200 steps per revolution. Gearboxes or micro-stepping drivers can be used to increase the position resolution. However gearboxes add cost, weight and can be a source of backlash.
The motors used in the fabricatable machines are often stepper motors (wikipedia link), these motors pass through a fixed number of electromagnetic positions per revolution called "steps" and most stepper motors have 200 "full steps" per revolution.
Example:
-
200 steps per revolution
-
Pcirc = 60 mm pinion effective circumference
Pcirc / 200 = 0.3mm Linear motion per full step
This is the smallest possible motion that can be made with full pushing force.
It is possible to make the motor move in smaller increments then the full steps, this is called "micro-stepping".
However the smaller the (micro)step that is attempted, the less force will be generated by the motor per step. If the generated torque is lower than the frictions and resistances in the motion system then the stepper will not move but rather apply a "magnetic spring force" in the desired direction. This generated force is not constant for all microsteps some will apply a little more and other a little less, depending on the motor and the stepper driver particular specifications.
Micro-stepping myths (Mirror) Discussion of the micro-stepping myth article
This is similar to what happens when you try to manually force the motor shaft away from it's held position. A microscopic turning of the axis generates almost no counter torque from the motor but as you push further the counter force builds up until you are fighting the full torque of the motor. It will feel like you are fighting a magnetic spring. If you press harder than the motor can resist then it will "slip" past the next full step and again feel springy, but in a new held position.
Example:
- 200 full steps per revolution stepper motor
- This assumes a theoretically perfect stepper motor and driver, a measured curve will have more irregularities Torque and micro-stepping for different stepper drivers
- Set to ANY microstepping level
- We force the motor half a (full)step out of position
Result: The counter force against our fingers will be ~71% of the full force the motor can generate.
The exact same amount of force will be generated if the shaft is held fixed (by you or by friction) and the stepper driver is instructed to:
- take a half step forward, or
- take 2 quarter steps forward, or
- 4 steps forward with 1/8 microstepping level set on the stepper driver
- etc...
Exception: Many stepper drivers are programmed to reduce the motor holding torque after they are left standing still for a short while, they do this to reduce unnecessary heat buildup in the motor since they assume that more force is needed to move the load than to hold it.
Here is a graph comparing three pinon sizes and the maximum theoretical pushing forces that they can generate using the same motor.
In this example the friction forces are 280 N and the motion system will not move if the force generated is below that. Meaning that for example the @60 mm pinion would not be able to make moves smaller than 0.2 mm, regardless of the micro-stepping level. Please note that these are theoretical max values so the actual smallest possible movement increment will be even less.
In summary No, the advantages outweigh the drawbacks in the case for micro-stepping. The biggest advantage is increased position resolution (finer step size) at a small cost in torque which is easily made up for by the other advantages of reduced noise and smoother motion.
Microstepping gives you a turning torque that is ~71% of the max hold torque
About ~10% torque loss from microstepping vs full step (Contradicts other sources)