Mk 2 preparation


Here I will keep track of pieces I am collecting for a new machine

Software will primarily be the same (except for changes re pnematic heads vs DVD steppers)

Thoughts on speed:

  • Ballscrews are too slow and introduce oscillation at higher speeds which cause parts to move on the tips
  • The X gantry is way too heavy
  • DVD steppers when half stepping have some vibration as well


  • Tradeoff between resolution & speed
    • What accuracy needed? 1 mil/thou sounds nice but perhaps 0.1mm is enough, if a 0.5mm pitch IC 0.1mm is probably ok so we will say that is the minimum, but 0.05 desirable = 20 steps per mm, ie 2 mil resolution
  • Belt drive
    • Some murmors on cnczone re resonance in belt drives, just have to try it
    • Do I need reduction or will microstepping do the job?  Assuming stepper is 1.8 deg per step = native 200 steps per revolution
    • Been some discussion on CNCZone about accuracy of microstepping, with some people seeing it as purely for smooth movement but not actual positioning.  Another view is that when there is no sideways load (as in this application) it will position ok. In my experience the positioning seems fine on the old machine with microstepping.  Ideally, need to trade off possible speed with positioning accuracy, so the least resolution acceptable..
      • If direct drive, assuming 20 spteps per mm is the target :
        • 10 tooth T5 pulley = 200/50 (5mm * 10 teeth) = 4, 32, 64 steps per mm (per each of 1/1, 1/8, 1/16 microstepping)
        • 16 tooth T5 pulley = 200/80 (5mm * 16 teeth) = 2.5, 20, 40 steps per mm (per each of 1/1, 1/8, 1/16 microstepping)
        • 20 tooth T5 pulley = 200/100 (5mm * 20 teeth) = 2, 16, 32 steps per mm (per each of 1/1, 1/8, 1/16 microstepping)
        • 32 tooth T5 pulley = 200/160 (5mm * 32 teeth) = 1.25, 10, 20 steps per mm (per each of 1/1, 1/8, 1/16 microstepping)
      • If need mechanical, for speed minimum reduction is desirable so assuming aiming at 0.05mm resolution needs 5-1 reduction.
  • Overhead gantry type to reduce X gantry mass
  • Overhead gantry also allows feeders from 4 directions
  • 2 gantries, X & Y, A & B with X & A on shared rails

Thoughts on heads:

  • 2-4 heads per gantry so possibly up to 8 heads again:
    • In my experience thus far, driving multiple heads is trivial, so to eliminate any need for convoluted tool swapping as I know nothing and dont have the engineering skills nor time, consider multiple heads with fixed different pickups for different parts
  • Pneumatic heads per MK1 machine as at Augus

Thoughts on rotation

  • Seen some head rotate RC servo solutions, but with multiple heads on a gantry and the goal of low mass as well additional control lines per head..
  • rotate the PCB
    • mount a stepper directly under the center of the center of the table with rotation done that way
    • or belt drive turntable:
      • Ok, tried belt drive turntable on existing MK1 machine but cannot get mechanism thin enough, have flat pancake stepper on order to try that

Thoughts on PCB arrays

  • I did this originally, but restarts were too hard, ie resetting component counts etc so now we run one PCB per time
  • I will leave this until we automate some feeders one day.. but as always will make allowances for it in Mach 3 variables / offset allocations etc

Thoughts on centering

  • I finally gave in and using bump centering on the current machine and we will continue with this on the next machine until a better solution presents itself
    • works most of the time so long as the initial pickup is good enough & vacuum is strong
    • has a significant speed impact hopefully offet by increased speed of new machine + 2 separate gantries
    • really dislike it as a solution but it is simple and it works thus far..


Parts so far:

Gantry parts:

2 x 8mm x 792mm shafts

  • mounts & linear bearings on hand as well

12mm x 600mm mounted rails

  • linear bearings on hand as well

20mm x 800mm mounted rails

  • linear bearings on hand as well

drive mechanisms

  • a random collection of T5-10 & T5-15 pulleys and belts for experimenting


  • 1 x Nema 23 76mm on hand
  • 3 x Nema 23 56mm
  • 3 x Nema 17
  • Tried for some Nema 8 as start at 50 grams for possible head rotation but too expensive


  • Lazy Susan bearings arrived to experiment with PCB rotation
  • Flat stepper also ordered to try to retro fit to MK1 machine (as cannot mount parts under the main base on the MK1)
    • Actually will just move onto the new machine now with a PCB holder mounted directly to a stepper
  • 10 siemens feeders arived, wow big change in direction

Head parts

Mini linear slides & pneumatic cylinders per current machine

Base parts:

Just cannot get hold of 8020 or similar so will build frame from MDF & less useful alloy extrusions

Other parts:

  • 2 x USB microscope on hand for now
  • 4 channel Toshiba Stepper PCB arrived
  • 12v, 24v & 36v PSU's arrived