Archive | September, 2012

Update on plans.

24 Sep

Been working at Kicad. Think I’m getting the hang of how to create component footprints, and convert schematics into PCB layouts. Seems the relaxation oscillator can be constructed on a single layer PCB, so looking for surface mount PUTs. The optocoupler is surface mount so if possible I may as well save the time/effort/cost of drilling holes.

Mulling over ideas for how to adapt my milling machine to CNC. First thought is to get a digital readout for each axis and then use an “elastic band” pulley between shaft of a DC motor and the handwheel. Feedback from the digital readouts goes to a controller, which then powers the motor in the appropriate direction. Problem for me is the scales for accurate digital readouts seem to have their own proprietary interfaces. and most aren’t cheep. I might have a shot at adapting a digital calliper for this. Failing that it’s gonna be position encoders driven directly by the axis motor, and reduction gears onto the handwheel shafts  to a) increase torque b) increase effective resolution of the position encoder. eg if there’s 30:1 speed reduction between motor and output,  the encoder will turn 30 times for each rev of the output. a really simple way to improve the effective resolution of an encoder. If the rpm weren’t too high, it would be possible to drive the gear off “speed up” gears – the encoder itself will have  a low inertia. Should work well to use the antibacklash technique where one gears pair is locked to their shafts , and another identical pair of gears is in parallel to the first, but only one gear of 2nd pair is locked to it’s shaft, the other gear is sprung loaded against it’s matching gear. the spring takes up the backlash.  Of course there’ll be extra backlash introduced by the gearing, between motor and axis, but I think that can easily be programmed out, by mimicking the antibacklash technique  used in manual machining. It’s only a problem when the direction along an axis changes sign. When backlash might occur, following is way to reduce it’s effects …

  1. the tool withdrawn from work.
  2. tool moved a short distance, 180 degrees against the direction it will move in after anti-backlash routine.
  3. Tool then moved back to where it was before it was withdrawn.
  4. machining continues.

Enough of ideas for now, so back to practical. Have an engine oil pump from a small 4 stroke motorbike on it’s way to me. Should turn up early next week, and when it does, it’ll be driven with an electric motor (or drill), for another attempt at getting my Babington waste oil burner design running . The last experiment certainly proved that the principle worked, even if the burner casing needs work to avoid flames outside of it, and had to be aborted when the automotive washer pump used to pump oil over the atomizing tube, cut out. After letting everything cool down, the washer pump began worked again as if nothing had happened to it. The veg oil had got pretty hot, so I’m guessing that the components of the gear pump heated up and expanded, and the extra friction then stalled the pump’s motor. Must have shut the motor down before it burnt its windings out. Should have more success with a pump that’s intended for moving hot oil around, and with a pulley so the motor can run faster without needing PWM to reduce the oil flow.


micro amp constant current source

14 Sep

1st project nearing completion and as I’d hoped output from photocouplers can charge capacitors at low constant current

optocoupler relaxation oscillator with 5nF cap
5mv/div 20us/div

perfect as uCurrent current sources perhaps?

Hello world!

9 Sep

Thought it to be about time I got my head around blogging, and began documenting some of my projects