[SML] lay out methods

[Bruce Cooper]

> the 1st, is the idea of a Roomba-like device (or actual 
> Roomba), with a marker installed, and a solenoid to lower the 
> marker.  The robot would need a way to accurately navigate. And 
> before anyone suggests it, counting revolutions of the wheel are 
> NOT anywhere near accurate enough.
>
> Such a device would require some sort of absolute positioning 
> system. Some years ago, I started on a project to design a 
> navigation sytem base on a fixed laser on a Roomba, and some 
> retro-reflective targets. The idea is that if you know the 
> locations of 3 reflectors, and then measure the angle between 
> them, you can calculate your position.  I did some initial proof 
> of concept, but then realized the laser really needs to 
> continuously navigate with a rotating laser and even then cannot 
> measure while moving. Though the idea may still be viable, if 
> you use wheel encoding to get to a point, and then use the laser 
> to fine tune the final point. E-mail me privately if you want to 
> know more.
> ---
> Another idea is to simply computerize a pair of lasers.  If a 
> compute can rotate a pair of laser line generators, then it can 
> position them to the correct position for a  particular point.  
> I calculate that to permit 1/4" accuracy, such a positioning 
> system would need 0.0298 Deg of accuracy. 
>
> One source (OMC-stepperonline.com) has a stepper/gearmotor 
> item: #0259 for $33, with 0.018 Deg step angle, but the gear box 
> has a backlash that is <1 Deg.  Perhaps if the angle of the 
> output were always approached from one direction, it would be 
> acceptable. Their motor controller for it, capable of 128 
> micro-steps, #0245 is only $20.


And while we're dreaming, I'd like a delta style 3-point high-speed 
pendulum installed.  Then I can change out the marking tip for an extruder 
tip and print the set in place.

:)

:Bruce
--
Overtired and covered in mud.  Never wash 300 sheets of DuraDeck in one 
day.