Summary: Used ball blower after several other tests. Works well but took ages. May add a one-ball-per-time selector. ___________________ Thanks for all the input. What I tried and put aside part way through is not among the more usual solutions but is lots of fun with lots of potential. What I did use (originally suggested by the "project manager" and obvious but not too usual) was also suggested in several variations by PICListers. I got a number of ideas from PICList suggestions and the many many many You-Tube marble-lift and marble machine pages (wow) and am keen to try some out "sometime". I started with a "flicker". Windscreen wiper motor has continual slow speed output with an arm with coupling knob on end. This is designed to drive a rod "which you do not have"* in to & fro reciprocating motion. Mount motor in box with shaft horizontal so arm end moves in vertical plane= .. Make a flicking arm from fibreglass PCB material (what else?) Use dexion bracket to allow length of arm (from mount support edge to flicker) to be adjusted. Make ball holder form adhesive foam tape (quick, to hand , works). Set motor to about 45 RPM. When mounted on lounge floor balls (plastic marbles and beads slightly smaller than std glass marbles) can be bounced off ceiling (8+ feet) with a gratifying thwack and showered llover lounge ( wife Val she jest watch de TV & lay low). Adding a guide tube ensures angular distribution not too wide. Final version would have fired balls up a rectangular tunnel in side of display and richocheted them off a top plate into a reservoir. The problems began with the ball feed. To get consistent and controlled flicking only one ball at a time can be on the flicker. Time is required between bll feed and flick action to allow the ball to settle. A ball feeder is not hard. In this case a vertical column could have a holed plate moved under it above a blind surface so that one ball drops in and the plate is then returned 'home' and ball is carried sideways over an open hole and away it goes. I trialled this with bits of drilled stuff and it worked fine,. However, time constraints killed it. Could use wiper motor cam to move plate. Need to mount clear of flicked ball. Then need a feed channel and .... . This is fine for winter workshop madness for people without enough to do but not good for "do it now' designs. Spent some time playing with a vacuum cleaner on suck. Quite fun. Balls sucked enthusiastically into a top receiver with screened vacuum line to prevent ball ingestion. (Screen got ingested sat one stage and balls followed :-) ). Using a coke bottle open neck down causes balls to spin happily in cyclone of bottle body till vacuum removed and they slowly spin down and drop put of neckl. Worked well. Main killer was that project manager wanted vsacuum mounted some way away. A large scale model train track runs over op of cabinet and he felt my tastefuil grey & purple Nilfisk [tm] vaccuum may detract therefrom. He was probably right. If I knew then what I know now I'd have probably used that system. So I used a "Blower". I had originally tried a 12v compressor (medium quality) but air flow/pressure/build rate was inadequate. Concept was "saved" [tm!] by using a 2V "LiLo" inflater which uses a centrifugal fan at high RPM with much noise to make highish volume st good head. This was mounted in blanked off compartment beneath ball-fall. BUT what was meant to be finished by lunchtime was instead done with 4 hours sleep over 2 days (along with a small amount of related stuff). In practice it worked in minutes under prototype conditions. Reducing it to practice in real world final form took hours and hours & ... . It has issues if too many balls are in feed line. Adding a solenoid controlled single ball feeder should be easy enough [tm] and it can go "on stage" so action can be seen. This allows one ball per button push and saves overloading ball fall with N balls - when it clocks, spits balls, stalls balls and generally shows that Murphy is to hand. _______________________ I played with many augur bits in a hardware superstore. Getting one which worked nicely was surprisingly hard. Only one Bosch one at about $85 for about 300 mm (1 foot) worked well. Angle of screw, slope of edges, depth of screw, sharpness , more ... matter. I acquired at 2.4m x 25mm wooden rod with possible expectation of cutting an augur / screw thread in it using a router and a jig to advance rod at desired pitch as rod was rotated. Didn't get to try it. Should work but may eat several rods before getting it right. Note that to work a vertical screw MUST trap the balls so that they cannot freely rotate through 360 degrees. This is usually done by having balls extend outside OD of screw and having an edge of some sort that balls are moved against by rotation. A true Archimedes screw must be angled and contents sit in a well or bucket inside the profile of the screw proper. A look at how such a screw moves water is enlightening as it's not intuitively obvious. Blow with pressure and suck with vacuum are obvious enough. Suck with pressure using bernoulli effect is cute. Must look at that. Many marble conveyors need a full column of marbles to work. Being able to move a sparse column or one at a time may be useful or even necessary. I liked my original idea of a single plate moving up and down with a matching static plate. This is similar to the various stair stepping system but not identical. If non return gates are not used it seems to need 3 channels (all move at once so only one moving part. eg Plates have slots as below. 1->2 means accepts ball from column 1 and rolls it into column 2. Moving fixed .... 2->3 1->2 3->1 2->3 1->2 3->1 2->3 1->2 1 - Ball enters MP at 1 MP moves up one level. FP accepts ball from 1 and moves it to 2 MP moves down. 2nd level 2->3 slot alligns with ball at 2. Ball rolls into MP and across to 3. MP up. Ball at 3 sees 3-> slot in FP and rolls into 3 hole and across to 1 .... Only complexity seems to be need for 1-3 slot which does not have access to 2. This is easy enough to achieve but slightly complicates the otherwise utter simplicity. Russell --=20 http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist .