I wasn’t able to get started until this morning, and had to leave for the week before the first parts were even printed. 
Oh well, next weekend will still be workable.
I have to find a donor table too, I might be able to do that on the way home Friday since I’ll be traveling through the big city and can peruse craigslist beforehand.
And they looked as good in person as in the listing. $50 for the pair and now I’m planning on two ZenXYs instead of 1. 
The first couple of parts are printing 3 hours to go.
Dang it, this new phone isn’t letting me rotate the picture right (or else I just am not getting it yet).
That looks very similar to the table I used. Very nice!
OK, let’s get started on this. Here are a few build pics from the week.
Looking at the fit[attachment file=42481]
OK, which way do they go[attachment file=42482]
Screws look at bit short[attachment file=42483]
Not going to gain much by tightening[attachment file=42484]
Starting to thread the belt[attachment file=42485]
Around the first idler[attachment file=42486]
Next the first motor[attachment file=42487]
Back across the roller spacers[attachment file=42488]
More to follow.
Onward we go…
Into the end piece, note which hole[attachment file=42491]
Back out and over to the other side[attachment file=42492]
Around the idler, note again which[attachment file=42493]
Around the next idler[attachment file=42494]
Across the center. I ended up reversing the zip-tie[attachment file=42495]
Back again to the side, but on a different idler[attachment file=42496]
Then into the second idler at the corner[attachment file=42497]
And out the other side[attachment file=42498]
And likely the last post with belt pictures.
Across the rollers[attachment file=42501]
To the second motor, note that we start on top this side[attachment file=42502]
Back to the next idler[attachment file=42503]
Now it’s time to hook up[attachment file=42504]
One end first[attachment file=42505]
All tied together![attachment file=42506]
Over she goes and the bearing clicks into place[attachment file=42507]
Great pics, Bill! Very helpful in visualizing everything, especially the belt routing. Thanks!
I somehow thought the captions would be visible near the embedded images, but at least they’re shown at the thumbnails. It makes a lot more sense with them than without.
OK, we’re a little farther along.
Set in place with caulk[attachment file=43629]
Just temporary, right?[attachment file=43630]
John’s cleaning excess caulk off[attachment file=43631]
Things should be dry enough to test with some soda tomorrow. Still have to decide on what to use for a decorative border to cover the areas the bearing can’t reach.
And it lives! Noisy as all get out though, sounds a bit like an air raid siren in miniature.
Just getting started on the first pattern [attachment file=43673]
A couple of minutes later and the rotation factor is more visible [attachment file=43674]
Second pattern and it skipped a layer when the magnet lost the ball. Picked it back up next time around [attachment file=43675]
Next steps involve the decorative border between the original table hole and the mechanisms limits, trimming and mounting the LED rope to fit that border (as you can see mine is three times too long), cleaning up the wiring and mounting a pullout ‘drawer’ for the display.
Eventually I want to control the LEDs from the Arduino (three MOSFETs and toggle it on in Marlin) and add Bluetooth to stream the gcode from Sandify to the Arduino.
It turns out if you put too much media down the ball can get lost during a session. I showed the symptoms above, but was able to verify that the issue was directly related to the depth of the baking soda. I printed a zen rake and used it to drag a bunch off to the side and patterns now complete without the skips. I’m using a half inch ball but haven’t tested to see how deep we can go without skipping. It’d be worthwhile to see just where the issue crops up, see how it differs with different sized balls, and see how much the magnet strength and distance affects it.
Stars from the outside in.[attachment file=44344]
Believe it or not we finally got around to ripping the boards to cover the difference between the area the ball could move and the actual open area on the table.
Table with light off:[attachment file=51567]
And with lights on…[attachment file=51568]
I somehow missed this whole thread. Looks great!
Dude, I had no idea!, Heffe I had no idea how you found this or why we didn’t get notified…
Bill, Thanks for sorry we missed it. Looks good, You might not be tap dancing right now but soon enough you’ll be able to get back to all of these projects.
Occasionally I find something in the recent replies that I am not subscribed to. Damn computers!
I’m going to come up with a sequence of patterns to program into it, trying to use the ‘load a file and execute it’ gcode command to chain them together. I’ll use Sandify patterns and possibly hand code something like an etch-a-sketch “BEER / HERE” logo. Then it goes down to be displayed in the window of Route 30 Bottles and Brews in The Dalles. I’m kind of hoping Brian will like it enough to swap all his tables over. 
I also have a WiFi/Bluetooth module I want to connect that would allow me to dump gcode via a virtual serial port using either protocol. Then I could potentially use Sandify to send a pattern directly from a computer or phone. And I want Sandify to expand to generating cursive text in the display area, so I don’t have to hand code it. ;}
I did a few crude DXF by hand in CAD and then let Estlcam make my gcode, I tweaked the first few and last few lines to get it to start and end where I wanted. I just used lines and curves in a sketch and never let them over lap but would get them super close (.1mm) where needed, that way it would generate gcode in the correct order in a continuous line. Export DXF.
The Arduino gcode sender Karl wrote and I worked on would work well on an esp8266 or esp32. Then it could host a webpage to upload gcode, or just open a port that you could stream a json file to. There are some neat esp32s for about $10 that have a built-in oled, and I’ve made a library to handle most of the BS for them. IIRC, they are called lolin32? Then just add an SD card for permanent storage. The front end could be sandify, or a page hosted on the esp.
Making sandify send to a serial, or some port would be pretty cool. The web version could even do it, as long as there was a way to specify where the zen XY was. Otherwise, you could run it locally, if the esps didn’t have internet.
I think the cursive wouldn’t be too hard, except for the art. The tricky thing is that every letter/symbol you would want needs to be well defined, and they need to be continuous, and start/end in specific places. If I had a list of x,y points for each letter, I can easily string them together in sandify. Anyone have any experience with fonts? Things like ‘t’ could retrotraverse to make the shape.
