2.5-3D thing

This is the initial-concept for a snap-together drum/driver-pulley... two "forks" cross-over each other and pass-through each layer of lasered-circles. (despite using transparency, the renderings don't show the forks passing-through the cylinders).

I'm trying to figure out how to account for material-thickness variances (they only guarantee 10% tolerance)... so I'm thinking "forks" coming from each side with serrated edges that when compressed together will ratchet into ever-tighter compression. It's frustrating, at this point.

Also trying to figure out how these pulleys will actually be mounted... so-far I'm thinking two additional small-circles (like the red ones) on either end (outside the green)... these'll ride inside a cutout mounting-plate on either-side. It's hokey, and the amount of drag...? yeah. But it's the best I can think of. Makes hand-driving a bit difficult, having to maneuver your fingers around the mounting-plates... but, again, it's the best I've come up with. Maybe a crank extending from one of the forks on the outside of the mounting-plate, but then there's torquing on one "fork" which might cause twisting... I'm not opposed to glue, but I'd *like* to not need it in the long-run.

relatively small radius and still have the cable physically attached to

the drum at one point, just a few (ten-ish?) turns are necessary for full-motion, so no need to worry about slippage...

Anyways, for attaching the motors, I'm visualizing yet another circle with a +-shaped cutout that slides onto the protruding part of the "forks"... But, the + doesn't really work with a motor-shaft attached in the center... so then a four-hole mate at the bottom (in the image) with longer fork-prongs...? But, then, there's the serrated/ratchetting thing that was discussed earlier... so, yeah... my brain's gone to mush again.

and, yahknow, what was the original purpose of all this, again? I think, somewhere in all this, was the idea "yahknow, it'd be nice to have access to a CNC-machine to make cases for things, etc..." which would certainly be a heck-of-a-lot easier to do off-the-bat with this lasering-order...

Oh, and, apparently the bit-banged UART receiver isn't working yet, which is surprising, to say the least, as I/O is clearly working, and precision-timing is clearly working, as well... Seems like that'd've worked pretty much right-off-the-bat. Is it the 3.6V signals feeding into the 3.3V Inputs? No, I tried feeding through a diode and a pull-up at 3.3V... so...? And the outcome is *really* weird... it seems to be echoing *several* 0xff's each time I transmit *one* byte to it. (and raw transmission from it, rather'n an echo, works perfectly fine). So, yeah.

(WAIT! Is it the fact this system's so fast...? I might've forgotten to increase those timer-comparison integer-sizes on the receiver, as I had to do on the transmitter.... TODO!)

Yeah, this is all as of several days ago, I'm just trying to remind myself the status.

I started out with the PIC32MX230F064B, for some reason... I also have the PIC32MX170F256B... I figured I'd continue with the one I started-with, but already ran into some (minor) difficulty: my USB->Serial converter uses 3.6V-TTL, whereas the PIC32 is running on 3.3V. Really, in this world, that should be acceptable, right...? 's just a hack at this point! Well, it ain't working, and I guess that's the first thing to check. Could it be causing "latching-up"? As far as I'm aware, never, in all my years, all my TTL and AVR circuits, actually ran into a case of that.

Anyways, the MX230 has a few 5V-tolerant pins... but apparently none that aren't already in-use. We've the JTAG pins, reset, and one dedicated to detection of the USB-Bus, that's it. Then there's the MX170, which appears to have 3 regular-ol' GPIO pins which're 5V-tolerant and have nothing to do with USB or JTAG. Nice.

...And, apparently I'm slow, because of course, I could just implement USB->serial in the MX230 itself, right? (That's *way* over my head, at this stage).

Anyways, my motor-driver communicates via 5V signals, so there's that.

------

The actual laser-cut mechanical design is somewhat on-hold. I completely burnt out on it a few days back. Though on-paper I think I came up with what to do about the "drum"/driver-pulleys and how to "snap" them to motors.

This was a lot easier, and the process more enjoyable, with legos... Especially the pulley-routing part.

Oh, and those wheels make contact with the bed for 1.5mm, hah!

Basically, if you're not looking at this on a phone-screen, the image is probably significantly larger than the real-thing. The file is nearly 3000 lines. This is absurd.

Hey, everything seems to scale nicely, at least... 'cause yahknow, a large pulley-driven 3-axis system makes a heck of a lot of sense.

Also, are those two diagonal pulleys going to cause weird torquing? Bah!

Those'll be nothing compared to the rest...

WTF...

Yesterday: (before I passed-out)

The "deadline" for a free first $20 order passes in something like 15 hours...

My brain has turned to mush on this project...

The latest:

Thought I'd switch to merely 2-axes for the $20 order, as a good proof-of-concept, and plausibly still functional as an etch-a-sketch sorta thing...

Glad I did; viewing it with less clutter made me realize several *more* design-flaws WEE!

Even the pulley "puzzle" idea is a bit difficult in this short time-span... Turns out some of my pulleys need to be rotated 90-degrees, which means they're no longer mounted on my surfaces, but require *additional* surfaces... Whelp, that's a bit of work, and adding those surfaces interferes with those that woulda been used in the "puzzle". So...

I kinda still dig the idea of a system that could be implemented in a mere 7.1in x 7.1in piece of material... But the monetary-crunch for it is dramatically-reduced with the $100 gift...

I've little doubt I *could* get a 3-axis, probably even 4-axis something within $100, and probably even a few other things, as well.

So... seeing as how this $20-free additional surprise was limited to ten days... well, let's just say I can't even remember the last time I bathed. 

[Note 2023: LO Friggin L! What was it three measly days? My how times've changed. I lost my beloved cat two years and six months ago in a few days. Last I bathed with anything other than lysol wipes was looking for him at a hotel! No joke. LOL I laughed so hard reading this I started to pass out thinking "well, this wouldn't be the worst way to go!" This ain't a call for pity. I'm fine. It's just incredibly funny to recall a time prior to "living in a van down by the river!"]

I think it's time to admit defeat.

It woulda been good to do a prototype before "the big order", though, since there's something to be said about the effects of tooling-widths and whatnot, that could have a dramatic effect on a system like this... Also, was kinda thinking about snap-together, which *definitely* would be effected by tolerances, tooling-widths, etc...

I mighta bit off more than I could chew... was trying to get a functional-prototype done in ten days (first order free up to $20!), but my brain just stopped (or started?) working today...

Found yet some more design-flaws, even in reverting to that which was well-established (on paper)... the one that doesn't lend itself well to 3D... And those issues just started piling up to an insurmountable level, today.

The pulley-system I was planning to use for the third axis causes torquing on that axis... OK, fixed, now there's four more pulleys and two more cables... per axis (woot!)

Then... aww, I can't remember...

Oh, Tried to keep the openSCAD model easily-swappable between the two designs (3d-friendly vs. not) but now neither of them are working... MATH. Gave up on the 3d-friendly model for the "deadline", and still can't get the 2.5d model back to where it was... Wee!

On the plus-side, I did some cleaning today... the cat knocked over a box full of books, greeting-cards, and playing-cards... it's been quite a mess for weeks. And she's been using my books as scratching-posts! Also the rolling-table/workbench had a larger surface about two months ago... I replaced that surface with a smaller one, but never retightened the screws, much to my buddy's shin's chagrin. So, that's fixed, too, ish.

The "deadline" fast-approaches, and I still have to figure out how to convert the 3D model to 2D SVG for cutting, which is another huge mental hurdle at the moment. I know how to do individual parts, but this thing is a 3D model of *all* the parts, put-together... So, we're about halfway through till the "deadline" and at this point I'd already planned to start the 3D->2D process, and I don't have a full system yet to even attempt to do that with.

So, now, I have this crazy idea that if it's gonna be a snap-together "kit" of sorts, then why not have a bunch of pulley-locations available (holes already "drilled"), and make it a bit of a puzzle (kinda like the Cordwood Puzzel), like the puzzle that's going-on in my head the past few days (and 10+ years)...?

This "kit" could be supplied with instructions to build the model itself, but leave the pulley-locating as a challenge... 'cause that's how it's gonna be if I attempt to meet this "deadline" ;)

Anyways, I know the pulley-system *can* work; I've *done* it... and, if done-right, they actually add to structural-stability as well as smooth linear-motion/reduced torquing... But, can it be "done-right" in this new orientation/layout? Remains to be seen. Pulley-Mount-Holes Everywhere!

As far as pulley-layouts... there's a couple options... The *easy* one, requiring the fewest number of pulleys, makes each axis dependent on the others... e.g. if X moves 5mm, then the Y motor must move 5mm to compensate... This sucks as far as "glorified etch-a-sketchs" goes... but isn't hard to do in software... The "easy" one, requiring the fewest number of software-hacks (and being compatible with a "glorified etch-a-sketch") requires nearly double the pulleys. On the plus side, it means that each axis is completely independent of the others. This concept I dig. And it lends itself well to the "glorified etch-a-sketch" idea. Again, software-wise, it's completely unnecessary... how hard is it to say "Y=X+y"? But, again, LOTS more pulleys and cable routings.

And, frankly, the harder-software-hack version lends itself well toward testing the precision of multiple-axes WRT each other... which is something that could be worthy of attempting...

So... maybe... the solution is what I came up with earlier... lots of mounting-holes for various pulley-configurations.

This one's a stretch...

It certainly clears up the 3rd-axis clearance-issue, but:

* may be quite a bit more difficult to run the cables/pulleys

* might not be stable-enough for merely two casters on the 'bed', should I use four?

(otoh, now that I think about it, part of the niceness-discovered with the pulley/lego system was that the cables, when positioned and tensioned well, actually add to overall stability quite-nicely)

* appears to use quite a bit more material for the same travel-distances...

* more difficult to add support left-to-right... may need longer bed-casters...

* can't add front-to-back support in the front

* kinda bulky/ugly...

Nicities:

* lots of vertical clearance for potential 3D-stuff....

* easier to make the Y/Z carriages larger, to mount larger things (dremmel-tool?)

* The Z-axis should be *quite a bit* stronger

* (maybe the Y, as well?)

* The Z-axis carriage and associated cable-tensions might actually add some stability

things to think about...

Ran into a few design-flaws along the way...

First was this:

Yeah, that's a 2.5ish-D system, and that third-axis... how the heck was I planning to transfer motion to Z for the pen-up/down...? Came up with some lever-system (the teal thing in the middle) but it seemed pretty wonky, to say the least. Also, this system requires pulleys to be *really high* off their attached-surfaces, which would cause some torquing on the material...

This is what I'm on, now:

Was pretty proud of that, have it even more completely-drawn-out on paper... including the pulley-layout which results in very little torque. But just realized another design-flaw... Always figured the Z-axis wouldn't be particularly strong, what with the number of pulleys it'd have to go through, but certainly good 'nough for a pen... but, I was kinda hoping to at least have the *option* for 3D-something... who knows what... Well, this doesn't lend itself to that well, at all... the Y-Axis carriage would hit whatever 3D thing was under it. Woot!

Early pen-and-paper sketches suggest swapping the Y and Z axes... have Z be the vertical plate attached directly to the X gliders, then slide Y left-to-right on that... Great! Except... very little stability, I'm sure. It'd require cutting out a significant amount of the structural-support provided by the Y (now Z) plate, and all that that was going to go behind it.

We'll see... Again, this was mostly just to get this darned pulley-thing out my system; a glorified etch-a-sketch would be a great start...

Here's the wheel-plates... I'm particularly proud of the red-wheel-mounts which are snap-in. Though, I'm still trying to figure out what to do about the green-wheel spacing, which would be offset by the laser's tooling-width as well as tolerances in the material-thickness...

Thinking about a diagonal notch and screws... it's vague in my mind, so far.

The whole system is designed to be easily-scaled, so I'm kinda thinking about the smallest-size sheets they have available 7in x 7in, and possibly trying to fit everything on a single sheet... A tiny little cartesian system for fun. Maybe even snap-together... held together still as a sheet which can be snapped-apart... I dunno.

The only things necessary besides the lasered-material, then, would be cable, pins, (glue?) and pulleys... and I'm contemplating how to cut a notch into laser-cut cylinders with the tools I have available...