Project Logs

Collapse

• Thicker Brass bis: Made a Thing

  Paul McClay • 02/20/2024 at 05:59 • 0 comments

  Made a thing from brass thicker than paper and that isn't just a flat cutout:

  It's a shanyrak[1], as stylized in the national emblem of Kazakhstan[2][3]. Because my wife is Kazakh.

  Yes, the paint wrinkled when I sprayed clear over it. A lesson in alkyd/acrylic/time issues. Call it a rough draft/proof-of-concept piece. Or maybe I should brag up my successfully executed wrinkle finish.

  In October last year (writing in Feb '24), in a press to maybe possibly get a competitive HaD Prize final entry together (failed but that's a different story), I tried cutting something more interesting than a backlash test pattern in a chunk of brass thicker than a thin sheet, which failed in a confidence-inspiring way.

  Months of other stuff happened.

  When opportunity to try again came around, I had the idea of (re)starting with a simple figure that my wife might like -- a shanyrak -- since she tolerates all this nonsense. And Valentine's day was imminent.

  That was supposed to be simple but turned into another scope creep fest... ...which could turn this into a dissertation that never gets finished so maybe I'll just add some notes to the [2] footnote in case I ever come back to write more about that.

  ---

  Process

  roughing + lowest flats

  I rough cut the face, outline, and the center pocket down to a level above any detail with a 1mm end mill. Then cut the deepest pockets to final shape & depth with a 0.38 mm (0.015 in) mill.

  I guess it doesn't take much metal to make a (relatively) big pile of glitter. 

  Here's an opportunistic illustration of the divided project architecture: 1) the CNC core, whether this project here or elder sibling #Minamil 2dc: a minimal CNC mill, and 2) a structure that can be something fancy like parallel work in progress at #"Desk Accessory" CNC Milling Machine but doesn't have to be.

  The CNC part doesn't care whether it's in a tidy little box or scattering chips to the winds. If you're here for the CNC you can build the CNC and carry on.

  This pic shows a concrete illustration of the realized practicality of having a "desktop" CNC on my desk. The mess stayed in the box and that little patch is trivial to suck up with a hand vac. So long as it's sitting on a flat surface, which I think is a reasonable expectation for a desk.

  (having written "vac" I probably should add "HEPA". and the opening in the box under the XY table is for an air filter that isn't there yet -- the double-stack fan (high static pressure) will pull air down and (filtered) out the back so fine dust doesn't float up out the open top of the enclosure. which it does. that arrangement worked pretty well in the previous frame build. that all could be replaced with a vacuum hose stuck in a hole in the same spot -- which could be interchangeable if the fan and vac hose were the same diameter. i'll have to write all this air/dust stuff in a log at some point and trade all these words for a link to that. after finishing the fan/filter box there.)

  ---

  paint - and lost messages from the future

  Paint's not my thing. What I learned from this encounter is to do pretty much everything differently.

  I think I cleaned the surface adequately. It was freshly cut and cut dry so it should have been pretty clean after vacuuming out chiplets anyhow. 

  I used a rattle can, which meant masking around the workpiece and covering a fairly large area to manage overspray.

  Lesson: dust again after masking & draping.

  This was weird: That's probably too much paint for a first/single coat. I ended up spraying a load of paint on the thing because the paint spray wasn't landing in the narrow pockets between the bars. At all. You don't believe me. I should have taken a picture. I don't doubt that surface...

  Read more »

• Build? Start here.

  Paul McClay • 02/06/2024 at 05:39 • 0 comments

  ---

  ---

  I've tried to make the CNC part of this (relatively) easy to reproduce from some 3d printed parts and a minimum of other stuff. It should go together by assembly with as near to zero fabrication as possible.

  Writing up how to actually get that done is a project in itself. What I have so far is pretty rough, but I think it conveys enough information to give you a fighting chance to build a working machine.  As of writing, I intend for this log entry to serve as a stable "start here" point with links to the latest work in progress as it evolves.

  Please let me know If you're thinking about actually building one of these. I've turned on "public chat" for this project (big orange button on the project page, or right here) and set up a Discord server. We'll see if either of those gets any traction.

  ---

  Here's what I've got so far...

  Read more »

• Jolly Wrencher 2023

  Paul McClay • 01/01/2024 at 05:57 • 0 comments

  ---

  Cut but didn't publish this back in October, so I better hustle it up here before 2023 expires (in my timezone)...
  

  Read more »

• Thicker Brass

  Paul McClay • 10/10/2023 at 00:26 • 0 comments

  ---

  Brass. More than 5 mil thick.

  If you squint, you might see a Jolly Wrencher in there. It didn't come out like I wanted, but the fault was mine.

  Read more »

• doors again: current and (maybe) future

  Paul McClay • 10/08/2023 at 05:20 • 0 comments

  In the previous log page I walked through the history of the enclosure part of "frame+enclosure". This entry describes the latest iteration. It's all kinda deep weeds and I don't know if anyone will ever actually read it. I wrote a lot about this because it's been one of the more convoluted parts of working out something that might work well enough to call "done" for this project.

  ---

  Many words and little proofreading follow, so I have no idea how this all differs from what I think I wrote...

  ---

  In all previous flavors of storage/working enclosure, the doors fold 180° to collapse for storage, then close rather conventionally like doors. Another possibility was to have each door fold 90 deg then each door closes around the two open sides, one over the other.  That would avoid having two edges meet at a corner -- which seemed like trouble before discovering that it worked quite well. Reasons I didn't do that in the first place included the hassle of making the hinges different between the two sides, and more significantly, with single-thickness door panels, that leaves no opportunity to attach anything to the inside of the outside door, I already wanted to avoid attaching anything to the outside.  So that was going to complicate closure.

  Then I looked for hinge options to get away from tape and decided to try these:

  They multiply the panel thickness. Folding them double also adds the width of the web between the two panels, so the fold would be that much more than twice as thick. Quite counter to my earlier obsession with minimizing footprint @@@link.

  That was a reason to reconsider wrapping the doors around the corner instead of folding double. Joining thin panels with thick hinges also relieved the problem being unable to attach anything to the inside of the outer door or outside of inner door (i.e. only to inside of inner door). So, giving that a try...

  Those hinges fit between panels. Their thickness sets a "free" thickness for hinges at the frame. After looking at a lot of unsuitable hinges, I figured out that was enough to space for bespoke hinges.

  The dark material in that pic is just tacked on to show where the solid side would end if it were cut for the thicker panels. The outside door edge should bear directly against the solid side panel.

  "bespoke hinge" really means three different hinges -- just for that side.

  On this side the hinges carry the inner door while the edge of the outer door should bear directly against the inside face of the solid side panel. For the inner door, the hinges also back up against the solid side panel, and clearance between the fixed half and the edge of the door goes to zero with the door closed. When pushed in from the corner, the inner door should bear directly against the fixed parts of the hinges which bear against the solid side panels. A few small plastic parts will be more squishable than the whole door edge meeting the side panel, but the hinge in the inner door will buckle and/or compress before sending much load into that door panel anyhow.

  Then there are more different hinges on the other side to keep the hinges behind the plane of the outer panel.

  Again for the hinged panel the clearance between the moving panel and fixed half of the hinge goes to zero with the door closed, to bear whatever load they get which will be moderated by the hinge between door panels on that side.

  In that corner, the inner door panel bears against the thicker fixed parts of the hinges, which again bear against the frame side panel.

  (This pic also shows how the enclosure panels extend below the hard box to the level of the bumper feet to sweep the work surface. That helps to contain the mess. And I think flat-ish smooth-ish surface is a reasonable requirement for what to have under this thing.)

  That all adds a lot of bulk, especially where the folding hinges double up at the...

  Read more »

• feature creep vol. {n+=1}: doors

  Paul McClay • 10/07/2023 at 05:34 • 0 comments

  Another log mostly about already-dead stuff that I took pictures of before breaking it down to make it different again.

  ---

  I tried to make this vulnerable corner at least not super fragile. It turned out to be super not fragile.

  
  

  It will take a much harder hit without damage. But the under-developed closure is just a short peg that sort of gets close enough to a small magnet to be weakly encouraged to not wander away, and hitting it much harder was bouncing the doors open, and that would confound the point of this visual demonstration. So just a tap for show.

  I think I'm going to miss that clean, durable arrangement. The problem is that it was hinged with packing tape and that's not a long-term solution. I've been noodling ideas for making hinges that should age better with least loss of compactness or acute durability.

  I've had another go at it, and maybe found a thing to try next. But first, a little about how we got here.

  ---

  ---

  Starting with some history from the preceding laser-cut designs up to what I liked about the last iteration before breaking it.

  Before beginning

  Way back in pre-history, a piece of paper on each side of the frame did the job.

  CDROM discards are a perilous gateway drug.

  ---

  One thing lead to another and I got started into the laser-cut precursor to this project, retronamed "Minamil 2dc".

  ---

  Zeroth: zero

  A couple of 2-axis test ⋅ cuts with just the first XY table encouraged me to:

  1. keep going with the idea
  2. rig up some sort of debris containment ...

  First: pepakura

  ... so before cutting anything, the first 3-axis frame had a 3+-sided paper shroud.

  The "wings" projecting forward from the sides kept debris pretty well corralled to the tabletop. That actually got a little bit elaborate where it folded around and projected beyond the frame, and passed moving wires through without opening a big hole. (cable management has been a thing from the get go)

  Aside: in this project I emphasize the distinction between the CNC mechanics and the fancy enclosure that you don't need to make the CNC part work. Here illustrated. You don't even need a road atlas if you don't mind a little mess.

  Made from garbage-grade wood scraps, used uncut as found.

  Second: win

  This actually works really well.

  Four-sided enclosure. That really helped to make table-top work more simply practical. Taking advantage of how the horizontal axes collapse to a footprint smaller than needed in operation, the walls of the debris corral fold in to make a smaller box that doesn't have to occupy a bunch of space that it's not using.

  Having the front panels cut before I had any idea what to use for hinges, I just taped the panels together to visualize the basic idea while I figured out what to do for hinges. That worked well enough to become the answer for hinges.

  Aside: the side panels were cut tall enough to be taller than the raised Z axis (sans tool) and protect it for mindless storage. Then the Z axis got longer and broke that feature. The counterweight and its mast and beam fit inside[a] the closed box for storage -- until the counterweight got killed off.

  [a] Sub-aside: that vid emphasized need to have a vacuum handy. That was self-inflicted by opening the doors outward then getting clingy debris up to my elbows after reaching in there. Later I got the clue that pushing the doors in, like the one on the left in the right half of the pic above, allows reaching in and doing stuff with much less hassle.

  Made from less crappy wood, HDF, and acrylic scraps, cut on a table saw.

  A relevant feature of this version is that the attached doors fold back flat against the side panels.

  That proved really helpful when flipping the box around to do stuff with the doors open, and when handling a removed side panel without either removing then re-taping the door panels, or dealing with an awkward big thing that's either awkwardly floppy or awkwardly levering...

  Read more »

• about runout

  Paul McClay • 10/02/2023 at 08:18 • 0 comments

  tl;dr: if you can see it you can tweak it

  Some examples of work from this little CNC and its precursor demonstrate pretty small runout from a generic faux-Dremel rotary tool. Repeating that will be part of replicating this project for anyone so inclined. It also seems like it could be useful to anyone trying to do fine work with a not-so-fine tool.

  The point, when I get to it, will be anticlimactic. It's pretty simple. But I haven't seen it elsewhere in my limited scope of reading. So: sharing. Akshully I've written this before but that was buried in another topic and without illustration. So: sharing with pictures this time.

  ---

  ---

  To reduce runout when using a cheap generic rotary tool and maybe some other small-cutter tools:
  

  • See the current runout using stable magnification that you're not trying to hold still in your hand
  • Tweak the bit toward center
  • Repeat until runout sufficiently reduced

  That's all <yawn/>.

  ---

  More verbosely, and with pictures...

  ---

  See

  The gif above shows a view through a microscope on a stand that can look sideways. But that's cheating. Here's a method more in line with the "low cost" theme here:

  That's my daily driver phone in a little pocket tripod (which I've cropped :/ but you've seen a tripod before).

  • that phone is not new but not too old to have a camera that can focus close enough to do this
  • manual focus in "pro" mode helps; otherwise you'll probably need to put something right behind the bit to appease the autofocus
  • the viewfinder grid provides a fixed reference to help perceive small deviations at the tool tip; otherwise put something with a vertical edge close behind the tool (which might also appease the autofocus if you're stuck with that)

   That looks like this:

  While less crisp than the gif above, you can see that you can visually resolve a small fraction of the bit diameter (0.5 mm in this case), which is the practical scale for runout.

  ---

  ---

  Tweak

  Established Brand general purpose rotary tool collets aren't fantasticly precise. Generic clone rotary tool collets are really poor (typically, in 2023Q4). They are basically poorly made ball joints. The good part of that is you can point the bit any direction you want.

  I'm using very small cutters here, which don't generate very large radial forces. So firmly finger tight on the collet is tight enough. A very tight collet makes it hard to make small tweaks. If the collet is not tight enough, the bit will "walk" out and drive down into the workpiece. So there's an element of "just right", but for small cutters that seems to be a pretty wide range and easy to hit.

  Very small cutters break easily at the tip, so take care to apply tweaking to the shank and not the thin end.

  I think you should probably not do this: I started out tweaking by hand -- with a finger around the bit and thumb against the collet for good leverage. That sometimes meant pulling pretty far down the narrowing neck of a short bit and perilously close to the skinny end. Nothing bad happened, but that just seemed way too close to blood to expect that it would never go badly. So I printed a tool:

  I've been using that, but It's just a first whack at the idea and not a great example. So I hope you'll make a better one.

  Applying tweak without slashing a finger. In the photo my finger is behind the bit for contrast, which makes this a poor example of applying tweak without slashing a finger, but you can see the idea.

  ---

  ---

  Repeat

  While you probably won't get as close as you want on the first try, this method seems to converge much faster than "open loop" fiddling for that once-in-many low-runout chuck-up that teases you with knowledge that it's possible but only happens when you're not trying to make it happen rant rant.

  ---

  ---

  And if you're still reading...

  Rolling a stiff cylinder around the collet nut to make a slip-on anti-pointy guard has saved some bits and probably some blood. Especially useful...

  Read more »

• to shrink, or not to shrink, that is the aggravation

  Paul McClay • 09/26/2023 at 02:07 • 0 comments

  This is just a rant so I can stop thinking about it.

  So I made another frame last year. The one on the left:

  I made it a little big considering uncertainty about how big it needed to be.

  Earlier this year, it looked like some of the packaging ideas were working and I got fired up about redoing pretty much everything less haphazardly and taking it to MRRF. Including shrinking the footprint to match shrinking uncertainty about how big it needs to be.

  To start with, I cut the mounting tabs from the bottom of the X+Y stack and switched to screws up from below into the interior of the tabless bottom part. That reduced the lower bound on footprint. And also slightly confounded my first draft build doc.

  But such a laundry list of things to (aspire to) do before the show, and finite time, and other life. And other aspirations on said laundry list -- mostly accomplishing or improving integrations of function into the frame and redoing the extra-janky first whack at the folding enclosure -- were bigger step changes than "see! it's a little smaller than that other one you won't ever see".

  So, reluctantly, I abandoned the idea of making a new frame. That was hard to swallow because most of everything else I wanted to do would be made to fit the frame and everything made to fit the oversize frame was a nagging reminder that instead of shrinking the frame I was sinking more sunk cost into the oversize frame that I still hoped to shrink "later".

  Later is now.

  Cutting the simple excess while allowing for thicker hinges shaves the square from 191 mm sides to 182 mm. Not very exciting. So I've spent yet another chunk of time pouring over sketches and CAD looking at things that make the footprint bigger than the basic square dimension of the X+Y axes and what I can shuffle to debigger them. Getting down to 175 mm wouldn't be too hard. 170 mm if I do something different with the rather thick X motor connector that I've just spent time neatly securing at the back of the X+Y stack. 

  So I could shrink the footprint by 21 mm each side. Less than an inch but a little over 10% or a little over 20% area. That serves the objective of minimizing the footprint i.e. maximizing the brag. And I don't think any hard constraints prevent shrinking other stuff in the box to match. But then that's time spent to build a new frame, plus time to more significantly re-arrange (vs. incrementally improve) all the sunk cost other stuff in the frame.

  To do, or to not do? To decide before sinking yet more time in the other integrated stuff. Aaaargh.

  Ok. I'm going to not. But keep plugging with the frame in hand, with all it's surplus bigness.

  BuT iT cOuLd Be SmAlLeR!

  But it's not. Again. Sigh.

• Z axis parts & assembly notes

  Paul McClay • 09/25/2023 at 22:21 • 0 comments

  I've just uploaded STLs for the Z axis to the files section.

  This Z axis works. It's also pretty crude. Unlike the X+Y stage which I've worked through several revisions, I've sliced and printed exactly one set of Z parts[a} and they've worked well enough to let me do other stuff -- like spin revisions of the X+Y parts.

  [a] except pulleys; I've printed more pulleys.

  If you've built the X+Y axes, then I don't think there's much in the Z axis that won't be self-evident.

  The major design deficiency relative to intent is that the tool clamp is not as quick and easy to use as I'd like -- to support the idea that you can use your everyday rotary tool instead of committing a unit to semi-permanent installation.

  Oh yeah - it's modeled for 6" zip ties. Almost missed that. 4" ties will probably work.

  About pulleys: I'm currently using one motor and running it too hot for PLA. Consequently, I've printed a pulley in PETG. It's working fine. I don't know if it's possible to run the motor cool enough to use PLA. Because I haven't tried. If not, then more likely it's possible to use two motors and run them cool enough for PLA.

  The design provides for using two motors. Mainly because the earlier laser-cut version of this basic configuration was uncomfortably vulnerable to dropping the spindle without warning so the redundant motor/pulley/cord was cheap insurance. I think that's less of an issue here because the hoist cord(s) don't pass close by sharp stuff and do run in plain sight. So you have a better chance of seeing trouble before it happens and less chance of provoking trouble. Also, by choice not necessity, I'm using Spectra® cord (UHMWPE) which seems to be practically indestructible.

  Non-printable parts:

  • see also here
  • 2 x rods: 6 mm x ≥165 mm (length not constrained)
  • 4 x bearings (3 x would work and might work fine)
  • 1 or 2 x 28BYJ-48 motors with
    • 5V windings
    • convert to bi-polar
  • 6 x m3 screws
    • 4 x 20 mm (15 ≤ l ≤ 25)
    • 2 x 16 mm (14 ≤ l ≤ 18)
    • so could be 6 x 15 ≤ l ≤ 18 mm
  • 6 x m3 washers feel like a good idea but probably don't make any real difference
  • (2 or 4) x m3 x 10-15 mm screws
  • string -- notes in "Hoist" part of "Step 4" in this 'ible. Or TMI.

  The "nose ring" part is threaded for common Dremel-like clones and might not fit real Dremel tools. It's not essential and I don't know how much difference it really makes.

  Tie motor end of hoist cord(s) and install with pulley like steps 14 & 15 here.

  Limit switch goes here:

  (the turns of hoist cord around the pulley should all be adjacent starting from the flange -- this photo was taken with the axis detached and handled randomly)

  Point of interest: the clamp parts & saddles are not circular but slightly "trilobular".

• XY stage assembly

  Paul McClay • 09/19/2023 at 03:11 • 0 comments

  <update today=2024Mar23: link to Discord server</update>

  This is a draft in progress, but if you want an early start here's something to start with. There is a lot of me describing what I do rather than saying what you should do because this is just the beginning of letting daylight into the path-dependent evolution of how I've been doing this, and of the "this" that I thought I was doing. Some stuff may be mid-edit nonsense. Or pre-proofread nonsense. Public chat is open for the project -- big orange button on the project page. Also a Discord server. So far the Discord has more traffic.

  ---

  Build yer X-Y table/stage/thing

  ---

  print

  plastic parts

  part checks & cleanup

  Check all the V-block surfaces -- the 45º flats within zip tie circuits. See bright blue highlights in the image below. Low spots are probably ok but high spots are trouble.

  Check some of the vertical flat surfaces, and probably trim some edges. See the yellow highlights in the image below. They include the inboard ends of the bearing V-blocks and both ends of bearing clearances. The outboard ends of the bearing V-blocks need to be clean enough to allow the bearings to sit level in the blocks but are not dimensionally relevant.  Also the perimeter end of the motor bay and the business end of the limit switch clearance. These surfaces should be flat and perpendicular to the top or bottom of the part. but likely have some distortion due to elephant's foot, top solid layer expansion or some such, and likely need a little trimming where the vertical meets the top/bottom face of the part.

  note: the bottom part you printed from the STLs uploaded here will have three attachment tabs around the edge that are not shown in the photos here (and fewer holes on the bottom)

  Here's what the bottom part uploaded here looks more like:

  I've uploaded the tabbed version because I expect that will be more practically easy to use in pretty much all cases other than obsessing about footprint. But I wasn't really thinking about that when I started taking pictures for this draft.
  

  The low locating feature at the other end of the motor bay needs to be clean too, but it always has been in my experience.

  Generally clean up stuff that doesn't look right.

  You can validate bearing clearance cleanup by dropping in a bearing and measuring the remaining clearance. That will be the hard limit to range of motion.

  With geometrically perfect parts, that would be 77.2 mm. If I continue assembly with that part as shown, and if that's the shortest clearance of the three in that axis, I'll end up with 0.6 mm less range of motion than the CAD model. That's still comfortably more than 75 mm, which I think is ok -- 3d printed PLA after all -- and is the reason why I say ">75 mm" instead of "77.2 mm". 
  

  Check that a zip tie fits through all the tie passages. It may not always be obvious which way a tie is supposed to go, so the next couple of photos try to show orientations for all the ties in the "middle" part.

  I think that part includes examples of all the ways ties are used in the "top" and "bottom" parts, so you can check those parts too.

  "Elephant's foot" or expanded top solid layers might constrict some of the tie head openings, so check that a tie head fits freely in any that look tight, or trim edges for a free fit if needed. It's unlikely that a head won't fit at all -- that would be really bad layer spread -- but any friction holding the head will make removal a little tedious.

  Where tie heads should not stick up, like on the top surface of the top part, make sure the bottom of the head clearance recess printed cleanly so that the head sits all the way down....

  Read more »

View all 18 project logs