Ed – Page 42 – The Smell of Molten Projects in the Morning

Subpixel Zoo: Capturing the Specimens

A Hacker News discussion led to the Subpixel Zoo, which led to thinking the patterns might make interesting layered “art”. After fetching the *.webp images and figuring out how to persuade Thunar to display them, the next step was converting them into paths suitable for laser cutting.

Although the images are algorithmically generated in a common layout, figuring out how to get the outlines as paths seemed to require a journey into the depths of the Pygame library and that would turn into a major digression.

Instead, start with one of the webp images:

The deliberate blurring apparently simulates what you see in real life.

Import the image into LightBurn, which converts it to grayscale under the plausible assumption you’re going to engrave the image on something. Then:

Create a rounded rectangle overlaying the lower-left-most subpixel to good eyeballometric accuracy
Turn it into a four-element rectangular array, twiddling the center-to-center spacing to match the subpixel layout
Duplicate those four upward in another array to create a subpixel block, as marked in the upper-left corner of the original image
Slam another array across the bottom row and upward, twiddling the spacing to match the subpixel block spacing along both axes

Which eventually looks like this:

SubPixels - LightBurn vector overlay — SubPixels – LightBurn vector overlay

I made the final array absurdly large, cropped it with a square to match the template I used for the layered paper patterns, resized the result to be 170 mm on a side, then dropped the square into the middle of the template:

Subpixel Zoo - Quattron RGBY - LightBurn black mask layer — Subpixel Zoo – Quattron RGBY – LightBurn black mask layer

One gotcha: crop the subpixels on a Fill layer so LightBurn will close the truncated edges, then put them on a Line layer for cutting. The doc explains why, although it’s not obvious at first, as is the fact that you must delete the group of shapes outside the square before it looks like anything happened during the cut operation.

The resulting layout contains all the subpixel rectangles, so it’s what you want for the top black mask layer. Duplicate the pattern and delete the subpixels corresponding to each color, until you have one template for each of the Red / Green / Blue layers:

Subpixel Zoo - Quattron RGBY - LightBurn layers — Subpixel Zoo – Quattron RGBY – LightBurn layers

The blank over on the right is the Yellow layer, which does get a quartet of layer ID holes cut in the lower right corner.

Then it’s just a matter of cutting the blanks, locating the fixture on the platform, dropping the appropriate color sheet in place, cutting it, then assembling the stack in the gluing fixture:

Subpixel Zoo - Quattron RGBY — Subpixel Zoo – Quattron RGBY

It’s kinda cute, in a techie way.

I did a bunch of layouts, just to see what they looked like:

Subpixel Zoo - 8x8 layouts — Subpixel Zoo – 8×8 layouts

In person, the RGBY patterns look bright and the RGB patterns seem dull by comparison. I’m using cardstock paper, rather than fancy art paper, which surely makes all the difference.

2025-03-07

Trivial Laser Projects

A nubbly knob on the M4 screws securing the honeycomb to the laser’s platform:

Leveling feet for the HQ Sixteen long-arm machine’s table for the high side of the floor:

HQ Sixteen - table leg leveler - short — HQ Sixteen – table leg leveler – short

And 12 mm taller on the low side:

HQ Sixteen - table leg leveler - tall — HQ Sixteen – table leg leveler – tall

Both of those “projects”, which may be too grand a term, went from “I need a thing” to having one in hand over the course of a few minutes yesterday. Neither required a great deal of thought, having previously worked out the proper speed / power settings to cut 3 mm MDF and 1 mm cork.

Other folks may lead you to believe lasers are all about fancy artwork and elaborate finished products. Being the type of guy who mostly fixes things, I’d say lasers are all about making small and generally simple parts, when and where they’re needed, to solve a problem nobody else has.

Perhaps I should devote more attention to using fancy wood with a hand-rubbed wax finish, but MDF fills my simple needs.

With a laser and a 3D printer, shop tools have definitely improved since the Bad Old Days!

2025-03-06

Medical Image CD: FAIL

I asked for the images from recent X-ray and MRI sessions, whereupon a CD arrived in the mail. Popping it into my desktop Linux box produced this directory listing:

ll /run/media/ed/Feb\ 21\ 2025/
total 146M
dr-xr-xr-x  2 ed   ed    136 Feb 21 13:14 ./
drwxr-x---+ 3 root root   60 Mar  2 13:40 ../
-r--r--r--  1 ed   ed   146M Feb 21 13:14 -NISLEY-DMBG8yMQcf8qXcVj.iso

It seems whoever / whatever produced the CD copied the ISO image to the CD, rather than burning the ISO directly to the CD. As a result, the CD has one file.

Raise your hand if you’ve never done that.

Well, I was going to save the CD as an ISO file anyway, so I just copied it to the file server.

Attempting to mount it produces an odd result:

sudo mount -o loop "-NISLEY-DMBG8yMQcf8qXcVj.iso" /mnt/loop/
[sudo] password for ed: <make up your own>
mount: failed to set target namespace to ISLEY-DMBG8yMQcf8qXcVj.iso: No such file or directory

Oh, right, starting a filename with a leading dash is never a Good Idea™.

Rename it:

mv -NISLEY-DMBG8yMQcf8qXcVj.iso NISLEY-DMBG8yMQcf8qXcVj.iso
mv: invalid option -- 'N'
Try 'mv --help' for more information.

Which is why leading dashes are a Terrible Idea™.

Force the rename to happen:

mv ./-NISLEY-DMBG8yMQcf8qXcVj.iso NISLEY-DMBG8yMQcf8qXcVj.iso

The same syntax works in the mount command, but it’s easier to solve the problem once and be done with it.

Now mount the file:

sudo mount NISLEY-DMBG8yMQcf8qXcVj.iso /mnt/loop
mount: /mnt/loop: WARNING: source write-protected, mounted read-only.

That’s entirely expected, because the whole filesystem is intended for a non-writeable CD.

What’s inside?

ll /mnt/loop/
ls: cannot open directory '/mnt/loop/': Permission denied

Why would that be?

ll /mnt
total 58K
drwxr-xr-x 15 root root 4.0K May 21  2023 ./
drwxr-xr-x 17 root root 4.0K Mar  2 13:43 ../
… omitted …
drwxrwx---  4  496  495 2.0K Feb 21 13:13 loop/
… omitted …

Maybe 496 and 495 are the UID and GID of whatever created the CD?

Force it to my UID:

sudo umount /mnt/loop
[ed@shiitake tmp]$ sudo mount -o uid=ed NISLEY-DMBG8yMQcf8qXcVj.iso /mnt/loop
mount: /mnt/loop: WARNING: source write-protected, mounted read-only.
[ed@shiitake tmp]$ ll /mnt/loop
total 16K
drwxrwx---  4 ed    495 2.0K Feb 21 13:13 ./
drwxr-xr-x 15 root root 4.0K May 21  2023 ../
drwxrwx---  4 ed    495 2.0K Feb 21 13:12 data/
drwxr-xr-x  5 ed    495 2.0K Feb 21 13:13 DICOM/
-rw-rw----  1 ed    495 1.7K Feb 21 13:12 README.txt
-rw-rw----  1 ed    495 3.2K Feb 21 13:12 view-studies.html

Now that’s more like it.

Finally, I can fire up Weasis to look at pretty DICOM images:

Spine - lateral T2 TSE SAG - 2025-02 - tweaked — Spine – lateral T2 TSE SAG – 2025-02 – tweaked

Apparently things looks suspicious around L4.

2025-03-05

HQ Sixteen: Table Leveling Blocks

The Handi-Quilter HQ Sixteen rides on two tracks along the 11 foot length of the table, with an unsupported 8 foot span between the legs on each end:

HQ Sixteen - remounted handlebars in use — HQ Sixteen – remounted handlebars in use

Contemporary versions of the table have support struts in the middle that our OG version lacks and, as a result, our table had a distinct sag in the middle. During the course of aligning the table top into a plane surface with tapered wood shims, I discovered the floor was half an inch out of level between the table legs.

Now that the whole thing has settled into place, I measured the shim thicknesses and made tidy blocks to replace them:

HQ Sixteen - table shims - finished — HQ Sixteen – table shims – finished

The OpenSCAD code has an array with the thickness and the number of blocks:

SHIM_THICK = 0;
SHIM_COUNT = 1;

Shims = [
    [3.5,1],
    [5.0,3],
    [6.0,2],
    [6.5,1],
    [7.0,1]
];

Yes, I call them “blocks” here and wrote “shims” in the code. A foolish consistency, etc.

The model is a chamfered block with a chunk removed to leave a tongue of the appropriate thickness:

HQ Sixteen - table shims - solid model - single — HQ Sixteen – table shims – solid model

Building them with the label against the platform produces a nice nubbly surface:

HQ Sixteen - table shims - solid model — HQ Sixteen – table shims – PrusaSlicer – bottom

The labels print first and look lonely out there by themselves:

HQ Sixteen - table shims - legends — HQ Sixteen – table shims – legends

The rest of the first layer fills in around the labels:

HQ Sixteen - table shims - first layer — HQ Sixteen – table shims – first layer

Putting the labels on the bottom makes the wipe tower only two layers tall and eliminates filament changes above those layers. Those eight blocks still took a little over three hours, because there’s a lot of perimeter wrapped around not much interior.

Having had the foresight to draw a sketch showing where each block would go, I slid one next to its wood shim, yanked the shim out, and declared victory:

HQ Sixteen - table shims - installed — HQ Sixteen – table shims – installed

The tension rod welded under the table rail prevents even more sag, but the struts under the new version of the table show other folks were unhappy with the sag of this one. Another leg or two seems appropriate.

With the table leveled and the surface aligned, the HQ Sixteen glides easily in all directions. The result isn’t perfect and Mary keeps the anchor block at hand, but the machine now displays much less enthusiasm for rolling toward the middle of the table.

The OpenSCAD source code as a GitHub Gist:

	// HQ Sixteen – table shims
	// Ed Nisley – KE4ZNU
	// 2025-02-27

	include <BOSL2/std.scad>

	Layout = "Show"; // [Show,Build]

	/* [Hidden] */

	SHIM_THICK = 0;
	SHIM_COUNT = 1;

	Shims = [
	[3.5,1],
	[5.0,3],
	[6.0,2],
	[6.5,1],
	[7.0,1]
	];

	Block = [40.0,20.0,15.0]; // overall shim size
	Grip = 10.0; // … handle length

	BlockRadius = 1.0; // corner rounding / chamfer

	LabelThick = 0.4;
	LabelSize = 5.5;
	LabelFont = "Arial:style:Bold";
	LabelColor = "Red";

	Protrusion = 0.1;
	Gap = 5.0;

	//———-
	// Define shim shape

	module ShimBlock(Height = Shims[0][SHIM_THICK],Part="All") {

	if (Part == "Block" \|\| Part == "All")
	difference() {
	left(Grip)
	cuboid(Block,anchor=BOTTOM + LEFT,chamfer=BlockRadius);
	up(Height)
	cube(Block + 2*[Protrusion,Protrusion,0],anchor=BOTTOM + LEFT);

	left(Grip/2 – BlockRadius/2) fwd(Block.y/2 – LabelThick) up(Block.z/2)
	xrot(90) zrot(-90)
	linear_extrude(height=LabelThick + Protrusion,convexity=20)
	text(text=format_fixed(Height,1),size=LabelSize,spacing=1.00,
	font=LabelFont,halign="center",valign="center");
	}

	if (Part == "Text" \|\| Part == "All")
	color(LabelColor)
	left(Grip/2 – BlockRadius/2) fwd(Block.y/2 – LabelThick) up(Block.z/2)
	xrot(90) zrot(-90)
	linear_extrude(height=LabelThick,convexity=20)
	text(text=format_fixed(Height,1),size=LabelSize,spacing=1.00,
	font=LabelFont,halign="center",valign="center");

	}




	//———-
	// Build them all

	if (Layout == "Show")
	ShimBlock();

	if (Layout == "Build") {
	for (j=[0:len(Shims)-1])
	back(j*(Block.z + Gap))
	for (i=[0:(Shims[j][SHIM_COUNT] – 1)])
	right(i*(Block.x + Gap))
	up(Block.y/2) xrot(90)
	ShimBlock(Shims[j][SHIM_THICK],Part="Block");
	for (j=[0:len(Shims)-1])
	back(j*(Block.z + Gap))
	for (i=[0:(Shims[j][SHIM_COUNT] – 1)])
	right(i*(Block.x + Gap))
	up(Block.y/2) xrot(90)
	ShimBlock(Shims[j][SHIM_THICK],Part="Text");
	}

view raw HQ Sixteen – table shims.scad hosted with ❤ by GitHub

2025-03-04

HQ Sixteen: Ball-mounted Stylus Laser

Installing the new ball-mount laser stylus on the HQ Sixteen’s electronics pod required nothing more than two strips of good foam tape:

HQ Sixteen - Stylus Laser - installed - overview — HQ Sixteen – Stylus Laser – installed – overview

In actual use, you would:

Lay down a “pantograph” pattern on a paper strip along the rear track under the machine’s carriage
Position the needle at the appropriate spot on the quilt
Aim the laser at the corresponding point on the pattern
Start the machine!
Move the laser spot along the pattern while the machine stitches that pattern in the quilt

Mary thinks free-motion quilting is easier and I’m not in a position to argue the point.

Anyhow, the key feature of my ball mount is that it’s completely out of the way:

HQ Sixteen - Stylus Laser - installed - front — HQ Sixteen – Stylus Laser – installed – front

Which looks comfortingly like the original solid model:

HQ Sixteen - Stylus Laser Mount - solid model - show — HQ Sixteen – Stylus Laser Mount – solid model – show

Minus the vivid red death ray and pew! pew! pew!

Power comes from a barrel jack in the back intended for the original stylus laser; all small lasers, unless otherwise noted, run from 5 VDC. The jack is 3.5×1.3 mm, but the Drawer o’ Weird Barrel Plugs disgorged a matching right-angle plug. Unsurprisingly, such things are readily available these days.

Splice the laser leads to the plug and cover the evidence with a braided loom + heatshrink tubing:

HQ Sixteen - Stylus Laser - installed - rear — HQ Sixteen – Stylus Laser – installed – rear

I considered a switch, but the anticipated low duty cycle suggested just unplugging it, so that’s that.

And It Just Worked™.

The backstory begins There and continues to now.

2025-03-03

HQ Sixteen: Stylus Laser Ball Drilling

With the ball mount in hand:

HQ Sixteen - Stylus Laser - ball clamp test fit — HQ Sixteen – Stylus Laser – ball clamp test fit

The next step is to drill a 12 mm hole for the red-dot laser module right through the middle of the 1 inch = 25.4 mm polypropylene ball.

I decided to use a more-or-less standard laser module, rather than the Genuine Handi-Quilter laser, because:

Cheap & readily available
Identical spares on hand
Two decades of red laser diode progress

Start by conjuring a lathe chuck fixture for a 1 inch ball from my OpenSCAD model and printing it in PETG-CF:

HQ Sixteen - Stylus Laser - center drilling — HQ Sixteen – Stylus Laser – center drilling

Run a few drills through the ball up to 15/32 inch = 0.469 inch = 11.9 mm:

HQ Sixteen - Stylus Laser - final drilling — HQ Sixteen – Stylus Laser – final drilling

Which looks terrifying and was no big deal.

The laser module didn’t quite fit until I peeled off the label, as setting up a boring bar seemed like too much hassle for too little gain. The ball is slick polypropylene and the laser module is chromed plastic, which means there’s not much friction involved and a stiff fit is a Good Thing™.

I did not realize the hazy white patches barely visible inside the ball were voids / bubbles:

HQ Sixteen - Stylus Laser - drilled ball — HQ Sixteen – Stylus Laser – drilled ball

Next time I’ll (try to) orient the patches toward the tailstock in hopes of simply drilling through them to leave solid plastic around the rim.

Ramming the laser in place makes it look like it grew there;

HQ Sixteen - Stylus Laser - laser test fit — HQ Sixteen – Stylus Laser – laser test fit

The alert reader will note the lens projects a line, due to my not ordering any dot modules back when I got a bunch of these things. After all, who wants a plain dot when you can light up a line or even a crosshair?

Next, wire it up and stick it on the machine …

2025-02-28

HQ Sixteen: Stylus Laser Ball Mount

My version of a mount for the HQ Sixteen’s “stylus laser” clamps a 1 inch polypropylene ball between two plates:

The plates have a sphere subtracted from them and a kerf sliced across the sphere’s equator for clamping room:

HQ Sixteen - Stylus Laser Mount - solid model — HQ Sixteen – Stylus Laser Mount – solid model

Given that this is a relatively low-stress situation, I embedded BOSL2 nuts to produce threads in the plate rather than use brass inserts.

The side plates start as simple rectangles:

HQ Sixteen - Stylus Laser Mount - solid model - mount sides — HQ Sixteen – Stylus Laser Mount – solid model – mount sides

Subtracting the electronics pod shape from those slabs matches them exactly to the curvalicious corner:

HQ Sixteen - Stylus Laser Mount - solid model - mount shaping — HQ Sixteen – Stylus Laser Mount – solid model – mount shaping

The weird angle comes from tilting the mount to aim the laser in roughly the right direction when perpendicular to the plates:

That angle can be 0° to 30°, although 25° seems about right. The slab sides neither stick out the top nor leave gaps in the corner over that range, after some cut-and-try ~~tinkering~~ sizing.

One of the M3 screws just did not want to go into its hole:

HQ Sixteen - Stylus Laser - threadless M3 screw — HQ Sixteen – Stylus Laser – threadless M3 screw

A bad day in the screw factory, I suppose.

The OpenSCAD source code as a GitHub Gist:

	// Handiquilter HQ Sixteen Stylus Laser Mount
	// Ed Nisley – KE4ZNU
	// 2025-02-23

	include <BOSL2/std.scad>
	include <BOSL2/threading.scad>

	Layout = "Pod"; // [Show,Build,Pod,Mount]

	/* [Hidden] */

	PodWidth = 110.0; // overall width of pod
	PodScrewClear = 50.0; // clear distance between pod screws
	PodRecenter = [0,0]; // pod trace upper corner to origin if not done in Inkscape

	BaseAngle = -25; // laser neutral angle

	BallOD = 25.4 + 0.2; // bearing ball + easy fit clearance
	BallOffset = [70.0,0,-35.0]; // upper corner to ball center

	LaserOD = 12.2; // laser module
	LaserLength = 38.0;

	Kerf = 1.0; // clamp gap
	Plate = [35.0,35.0,8.0 + Kerf]; // basic mount plate

	WallThick = 5.0; // upright walls: plate to pod

	WasherOD = 7.0;

	ScrewPitch = 0.5;
	ScrewNomOD = 3.0;
	ScrewNomID = ScrewNomOD – ScrewPitch;

	ScrewOC = Plate – [WasherOD,WasherOD,0];

	Gap = 5.0; // build spacing

	//———-
	// HQ Sixteen electronics pod

	module Pod() {

	xrot(90)
	down(PodWidth/2)
	linear_extrude(height=PodWidth,convexity=5)
	translate(PodRecenter)
	import("HQ Sixteeen – pod profile.svg",
	layer="Pod Profile");
	}

	module LaserPointer() {
	cylinder(d=LaserOD,h=LaserLength,center=true);
	}

	module Ball() {

	union() {
	sphere(d=BallOD,$fn=4*12);
	down(0.25*LaserLength)
	LaserPointer();
	}
	}

	module Mount() {

	union() {
	difference() {
	union() {
	cuboid(Plate,anchor=CENTER);
	for (j=[-1,1])
	translate([-(BallOffset.x – Plate.x)/2,j*(Plate.y + WallThick)/2,Kerf/2])
	cuboid([BallOffset.x,WallThick,-0.75*BallOffset.z],anchor=BOTTOM);
	}
	cuboid([4Plate.x,4Plate.y,Kerf],anchor=CENTER);
	Ball();
	for (i=[-1,1], j=[-1,1])
	translate([iScrewOC.x/2,jScrewOC.y/2,0])
	cylinder(d=1.2ScrewNomOD,h=2Plate.z,anchor=CENTER,$fn=6);

	yrot(-BaseAngle)
	translate(-BallOffset)
	Pod();

	}


	for (i=[-1,1], j=[-1,1])
	translate([iScrewOC.x/2,jScrewOC.y/2,Kerf/2])
	// flat size root dia height pitch
	threaded_nut(1.5*ScrewNomOD,ScrewNomID,(Plate.z – Kerf)/2,ScrewPitch,$slop=0.10,
	bevel=false,ibevel=false,anchor=BOTTOM);
	}
	}

	//———-
	// Build things

	if (Layout == "Pod")
	Pod();

	if (Layout == "Mount")
	Mount();


	if (Layout == "Show") {
	yrot(BaseAngle) {
	color("SteelBlue")
	Mount();
	color("Magenta",0.5)
	Ball();
	color("Red")
	yrot(180)
	cylinder(d=2,h=-2*BallOffset.z,$fn=12);

	}
	translate(-BallOffset)
	color("Silver",0.8)
	Pod();
	}

	if (Layout == "Build") {
	left(Plate.x/2 + Gap/2)
	intersection() {
	cuboid([4Plate.x,4Plate.y,-BallOffset.z],anchor=DOWN);
	down(Kerf/2)
	Mount();
	}

	right(Plate.x/2 + Gap/2)
	intersection() {
	cuboid([4Plate.x,4Plate.y,Plate.z/2],anchor=DOWN);
	up(Plate.z/2)
	Mount();
	}
	}

view raw HQ Sixteen – Stylus Laser Mount.scad hosted with ❤ by GitHub

2025-02-27