The Smell of Molten Projects in the Morning
Ed Nisley's Blog: Shop notes, electronics, firmware, machinery, 3D printing, laser cuttery, and curiosities. Contents: 100% human thinking, 0% AI slop.
Making the world a better place, one piece at a time
The weekly battery swap revealed the dismal state of the headset wires on Mary’s radio:
That’s after sorting & disentangling loose ends, ramming cables under their ties, and generally tidying things up.
Which suggested an improvement I should have done long ago:
Verily, it is written: the bigger the blob, the better the job.
Gotta glob my bike the next time around.
Update: It’s hot melt glue!
My ancient fluorescent magnifying desk lamp emerged from a box and cried out to be used, but the equally ancient 22 W fluorescent ring light was long past its prime and cried out to be replaced with something from the current millennium.
So I removed the fluorescent ballast / choke from the junction box at the lamp base:
That’s a grounded outlet in the cover plate serving as a wire termination block. The red crimp connector joins a white wire that formerly went to the ballast with the black wire going to the lamp head; you’ll note the black wire from the line cord going into the same heatstink tubing at the outlet.
The lamp head had a push-to-start switch, presumably with an internal starting capacitor or some such, but also sporting a pair of terminals behaving like a single-pole push-on / push-off switch. A bit of rewiring, of which there are no pictures, made it work perfectly with the new 13 W LED ring light:
It now sits on a bit of laboratory ironmongery weighing about as much as a small child:
Although the base has four feet, it sits perfectly flat on my (admittedly battered) surface plate because all four feet have been ground to make that happen:
Those feet will be hostile to any table / bench top outside their intended laboratory environment. Fortunately, the geometry is simple enough to build directly in LightBurn and cut from a cork disk with PSA backing suited to become a coaster:
Which fit well enough, although all four feet are just slightly different:
The new Basement Shop™ is coming together and this stuff is getting easier …
The WordPress AI came up with a plausible steampunk build:
Love those flowy feet, although the vertical rod in the back seems misplaced.
Adding “one-piece base” to the prompt produces contemporary style:
Dunno what the dingus on the lower arm might be (perhaps a spring?), but it’s got the right general idea.
If I’m making something with the laser cutter during a Zoomed SquidWrench meeting, switching the view to a camera inside the cabinet is always a crowd-pleaser. Having tried several locations with various degrees of success, the camera now sits atop a small chipboard box holding it as high as it can get over the front left corner of the platform:
It juuuust barely kisses the head in that corner:
The box is a simple ClosedBox from boxes.py with a jawbreaker URL full of parameters, minus two holes, plus a quartet of shims, all cut from chipboard:
The holes fit a 1/4-20 button head screw for the Logitch C615 camera’s tripod mount and the hex wrench needed to tighten it:
The box is held together with Genuine Scotch Tape, because I want it to fall apart if it gets hit by the laser head. It’s held to the cabinet with a finger-crushingly strong bar magnet.
The camera has a reasonably good view of the entire platform:
The camera’s closest focus point sits about halfway across the platform, roughly corresponding to the typical monitor-to-face distance the camera was intended for, but it’s Good Enough™ for the purpose.
After only a dozen years, one of the 3D printed replacement handles for my Harbor Freight bar clamps snapped exactly where you’d expect:
The replacement this time around is laser-cut plywood, with a pair of 3 mm sheets glued together to just about match the original thickness:
I hacked the OpenSCAD code to use its projection() operation to export the outline of the solid model on the XY plane, inhaled the SVG into LightBurn, replaced the original chunky hole with a Real Circle, cut a pair of them, discovered I messed up the diameter, tweaked that, cut a pair that fit perfectly, and that was that.
Flushed with success, I cut another pair to replace the (not yet failed) handle in the other HF bar clamp and restarted the failure clock.
Not as fancy as something milled on the Sherline, but way easier and, if it lasts another decade, I’ll call it a win.
The WordPress AI had fun with this post:
The thing over on the left must be a 3D printer, but what’s floating in the middle? Those hand tools look downright scary.
Having had considerable success re-stringing the seat on Mary’s Tour Easy with Kevlar line, this seemed like a good idea:
They’re a pair of plumb bobs I’ve had forever, with ten feet of 100 pound Kevlar line in place of their original (and well-worn) string. The key parameter is the 1 mm diameter to fit the holes in the bobs, with the 20 pound working strength being grossly overqualified.
Because it’s braided line, the bobs no longer spin merrily when deployed, which is a major win.
They hibernate in pill bottles during their downtime, where the line doesn’t get nearly as tangled as you might imagine.
Admittedly, I don’t deploy plumb bobs all that often, but in recent months I have needed to know what was directly below something else often enough to remind me to get this done.
The WordPress AI image for this post is, as usual, disturbing:
I like the perfectly rendered gibberish text. Is that a bizarre spade drill in the lower left? So many pills!
Mary’s Juki TL-2010Q sewing machine sits in an Arrow Gidget II sewing table with a clear acrylic insert filling the opening:
Before the insert arrived (it had month of leadtime), I hacked out a temporary cardboard insert:
Although it may not be obvious from the picture, unlike my cardboard insert, the acrylic insert does not fill the tabletop hole to the immediate right of the machine:
Custom Inserts are U-shaped, designed to fit around all 3 sides of your sewing machine
Shortly after the insert arrived I hacked a temporary filler, for which no pictures survive, to keep pins / tools / whatever from falling to their doom. This turned out to be a blessing in disguise, because she wanted the machine positioned an inch to the right of its intended spot to leave enough space for a finger to reach the bobbin hatch latch.
I then promised to replace the ugly cardboard filler with a less awful acrylic filler and finally got it done:
The stack of cardboard prototypes show iterative fit-and-finish improvements, with the odd shape on the top serving to measure the machine’s 25 mm corner radius by comparison with known circles.
The insert filler is made from smoked gray acrylic, because I have yet to unpack the acrylic stockpile and may not, in fact, have any clear 6 mm acrylic, so we’ll regard this as a final prototype pending further developments. It did, however, confirm the laser survived the move, which was pretty much the whole point.
The end of the machine is not a straight line. Part of the iteration was measuring the curve’s chord height to calculate the circle’s radius, which turned out to be 760 mm:
With that in hand, a few Boolean operations produced the filler shape:
A pair of silicone bumper feet stuck to the side of the Juki hold the left edge of the filler at the proper level.
For the record, the smoked acrylic came from a fragment of a Genuine IBM Printer stand I’ve had in the scrap pile since The Good Old Days:
By a quirk of fate, the Chamberlain garage door opener in our new house has the same “purple learn button” as the Sears opener in our old house, so I introduced it to our remotes and they work just fine.
I then replaced the four-button remote in my bike pack with a new single-button remote to reduce the dexterity required to hit the button:
Alas, the opener only responded when the remote was immediately outside the aluminum garage door. Checking the battery (because sometimes “new” does not mean what you think it means) reminded me we live in an age when hardware is free compared with bookkeeping:
Maybe the second button doesn’t work and this is how they monetize their QC reject pile?
I want the door to start moving when I’m at the end of the driveway, giving it enough time to get all the way up so I can bike right in. You can actually buy remote / extension antennas, although for fancier openers with SMA antenna connectors, but sometimes a little RF black magic will suffice:
The wavy wire hanging down from the opener’s rear panel is the original antenna, which might be kinda-sorta omnidirectional. The opener operates around 433 MHz= 69 cm, so a quarter-wave antenna will be 17 cm = 7 inch long; the (unbent) wire is maybe 10 inches long from the hole in the panel.
So I taped 11 inches of wire to the opener to form a very very crude Yagi-Uda antenna. It’s too long to be a director element, it’s about right (albeit in the wrong place) to be a reflector element, it might be neither.
What it does do is warp the antenna’s pattern just enough to let the remote reliably trigger the opener as I approach the end of the driveway.
Do not even begin to think about polarization mismatch from what looks like the tiny loop antenna on the remote’s PCB.
The Smell of Molten Projects in the Morning 