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.

Author: Ed

LightBurn: Nesting Shapes

A question on the LightBurn forum about packing shapes onto an irregular piece of leather let me work out the details of a LightBurn feature I thought I understood but had trouble explaining.

Start with an irregular shape:

Random fabric – overview

That’s made of rags from the box o’ wipes out of view on the right, laid out in no particular order, on a contrasting background to simplify the next step.

The camera tucked into the lid shows the view from above:

Random fabric – LB camera overlay

Tap the Trace button and fiddle with the sliders to get a nice solid outline, along with other junk off the edge of the cardboard:

Random fabric – LB trace

All of the traced vectors will be in a group:

Random fabric – LB shapes

Ungroup them, select the outline in the middle, invert the selection, and mass-delete the junk around the edges.

If you don’t move anything, the outline will be exactly over the shape on the platform. This will come in handy later.

Import all the shapes you want nested inside the outline, group them with the outline, and hand them to the Arrange → Nest Selected tool:

Random fabric – LB nesting setup

LightBurn saves the selected objects as an SVG file with the file name in the clipboard and fires up a browser tab at https://svgnest.com/. Upload the SVG and let the nesting algorithm chew away for a while:

Random fabric – LB nested

The weird triangles come from the Dot Mode perforations that ought not be there; inner shapes get subtracted from outer ones, which makes perfect sense. Your shapes will differ.

Download the nested shape SVG, load it back into LightBurn at the prompt generated after exporting the shapes, and LightBurn will apply the transforms to the original shapes. Put the outer shape on a tool layer and the inner shapes on whatever cutting layer you like, snap the outer shape (with the nested shapes inside) to the previously undisturbed outline of the stuff on the platform, and Fire The Laser!

Now there’s a pretty good chance I can do that again …

2023-07-26
Thanks for the Notification

This year’s MVP health plan has a different “OTC Benefit” than last year, even though MVP is contracting with the same company to provide what seems to be essentially the same benefit.

This arrived half a year after the new OTC benefit card showed up:

MVP OTC Card Expiry

I suppose somebody noticed MVP hadn’t gotten around to telling us they were cancelling the old card, despite its Valid Thru 12/26 notation. Well, the card isn’t exactly cancelled, it just stopped working when all the money evaporated.

This not being my first ride in this particular rodeo, I spent all those sweet OTC benny bucks days after they become valid on the first day of every quarter-year, buying up my stock of overpriced OTC stuff.

In theory, you could buy the stuff elsewhere, but you had to scan each item in the retail store using the worst app imaginable to determine its eligibility and coverage. If the store was in a no-wireless-data phone zone: too bad, so sad.

This year’s program is simpler: you must buy everything from the sole-source supplier, even though it costs four times more than the comparable item at, say, Walmart or even Amazon.

2023-07-25
Manjaro XFCE Slow File Loading
A month or so ago a Manjaro update caused all file loading to take minutes, rather than seconds. This sort of breakage seems endemic to rolling update distros, although most glitches vanish within a few days as more knowledgeable users track down the problems and apply the fixes.

File loads and program startups continued to be achingly slow, so I trawled the Interwebs in search of a resolution, tried various suggestions, and had no success until:
```
sudo pacman --remove xdg-desktop-portal-gnome
```
Some background information:
A description of what a desktop-portal is all about:
- https://wiki.archlinux.org/title/XDG_Desktop_Portal#List_of_backends_and_interfaces
When using Free Software, when it breaks you get to keep all the pieces. In this case, I do not profess to understand how the pieces fit together.
2023-07-24
Onion Maggot Fly vs. Sticky Traps: Season 3 Round 1
Six sticky traps have been out in Mary’s Vassar Farm onion bed from mid-April through mid-July, collecting onion maggot flies, other flying insects, and a bunch of shredded leaf mulch. Having just replaced all the sticky sheets, these are the results so far:
PXL_20230711_215255180 – VCCG Onion Maggot Trap F
PXL_20230711_215229538 – VCCG Onion Maggot Trap E
PXL_20230711_215159950 – VCCG Onion Maggot Trap D
PXL_20230711_215129817 – VCCG Onion Maggot Trap C
PXL_20230711_215041012 – VCCG Onion Maggot Trap B
PXL_20230711_215002214 – VCCG Onion Maggot Trap A
Each image is the front and back of a single sticky sheet flipped over left-to-right; I did not keep track of the original trap locations.

If you need the original camera images to get enough pixels for itemizing the smaller dots, let me know.
2023-07-21
Laser Perforations
A discussion on the LightBurn forum produced a hacky way to laser-cut pinholes at precise locations:

Laser-cut pinhole – aligned exit

That’s the 0.3 mm exit wound in 3 mm acrylic, one of the mini-lathe chuck stops, carefully hand-held to align the channel.

Squinting at similar holes through clear acrylic shows they’re smoothly melted (as you’d expect), but not exactly perpendicular to the surface. I’m sure the acrylic gas pushes the beam around and erodes the sides of the channel as it boils out of the progressively deepening hole.

The entry wound is about half a millimeter:

Laser-cut pinhole – entry

The heat-distorted strip around the perimeter is less obvious in real life without magnification. The protective plastic film over the surface melts easily and, although it does keep the fumes from condensing, causes a bit of damage.

Each pinhole comes from a single dot in LightBurn’s Dot Mode, so you must arrange the dot spacing to match the path:

Lathe Chuck Stop – Pinhole distance

The pockets are on a 40 mm BCD, so they’re out 20 mm from the center and the hole-to-hole distance is:
```
34.64 mm = 2 × 20 mm × cos(30°)
```
Set the dot distance to that exact number and It Just Works.

The laser turns on for a specific number of milliseconds at each dot. In this case, I used 50 ms with the layer set to 70% PWM. You could surely optimize the values.

The starting pinhole gets drilled twice, which happens because Dot Mode expects to make a line of perforations with one dot at each end. In this case, the end of the last line overlaps the start of the first line; two lines would work better than a triangle.

You could make a square array from a single line with (many) dots at the desired spacing, separating the lines by the same spacing.

A circular array might work, too, with a straight line joining successive holes.

Undo would definitely be my copilot while figuring those out.

This could make an easily clogged trash strainer or a filter for small chunks.
2023-07-20
Laser Power Measurement: Geometric Beam Absorber
CO₂ laser power meters seem to depend on a flat-black absorbing surface to soak up a (typically unfocused) beam pulse, backed by a known metal mass with a thermocouple to measure the temperature rise above ambient. Knowing the pulse width, the temperature rise, the absorber mass and specific heat capacity, you can compute the pulse energy and average power during the pulse.

Previous tinkering with an old Gentec ED-200 showed this works well, although the absorber surface took something of a beating because it was definitely not rated for the OMTech’s 60 W (claimed) beam power.

Rather than using a spendy absorber surface with a durable coating, perhaps a geometric absorber using reflective surfaces arranged to channel the energy into the material, rather than away from it, might suffice.

Consider a pack of ordinary utility knife blades:

Beam absorber – utility blades – overview

Seen kinda-sorta perpendicular to the sharpened side of the blade edge, they’re wonderfully reflective:

Beam absorber – utility blades – edge flat

Seen perpendicular to the edge itself, they’re dead black:

Beam absorber – utility blades – edge-on

Well, pretty close to dead black. It’s darker in real life, with glimmers along the edge and the rest of it a deep black. The edges are sharp, but utility knife blades will lead a rough life and they don’t start out Scary Sharp.

Xacto blades come closer to an ideal razor edge:

Beam absorber – Xacto 11 blades – edge-on

The only things you (well, I) see is dust on the edges. The rest is dead black, because light hitting any shiny surface is reflected deeper into the notch between two blades and eventually absorbed.

Double-edge razor blades are sharper and would likely be even blacker, particularly cheap ones without fancy lubricating coatings.

Bonus: the wavelength of CO₂ laser IR light is 10-20× that of visible light, which makes the surfaces that much more reflective. The geometry still channels the reflections into the block and nothing comes out.

There are some fairly obvious reasons why nobody uses a stack of razor blades as a beam absorber in real life:
- Lethally sharp cutting hazard
- Impossible to clean without wrecking the edge
But for personal use, why not?

Some doodles:

Steel has a specific heat around 0.47 J/g·K and a stack of utility blades weighing 140 g is 23 mm across. Soaking up a 60 W beam will raise the temperature of the stack by:
```
0.91 K/s = 60 J/s / (0.47 J/g·K × 140 g)
```
Which seems reasonable: fire a 10 s burst, measure the temperature rise, and multiply by 0.91.

Similarly, a stack of Xacto #11 weighing 15 g is 11 mm across and the temperature will rise 8.5 °C/s. You’d use that for lower power beams.

You could clamp the blades into a larger heatsink, perhaps with a thermocouple / thermistor in a hole drilled into the block.

Calibrate the stack / heatsink with an embedded cartridge heater: voltage × current × pulse width gives the power dumped into the block, so measuring the temperature rise gives you the temperature-power relation.

This feels like a great Arduino project, although it’s nowhere near getting started.

At least I got that scrap of paper off my desk …
2023-07-19
Kenmore 362.75581890 Oven Igniter: Third Contestant

Although the oven igniter I just installed worked, its 3.0 A current fell below the gas valve’s minimum 3.3 A, which, based on past experience, suggested it would fail in short order. Just to see what happened, I sent a note to the seller, who offered a warranty swap and, after a bit of fiddling, the replacement arrived:

Oven Igniter B – 3.3 A initial current

This one draws exactly 3.3 A, so it just barely meets both its product description and the gas valve’s minimum current.

We’ll see how long this lasts …

2023-07-18