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.
A pair of Barred Owls set up housekeeping nearby and we’ve watched them swoop down on rodents in the yard. We hope they succeed in raising their owlets!
This one kept watch on us as we cleared out a dead holly bush near the garden:
I took another picture at +1 EV to show more of the owl’s face; while blowing out the highlights:
Go, owls, go!
Taken with the Sony DSC-H5 and 1.7× teleadapter, because the Pixel phone’s camera just doesn’t have enough moxie at full zoom.
We have several high-intensity / long-attention-span home projects scheduled this summer, all of which will keep me away from the Basement Laboratory.
We’re OK, all is right with our world, but painting rooms and yard maintenance always take way more time than they should, while having close to zero intellectual content.
Like, for example, the result of a strenuous morning devoted to removing a severely overgrown holly bush:
I’ll post odd & ends a few times a week until maybe mid-August, whereupon I should get back to more usual pursuits.
Enjoy your downtime …
Prompted by RCP’s battery misadventure, I replaced a handful of the Forester’s incandescent bulbs:
Despite what look like “squeeze here” markings, you must push the license plate bulb holders toward the center of the car:
They were both stuck firmly to the trim plate, so I braced a screwdriver against the outboard edge of the trim panel, after which it becomes obvious how pressing inward compresses the (plastic) spring clip so you can pull the outward side of the holder away from the hatch.
Casual searching turned up a bunch of exceedingly helpful advice for anyone DIY-ing through a Forester.
The bulbs with conical ends, known as “festoon” lamps, (unsurprisingly) come in several lengths. The Forester bulbs are about 25 mm long, (unsurprisingly) much shorter than the 31 mm LEDs that seem to be the smallest available replacements, but (surprisingly) the socket tabs have barely enough compliance for the extra half dozen millimeters:
The LEDs are much much much brighter than the incandescents, although I’d prefer warm white to cool white. The cargo compartment lamp in the back is still way too dim; I don’t understand how Subaru decided on a plastic cover tinted dark smoke gray.
All in all, a worthwhile upgrade!
I wonder how long they’ll last? I have one spare of each type …
Spotted another big turtle ready to cross the Dutchess Rail Trail along Daley Rd:
This must be the best season ever for turtles crossing vast expanses of asphalt, because I don’t recall seeing this many turtles during any previous spring. Confirmation bias in full effect, to be sure.
The picture is a dot-for-dot crop from the Sony HDR-AS30V helmet camera, demonstrating why image compression has more to do with resolution than the number of dots.
The dark spot in the grass, barely visible over on the left, is a dinner-plate-size snapping turtle recently teleported from the middle of Rt 376 just north of Robinson Lane:
The driver of the white van managed to stop both lanes during the rescue and, judging from the lack of gore, handled the snapper without incurring organic damage.
Color me impressed!
In addition to sawing through the side of the cable ferrule, the front derailleur cable began breaking at the edge of the derailleur arm:
It wouldn’t have survived another ride!
Dan pointed out CNC machined aluminum cable clamps are a thing, but those are sized for larger frame tubes than the 1.0 inch steel used on our Tour Easy ‘bents and, although I’ve shimmed everything else on the frame, I wanted to tweak the cable angle to match the arm on the derailleur.
A bit of OpenSCAD wrangling produces a likely candidate:
That’s a bulked-up revision of the prototype:
Done up in orange PETG, it demonstrated the idea worked, but two perimeter threads wrapped around 15% infill isn’t quite up to the task. Note the split along the screw on the far half and various irregularities around the ferrule.
The cable angle isn’t quite right, either, as the proper compound angle would, alas, aim the cable into the pedal crank. The bulky bushings get in the way of putting the ferrule where it should be with the screws aligned in a tidy manner, so I must get used to the jaunty angle.
The bulkier version, done with 50% infill and four perimeter threads, has the same tilt angle, but the ferrule sits further from the screws:
The view from the left side shows the cable angles slightly to the rear, but the smaller angle should make it happier:
Probably should have used black PETG. Next time, for sure!
The OpenSCAD source code as a GitHub Gist:
| // Tour Easy Derailleur Cable Clamp | |
| // Ed Nisley KE4ZNU – June 2017 | |
| /* [Build Options] */ | |
| Layout = "Build"; // [Build, Show] | |
| /* [Extrusion] */ | |
| ThreadThick = 0.25; // [0.20, 0.25] | |
| ThreadWidth = 0.40; // [0.40] | |
| function IntegerMultiple(Size,Unit) = Unit * ceil(Size / Unit); | |
| /* [Hidden] */ | |
| Protrusion = 0.01; // [0.01, 0.1] | |
| HoleWindage = 0.2; | |
| ID = 0; | |
| OD = 1; | |
| LENGTH = 2; | |
| /* [Cable Clamp] */ | |
| FrameOD = 25.7; // Tour Easy has hard inch tubing + paint | |
| Ferrule = [1.5,5.1,12.0]; // cable ferrule | |
| EntryPoint = [0,13,60]; // cable entry to derailleur, +Y to rear of bike | |
| CableTilt = -20; // tilt from parallel to frame tube | |
| CableTheta = 0; // rotation around clamp from +X axis | |
| /* [Screws and Inserts] */ | |
| ClampScrew = [3.0,5.5,35.0]; // M3 button / socket head cap screw | |
| ClampWasher = [3.7,7.0,0.7]; // M3 washer | |
| ClampNut = [3.0,6.0,4.0]; // M3 nylock nut | |
| /* | |
| ClampScrew = [4.0,7.0,25.0]; // M4 button head cap screw | |
| ClampWasher = [4.5,9.0,0.8]; // M4 washer | |
| ClampNut = [4.0,8.0,5.0]; // M4 nylock nut | |
| */ | |
| NutShift = -0; // slide bushing toward nut for clearance | |
| //- Set clamp ring dimensions | |
| WallThick = 10.0; | |
| BushingSides = 8; | |
| Bushing = [ClampScrew[ID], | |
| // ClampWasher[OD]/cos(180/8) + 4*ThreadWidth, | |
| Ferrule[LENGTH]/cos(180/BushingSides), | |
| ClampScrew[LENGTH] – 2*ClampWasher[LENGTH] – ClampNut[LENGTH]]; | |
| Ring = [FrameOD + HoleWindage,FrameOD + 2*WallThick,Ferrule[LENGTH]]; | |
| ClampScrewOC = IntegerMultiple(FrameOD + ClampWasher[OD],1); | |
| echo(str(" screw OC: ",ClampScrewOC)); | |
| ClampKerf = 0.75; // kerf between separated halves | |
| NumSides = 8*4; | |
| //- Adjust hole diameter to make the size come out right | |
| module PolyCyl(Dia,Height,ForceSides=0) { // based on nophead's polyholes | |
| Sides = (ForceSides != 0) ? ForceSides : (ceil(Dia) + 2); | |
| FixDia = Dia / cos(180/Sides); | |
| cylinder(r=(FixDia + HoleWindage)/2,h=Height,$fn=Sides); | |
| } | |
| // Construct things | |
| module ClampRing() { | |
| difference() { | |
| union() { | |
| cylinder(d=Ring[OD],h=Ring[LENGTH],$fn=NumSides); // basic ring | |
| for (j=[-1,1]) // screw bushings | |
| translate([Bushing[LENGTH]/2 + NutShift,j*ClampScrewOC/2,Ring[LENGTH]/2]) | |
| rotate([0,-90,0]) rotate(180/BushingSides) | |
| cylinder(d=Bushing[OD],h=Bushing[LENGTH],$fn=BushingSides); | |
| intersection() { | |
| rotate([CableTilt,0,CableTheta]) // reinforce cable ferrule | |
| translate([(Ring[ID] + Ring[OD])/4,0,Ferrule[LENGTH]/2]) | |
| rotate(180/8) | |
| cylinder(d=3*Ferrule[OD] + 0*ThreadWidth,2*Ferrule[LENGTH],center=true,$fn=8); | |
| cylinder(d=2*Ring[OD],h=Ring[LENGTH],$fn=NumSides); // basic ring | |
| } | |
| } | |
| translate([0,0,-Protrusion]) // frame tube | |
| cylinder(d=Ring[ID],h=Ring[LENGTH] + 2*Protrusion,$fn=NumSides); | |
| rotate([CableTilt,0,CableTheta]) // cable ferrule | |
| translate([(Ring[ID] + Ring[OD])/4,0,-0.25*Ferrule[LENGTH]]) { | |
| rotate(180/8) | |
| PolyCyl(Ferrule[OD],Ferrule[LENGTH],8); | |
| rotate(-22.5) | |
| PolyCyl(Ferrule[ID],2*Ferrule[LENGTH],4); | |
| } | |
| for (j=[-1,1]) // screw holes | |
| translate([Ring[OD]/2,j*ClampScrewOC/2,Ring[LENGTH]/2]) | |
| rotate([0,-90,0]) rotate(180/6) | |
| PolyCyl(Bushing[ID],Ring[OD],6); | |
| for (i=[-1,1], j=[-1,1]) // screw & nut seats | |
| translate([i*(Bushing[LENGTH]/2) + NutShift,j*ClampScrewOC/2,Ring[LENGTH]/2]) | |
| rotate([0,i*90,0]) rotate(180/BushingSides) | |
| cylinder(d=Bushing[OD],h=Bushing[LENGTH],$fn=BushingSides); | |
| translate([0,0,Ring[LENGTH]/2]) // slice it apart | |
| cube([ClampKerf,2*Ring[OD],2*Ring[LENGTH]],center=true); | |
| } | |
| } | |
| //- Build things | |
| if (Layout == "Show") { | |
| translate(EntryPoint) | |
| cube(1,center=true); | |
| ClampRing(); | |
| } | |
| if (Layout == "Build") { | |
| ClampRing(); | |
| } |
Mary harvested a great bunch of spearmint from a place where it wouldn’t be missed and, after rinsing, plucking, and chopping, we now have a liter of Mint Extract in the making:
The big jars got 3 oz of coarse-chopped leaves apiece, the smaller jar 1 oz, and the (removed) stems added up to 3.5 oz, so call it 1/3 waste. Not that this is an exact science, but I’d say 3/4 pound of just-picked mint, packed slightly tighter than those jars, would produce a liter of extract.
Because we started with fresh-picked leaves, a liter of 190 proof = 95% ethanol Everclear (*) will extract the oil better than the 80 proof = 40% ethanol vodka I used for dried vanilla beans.
A day later, the leaves definitely look dehydrated:
Those bottles are lying on their sides with the camera above, looking through the air bubble to the leaves. Unlike commercial mint extract, this stuff is green!
It’ll be finished after a month of daily agitation, but surely it’s an exponential process: a few hundred μl already pep up a mug o’ cocoa just fine.
In very round numbers, I get 10 drops / 0.1 ml, so 1 drop = 10 μl.
Bonus: the cutting board smells wonderful.
(*) It may be Olde White Guy Privilege, but clerks don’t even blink when I stagger up to the counter clutching a bottle of high-octane hooch; they don’t even card my age!
The Smell of Molten Projects in the Morning 