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
Another stack of proto boards arrived, this time 80×120 mm, and I ran off another pair of holders:
Not wanting to, ahem, screw around with the lathe, the screws got themselves shortened the old-fashioned way: by hand, with the screw cutter, then filed and passed through a 4-40 die to clean up the threads.
Bah!
Having hacked back the end of the USB gooseneck extension, a tweak of the COB LED heatsink mount for my desk lamp produces a smaller version for a 1.8 W LED:
That fits half of a random heatsink, bandsawed just to the far side of the middle fin and milled flat.
Ream out the 5 mm hole with a #8 drill for a snug fit around the gooseneck, jam gooseneck in place, dab epoxy on the corners of the recess, mash the heatsink in place, solder wires to LED, smear epoxy on the aluminum backplate, clamp while curing:
And it looks pretty good, if I do say so myself:
The hook-n-loop tape holding the cable to the bandsaw gotta go, but should suffice until I conjure a better mount.
The alert reader may wonder how a 9 V COB LED runs from a 5 V USB cable with nary a trace of a voltage booster to be seen. Well, that’s not really a USB cable any more; I paralleled the red+white and black+green wires for lower resistance, then hacked a 9 VDC power supply into an old USB hub:
I ripped out the upstream USB plug, hotwired the 9 V supply where the 5 V USB wires used to be, soldered jumpers on the downstream sockets to short the outer two pin pairs together, razor-knifed the power leads going into the epoxy-blobbed USB controller, and declared victory:
Admittedly, that “In Use” LED runs a bit brighter now.
I have a few other tools on that bench in need of LED lights; when I build ’em, they can all plug into this hub. No reason to invent new connectors & cables & all that. It may need a power switch.
Turns your stomach, eh?
The OpenSCAD source code as a GitHub Gist:
| // Chip-on-board LED light heatsink mount for desk lamp | |
| // Ed Nisley KE4ZNU December 2015 | |
| // February 2017 – rectangular COB, smaller heatsink | |
| Layout = "Show"; // Show Build | |
| //- Extrusion parameters must match reality! | |
| ThreadThick = 0.25; | |
| ThreadWidth = 0.40; | |
| HoleWindage = 0.2; | |
| Protrusion = 0.1; // make holes end cleanly | |
| inch = 25.4; | |
| function IntegerMultiple(Size,Unit) = Unit * ceil(Size / Unit); | |
| //———————- | |
| // Dimensions | |
| ID = 0; // for round things | |
| OD = 1; | |
| LENGTH = 2; | |
| Gooseneck = [3.0,5.0,15.0]; // anchor for end of gooseneck | |
| COB = [30.0,11.0,2.5]; // Chip-on-board LED module | |
| Heatsink = [37.1,19.2,10.0]; // overall | |
| HeatsinkBase = 2.0; // solid base below fins | |
| HSLip = 1.0; // width of lip under heatsink | |
| BaseMargin = 2*2*ThreadWidth; | |
| BaseRadius = 3*ThreadThick + Gooseneck[OD]/2; // defines slab thickness | |
| BaseSides = 2*4; | |
| Base = [(Gooseneck[LENGTH] + Gooseneck[OD] + Heatsink[0] + 2*BaseRadius + BaseMargin), | |
| (Heatsink[1] + 2*BaseRadius + 2*BaseMargin), | |
| 2*BaseRadius]; | |
| echo(str("Slab thickness: ",Base[2])); | |
| //———————- | |
| // Useful routines | |
| 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, | |
| $fn=Sides); | |
| } | |
| //– Lamp heatsink mount | |
| module Lamp() { | |
| difference() { | |
| translate([(Base[0]/2 – BaseRadius – Gooseneck[LENGTH]),0,0]) | |
| hull() | |
| for (i=[-1,1], j=[-1,1]) | |
| translate([i*(Base[0]/2 – BaseRadius),j*(Base[1]/2 – BaseRadius),Base[2]/2]) | |
| sphere(r=BaseRadius/cos(180/BaseSides),$fn=BaseSides); | |
| translate([(Heatsink[0]/2 + Gooseneck[OD]), // main heatsink recess | |
| 0, | |
| (Base[2] + Heatsink[2]/2 – HeatsinkBase)]) | |
| cube((Heatsink + [HoleWindage,HoleWindage,0.0]),center=true); | |
| translate([(Heatsink[0]/2 + Gooseneck[OD]),0,HeatsinkBase]) // lower lip to shade lamp module | |
| scale([1,1,2]) | |
| cube(Heatsink – [2*HSLip,2*HSLip,0],center=true); | |
| translate([0,0,Base[2]/2]) // goooseneck insertion | |
| rotate([0,-90,0]) rotate(180/8) | |
| PolyCyl(Gooseneck[OD],Base[0],8); | |
| translate([0,0,Base[2]/2 + Gooseneck[ID]/2]) // wire exit | |
| rotate([180,0,0]) | |
| PolyCyl(Gooseneck[ID],Base[2],6); | |
| translate([Gooseneck[OD],0,(Base[2] – HeatsinkBase – Protrusion)/2]) // wire slot | |
| rotate([180,0,0]) | |
| cube([2*Gooseneck[OD],Gooseneck[ID],(Base[2] – HeatsinkBase + Protrusion)],center=true); | |
| } | |
| } | |
| //———————- | |
| // Build it | |
| if (Layout == "Show") { | |
| Lamp(); | |
| } | |
| if (Layout == "Build") { | |
| } |
A stack of PCB proto boards arrived from halfway around the planet and, in a rare preemptive strike, I made some holders before I need them:
That called for trimming eight 4-40 screws, so I also tried a first pass at a screw cutting fixture:
That’s an ordinary 4-40 brass insert epoxied inside a drilled hole, with a snug 4-40 clearance hole on the other side.
If I needed slightly longer screws, they’d get a jam nut on the inside, but in this case I just tightened it firmly, ran the lathe in reverse, and cut against the far side to keep the screw from working loose:
The insert is on the left side of the fixture, just under the screw head.
Unfortunately, my faith in epoxy bonding led me astray: there’s just not enough griptivity to lock the insert inside that drilled aluminum hole. Despite my taking sissy cuts, the cutting forces pushed the insert out of its hole. I completed the mission by cutting the last four screws by hand.
The original idea for the fixture would have me turning the fixture from steel and tapping the screw hole. That puts a lot of labor into something that may get chewed up fairly quickly, so I wondered if a brass insert would suffice.
Not in that orientation, it doesn’t. Putting the insert on the other side of the fixture (to the right, away from the chuck) would have the cutting forces pushing it into the fixture, which should have been obvious from the start. So it goes.
When I must cut a larger screw, I’ll redrill that fixture, put the insert on the other side, and see how that plays.
If I turned the fixture from steel with similar drilled similar holes, I could braze / silver solder the insert into the hole: that would prevent it from turning, even if the jam nut wasn’t quite up to the task.
After a week of use, Mary decided the single additional graphite disk in each stack produced a too-high initial speed when the sewing machine started up; this being a matter of how it feels injects some of trial-and-error into the repair.
Shaving a graphite disk down from 0.8 to 0.4 mm seemed entirely too messy, so I snipped squares from 0.40 mm = 16 mil brass shim stock, nibbled the edges into a polygon, and filed the resulting vertexes to produce a (rough) circle:
Each stack looks like this:
Protip: dump those shards onto a strip of wide masking tape, fold gently until it’s all corners, and drop in the trash. Otherwise, you’ll pull those things out of your shoes and fingers for months…
You can get cheaper nibbling tools nowadays; I’ve had mine for decades.
A cleaned up version of my trusty circuit board holder now keeps the 60 kHz preamp off what passes for a floor in the attic:
The solid model became slightly taller than before, due to a serious tangle of wiring below the board, with a narrower flange that fits just as well in the benchtop gripper:
Tidy brass inserts epoxied in the corners replace the previous raw screw holes in the plastic:
The screws standing on their heads have washers epoxied in place, although that’s certainly not necessary; the dab of left-over epoxy called out for something. The screws got cut down to 7 mm after curing.
The preamp attaches to a lumpy circle of loop antenna hung from the rafters and returns reasonable results:
The OpenSCAD source code as a GitHub Gist:
| // Test support frame for proto boards | |
| // Ed Nisley KE4ZNU – Jan 2017 | |
| ClampFlange = true; | |
| Channel = false; | |
| //- Extrusion parameters – must match reality! | |
| ThreadThick = 0.25; | |
| ThreadWidth = 0.40; | |
| function IntegerMultiple(Size,Unit) = Unit * ceil(Size / Unit); | |
| Protrusion = 0.1; | |
| HoleWindage = 0.2; | |
| //- Screw sizes | |
| inch = 25.4; | |
| Tap4_40 = 0.089 * inch; | |
| Clear4_40 = 0.110 * inch; | |
| Head4_40 = 0.211 * inch; | |
| Head4_40Thick = 0.065 * inch; | |
| Nut4_40Dia = 0.228 * inch; | |
| Nut4_40Thick = 0.086 * inch; | |
| Washer4_40OD = 0.270 * inch; | |
| Washer4_40ID = 0.123 * inch; | |
| ID = 0; | |
| OD = 1; | |
| LENGTH = 2; | |
| Insert = [3.9,4.6,5.8]; | |
| //- PCB sizes | |
| PCBSize = [110.0,80.0,1.5]; | |
| PCBShelf = 2.0; | |
| Clearance = 2*[ThreadWidth,ThreadWidth,0]; | |
| WallThick = 5.0; | |
| FrameHeight = 10.0; | |
| ScrewOffset = 0.0 + Clear4_40/2; | |
| ScrewSites = [[-1,1],[-1,1]]; // -1/0/+1 = left/mid/right and bottom/mid/top | |
| OAHeight = FrameHeight + Clearance[2] + PCBSize[2]; | |
| FlangeExtension = 3.0; | |
| FlangeThick = IntegerMultiple(2.0,ThreadThick); | |
| Flange = PCBSize | |
| + 2*[ScrewOffset,ScrewOffset,0] | |
| + 2*[Washer4_40OD,Washer4_40OD,0] | |
| + [2*FlangeExtension,2*FlangeExtension,(FlangeThick – PCBSize[2])] | |
| ; | |
| echo("Flange: ",Flange); | |
| NumSides = 4*5; | |
| WireChannel = [Flange[0],15.0,3.0 + PCBSize[2]]; | |
| WireChannelOffset = [Flange[0]/2,25.0,(FrameHeight + PCBSize[2] – WireChannel[2]/2)]; | |
| //- 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,$fn=Sides); | |
| } | |
| //- Build it | |
| difference() { | |
| union() { // body block | |
| translate([0,0,OAHeight/2]) | |
| cube(PCBSize + Clearance + [2*WallThick,2*WallThick,FrameHeight],center=true); | |
| for (x=[-1,1], y=[-1,1]) { // screw bosses | |
| translate([x*(PCBSize[0]/2 + ScrewOffset), | |
| y*(PCBSize[1]/2 + ScrewOffset), | |
| 0]) | |
| cylinder(r=Washer4_40OD,h=OAHeight,$fn=NumSides); | |
| } | |
| if (ClampFlange) // flange for work holder | |
| linear_extrude(height=Flange[2]) | |
| hull() | |
| for (i=[-1,1], j=[-1,1]) { | |
| translate([i*(Flange[0]/2 – Washer4_40OD/2),j*(Flange[1]/2 – Washer4_40OD/2)]) | |
| circle(d=Washer4_40OD,$fn=NumSides); | |
| } | |
| } | |
| for (x=[-1,1], y=[-1,1]) { // screw position indexes | |
| translate([x*(PCBSize[0]/2 + ScrewOffset), | |
| y*(PCBSize[1]/2 + ScrewOffset), | |
| -Protrusion]) | |
| rotate(x*y*180/(2*6)) | |
| PolyCyl(Clear4_40,(OAHeight + 2*Protrusion),6); // screw clearance holes | |
| translate([x*(PCBSize[0]/2 + ScrewOffset), | |
| y*(PCBSize[1]/2 + ScrewOffset), | |
| OAHeight – PCBSize[2] – Insert[LENGTH]]) | |
| rotate(x*y*180/(2*6)) | |
| PolyCyl(Insert[OD],Insert[LENGTH] + Protrusion,6); // inserts | |
| translate([x*(PCBSize[0]/2 + ScrewOffset), | |
| y*(PCBSize[1]/2 + ScrewOffset), | |
| OAHeight – PCBSize[2]]) | |
| PolyCyl(1.2*Washer4_40OD,(PCBSize[2] + Protrusion),NumSides); // washers | |
| } | |
| translate([0,0,OAHeight/2]) // through hole below PCB | |
| cube(PCBSize – 2*[PCBShelf,PCBShelf,0] + [0,0,2*OAHeight],center=true); | |
| translate([0,0,(OAHeight – (PCBSize[2] + Clearance[2])/2 + Protrusion/2)]) // PCB pocket on top | |
| cube(PCBSize + Clearance + [0,0,Protrusion],center=true); | |
| if (Channel) | |
| translate(WireChannelOffset) // opening for wires from bottom side | |
| cube(WireChannel + [0,0,Protrusion],center=true); | |
| } | |
After watching Mary fiddle with the shrunken presser foot screw, I tapered the tip as a guide into the hole:
A hint-and-tip (which I cannot, alas, find again) suggested making bushings to simplify trimming screws in the lathe. A rim on the bushing aligns it with the front of the jaws, the screw threads into the central hole with a jam nut locking it in place, then you can turn / shape / file the end of the screw just beyond bushing with great support and a total lack of drama.
For the moment, I just aligned the screw in the tailstock drill chuck, crunched the three-jaw spindle chuck on the screw head, backed off the tailstock, took unsupported sissy cuts, and it was all good:
Gotta make those bushings!
It turns out that the various Avahi daemons performing the magick between whatever.local names and dotted-quad 192.168.1.101 addresses for Raspberry Pi descend into gibbering madness when confronted with:
The least of the confusion involved an incorrect IP address linked to a familiar name pulled from deep history by a baffled daemon doing the best it can with what it thinks it knows. Despite what I concluded, rather early in the process, there’s no real error, other than my performing what amounted to a self-inflicted fast-flux nameserver attack.
Anyhow, I devoted the better part of an afternoon to sorting out the mess, which involved labeling all the streaming radio players with their MAC addresses and rebooting them one-by-one to allow all the daemons time to recognize the current situation:
That label corresponds to the Pi 3’s on-board WiFi adapter.
For Pi 2 boxen, the MAC address travels with the WiFi adapter jammed into a USB port:
I didn’t label the (unused) Ethernet jacks, figuring I’d solve that problem after it trips me up.
The Smell of Molten Projects in the Morning 