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
It never ceases to amaze me that these capacitors appear in the AC power line circuits inside old-school fluorescent shop lights:
It really is a capacitor:
Its sibling from the other end of the fixture had more ESR:
Both were likely within spec, whatever that means.
I have no idea what’s lurking inside the tidy LED tubes now living in that same fixture, of course.
Found inside a fluorescent desk lamp being salvaged for possible use as an LED task lamp:
It’s one of the few Underwriter’s Knots I’ve ever seen in the wild. Many recent (i.e., built in the last half-century) lamps pass the cords through a plastic clamp or depend on simple bushings, with some just ignoring the problem.
This anonymous lamp sports the usual Made in China sticker, but also features a genuine-looking UL sticker complete with elaborate holograms, so it may well have been sold by a reputable company. IIRC, it came from a trash can in a Vassar College hallway, back when in-person meetings were a thing; perhaps Vassar required known-good electrical hardware.
A few months of inactivity left the CNC-3018XL table parked in its homed position where the gentle-but-inexorable pressure of the switch lever displaced the foam holding the plastic actuator tab on the X-axis bearing enough that it would no longer operate reliably:
Putting foam tape in a highly leveraged position produces the same poor results as in finance.
The fix requires reorienting the switch so a solid block on the bearing can push directly on the actuator lever:
The block must curve around the bearing to give the tape enough surface area for a good grip:
The solid model for the new X-axis mount looks about like you’d expect:
I increased the home switch pulloff to 2 mm, although it’s not clear that will make any difference in the current orientation.
The OpenSCAD source code as a GitHub Gist:
| // 3018-Pro Mount for Makerbot Endstop PCB | |
| // Ed Nisley KE4ZNU – 2019-07 (using OEM machine axes) | |
| // 2022-02-02 rotate X block (after renaming axes to match new layout) | |
| /* [Build Options] */ | |
| Layout = "Show"; // [Build, Show] | |
| /* [Hidden] */ | |
| ThreadThick = 0.25; // [0.20, 0.25] | |
| ThreadWidth = 0.40; // [0.40] | |
| function IntegerMultiple(Size,Unit) = Unit * ceil(Size / Unit); | |
| Protrusion = 0.01; // [0.01, 0.1] | |
| HoleWindage = 0.2; | |
| ID = 0; | |
| OD = 1; | |
| LENGTH = 2; | |
| //- Shapes | |
| // Basic PCB with hole for switch pins | |
| // origin at switch actuator corner, as seen looking at component side | |
| SwitchClear = [15.0,5.0,2.0]; // clearance around switch pins | |
| SwitchOffset = [12.5,9.0,0.0]; // center of switch pins from actuator corner | |
| PCB = [26.0,16.4,2*SwitchClear.z]; // switch PCB beyond connector, pin height | |
| //XBlock = [PCB.x + 10.0,PCB.y,20.0]; | |
| XBlock = [PCB.x,PCB.y,10.0]; | |
| XBearing = [10.0,26.5,28.5]; | |
| XPin = [10.0,20.0,10.0]; | |
| module XMount() { | |
| if (false) // side-push switch tended to slip | |
| difference() { | |
| translate([-10.0,0,0]) | |
| cube(XBlock,center=false); | |
| translate([0,-Protrusion,10.0]) | |
| cube(XBlock + [0,2*Protrusion,0],center=false); | |
| translate(SwitchOffset + [0,0,10.0 – SwitchClear.z/2]) | |
| cube(SwitchClear + [0,0,Protrusion],center=true); | |
| } | |
| else { | |
| difference() { | |
| cube(XBlock,center=false); | |
| translate(SwitchOffset + [0,0,XBlock.z – SwitchClear.z/2]) | |
| cube(SwitchClear + [0,0,Protrusion],center=true); | |
| } | |
| translate([1.25*XBlock.x,0,0]) | |
| difference() { | |
| cube(XPin + [0,0,XBearing[OD]/4],center=false); | |
| translate([-Protrusion,XPin.y/2,XPin.z + XBearing[OD]/2]) | |
| rotate([0,90,0]) | |
| cylinder(d=XBearing[OD],h=XPin.x + 2*Protrusion,center=false); | |
| translate([-Protrusion,-XPin.y/2,XPin.z]) | |
| cube(XPin + [2*Protrusion,0,0],center=false); | |
| } | |
| } | |
| } | |
| YBlock = [PCB.x,PCB.y,5.0]; | |
| module YMount() { | |
| difference() { | |
| cube(YBlock,center=false); | |
| translate(SwitchOffset + [0,0,YBlock.z – SwitchClear.z/2]) | |
| cube(SwitchClear + [0,0,Protrusion],center=true); | |
| } | |
| } | |
| ZBlock = [PCB.x,PCB.y,6.0]; | |
| ZPin = [20.0,10.0,5.5]; | |
| module ZMount() { | |
| difference() { | |
| cube(ZBlock,center=false); | |
| translate(SwitchOffset + [0,0,ZBlock.z – SwitchClear.z/2]) | |
| cube(SwitchClear + [0,0,Protrusion],center=true); | |
| } | |
| translate([1.25*ZBlock.x,0,0]) | |
| difference() { | |
| cube(ZPin,center=false); | |
| translate([ZPin.x/2,-Protrusion,4.0]) | |
| cube(ZPin + [0,2*Protrusion,0],center=false); | |
| } | |
| } | |
| //- Build things | |
| if (Layout == "Show") { | |
| translate([0,XBlock.y,0]) | |
| YMount(); | |
| translate([0,-XBlock.y/2]) | |
| XMount(); | |
| translate([0,-(ZBlock.y + XBlock.y)]) | |
| ZMount(); | |
| } |
Just before the turn of the millennium, I bought what turned out to be a never-sufficiently-to-be-damned HP 2000C inkjet printer that served as my introduction to refilling inkjet cartridges. A few years later, a Canon S630 printer joined the stable and worked fine for perhaps five years before succumbing to a printhead death. An Epson R380 that might have cost fifteen bucks after rebate took over, drank maybe a gallon of knockoff ink through a continuous ink supply system during the next thirteen years, and finally suffered progressive printhead failure during the last year.
Something recently changed in the inkjet market: Epson (among others) now touts their “Ecotank” printers featuring large internal reservoirs refilled by 70 ml bottles of color ink priced at perhaps 20¢/ml, obtained direct from Epson via Amazon. They proudly note you can save 90% off the cost of cartridges (“Kiss Expensive Cartridges Goodbye”), without mentioning how their previous extortionate cartridge business made that possible. Of course, Ecotank printers cost far more than cartridge-based printers, but that seems reasonable to me.
Because the ink bottles fit neatly into the printer through a push-to-flow valve interlock, I can finally retire this relic:
That’s maybe fifteen years of accumulated splotches.
I hope my refusal to buy their cartridges helped immanentize their eschaton, just a little.
Good riddance.
The CNC-3018XL and MPCNC machines each have a Raspberry Pi feeding G-Code into an Arduino clone controlling the stepper motors. The former grew a USB WiFi interface in place of its internal WiFi hardware when it seemed to have difficulty connecting to the house router, while the latter pretty much worked. Of late, however, I’ve been trying to reduce the number of WiFi devices cluttering the airwaves, with the result of wiring both machines to an old Ethernet switch from the Box o’ Network Stuff:
The blue puck is the KVM button to select one of the machines for the keyboard / mouse / monitor on the bench.
One key point I generally screw up: the WiFi IP address cannot become the wired IP address without rebooting everything else on the network. Instead, just change the IP addresses and be done with it.
Collecting all the pieces in one place:
Disable the both internal WiFi hardware and Bluetooth in /boot/config.txt, thereby eliminating the need to force the WiFi down in /etc/rc.local:
dtoverlay=pi3-disable-wifi
dtoverlay=pi3-disable-bt
Define the static IP address in /etc/dhcpcd.conf:
interface eth0
static ip_address=192.168.1.34/24
static routers=192.168.1.1
static domain_name_servers=192.168.1.2
Kill IPV6 activity in /etc/sysctl.conf:
net.ipv6.conf.all.disable_ipv6=1
I very much doubt this information is either complete or correct, but it serves the purpose as of early 2022.
A new floor lamp similar to the one I adjusted to suit my chair appeared next to Mary’s chair. It was, as I expected, much too tall, but shortening it required just removing one of the vertical tube sections (exactly one foot long!), as Mary was content with the flexy arm’s reach. Perhaps as a nod to the current chip shortage, this version of the lamp has a control consisting of a mechanical knob in a lump just under the flexy arm: push to turn on, rotate for intensity, tap for color, push-and-hold for off. This is much more usable than the finicky proximity pads on my lamp (and the slightly more expensive version of this one), which is why I picked it.
Because the coaxial power connector doesn’t fit through the bushing in the base of the vertical tubes and didn’t have a connector at the control lump, I had to dismantle the lump to disconnect the power cable to remove the pipe section, an operation deep in warranty violation territory.
So, we begin.
Loosen the screw clamping the power cord to the tube just below the control lump:
Remove the two screws holding the control lump together:
Pull the front of the lump off the tube and peel off a protective foam sheet to expose the circuitry:
Power comes from a 12 VDC 400 mA wall wart, so note the wire markings:
In this case, the marked wire (with the dashed lines) is the positive conductor:
Unsolder the cable and pull it out of the entire collection of tubing. The topmost section has two inner threads, so remove one of the other sections (with inner and outer threads) and reassemble the rest. Poke the cable through the tubes, solder to PCB, tighten clamp screw, reassemble lump in reverse order, then declare victory:
The business end now hovers 39 inches (a neat 1 m) over the floor, just below her eye level, where it belongs.
An ancient Ottlite fluorescent floor lamp (one of a pair bought during a closeout sale at a minute fraction of their absurd sticker price) finally aged out. Pondering what to do with the carcass led to this discovery:
Half of a Samsung (!) LED panel (presumably sheared by the surplus supplier) fit so perfectly in place of the fluorescent tube that I just had to make it happen.
The original fluorescent ballast mounted in the smaller compartment:
I like the air-cooled triac sticking off the side of the PCB.
The lamp originally mounted parallel to the flex arm, but I wanted it at a right angle, so the molded bracket had to go:
Which required a few minutes of manual jogging:
Some coordinate drilling on the Sherline converted a rectangle of aluminum sheet into a backing plate inside the base (visible through the original holes) to spread the stress over a larger area:
The new 24 V 1 A power supply mounts pretty much where the OEM ballast came from, although I had to hack out the molded screw bosses and perch the PCB atop four aluminum standoffs anchored in globs of high-temperature hot-melt glue:
You might think the white and black wires on the right are interchanged, because you’re not supposed to switch the neutral, but only if you also insist anybody cares about the colors of wires inside a molded cord. This one came from a nominally good-quality cord with an IEC connector now in the e-waste box: trust yet always always verify.
The LED panel sticks to the aluminum sheet with thermal tape and is clamped in place with a quartet of M2.5 standoffs:
I’ll eventually make a better cover than a strip of overhead projector film (remember overhead projectors?), as spattering the LEDs with cutting oil and random conductive swarf is Bad Practice™.
A little more cutting and drilling produced an angle bracket for the lathe backsplash panel:
Thing looks like it grew there, doesn’t it?
The end of the backsplash might need a 3D printed bracket to stabilize its right-angle bends and prevent wobbulation, although I’ll wait until that becomes a real problem before solving it.
The top of that stylin’ lamp shade tapers along its length and, unfortunately, appears directly in front of the MPCNC bench across the basement (out of sight at the top) as I stand at the lathe. Having the shade not align exactly parallel to the bench is more annoying than it really should be; perhaps I can get used to it after spending more time at the lathe.
I loves me some good LED lighting …
The Smell of Molten Projects in the Morning 