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.
Things around the home & hearth
Surprisingly, the flashlight holding these cells wasn’t damaged.
Judging from the position of the switch, my mother tried to turn the thing on, it didn’t light up, and she just dumped it back in the drawer. Time passed, corrosion never sleeps, and the weak link (fortunately, between the two cells) let the alkaline nastiness out.
I gotta collect all these pix in one big pile…
(These are not the D-cells from the Maglite adventure.)
Just refilled the continuous-flow ink tanks on my Epson R380 for the first time this year.
In milliliters:
Done in that order, the whole process requires only three syringe cleanings.
I haven’t checked lately, but the last time I worked it out, printer ink cost $1.80/ml. Let’s call it $2 nowadays, which means I just squirted nigh onto 500 bucks worth of ink into those tanks.
Much of that seems to go into head cleaning, as I don’t print that many photo-quality images on it. Nevertheless, they don’t charge you any less for ink that winds up in the diaper inside the printer.
I don’t keep track of the ink going into the refillable tanks on my Canon S630, but it’s the same order of magnitude. I have four trios of color tanks and six black tanks for that printer and refill them all when the last set runs low. I’ve been doing that for years with no printhead issues.
Sooo, for not spending a kilobuck a year on ink cartridges, I’m willing to spend 100 bucks for bulk ink, undergo some hassle, and endure the occasional oddly colored thumb…
For reasons that shouldn’t require the least bit of explanation by now, I had to dismantle(*) an old 2-D-cell Maglite. The operative word here is old, because you can find plenty of instructions & pix telling you how to dismantle the newer (post-2001, evidently), cheapnified Maglites. Mine dates back to the early days.
Unlike new(er) Maglites, the switch assembly in this one comes out through the front. An aluminum retaining nut holds it in place, as shown in the first picture. You’ll find directions telling you to unscrew the nut by jamming a pair of needle-nose pliers into the holes, but that’s not how it’s done.
The job calls for a pin wrench!
Measuring the dimensions is no BFD after you’ve got the damned thing apart, but I didn’t have that luxury. Given this was an American product from back in the Olde Days, I assumed everything was denominated in inches, which turned out to be close enough.
The “Max” dimensions at the bottom are the actual ID measurements from the housing after disassembly, using telescoping gages. I made the wrench to the dimensions on the line just above and they worked fine.
Believe it or not, I found a steel cylinder in my scrap heap that was just exactly what I needed, right down to the 7/8″ bore in the middle. Not only that, it was free-machining steel. Whew!
The inner bore must clear the brass screw head sticking out of the lamp tower in the middle (which rides in a slot as part of the sliding focus mechanism). Once you’ve extricated the switch assembly, you remove that screw with a 2 mm (so much for hard inch dimensions) hex key. If you’re desperate, you can probably worry the screw out by goobering it with the aforementioned needle-nose pliers; it has an ordinary right-hand thread.
I turned the cylinder down in the lathe, then drilled the pin holes. That’s a mistake: the outside edge of the pins is exactly even with the OD of the wrench nose. If you do this, clean up the stock OD & face the ends to get a nice cylinder, drill the pin holes, then turn down the barrel clearance and nose. It need not be perfectly concentric, so stop worrying.
I did the drilling using manual CNC on the Sherline mill, mostly because that’s the only way I could poke the holes in the right spots. The mill doesn’t have a lot of vertical headroom, so I clamped the wrench directly to the table and touched off the X and Y axes to put the origin in the center.
I got it all clamped down, removed the right-hand clamp to touch off on the +X side, then re-clamped it.
Center drill to fix the hole location. Drill 1/8″ about 0.250 deep: 3000 rpm, 10 ipm feed, use a little cutting lube. Do those both in sequence at each hole.
I sliced two overly long stubs from some 1/8″ drill rod with a Dremel cutoff wheel, dabbed JB Weld in the holes, and poked them in. The next morning I sliced them down to about the right length, cleaned up the ends with a file, broke the edges, and the wrench was good to go. The pin length in the drawing was what I’d have used if I could have measured the holes before taking it apart.
The pins were actually on the long side of 60 mils, just an itsy too much to keep the wrench flat on the nut. The next picture shows some gouging on one of the holes, due entirely to not engaging the wrench quite enough at first.
I thought about putting flats on the wrench, but simply grabbed it in the bench vise, swallowed it with the flashlight, engaged pins with holes, leaned into the wrench, and unscrewed the ring. It took a lot more force to get those threads turning than I expected, but the ring eventually spun out easily. Right-hand threads, of course; obvious after the fact.
Before you can remove the switch assembly, you must pry off the rubber switch cover, stick that 2 mm hex wrench down the hole thus revealed, and unscrew the setscrew ‘way down inside there. That backs the setscrew out of a recess in the housing that makes electrical contact with the negative end of the bottom D cell. Do that before you remove the ring, lest you forget.
Surprisingly, the blue plastic switch housing seems to be slightly soluble in potassium hydroxide. Who knew?
With the switch assembly out, you (well, I) can proceed to beat the corroded cells out by chucking the housing in the lathe (it exactly seats on the three-jaw chuck’s front face!) and ramming a fat dowel up its snout with a two-pound hammer.
Yeah, genuine Ray-O-Vac Maximum D cells: they all leak if you leave ’em in there long enough. This flashlight worked fine, right up to the point where I checked inside to see how long the cells had been in there. Oops.
I’m thinking of rebuilding it with some killer LED clusters up front; scrap the reflector, rework the switch assembly. Certainly that’d have better heatsinking than those absurd 3-watt LED bulb-like thingies.
(*) Yes, Maglite has a lifetime replacement warranty that even covers death due to battery corrosion. Now, I ask you, what’s the fun in that?
You’ve probably seen “ultimate geek clocks” floating around on the web, which seem to be Nixie tubes, binary readouts, or analog clocks with lightly encoded markings.
Poseurs, all of them!
If you’re an alpha geek, this is how you tell time…
It’s a WWVB receiver wired up to a CR123A primary lithium cell. The time display is a single red LED, driven by a low-threshold FET. Yeah, you can package it up in a cute little box (which is the picture on hackaday.com), but this is the essence of the thing.
Over the course of a minute, the LED blinks out the hour, minute, year, day-of-year, Daylight Saving Time, leap year, leap second, and some other stuff in binary-coded decimal.
The key to the format is there and the bit format is straightforward:
You just watch the LED, catch the frame marker, decode BCD data on the fly, convert from UTC to local time, and that’s all there is to it.
Sheesh, it’s only one bit a second: anybody can handle that, right?
Truth to tell, I can hang on long enough to get the minute, but I taper off pretty quickly after that.
Tech detail…
Basically, you get the receiver and CR123 cell holder from DigiKey for maybe fifteen bucks. Wire up a FET (ZVNL110A or some such) to the receiver’s inverted-polarity output, so the LED is ON during the data bit’s active time (carrier drops 10 dB). I blobbed on a 300 ohm SMD resistor, so the total current is maybe 250 µA with the LED on. If you’re going crude, you can probably wire the LED & resistor directly to the receiver’s positive-polarity output.
A primary CR123A is good for 1500 mAh and the average current is maybe 150 µA, so the clock will run for nearly a year. The LED is pretty dim, but perfect for late-night viewing.
Reception is iffy during the day here in the Hudson Valley. At night it’s just fine. Interference from LCD panels with near-60-kHz refresh is a real problem, so it doesn’t play well near PCs.
I put the clock on a shelf where I can watch it when I wake up in the middle of the night: it knocks me out again pretty quickly.
In real life, I put this together to verify my WWVB simulator… but I might just box up a spare for the shelf, too.
This is depressing …
We got a boom box so Mom could have background music; the Olde Family Tube Radio was far beyond its Best Used By date.
Prompted by recent events around here, I checked it on a recent visit and, yup, more corrosion. In all fairness, the cells suggest “Best If Installed By Jan 99”, so they’re well past their date, too.
This used to be a whole lot less of a problem when flashlights and radios (without clocks!) were the only things using “dry cells”: when the battery went dead, the thing didn’t work and you replaced the cells.
Nowadays, we expect alkaline cells to supply keep-alive trickle current for memory backup; even after the cell corrodes, it still supplies that tiny current and we never notice what’s happening inside.
I’m beginning to loathe alkaline cells just like I loathe the small internal combustion engines in yard equipment.
Must be something going around…
The outdoor thermometer over my desk (which also displays UTC so I don’t have to reset the mumble clock twice a year) started blinking. That’s the usual sign of a dead battery and, yup, when I opened it up, that “leakproof” Eveready was pretty far gone.
Surprisingly, at least to me, the cell hovered around 1.1 V open-circuit and 800 mV under the meter’s “battery test” load. Given the amount of corrosion, I thought it would be flat dead.
The corrosion had crawled out of the compartment along the negative terminal and coated the entire metal tab with bluish-green crystals. Some protracted dabbing with vinegar, rinsing with wet cotton swabs, and drying put things pretty much back in order.
I usually scrawl the date on each cell when I install it, but either I didn’t do that here or the corrosion ate the ink. All I know is that it’s been up there for quite a few years; look at the discoloration where it faces the sun through the window!
The thing was a surplus freebie to begin with and has long since been fully depreciated…
Mary has been quilting up a storm lately and wanted a larger surface to handle a bed-sized quilt. A table in the basement was big enough, but she wanted a larger flat surface around the sewing machine adjacent to the table.
I converted the typing return (*) from her upstairs desk into a table, then cut a piece of aluminum-clad 1-inch foam insulation board to fit. It’s 4 feet long, a convenient length to cut from the 4×8-foot insulation board, and slightly narrower than the typing return. Cutting it required a long X-Acto knife blade, but a really sharp utility knife would work as well.
Some stainless-steel tape finished off the edges. The tape itself is lethally sharp-edged, but it’s perfectly harmless if you do a good job of smoothing it against the foam board…
A pair of closed-cell rigid foam blocks held one end of the board at the proper height around the sewing machine, while a pair of cutoffs from the wood pile were just the right thickness & length to extend under the other end. It turns out that precise height isn’t nearly as vital as we expected; close enough is fine.
I cannibalized a pair of table-saw feed roller stands for this project; they had just the right height adjustment and shape to support the typing return and the foam board.
The end result aligns the surface of the sewing machine with both the top of the table and the surface of the foam board. The quilt slides easily over the whole affair and doesn’t bunch up like it did before. Success!
(*) A “typing return” is the little table that sticks out from a desk, upon which you put a typewriter, back in the day when typewriters ruled the land. Nowadays, she uses it for her sewing machine, which normally lives at her desk, because there’s no practical way to type at right angles to one’s desk.
That’s the sort of item you can’t do web searches for, because all the terms are so heavily overloaded. Give it a try; you’ll find one or two useful hits. There’s a difference between syntax and semantics; we’re not in the semantic web yet by long yardage.
The Smell of Molten Projects in the Morning 