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.
If it used to work, it can work again
Having run out of dead CR123A lithium cells, I soldered up another holder to work through a handful of dead AA alkalines:
It’s amazing how much light comes from a single wearable LED powered by a dead battery in a dark room!
Stipulated: I should have 3D printed a CR123A cell adapter. Maybe next time …
Our long-suffering and much-repaired Kenmore clothes dryer didn’t shut off, with the heat on and the timer failing to advance from whatever position we set it to; the clothes were plenty dry and scorching hot. I tried “timed air dry” to eliminate the heater from the problem and found the timer still didn’t advance.
Referring to the wiring diagram may be of some help:
The timer motor is in the next-to-bottom ladder rung. Its BK terminal on the left connects to one side of the 240 VAC supply and the switch just to its right connects terminal TM to either the neutral AC line (thus unbalancing the 240 VAC line by a smidge) or to through a 4-ish kΩ resistor and the heater element (essentially zero, on this scale) to the other hot line; the resistor thus dropping 120-ish VAC.
The various switches around the timer collect nearly all the wiring in the dryer:
A closer look at the back, minus all the wiring:
The motor comes off easily enough, revealing the fact that it’s not just an ordinary (i.e., cheap & readily available) timer motor:
Hotwiring the motor through a widowmaker zip cord showed it worked just fine. For reference, the upper pinion rotates at about 45 sec/rev, the lower pinion takes maybe 1 hr/rev, both counterclockwise.
Reassembling and hotwiring the complete timer showed it worked just fine, too.
Poring over the wiring diagram suggested the power resistor might be open and, indeed, it was:
The raised zit near the front shouldn’t be there:
Apparently, the resistive element broke at that spot, burned through the thermoset plastic case, and failed safe.
Introducing it to Mr Disk Sander revealed a cavity below the zit, surrounded by the remains of the resistive element:
You can get a replacement resistor from the usual suspects for prices between $20 and $40, plus or minus shipping, but their pictures look a lot like an ordinary power resistor inside a length of heatshrink tubing, rather than the molded OEM part. I don’t put much stock in reviews & comments, although they seemed to suggest you get, indeed, an ordinary power resistor.
I didn’t have a 4.7 kΩ power resistor in my (diminished) collection, so I soldered a giant 1.5 kΩ cylindrical resistor in series with a small 3.5 kΩ sandbox, wrapped them up, and tucked them under the front panel’s ground wire:
A small box of resistors should arrive in the next month and I’ll re-do the repair with a bit more attention to permanency.
Once is happenstance, twice is coincidence:
Three times is enemy action, but we’re not there yet. I was willing to believe something I’d done had killed both of the radios, even though it seemed unlikely for them to last five years and fail almost simultaneously.
So I dismantled this one to see what’s inside. Pull off both knobs, remove the two screws at the bottom of the battery compartment, pry gently with a small screwdriver, and the whole PCB pulls out:
A bit more prying separates the big pieces:
Looking closely at the main PCB showed some problems I definitely didn’t cause:
Although it’s been riding around on my bike, the white blotches on the PCB came from inadequate flux removal after hand soldering.
A collection of images taken through the microscope reveals the problems:
I swabbed off the crud with denatured alcohol to no avail. The bottom side of the PCB has even more components and, I’m sure, even more crud, but I didn’t bother removing all the screws required to expose it, nor did I dismantle the other failed HT.
I doubt Wouxun’s QC improved over the last few years, which means the two replacement KG-UV3D radios I just bought are already on their last legs, despite my paying top dollar to the same reputable source that sold me the first pair.
We’ll be ready for new radios on new bikes by the time these fail.
The idea behind this gadget surfaced while I was looking for something else and, although the front panel makes my skin crawl, it’s just an adjustable DC power supply:
This is apparently the cheapnified version of the item appearing in the eBay listing:
Let’s say it has the potential to be a DC power supply, although we might quibble about the “Precision” part.
As delivered, it’s a deathtrap. Of course, it’s not UL listed and I didn’t expect it to be.
How many lethal problems do you see?
For starters, it has a three-wire AC line cord with the green-and-yellow conductor chopped off flush with the outer insulation inside the heatshrink tubing just behind the transformer:
The blue wire is AC neutral, but it really shouldn’t be connected to the finger-reachable outer fuse terminal.
The brown wire is AC line, which goes directly to one power switch terminal. In the event of a hot wiring fault, an unfused conductor touching the case will test the GFI you should have on your bench wiring.
The AC line cord uses some mysterious copper-colored metallic substance that’s about as stiff as music wire:
The strands cannot be twisted together like ordinary copper wire, although they can be soldered. They may be copper-plated aluminum, because a magnet ignores them.
After soldering the strands together, they snap when bent:
Generous strain relief is not just a good idea, it’s mandatory.
After some Quality Shop Time, the ground wire now connects to the case through the transformer’s rear mounting screw, the neutral AC wire connects to the transformer, the hot AC wire goes to the tip of the line fuse, and the fuse cap terminal goes to the switch:
I relocated the white LED to the middle of the meter, where it looks a bit less weird:
I have no idea what “Porket indicate” might mean. Perhaps “Precision indicator”?
The right 1/4 inch jack, labeled “Foot”, normally goes to a foot switch you don’t need for a bench power supply, so I converted a length of drill rod into a dummy plug to short the jack contacts:
The tip comes from a bit of lathe and file work and the white cap comes from a bag of wire shelf hardware.
A genuine hologram sticker (!) on the back panel proclaims “1.5 – 15 VDC 2 A”, which seemed optimistic. Some fiddling with power resistors suggests tattoo liners (I learned a new word!) don’t draw much current:
It can reach a bit over 18 V (pegging the meter) at lower current, so it’s Good Enough for small projects with un-fussy power requirements.
Caveat emptor, indeed.
When I wired up the MPCNC’s tool length probe, I planned to reinforce the wiring with a dab of epoxy. What I didn’t notice in my enthusiasm, alas, was the opening from the rear to the front in each pin slot:
Which let the epoxy flow completely through the connector:
So I cut the mess off and applied heatstink tubing on each wire, just like I should have in the first place.
Now you know the rest of the story …
I really dislike pin headers as cable connectors, but that’s what the Protoneer CNC board uses:
It’ll be Good Enough if I don’t do anything else particularly stupid.
After three years, the temple screw on Mary’s oldest and most-battered “reading” glasses worked loose. A dab of low-strength Loctite should hold it in place forever more:
That brass stake pin certainly adds a Steampunk flair to the proceedings …
A high energy collision / accident / mishap in front of Adams Fairacre Farms (a.k.a., the grocery store) demolished 20 feet of their dry laid stone wall along Rt 44, flattened several bushes, gouged trenches in the grass, and scattered plastic debris into the parking lot. The remains of a headlight eyebrow running light emerged from a snow pile:
From the back:
Contrary to what I expected, it has one white LED at each end of the chromed reflecting channel, topped with a shaped plastic lens collecting the light:
The LED PCBs are in series, which produced a backwards wire color code on one end:
The other end looked more reasonable:
The white SMD LEDs draw 300+ mA at 3.6 V, so they’re obviously depending on external current limiting provided by the regulator PCB, sporting a TLE4242 linear current regulator and a handful of passives:
AFAICT, they didn’t use the chip’s PWM control input or its LED failure status output.
Extracting the various PCBs from the wreckage and reconnecting the wires produced a satisfactory result:
The regulator limits the LED current to 120 mA at any input from a bit over 7 V to well past 12 V, with each LED dropping 3.0 V.
Dunno what I’ll use this junk for, but at least I know a bit more about eyebrow lights. The chip date codes suggest 2010 and 2012; perhaps linear regulators have become passe by now.
The Smell of Molten Projects in the Morning 