i should offload some random old pcb photos

zero FPS and zero RGB = not gamer approved

@mhoye and hooking your website into the modern advertising industry immediately rules out that user. Their system can't handle all the popups and animations and JS.

Are you sure you want
computer now?

Station construction is all complete but for a couple of small details. I do then need to add posters, platform furniture, and people - as a minimum - and finish programming the signals. Then I can think about ballast!

#ModelRailway #Metcalfe #StationBuildings #Progress

happy midnight to all who celebrate

-carrie

i know im not the only one
-sam smith

After fixing that solder joint and reassembling everything with a little glue, it's good to go.

9/

I'm pretty sure this failed because the center board section closest to the battery is not attached solidly to anything; it's entirely supported by flex. It floats freely between the plastic posts that the upper stacked section is glued atop. (See the gap between it and the battery.)

When dropped, it probably oscillates up and down and stresses the soldered battery-tab connections. (Especially if they're cold joints :)

8/

The battery terminals are soldered to a flex section that descends from the center rigid section. They've put this row of three vias (on the right side of the photo, center one filled with solder) to try to accommodate wherever the battery's positive tab is.

Fortunately, the problem was simply that this solder connection broke!

You can also see the two pass-through charging studs where the case's pogo pins make contact on the outside of the enclosure. (They're not directly across the battery; the charging supply has its own trace on the flex.)

7/

The third test pad from the left is ground, and I found an always-on ~3.55V (iirc) on this via... except not always: the problem seems to be an intermittent battery connection.

6/

The outward-facing side of the PIC board has the button. Is that Rev Q?

5/

Opening the flex like a book, we are greeted by a PIC16F18326. This will be handling the on/off/mode/low-battery functionality and generating the corresponding beeps. The PIC also changes the attenuation behavior on the low/high mode settings, so there would certainly be signals fed from here into the audio circuitry.

The top semicircular section with test pads seems to have a reverse-mounted microphone, with that foam gasket to bridge the gap to the other half. The two tiny red and green wires disappear from the left side to the earpiece below.

The bottom perpendicular board section is blobbed with epoxy, potentially because it's near an exposed battery terminal when assembled. (Or maybe for a more interesting reason. I did not currently feel like destroying this unit to learn what's underneath, but that could be done sometime. I suppose the flex could also be probed :)

Two thin rods seem to be inserted to locate the enclosure halves during assembly. (Maybe the bottom one was separating the perpendicular board section and that metal plate?)

4/

The upper half of the enclosure just contains the rubber pushbutton surface and two microphone ports at the top.

On the lower half is a rigid-flex assembly with four rigid sections: a semicircular one near the earpiece, two stacked sections above the battery, and an extra perpendicular section at the bottom. The top stacked section is glued at its corners to four plastic posts.

Against the bottom perpendicular section, a metal plate also sits in the enclosure to magnetically hold the earplug in the charging case. It's odd that this rigid section has exposed vias.

3/

After carefully cutting/prying around the perimeter, the two halves come apart. The electronics are affixed to the lower half, with the yellow lithium-ion cell at the bottom.

2/

I have some Etymotic "Music Pro Elite" earplugs, which are a pretty cool idea: essentially a compressor between a microphone on the outside and an earpiece driver on the inside, so that loud sounds can be dynamically attenuated.

I quite liked them until I dropped one again recently (apparently one too many times) and it would no longer turn on or charge. So of course I had to pause my day to tear apart this tiny thing.

1/

I was looking through the symbol footprints in KiCad and discovered a Blåhaj.

one step closer to forming The Semiconductor Company

https://www.ti.com/about-ti/newsroom/news-releases/2026/2026-02-04-texas-instruments-to-acquire-silicon-labs.html

interestingly this means TI will be competing with itself in Wi-Fi/BT: it'll have both its existing SimpleLink line (which traces back to its acquisition of Chipcon) and now the Redpine Signals stuff which comes with SiLabs (RS91x/SiWx917).

@grumble209 The bug-finding rate was remarkably linear on a semi-log graph: for every 10 times more tests we ran, we found ~5 more bugs, across seven orders of magnitude. I wasn’t allowed to say so in the 1990s, but I am allowed to now, that the test sequences were interestingly more sophisticated than just pseudo-random. I’m currently working on a paper that will explain the technique. One of the surprising conclusions in the paper is that the test generator, though it was vastly better than what we’d used before, must not have correctly implemented the intended technique—if it had, it would have found bugs even faster than it did, and it’s possible that when verifying a late-1970s processor, it *would* have found all the bugs. But on designs from ten years later, no way.

who up single-stepping they windbg
@who_up_x_they_y