Keith's Electronics Blog

Today my boss mentioned a gas leak scare at his home (the service company found a slightly loose exterior fitting that happened to be near a vent in the side of the house), and I took the opportunity to ask whether he was still interested in a gas detector for the trailer at the farm.

Nay. They have a new trailer, and the old one leaks water (as well, presumably, as gas), and they don’t use the old one any more.

So I probably won’t go any further with my flammable gas detector project. But it was still enjoyable figuring out as much as I did.

This is the back side of an aging SGI cluster that we host in our data center at work for the high-performance computing group.

For a sense of scale, this is a 23″ rack, and that’s a grey CAT5 patch cord stretched down the left portion. The white cables are close to an inch thick.

The fat cables with the beautiful, tough braided jackets seem almost synthetic-organic to me. They always make me think of Bishop after he’s been torn in half by the alien.

Back in September, I was working on something and got tired of fishing the right pliers out of the pile on my workbench — flat nose for straightening bent component leads, needlenose for shaping leads, wire strippers, cutters, etc. So I made a dirt simple tool holder.

It’s just a 2×4 block with 1/2″ holes drilled in at regular intervals (Forstner bit makes nice flat bottoms), spaced for the at-rest width of my pliers handles and as close together as is easy to reach in and pluck out the one I want.

It’s cleared up a significant mess on my bench, and it’s so fast and easy to access that I tend to put tools back between each use rather than each session. It’s ugly as can be, but I wanted to know how well it would work before making a pretty one. Problem is, now that enough time has passed for me to know it works well, I’ve lost some of my motivation to make it pretty.

No LEDs. Nothing to do with electronics at all. Just wood.

This is for Lawrence’s wife. I made a cherry rolling pin for Cort a couple of years ago; and when I described it to Gail, she was very interested. She said she preferred hers with a maple stripe in it, and I finished it yesterday.

If you think cherry wood should be redder because that’s what you’ve seen in furniture stores . . . let me tell you that what you’ve seen bears as much resemblance to real cherry wood as a maraschino does to a real cherry: They’re both full of dye and completely unnatural. With exposure to light, this cherry will darken to a lovely, deeper brown shade within a few weeks, further highlighting the difference between the cherry and the maple. No bright red about it.

Glue-Up

I started a while back by gluing up the block. I have a large supply of short, narrow cherry and maple pieces from a long-lost high school classmate (who turns out to be Ron-the-TV-guy’s brother-in-law; lots of connections in a small town) who worked at a cabinet shop and got all the rails and stiles when they closed, then got tired of storing them and cheerfully sold them to me for a song.

The cherry and maple were both flat-sawn, so they showed broad grain patterns on their faces and tight (boring) grain on the edges. I didn’t like the way the maple edge grain looked sandwiched between the cherry; so I sliced the maple into thirds, rotated the slender sticks 90°, edge-glued them back together to put face grain on the outside, and planed it flat again. Then I glued up the block — from the end, you can still see three maple squares sandwiched between two solid cherry pieces. And (like many of my projects), it then sat a while.

Turning

This weekend I returned home from a conference, played a rock concert to benefit the high school robotics team, finished my obligations to the world, and finally had some time to myself. Electronics is great, but it’s awfully cerebral; and yesterday afternoon it was nice to give the mind a rest and do something visceral.

I’m not a particularly good wood turner, and my results are due entirely to patience rather than skill, but I enjoy the process. I finally got to use the new roughing gouge that I got last Christmas, and boy, was it a treat. The little picture doesn’t do it justice — the gouge is huge, with a 2″ scoop shown in the picture. It cleaned up the blank quickly and with minimal chatter, and I was actually able to use it for the whole turning. I should probably have used a sharper tool at the end, because I had some slightly torn grain on the figured portions of the cherry, but it sanded out nicely.

Finishing

After sanding down to 1600 grit, I cut the ends free, removed it from the lathe, and started coating it with food-safe oil. I kept coming back to it all evening to add more oil, and I think I’ve put on about as much as it’s going to take right now.

This morning, I got to re-experience one of the pleasures of woodworking with which non-woodworkers are probably unfamiliar: the wonderful smell of fresh-cut cherry still permeating my garage as I left for work. And yes, the smell of cut cherry is distinctly different than the smell of fresh-cut maple, oak, and walnut, and completely different than the smell of pitch from softwood (pine, fir).

Schedule permitting, I’ll be able to take it to Gail tomorrow night.

I’ve been swamped at work lately and it’s been spilling over to home, so I haven’t done much with CNC (or other projects) for several weeks. But I bought this Diebold Transaction Number Generator on eBay for $10, out of sheer curiosity to see what was inside. I have no idea what it’s supposed to do, and can’t find any information about it online, but industrial security and crypto equipment fascinates me.

My goal was not to reverse-engineer the whole thing, but to get a rough idea how it worked and look over the interesting bits. I think I’m not alone in that when I get a new electronic device, I like to look inside to see what they used to make it work.

The back panel has video in, video out, and screw-clamp wire connectors:

(Nearly) all of my blog pictures are links to the full-resolution version, so you can click to zoom on this, but I’ll save you the trouble. Here are the labels over all the wire connectors:

Okay, so I get what “AC” is, and “EN” might be enable, and “ALM” looks a lot like alarm, and there sure are a lot of ground lines, but the rest of the connections pretty much escape me. The DA-GND-ST triples make me think of some kind of balanced transmission, but wouldn’t a balanced transmission line have a shielded connector instead of a terminal strip?

And why video? Does this box generate transaction numbers and superimpose them on a video feed? I dunno. But it sure is pretty inside. Let’s have a look.

I really dig the red circuit boards (red solder mask). Green solder mask predominates these days, and I’ve seen blue and other colors, but I’m pretty partial to red.

The small board in the upper left appears to be mainly the power supply. The medium board in the upper right interfaces to the video feed, and the large board in the middle has some interesting digital logic on it.

Power Supply Board

Here’s the power supply board. The bridge rectifier, CR201, connects to the main AC terminals on the back panel and feeds the LM317 regulator U201. U201 uses R202 (150Ω) and R201 (1.3kΩ) to set its regulated voltage.

The LM317 maintains a nominal 1.25V voltage across its reference terminals, across which R₁ thus sets the reference current I₁. The reference current also flows through R₂, determining the regulated voltage. From the LM317 datasheet,

V_OUT = V_REF (1 + R₂ / R₁) + I₁ R₂

where

I₁ = V_REF / R₁ = 1.25V / 150Ω ≈ 8.33mA

So

V_OUT = 1.25V (1 + 1.3kΩ / 150Ω) + 8.33mA 1.3kΩ

    ≈ 1.25V * 1.87 + 10.8V ≈ 13.1V

With a bridge rectifier and capacitor-input filter, output voltage is approximately equal to peak input voltage, which is V_AC / √2, so the box is expecting an AC supply voltage of not less than 18.6V; compensating for diode voltage drops gives about 20VAC. Knowing that, I could (if I cared enough) now power it up.

Note the bridge rectifier made of discrete diodes CR202-207 connected to external pins T2, T2F, and ground. CR206 and 207 double in two bridge circuits with CR202-203 and CR204-205, making me suspect that T2 and T2F are not meant to be used simultaneously, but as alternates. On the other hand, T2 heads off to the OUT1 (common) pin of the MC14066 quad analog mux chip. Kind of looks like it’s being switched to one of two different destinations.

Video Board

That’s about as much effort as I care to put into the power board, so let’s move on to the video board.

The four chips in the lower left are MC14538 oscillators, there’s another LM317 voltage regulator in the lower center, and the board has a mechanical relay up at the top. But the really interesting part is the two large chips in the lower right: MM58146 on-screen display drivers for television channel indicators. So this box is built to superimpose some kind of transaction numbers onto video. That’s interesting to know.

Logic Board

On to the logic board in the center, which I’ve rotated here to make the part numbers easier to read. The four chips U5-8 near the empty sockets on the right are MC14495 hex to seven-segment decoders. Whaaaa? I assume the on-screen display driver chips do their own character generation, so why decode binary to seven-segment? Maybe the four empty sockets U1-4 are for DIP seven-segment displays, to show the transaction numbers internally during testing or troubleshooting.

The four large chips in the middle, U11-12 and U14-15, are even more interesting. They’re MC14034 universal bus registers, permitting bidirectional transfer of data between two parallel buses, serial-to-parallel, and parallel-to-serial conversion.

The datasheet tantalizingly offers, “Other useful applications for this device include pseudo–random code generation,” but says nothing further about it. My guess is you’d wire them to shift and output bits, possibly with external inverters to shuffle things around at different stages, to make a hardware implementation of a PRNG algorithm.

You might even be able to wire them up to make a digital version of something like the rotors from a German Enigma cipher machine — but without spending a lot of time tracing pins on the bottom side of the board, I don’t know whether that’s how they’re being used here, or merely for more mundane purposes. It’s not quite worth it to me to figure out exactly how the MC14034s are used, so I’m pretty well done with my exploration.

Conclusions

I didn’t figure out a whole lot about the device, but it’s pretty clear that it takes a video input and superimposes numbers onto it. (One thing I wondered initially was whether it might be using video as a source of random noise to generate transaction numbers.) And it has some interesting bits inside, with logic chips I might be able to reuse for something else. That’s enough to satisfy my curiosity for now.

A reader named Dave posted a very well-informed comment that what I’ve been referring to as an anisotropic strip (connecting an LCD to a PCB) is more specifically called an elastomeric connector. I thought it was interesting enough to be worth pulling up into a new post:

The anisotropic strip on the LCD is technically known as an elastromeric connector, but usually referred to as a zebra strip. If you examine it closely under a strong magnifying glass or a low power microscope, you’ll see alternating bands of conductive and insulating material, usually something like graphite loaded rubber for the conductive material and unloaded rubber for the insulating material. The graphite loaded rubber gives that portion of the strip a darker colour. Thus, you have alternating light and dark bands, similar to what a zebra looks like.

Dave

Thanks for the detail!

Wednesday afternoon:

I never expected this kind of web traffic to my blog, and my current image hosting solution isn’t up to the job.

I’m mirroring right now; images should be back up within a couple of hours.

Wednesday evening:

Images for recent posts should be back up. The remaining images from 2007 are almost done mirroring to the new server, and earlier images will get mirrored if there’s still time while traffic load is still high.

Thursday morning:

Everything seems to be mirrored and working!