Keith's Electronics Blog

A Theremin is an electronic instrument played by waving your hands in the air near a couple of antennae. It has matched pairs of high-frequency oscillators, and the change in capacitance due to the proximity of your hand changes the frequency one oscillator in the pair, causing “beats” (heterodyning) between the pair that form audible wavelengths. The Theremin was invented early in the 20th century by Leon Theremin, a Russian physicist. It’s the spooky sound you hear in the Beach Boy’s “Good Vibrations”–although that was actually recorded with an electro-theremin, a different instrument designed to mimic the Theremin’s sound.

I don’t know when I first became aware of Theremins, but I know I’ve wanted one at least since February of 1996, when Electronics Now published “Build this Theremin,” by John Simonton of PAiA Corporation. The full PCB layout and parts were published in the magazine, and PAiA has also sold their “Theremax” kit continuously since then.

And I finally ordered one.

Two, actually; my friend Cort wants one as well.

They arrived about a week later.

Board Assembly

I bought the kit with the PCB and all components, and the separate partial case kit with the antennae and a nicely screened front panel. The full case kit was another $60, and I figured I could make a case that I’d be happy with for less than that. I think it says a lot about PAiA and their understanding of the hobbyist marketplace that they offer so many options, instead of forcing everyone into the same cookie cutter package.

The first thing I did was sort out all of the components. There are dozens of resistors and capacitors, a handful of transistors, four tunable inductors for the oscillators, pots and jacks for the panel (included in the parts kit, even if you don’t buy the panel), and miscellaneous bits like LEDs and ICs. They even threw in hookup wire to connect the PCB to the control panel, and a gator jumper for bypassing the volume control during one stage of hookup–literally every part you need to get up and going, except solder.

Next, I took a look at the PCB. It’s a nicely-made single-sided board, with silk-screening on the top side for component identification only. I have only two small complaints:

• The identifying marks for resistors and capacitors are under the component, so they’re no longer visible once the component is installed. I know there’s not a lot of extra room on the board for labels outside of the component footprint, but having to refer to a separate parts placement diagram really makes troubleshooting more difficult.
• The resistors and capacitors are roughly numbered from upper left to lower right; but some of the numbers stray across the board, making it hard to locate the correct part. Maybe I’m too picky, but I really like an orderly numbering system during the assembly and troubleshooting stages.

The assembly instructions give a reasonable order for soldering in the components, which I mostly followed–jumper wires and resistors first,

then diodes. Note the blue diodes near the center of the board–germanium, with a lower voltage drop for increased sensitivity. I’ve read a nice description of how that aids the Theremin circuit, although I don’t recall it at the moment.

Here I deviated slightly from the instructions’ assembly order, and installed the ICs and inductors next. Because I don’t have a cool PCB holder, I solder with the board upside down on my workbench; and because of that, I prefer to install the shortest components first and work up to the tallest,

which means doing the inductors before the ceramic capacitors,

and the electrolytic capacitors last.

The transistors are actually shorter than the electrolytics, but they cling to the board pretty well, thanks to the hole spacing being considerably wider than their lead spacing.

Control Panel

PAiA’s professionalism really shines here, and my photo doesn’t do it justice. This is a very nicely lettered control panel, as good as those on most of the equipment I own. I like the retro feel of the white paint on black brushed aluminum, rather than some glitzy decals on colored plastic that you might see these days.

I installed all of the potentiometers and jacks, twisting them as needed to try to get the hex nuts on the front to line up evenly. (I like to use the phrase detail-oriented.)

The instructions say to run bare wire from pot to pot for the ground connection; but I’m not wild about bare wire (especially when it has to loop and could be snagged and pulled out to contact something else), so I put some clear heatshrink over it for sleeving. I also built up the LED assemblies–replacing the provided dark red LEDs with my own green LED for the power light, and a matching red LED for the gate trigger indicator.

Then I soldered all the fly wires to the PCB. The assembly guide gave very specific lengths; but I think things must have shifted since that was first printed, because the wires had completely different amounts of slack once connected. I shortened some wires and should have lengthened others, and made notes about all the changes I made and further changes I should have made. Here are my notes about the colors I used and the lengths I’d recommend:

The last thing I did in that batch of soldering was run signal wires within the back of the panel. The instructions gave this step at the same time as the ground wiring, but I wanted to wait until I had the fly wires chosen so I could maintain the same color wire everywhere a signal traveled.

Next Steps

By now, I’ve connected the fly wires to the control panel, built a case, and semi-mounted the kit in the case. I’ll post those pictures soon.

Last fall, my friend Jeremy and I were killing time at Best Buy when we ran across this blue gel keyboard wrist rest. He was looking for a wrist rest anyway; and since we’re obsessed with blue LEDs, I offered that I could rig it up to glow. He loved the idea, and his wrist rest has been sitting at my house begging for attention ever since.

I’ve been tinkering around with ideas for a month or two, but finally yesterday I cleared some space on my workbench and got at it. There were four main components to the project:

• The translucent blue wrist rest, which I hoped would have enough internal reflection to glow nicely
• Blue LEDs, which I harvested from a string of Phillips Christmas lights
• A USB cable, to supply power to the LEDs (Jeremy preferred USB so the lights would power on and off with the computer)
• Some means of holding the LEDs in place

Glow Test

I’d been thinking I’d drill some holes into the wrist rest to embed the LEDs, but I wasn’t sure quite what the effect would be. I started testing by putting a couple of LEDs in my LED tester and shining them into the gel pad from different locations and angles, with disappointing results–they didn’t make any part of the pad glow like the edges of plexiglas (high internal reflections from the parallel surfaces) or milky plastic (high diffusion of the light beam).

So they kinda shine through instead of making the whole thing glow, and I just have to deal with that. On the other hand, that meant I didn’t need to bother drilling holes into the gel–I could butt the LEDs up against it and they’d work just as well. So I started thinking about etching a long, skinny circuit board to go across the back . . .

USB Cable

And meanwhile, tried to figure out where to get a USB cable to sacrifice. USB cables cost more than it seems like they ought to, so I was asking around for a dead USB mouse that I could salvage. I really wanted a black cable, to disappear on a dark desktop; and I really wanted a skinny and supple cable, to route well.

Then a couple of guys at work gave me a USB hot plate, I kid you not. It was from a trade show and was supposed to keep a coffee mug warm–which, no surprise, didn’t work out so well, given the limited current USB can supply. It had a skinny cable with silver foil inside–not black, but it still sorta fit with the theme. Perfect. Well, good enough.

Light Bar

Since it would have taken me a while to rig up a PC board for the LEDs, I was thinking about other ways to hold them in place and wire them together. I have some black plastic U-channel pieces about a foot long, which were supposed to be magazine or books supports for some wire cube shelves I bought, but which I’ve never used. Yesterday I cut the brackets off one, drilled holes in the back to poke leads through, and stuck four LEDs into it.

Several of the LEDs in the Christmas tree lights had 250Ω resistors already soldered to one leg, and that turned out to be a nice current-limiting value for a 5V supply, so I just left them in place. I daisy-chain soldered wires to the leads, connecting the LED/resistor assemblies in parallel, and attached the USB cable to one end. [Side note: In the USB cable, the black wire was positive and the white wire was ground. Good thing I checked first.]

Yup, I was soldering wires onto leads that were poking through (and touching) plastic. Surprisingly, I didn’t have much trouble with it–the wires were pretty clean, so they all soldered quickly, and apparently didn’t heat up much. I didn’t notice any melting plastic at all.

I used some big heatshrink to anchor the end of the USB cable in place and provide some strain relief (lower right end of the above pic), and the light bar was ready to go.

The LEDs are on the inside of the bar pointing out. That does a reasonable job of shielding the (annoyingly bright) light sources from direct line of sight, and plenty of light still makes it into the gel pad when the light bar is placed up against it.

This isn’t just a trick of the camera–it really does look like four beams of light aimed through the wrist rest. It’s not quite what I was aiming for, but isn’t a bad effect in itself.

The digital pic exaggerates the effect, but the four LEDs really are two different colors–two slightly more aqua, two slightly more violet. They’re all from the same string of LEDs, and I’d never noticed the effect before–I assume they’re from different manufacturing batches, with not particularly high quality control.

Credits

It wasn’t much of a project, but (1) it provided motivation to get some things cleaned off of my workbench; (2) it was itself part of the clutter on my workbench; and (3) it helped build momentum to get me back into tinkering. I’ve already started on my next project, which should appear here in a few more days.

Thanks to Jeremy Burkey for the photos. I forgot to take pics as I was working on it; and his camera did a much better job with the dark shot than mine would have anyway.

Based on a reader request, I’ve posted my EAGLE to PADS netlist conversion script, for drawing circuits in EAGLE and exporting them to produce in FreePCB (if the board is too big to draw in free EAGLE).

If you have questions, suggestions, or package translation contributions, please submit them as a comment attached to this post. If you’re viewing this post alone on a page, the comment box should be visible immediately below. If you’re viewing this inline in my blog, you should be able to click the “Comments” or “No Comments” link below to get to the comment box.

A logic analyzer is a device somewhat like an oscilloscope, in that it displays the waveforms of electronic signals on its screen. It differs from a traditional scope in that it captures data in memory for longer study of ephemeral events rather than only displaying what’s happening in realtime (modern capture scopes do this, too); it typically has many channels (8, 16, or more plus clock and other “meta” inputs) rather than one, two, or four; and it can be configured for complex trigger events rather than only the level of a particular signal. In short, a logic analyzer does for digital system troubleshooting what a scope does for analog signal work.

I’ve never had a logic analyzer, and last spring when I started working more with the LogoChip and other digital parts, I started hankering for one. I found and bought this Tektronix on eBay for only $30 plus shipping:

It has text-based menus, online help, and a charming monochrome green screen. It’s built almost like general-purpose 6502 luggable microcomputer, with a big stack of signal handling boards in it.

The only catch was, it didn’t come with any probes. I was hoping at first that all I needed was ribbon cables to fan out from the dual-row headers on the front panel, but I couldn’t find pinouts anywhere (including online). I emailed Tek’s tech support asking whether they might still have pinouts available that I could download (since they no longer sell or support this analyzer and wouldn’t be losing any money on me), and I got an incredibly helpful phone call back from them.

Their tech support representative indicated that the probes have active CMOS circuitry in them, so you need more than just pinouts; and he gave me the model numbers of three different probes that should work with that analyzer. I saved an eBay searched and waited, until a couple of weeks ago when one of them popped up and I bought it for about $25.

There’s obviously a fair bit of circuitry in there that I wouldn’t expect to be able to duplicate, so I’m glad to have bought it. Now I still need cables to go from the probe module to the circuit being analyzed. And after I received the probe module, the seller emailed me to say that in the same box as the module, he also had these connectors that went with it; did I want them for another $25?

Okay, if he’d included them in the auction with the module they go to, he probably wouldn’t have got any higher bids. It’s his job to maximize his profit, so I suppose he’s doing what he ought to. But I feel like a guy who just negotiated to buy an antique auto out of a farmer’s field, only to be asked whether I’d also like to buy the engine that goes with it that I have over here in the barn. Yes, sure I’d like the wires that go with the probe module you already sold me; and at the same time, no, I really don’t feel like sending you any more money for something that really belongs with what I bought the first time.

Furthermore, most of the time I’ll be using the analyzer, it’ll be with circuits on a breadboard, where I’m going to have to make my own cables terminating in pins anyway. These probe tips have really cute grippers on them that are perfect for grabbing onto classic DIP leads, but useless on a breadboard.

But I was still wavering, since I don’t know how to describe these tips well enough to save an eBay search that would pick them up, making this a rare opportunity to get something I might eventually want.

I called Cort last night to ask his opinion, and the moment he said “troubleshooting your arcade games,” I knew what the answer was.

I just sent the PayPal.

Independent of the LED clock project, I’ve been thinking of building myself a new alarm clock. . . . More on that in a while, after I receive some prototyping materials I’ve ordered.

But tonight, watching The Animatrix, I got the wacky idea of making an alarm clock that uses a rotary phone dial for input to set the time. Pick up the receiver, dial the four digits for which you want to set the alarm, and bingo! Good to go.

It kind of escalated from there, to a fairly-well fleshed out rotary telephone alarm clock. Consider:

• Rotary dialing to set time and alarm.
• Mechanical bell ringer for alarm–you won’t oversleep that baby!
• Pick up the handset for alarm functions. Take the phone offhook and dial a number to set the alarm. Alarm rings, lift the handset to silence the alarm.

Primo! Of course, I still have to work out a few details:

• How do you tell the phone that you’re setting the time? It could accept dialing while the handset is on hook; but that just begs for people who walk up and twiddle with it to be randomly resetting the time. Maybe indicate it with a special dialing sequence like starting with 0 or 411?
• How do you display the time? Gotta have a four-digit display on there somewhere, and that’s totally gonna mess with the aesthetic. Oh well. Probably have to mount it through the lower front of the phone.
• How do you snooze versus shutting off the alarm for the day? Leave the handset offhook versus putting it back on?

A rotary multi-line phone with lighted buttons across the bottom could be fun, as would a speech synthesizer to read the numbers to you in the handset as you dial.

Spark Fun has done a rotary cell phone, and they wrote up the processes of interfacing to the rotary dial and switchhook and driving the ringer. Of course I’d write to ask permission first, but I’m sure I could use their work to build the pertinent sections of my clock.

Now I just have to get my hands on a couple of old rotary phones . . .

My uncle Adolf is probably most directly responsible for my interest in electronics. When I was growing up, he and Aunt Wanda ran A&W Electronics, retailing and repairing TVs and other equipment. This was back in the days of chassis construction, before circuit boards, when everything was built onto a metal chassis with insulating standoffs, vacuum tubes, lots of wire, and lots of solder.

When they received TVs that were irreparable, he’d often remove the dangerous parts (picture tube, flyback transformer, and huge capacitors) and give the remaining pieces to my brother and me to take apart. Which we did with great zeal, and a couple of wirecutters. I still have boxes of components that we salvaged out of those TVs; and man, they don’t make resistors and capacitors like they used to. And that’s mostly a good thing.

So when they had a household auction recently, in preparation for moving to a retirement community, I was pleased and proud to purchase two boxes of soldering equipment.

I could see the Weller/Ungar pencil irons, and knew that any one of them was worth the $16 I paid for the box. It wasn’t until I got home that I could really see what all I got:

There’s just an amazing amount of stuff crammed into the boxes. These irons have replaceable, screw-in heating elements, and the boxes have several spares of different wattages. One or two of the elements have replaceable tips like we’re used to using these days, and there are plenty of spare tips, too. Eventually, I plan to go through everything, sort out what goes with what, and do a little cleaning on the pieces that do fit together–replace brittle power cords, clean iron handles, polish heating elements, etc.

So?

I desolder a lot of components from dead electronic equipment that people give me. I really prefer older stuff with through-hole components, because SMT parts are a pain to identify and sort, and through-hole stuff is easy to prototype and breadboard with. The fastest method I’ve come up with is to pry up crimped leads with a $2 chisel, heat the solder side with a heat gun, and pull the components out with a chip puller or needlenose and drop them into a tub.

Generally.

Except (he finally gets to the point) for the TO-220 MOSFETs on the printer boards. For whatever reason, their through-holes were plated only barely larger than their leads, and I was having a dickens of a time desoldering them with the heat gun. The board was scorching, I was breathing really noxious fumes (and I like the smell of solder flux, but not so much burning fiberglass epoxy), I was bubbling and flexing the board as I pulled on the FETs’ tabs with a pliers, and still no joy. I ended up pulling the entire plated through-holes clean out of the board (um, maybe “clean” isn’t the best word to use here) and sliding them off the transistors afterward with my soldering iron.

And the Iron?

I’ve previously desoldered with soldering irons and guns, and under some circumstances, direct heat can be very effective at removing stubborn leads. So a couple of nights ago, I got around to digging through the auction boxes to see whether I could find a wide tip. I was looking for something that would hit all three transistor leads at once, and I found exactly that–a gimongous .3″ chisel-tip 40W heating element, still sealed in its original bubble package.

I opened the package, cleaned the oxidation off the tip, screwed it into one of the handles, and plugged it in. When it started getting up to temperature, I smiled in appreciation at the ceremonial first tinning of the new tip. I grinned at its extreme shininess as I wiped off the solder on the damp sponge. I dropped my jaw in amazement as the transistor slipped right out of the board in about a second of applying the tip to the leads. I desoldered the remaining thirty or so transistors in a matter of brief minutes.

The tip kept heating up the whole time. As I was finishing, I would occasionally cast shadows on the iron and see that the tip was now red-hot. I’d wipe it on the damp sponge and see little embers rise into the air as it tore off and burned rough parts of the sponge surface. I was rather unnerved to see this from an element rated at 40W.

But boy, did it desolder them transistors! Yee-haw!

Obligatory FPS Reference

I’m somehow reminded of the differing effectiveness of various weapons against various monsters in Doom. When you run into a cacodemon, you better make sure you have the nailgun handy–even though it’s utterly useless against several other species.

Okay, I’m not much for filling blogs with links to blogs that link to other blogs ad infinitum . . . but this is just too cool. Ted Johnson built a clock out of acrylic sheets with runes scratched into them, edge-lit with LEDs, and stacked to show multiple different characters per position.

Ted Johnson’s Project