This is my Arizona test bench, where I put a few of my more recent devices together. What's shown here is a PC-board lashup of a 2-stage CW transmitter, with a 6AG7 crystal oscillator driving a 6DQ6 sweep tube final. I got almost 18 watts output on 80 meters, which is more than I expected from that tube, admitting that I drove it pretty hard. The circuit was of my own devising, but it was based on several of the classic 6AG7-6L6 circuits that were handed around in the 50's. B+ was 450V, from the home-brew power supply to the left of the lashup. I hadn't really expected to build a finished transmitter—I was just trying to see what I could coax from the 6DQ6.

Here's a better view of the main tube-project power supply that I use on the bench, after I had finished it. I bought it at a hamfest for $5, apparently an abandoned project with only the power transformer, the two chokes, and a socket for a 5U4 on it. It hadn't even been wired, and I was really after the chokes. However, the galvanized iron chassis was very well done, so I decided to finish wiring it, add a few gas regulator tubes, and bring everything out to the front panel. Two 0D3 regulator tubes in series give me 300V regulated, with a tap in the middle for 150V. An 0C3 gives me an additional 75V regulated output for working on battery tube (3V4 etc.) designs. The two big chokes and a fat NOS filter cap make the DC output extremely smooth—and inrush current isn't an issue with tube rectifiers. The schematic as I built it is here.

Here's a better view of the 6DQ6 lashup. I breadboard tube circuits the same way I breadboard solid-state: On pieces of double-sided PC board. Parts are soldered either to the ground plane or to islands I isolate using a Dremel tool with a 1/8" round burr. I mount tube sockets on 2 1/2" squares of PC stock which I mark into 8 equal segments and then divide the segments with the burr. The socket module then mounts to the main board with a screw at its center, on a 1/8" spacer. I have a selection of socket modules made up for most common tube bases, and I can swap them out easily by unsoldering any parts and unfastening the central screw. Works great!

Here are two of the socket modules as used in the above lashup. They were easy to make, since in both cases (octal and 7-pin miniture) the pins are on 8 equally spaced sectors. (The 7-pin socket has a pin-sized gap to bring it up to 8.) It's a little trickier laying out the sectors for 9-pin miniature or 12-pin Compactron, since you have to divide a circle into 10 and 13 equal parts, respectively.

In the fall of 2005 I decided I wanted to (finally!) build an all-tube stereo amp so that I can MP3s down in my workshop. I had had the basic circuit in my library for many years, in the GE Hobby Manual, Second Edition (1965) which also contained the schematic for the 1-Compactron shortwave receiver (see the photo later in this page) I had built in seventh grade. I did some fair amount of redesign on the circuit, adding a balance control and tweaking the tone control, and establishing a single-point chassis ground to eliminate ground loops. The circuit as I built it is here. It's not as "hi-fi" as true tube audio fanatics would demand, but it wasn't difficult to build and didn't cost me a fortune. Certainly it's good enough to play that scratchy rock'n'roll while I'm gouging out circuit board prototypes with my Dremel tool.

The stereo amp circuit is fairly simple and wiring it was not brutally difficult. One thing I discovered is that "simplifying" a circuit by using multisection 12-pin tubes complicates wiring by making a lot of components connect in a very small area of the chassis. I had to pay close attention to the order that I was soldering in components so as not to block easy access to adjacent socket pins. In the photo above, the unit is almost finished. The only part remaining to be wired was the tone control pot, abobe left.

Things got a little dense behind the front panel as construction drew to a close. Here's a closeup of the tone control wiring, which gave me a lot of trouble due to my use of RG-174 coax as miniature audio cable inside the box. The insulation melted very easily, and at one joint the center conductor shorted to ground after its insulation got a little too soft. Overall, it was an excellent project, but before you attempt it you should have considerable experience with tight chassis wiring. There was a lot of very touchy needle-nose pliers work in there!

This is a PC board lashup of a two-tube receiver that I was designing some time in 2004 and set aside incomplete—I wasn't satisfied with the sensitivity, and I still intuit that there's something in the circuit I'm just not seeing.

Another thing I've tried is to cut the tube socket pads right into the main PC board with a Dremel tool. This actually works quite well, though it's a little more work with the Dremel and not quite as versatile as using removable socket modules. Shown here is what I hope will eventually become a 3-tube FM—not AM—receiver. I'm cribbing from the sound circuits of cheap b/w TV receivers from the early 60s, working at a 4.5 MHz IF and using a 6BN6 gated beam tube for the FM detector. Amazingly, by changing the bias voltage on the 6BN6, the detector can be made to detect AM as well as FM. (The trimpot at lower right adjusts the bias.) The IF and detector work fine, and I'm still designing a converter stage to bring 100 MHz broadcast FM down to 4.5 MHz. Images may be a problem, but this radio is a sort of a stunt, and not intended to be stereo Hi-Fi. I'll post the whole schematic on the Web here when I perfect it—assuming that I perfect it. Not all designs work out, but all you engineers out there know that, right?

This is a little 1-tube AM receiver for the AM broadcast band. The tube can be either a 3V4 or 3Q4, with allowances for differences in the pinouts. (The two tubes are electrically identical but have different pin assignments.) This was a nostalgia trip for me: The circuit was something published in Harry Zarchy's juvenile electronics book Using Electronics from the late 1950s, and it was the first radio I built that worked really well.

The downside to the circuit is that it needs at least 45 volts to do anything at all, and doesn't pull in the weak ones without 60 or 70 volts on the plate. 90V is actually too much given the coil I wound for it. 45V and 67.5V batteries were a commonplace in 1963, but no more—I run it off an adjustable laboratory power supply. (I had a military BA-63 45V battery for it but it died out in the heat in the garage over the past year.)

The underside of the MC3362 receiver. It's built in a great little 2-piece aluminum box that I got for a buck at a hamfest. I wish I knew if it were still available, as it's extremely sturdy and snaps together so tightly you almost need a screwdriver to pry it apart. The circuit itself is from an article in the July 1988 issue of Ham Radio Magazine. The author is Rodney A. Kreuter WA3ENK. I learned a lot building it and poking at it, which was good practice for my own 6M MC3362 receivers that I built afterward.

This is the matching power supply for the Tinderbox. I broke with my longstanding tradition of not labeling things, since I didn't want somebody who might inherit this after I'm gone getting the power jacks mixed up. It's the same width and depth dimensions as the transmitter, though it's completely enclosed. I tried something interesting here: When you plug in the supply (even though the front panel switch is off) the transformer and rectifier are energized, but feed DC to the filter caps through a high-value "bleed-in" resistor, to keep inrush current from frying the bridge rectifier. When you flip the front panel switch on, the resistor is cut out of the circuit. There's a "real" on-off switch on the back panel that cuts power to the circuit completely.

This is the oldest homebrew radio I made that still survives. It was the first one I made that wasn't tacked together with Fahenstock clips on a piece of scrap lumber. My dad spent hugely on getting me the parts ($20 I think—don't forget that this was in 1964!) but otherwise he had no part in its construction. I did all the metalwork and soldering myself, and most of the parts were gouged out of junker radios and TVs. The power supply rectifier diodes, in fact, are still caked with mud from the stinky Des Plaines river, having come from a TV chassis I had hauled out of the river.

Here's the back of the receiver. The circuit never worked correctly in 1964 because I had reversed the sense of the tickler coil, and the article (in the January 1963 issue of Popular Electronics) didn't warn against that. It was ten years later, when I got my novice ham license, that I fixed it. It works...ok. The circuit is idiotic: A 6AF11 Compactron triple triode, with a single-stage regenerative detector followed by two stages of audio. It was deafening—but so "dead" at RF that mostly what you heard out of the speaker was 60 Hertz buzz. It has other problems due to bad caps and noisy pots, but I won't try to fix it because it's a testament to what I could do at age 12. It wasn't great—but it was the first chassis-based radio I ever tried.

This is my other "first": The first ham transmitter I ever had, and built myself in 1973 from an article, "The Mini-Mitter", in Electronics Illustrated. Again, the circuit was idiotic: A 50HC5 oscillator and 50L6 final, with their filaments in series (plus a dropping resistor) across the AC line, with B+ coming from a voltage doubler working on the AC right from the wall. No transformer, no isolation. The key jack (note the cork washer) was connected right to the hot side of the wall! I called it "The Shockbox" for obvious reasons, but I worked 10 states with it before the local guys convinced me to buy a "real" transmitter. AC buzz haunted the note, and I routinely got T5 signal reports. Nonetheless, as with the receiver, I'll keep it forever to show how far I've come.