Building and Operating a Transmitter from the Roaring 20s

Radio was booming in the 1920s: transmitters had started using the pure tone of 'continuous wave' signals, replacing signals sent with sparks, and increasingly selective and sensitive receiver circuits offered greater range seemingly by the month. Amateurs, not yet able to tame their simple transmitters with crystal-controlled oscillators, constructed them on simple wooden boards with heavy wiring and large inductor coils in order to keep heating from causing their signals to drift. These homebrew transmitters were ongoing experiments in electronics and engineering. The position of a coil might increase output power but cause the signal to sound like a buzzsaw. If you could find a higher voltage power supply, that too would raise its power, but the increased currents in the rig would add more drift: if the other guy can't find your signal wandering up and down the band, what use is that extra 3 watts?

The 1929 QSO Party

Every year the Antique Wireless Association holds a two-weekend event dedicated to winding back the clock to this age. During this Bruce Kelly 1929 QSO Party, dozens of transmitters home-built according to pre-1930 techniques contact each other on the160m, 80m, 40m and, for the very brave, 20m bands. Some actual antiques join the fun, too. The rules are reasonable: you don't need to use parts manufactured before 1930, but they -- and the accompanying circuit -- have to be of a type that was developed in that period. This keeps the price of building a transmitter affordable. Indeed, a favoured one-tube 'Hartley' design uses the modestly priced and widely available type 27 tube.

I heard about this from Mike WU2D's excellent YouTube channel back in April or May, and I was intrigued, partly because my father, who was visiting my town for the summer, was born in 1929, and as a practiced woodworker he'd be game to help me with the breadboard design. Even better, the AWA had a simple design offered on its website.

Putting the Transmitter Together

But how would I find a power supply? That problem always seemed to complicate tube transmitter projects: sure the transmitter itself is only 10 parts, but it's at least as much work to get the 300 volts of DC at 25 watts to supply it. Only days later, Aliexpress offered an exciting solution, a DC 12V to DC 200-450V boost converter that boasts 70W! At under $10 I ordered it half expecting it to be too noisy or unstable to use. I didn't like its open design, so I designed and 3D-printed a slide-over cover for it to keep me from zapping my fingers. It was surprisingly 'stiff' under a load, had only a bit of ripple and didn't overload my receiver with hash. Sure, some day it would be great to build a tube-based power supply for these tube projects, but for me the project really got going when I knew I had the power worked out.

From there on, it was a matter of either finding parts in my junk box, ordering them online or buying at the hardware store. Brackets, adapter plates and a tuning dial were printed in 3D.  (I meant to replace these garish orange parts with ones printed in a wood-based material, but I ran out of time.) Brass screws were used as soldering posts and ordinary housing wire used to hook everything up. I was able to find a suitable 2.5mH inductor at Digikey, and I ordered a bunch of 1000v capacitors to design the tuned circuit. I used my nanovna to verify that the 12 turns of iconic 1/4" copper tubing and the capacitor network were tuned in the 80m band.

Next, the antenna. I had a 40m end-fed halfwave antenna that needed an upgrade, so I replaced its 20m wire with 40m strung up about 30' in the air. I made a few contacts with New England using the 2 watts of energy I could expect from the homemade transmitter, and all seemed good. (An antenna for this band at this height is expected only to operate relatively local stations, say up to 1000km, but that's preferable for such a weak signal.)

A day before the first weekend of the event, I fired the transmitter up and ... nothing. No radio frequency radiation, anyway: the circuit was pulling power, but not resonating. I tried changing the grid-leak resistor and capacitor. Then I just took out of the circuit all the fixed capacitors that I'd soldered in to lower the tuning frequency, figuring they were the greatest number of components in the whole circuit. Voila! It turns out one of them wasn't up to the task. Only one problem: without those capacitors, I was making radio signals 2 MHz away from the 3.5 MHz 80m band. I gave up on transmitting during the event's first weekend, and vowed to get everything in place for weekend two.

Finishing Touches

Junk box capacitors were pressed into service to pull the transmitter down to the 80m band and then I made a coil out of house wire to pick up the generated signal and send it off to the ether. At this point the output power could be measured: 0.7 watts! Not great. The size and position of the pick up coil would vary the amount of energy coupled into it, so I experimented with that a bit and got up to 1.5 watts with 250v power supply. 

By mid-week it was time to hook up the antenna and see if this transmitter could transmit. I used a web-based radio receiver from http://websdr.org/, one located in my target area of New England, to listen to my own signals, waiting until evening (because the 80m band radiates better then). Yep, there I was. Thursday night, I pushed the voltage of the power supply a bit to reach almost 2 watts. In the end, I was putting 11 watts into the transmitter: not great efficiency! (The event allows up to 25 watts.) I double-checked the oscilloscope for key-clicks on Friday, and awaited Saturday night.

QSO, 1929 Style

It didn't take long to start to work my first contact, with Mike W1JAS in Maine. To make an 'official' contact, it had to be with another radio built for the occasion and the following information had to be exchanged: RST, the design of your transmitter, the year that design was published (mine was a 1929 Hartley) input power, name and state (or province). This proved difficult because my transmitter was warbling over hundreds of hertz, rising in frequency as more heat went through those junk box capacitors. I tried lots of tricks, but none really helped: as soon as I got to the part of my call sign with lots of dashes, VE9QR, the signal whooped away and I pictured Mike chasing me with his receiver like a hound after a deer.

These AWA operators are nothing but patient and kind. They'd ask for me to repeat my name multiple times, and then my power, and then my rig year, and so on. After it all, they'd add, "TNX AWA QSO OM". I eventually tamed my transmitter every so slightly by keying with my left hand and using my right to wave close to the tuning capacitor, causing the frequency to drop, or at least not rise so much. I was playing a theremin on morse code.

I quickly learned the old habits of radio morse code. Call with the other guy's call sign at least twice, so they can find your signal, which has likely shifted from where they were tuned before. Slower speed is probably better. Maybe use a couple of 'v's to warm up the transmitter as you're tuning it. I remember all this from the 70's when I was listening to hams as a teenager on a tube transceiver (whose transmit function I steadfastly avoided until licensed).

I worked 6 stations in the two nights. The furthest was N2BE in New Jersey, about 850km from me. The contacts even included Mike WU2D, whose YouTube videos got me hooked in the first place. In fact, Mike came back to my CQ on the second night. I was thinking that we might get penalized for duplicate contacts on the same band, so I told him we'd worked before. I've since found out that in the day of computer cross-checking those kinds of penalties have been dropped. 

Next Steps

Well, I have to improve that transmitter's stability before I use it again. I sprayed some canned air on the various components and found that the junk box capacitors were the most thermally unstable. Probably an antique mica one would be much better, and I have a few of those around here somewhere. 

But to really  experience early radio, I need to pair my transmitter with a old time receiver, not a 21st century model. In fact, the AWA's Linc Cundall memorial QSO Party is in February, and it features transmitters and receivers built or based on a design from 1949 and earlier. Can I get a one- or two-tube regenerative built in time? Maybe a design from my birth year, 1966?