1924: A Five Watt Sending Set for $25

The first question about a transmitting set that a fellow wants answered, —or maybe it is his dad who wants answered—is, “How much is this thing going to cost?” Well, that all depends. You can buy all of the fixin’s and do-gadjits and make a set that cost a couple of hundred dollars and still have only a 5-watt set, or you can buy only a few things, make the rest of the parts yourself, hook ’em up in a good old circuit, and talk to amateurs several hundred miles away for $25. The latter sounds the best.

Mason, H.F., “A Five Watt Sending Set for $25” QST, Vol 8 No. 2 (Sept 1924), p. 56.

Despite the changes in wireless technology over the last century, there is something timeless in the attraction of “DX”.  In my experience it was first the listening to distance voices, and then at the age of eleven the idea that I could talk back.  It was indeed wireless—not just in the sense of being without physical wires, but also entailing the independence of human infrastructure between here and there; no tolls to pay.   The idea that I myself could induce an infinitesimal amount of energy in a wire thousands of miles away, and that this energy could be used to fill someone’s headphones–or a room or a stadium–has timelessness transcends all the modern conveniences we have at our disposal. It was and still is magic.

It was not expected or even encouraged that a modern Novice licensee in 1991 would build his first transmitter (I was one of the dwindling few who did not start with the “no-code Tech”). In prior generations, however, (and likely for some reading this) it was an economically required part of the journey.  It’s seems that hams have always had a tendency to be cheap stingy economical. The opening paragraph above makes me believe that this has always been so.

That said, one learns much by imitation, and the experience of building a 1924 sending set with period techniques and materials has given me a taste of what this might have been like for the teenager that Mr. Mason is addressing.  To be frank, building this was work—some of it sharp, hot, and messy as the photographs show. I had never cut glass or baked condensers in an oven, but likely neither had the original reader in 1924.  Thus, whether the young Mr. Mason could have anticipated this or not, the unnamed

teenager in 1924 to which he refers, the archetypal old guy pictured on the cover of the September issue in which the article appeared, and I myself are strangely one and the same audience….and somewhere in that experience is a lesson as to just how excited these early amateurs were to get that little “5-watter” on the air to see what might happen.

On to the transmitter…

A transmitter for new audiences

The transmitter itself looks strange even to those familiar with 1920’s transmitter construction, but most of this can be attributed to the combination of economy and our own modern sensibilities.  While thrift is implied in the title, it is the familiarity to the intended audience that explains many of the unusual visual features of the set.  

It is instructive to see it through the eyes of both the “BCLs” and the “old-timers”, the two audiences that the ARRL was trying to appeal to at this time.  The spider-web inductances, while less common in a tube transmitter at the time, would be familiar to any broadcast listener who had built a Reinartz tuner (which is also a Hartley oscillator like this transmitter) or looked inside a brand new Crosley model 51 receiver. The returning old-timer from the spark days would see in it something of a smaller version of a spiral oscillation transformer that is likewise tuned by moving taps.  To both parties the construction is a different size but familiar.  Additionally, building condensers from raw materials would have been squarely in the comfort zone of a pre-war amateur.

The “latest kinks” to which the lead refers includes loose-coupling of the antenna circuit, a variable grid leak, center-tap keying, and a variable antenna series condenser.  The “loose-coupled” (i.e., inductively-coupled) antenna was required by the new radio law to reduce “key thumping” in the broadcast receivers, and was something the ARRL had been evangelizing for a while.  The variable grid leak was made possible by the Bradley-Ohm compression resistor that was newly available in higher resistances appropriate for the grid bias in transmitters.  Center-tap keying was a relatively safe method of keying the transmitter didn’t involve RF of the full plate voltage present at the key contacts. The antenna series condenser is also a modern touch that was often neglected because of misguided belief that it was lossy and because of the added expense (cleverly avoid here–see below).   Electrically, however, the circuit is and ordinary shunt-fed Hartley oscillator, a mainstay of one-tube transmitters in the 1920’s.

QST, August 1924, p. 55

As for economy, the article encourages the builder to make everything he can, and if he can’t make it to find cheaper substitutes for the required function.  We must remember that some parts of a transmitting set were rarely purchased new even in the most opulent sets because they were cheaply and quickly constructed, or because they needed adjustment after the fact (e.g., inductances and RF chokes).  Some parts could be constructed from raw materials if one had significant time to spend and the available raw materials available (e.g., fixed condensers, transformers).  Finally, some components were beyond the ability of the average workshop.  This last group includes meters, reliable tuning condensers, and most obviously, the vacuum tube.  Even here the author suggests some ways of pinching pennies.

Building the set

I’ve shared the construction of the fixed condensers in another post, so I won’t dwell on on that here. Needless to say, you can make pretty good RF capacitors with wax, glass, mica, and tinfoil. Cutting glass hasn’t changed since 1924–oil and a wheel, and it’s not something I relish doing!


The set was designed with the old 200-150m band in mind.  The tank coil is very large—25 turns on a 7″ diameter form, which measures out to around 60uH of total inductance.   The intention was that you only used a fraction of the coil by choosing the appropriate taps. 

How to make the antenna and tank inductance (Sept. 1924 QST, p. 58). Most home-made coil forms like this that I have encountered in broadcast sets also have 15 divisions. It’s an ingenious way to divide a circle into an odd number of equal divisions without a protractor!

To illustrate just how oversized this is, the whole coil with the specified 500pf condenser resonates at 900 kHz!  It turns out that only 1/3 of the coil is required for 1500 kHz (200 meters)—the lowest frequency amateurs could operate in 1924.  If you want to tune to the other end of the band at 2000 kHz (150 meters) you only needed about 12μH, or one fifth of the coil.  Most of the coil is unused! This ends up being realatively easy to tune at 1500 kHz, but one starts to run into dead-end effects trying to steal the resonant frequency when tuning to the modern 160-meter band at 1800 kHz.

The antenna inductance is the same size as the tank coil, but it is only tapped every five turns.  By making this inductance variable both in its inductance and coupling, along with the series condenser, it could match any reasonable wire antenna a beginner could put up.

I wound my inductances on 1/16” Bakelite sheets instead of “hard fiber” suggested in the article.  You can get “hard” or “vulcanized” fiber (or fibre) today, but it usually is sold in thinner sheets as electrical “fish paper”.  I was unable to source any of this material that was thick enough to be a sturdy coil support. The Bakelite form is fine from an RF perspective, espeically since this tank condenser is in parallel with a wooden condenser!

The article specifies #16 d.c.c. (double cotton-covered copper) wound with a string to space the turns.  This wire is almost unobtainable on the surplus market in good condition and in the length required, but fortunately you can get newly manufactured d.c.c. wire from England! You’ll have to do the math on millimeters to AWG conversion, but if you need pristine stuff for a really special project, this is a good source. Given the dimensions, it is clear that the string is intended to be about the spacing of the wire itself.  The article does not say whether the string is to be removed after winding, but it looks as if it was left on the form in the photograph.  Mason may have found, as I did, that it is impossible to remove after winding! 

Forming the taps–one on each turn of the spiral.

Interestingly, he specifies that about 120 feet of wire is required for winding these two coils (p.58).  Unlike Mr. Mason, I could not purchase #16 d.c.c. in arbitrary lengths from my local discount wireless dealer, so I thought it would be wise to calculate the actual length so I could budget appropriately.   I am glad I did. It turns out that only 60 feet is required, 30 ft. in each coil!  It’s my guess that that Mr. Mason (or the editor) calculated the required length after building the set  and used the well-known formula for an Archimedean spiral given inner and outer radii without accounting for the spacing of the string.  This would double the amount of wire and explain the error.

Finished taps on the tank inductance.

Since this transmitter (somewhat unusually) uses a fixed-value tank condenser, the only method of frequency adjustment is by moving taps on the inductance. Consequently, there is a tap on every single turn to provide the needed frequency granularity. After winding, I coated the tap points with some clear shellac, let it it dry, and then cut the insulation off with a knife. (If you don’t put some kind of binder on there you’ll find the cotton covering just makes a mess. If you don’t like or don’t have shellac, get a small bottle of “Fray check” that sewers use to keep threads from fraying. It’s the same idea, and it uses a nylon binder so it probably is both more RF friendly and entirely vegetarian.) After that, they can be pinched together and tinned to make a tap that can be grabbed by an alligator clip.

The RF Choke

The RF choke is another one of those things that looks funny to us, but was pretty much standard in the 1920’s. The late 1920’s saw a move toward longer and much thinner chokes, usually wound on a wooden dowel, but 2-3″ diameter forms were commonplace in the early and mid 1920’s. There was no ferrite available then, so the familiar and much-more-convenient pi-wound choke is still a few years away. I wound this one in 1924-fashion: with a vise holding a hand drill. Brass screws and thumb nuts make the connections, and the whole thing is screwed down to baseboard with a wooden insert.

Antenna series condenser

QST, Sept 1924, p. 57

Okay…if you that the inductances and homemade capacitors are cool, this is even better! It’s a four-plate glass condenser in which the stationary plates are glass-and-wax sandwiches and the movable plates are pieces of roofing tin attached to a pivoting handle (the article suggests an old gas is also acceptable). The stationary plates were made in the oven along with the glass tank condenser (two glass-tin-glass sandwiches) and held in place by some finishing nails to restrain their movement. They rattle around a little bit, so care must be exercised when transporting the set, but boy does it look feel cool when you move a wooden handle to maximize the antenna current! Who needs knobs?

Meters…and a Faustian bargain

Okay, now we get to the scary stuff. There are a three meters on the this transmitter, which at first seems somewhat decadent for an inexpensive transmitter. We’ll take this from left to right.

The antenna current meter is a Radio Corp. UM-530. While not specified in the article, this is surely the meter in the picture. These were low-current hot-wire ammeters, and weren’t terribly accurate. The good news is that accuracy is not needed–one only needs relative indication to peak the current (and if one wanted a stead note, back off a little).  The following advertisement in the classified section of the August 1924 QST gives a rough sense for the cost of meters on the surplus market, including a UM-530 like the one used in this set.   Note that all of these RCA (GE-manufactured) meters were discontinued at the time and represent liquidation pricing.

Advertisement, QST, Aug. 1924, p. 92

The middle meter is the filament voltage meter. It is actually a “pocket watch” style battery-test meter intended for testing either 6V automotive batteries or the filament “A” batteries for a radio receiver. The author has realized that these are just iron-vane “plunger” type meters, so they work on both AC and DC. The case is one terminal and the plunger tip at the bottom is the other. The real trick was to get the connections soldered to these without damaging anything. A butane torch was needed for the top connection.

The meter on the right is the “plate current” meter. The author realized that the difference in the “ammeter” and “voltmeter” versions of these pocket-watch meters is just how many windings are on the coil and how the face of the meter is calibrated. In some sense, they are all really ammeters. He found that a B-battery voltmeter (50- or 60 volts F.S.) worked just fine for a relative indication of the plate current in the range of a few tens of milliamps.

Western Electric Company Inc., Telephone Apparatus and Supplies No. 3, New York (1916), p. 30


There’s just one small caveat: the meter is in series with the B+ supply, and this means that the case of the meter is at the same voltage as the plate of the tube! Yikes! While this sounds like a bad idea (and it is), we should remember that standard practice was to place transmitters out of the way and once adjusted were not supposed to be touched. It’s not like the operator would be reaching for the panel like they would on an IC-7800, but it is what it is.

The meter I used for this happened to be used as a “freebie” along with a battery purchase in 1924. Just the thing a cheapskate would use!

Advertisement for the Shur-Test meter with B-Battery purchase. (The Quaker, Salem Ohio, 19 Dec. 1924, p. 2)

There is easily a savings of $10 in these two plunger-type meters compared to using quality meters designed for these purposes.  This is probably the most significant cost-cutting gimmick in the set. Maybe that was worth the chance of getting your bell rung (or worse) on the front panel.

The tube

QST, Sept. 1924, p. 70

While constructing your own transmitting tube was out of the question, the author offers some money saving advice in its purchase.  The standard “5-watt” tube was the RCA UV-202 (or Cunningham C-302) and it cost $8.  The author suggests that you could roll the dice with an off-brand such as the “Rolls Royce 202” (shortly thereafter “Roice” after presumably being litigated by a company with a lot more money). It was advertised for $3 on p. 70 of the same issue as the transmitter article.  The author also mentions surplus military tubes such as the VT-2 if one were willing to either modify a General Radio socket (of the type used in this set) or remove the locking pin on the tube base.

Radio, January 1922, p. 44. The “cheap” sockets cost $1, but the GR socket really was better for the reasons stated here.

Other parts

The connections for plate, filament, and key are all placed on a row of Eby binding posts. Not having any thick brass strip around, I hammered out some copper supports, which I also used for mounting the mica condensers.

Binding posts and GR No. 156 socket. You can also see the mounting of the homebrew mica condensers on the left and the RF choke in the back.

The grid leak resistors is something interesting. It’s a “Bradley-ohm” 20K ohm variable carbon-compression resistor. The usual grid resistance was usually in the neighborhood for 5-10K ohms, so this provided the range to get things dialed in optimally. This was a relatively new product, and had a redesign to the version used here right at the time the construction article was published.

Radio, November 1924, p. 112

One of the little things you might not think about are the aligator clips used for connecting to the inductances. Although the familiar copper Mueller clips were available at this time, the ones shown in the construction article had a distinctive shape and are probably the Morse Eureka clips. These were advertised from the mid teens until around 1925. They were cheap at the time: about 10 cents each.

Electrical Record and Buyer’s Reference, Buyer’s Reference Company, November 1908, p. 229

The wiring was done with 14 AWG bare copper that I had straightened to make bus wire–a common practice then. The flexible leads were from an partially mouse-chewed lamp cord. Fortunately there were long sections in good shape that could be salvaged for this project!

The parts were mounted on a baseboard of locally-sourced upstate-NY Eastern Hemlock (after checking for shakes), and covered with a few coats of amber shellac. The front panel is shellacked basswood.

The completed transmitter

How much did it really cost?

One of the first questions people asked me in regard to this transmitter is, “How much money was $25 back then?” This is a harder question to answer than it appears. One can look at the price of gold, the inflation rate, the price of gasoline, or the stock market. All of these may be useful for tracing long term investments or the health of nations’ economies, but none of them really help you to understand what this meant to a person who spent it in 1924. Perhaps a better way is to look at the potential audience and see what the opportunity cost was. In other words, what would a working-or middle-class family budget spend $25 on and what would the equivalent stuff cost today? Economists call this a “market basket” measure. Without getting into fine details, you come up with a number that would be something like $200-400 today. Think of a 15-year-old today, and consider that this is like purchasing a bicycle or maybe an X-box or iPad on the high side. That gives you an idea of what it was like.

Can you build this for the 1924 equivalent of $25 today? Probably, even if one has to purchase the materials. The only parts that have more-or-less “retained value” compared to their 1924 prices are the UV-202 tube and possible the GR socket ($8.00 and $1.50 in 1924, respectively) as these tend to be desired on the collectible market today. Everything else can be found on the surplus market or in junk boxes. The real “investment” in the set is the time it takes to build it; elbow grease as has also retained its value.

Transmitter in operation during the 2017 Bruce Kelley Memorial QSO Party, sporting a VT-14 tube. This sat way out of reach of the operating position, and I have removed a plastic barrier from the plate meter for the photograph. This transmitter was the recipient of the Matlack Transmitter award at the 2017 Antique Wireless Association convention.

Leave a Reply

Your email address will not be published. Required fields are marked *