Here is KB8WFH’s truly superb replica station, showing what a rotary spark gap transmitter would have sounded like for the man at the key.
What about how it sounded to listeners on the air? For that, here is VK7RO’s tape recording of another re-creation project, this one probably from the 1960’s, under the aegis of the Antique Wireless Association. It is meant to be what people heard on 600 meters in the early 1920’s.
This is a magnificent replica of an amateur transmitter from the late 1920’s. It electronically manufactures a signal (by what’s called an ‘oscillator’ circuit) which is interrupted by a Morse code key. This is ‘continuous wave’ transmission, which is different from the mechanical method of signal generation just going into eclipse in the same decade. To make a signal by machine, you’d spin a disc-shaped armature in a magnetic field — a dynamo, really, and use your key to interrupt this. It sounded like a jamming signal, and it splattered widely across the band. CW, by contrast, was a narrow, tidy way of transmitting. This was important because the spectrum was crowded beyond capacity.
QRM, or unintentional interference, in most cities was simply intolerable. Noise from raucous spark sets was even creeping into the telephone. Boston operators complained of being pushed out by experimenters, research workers, long waver and short wavers, ‘500 cyclers’ and ’25 cyclers,’ wireless phone owners, and worst of all, those wretched “squeak boxers”—owners of the old and splattery spark coil sets. There was also a thing called “ICW,” for interrupted continuous wave, much like what you see here, only it got its plate voltage (a plate is one of the glowing bits inside the vacuum tube) from an ordinary spark coil. It was cheap, effective, and electrically noisy. Los Angeles complained of being shut down by 10:30 each night because of the thousands of spark-coil variants ranged across the neighborhoods. It wasn’t just the number of signals that offended, it was the primitive tuning. (Notice how hard it is to keep this set stable. The two coils on the breadboard here were positioned relatively to each other with care: this was a tuning method called ‘loose coupling’.) We were all still sloppy enough (or lazy enough) with this particular art that many of us hadn’t even gotten down to 200 meters, where the government wanted us, as late as 1921. Cities like Springfield, Massachusetts, and San Francisco desperately set up evening rotas of when different types of transmission—spark, c.w., and also radiotelephone, in all their variations—were allowed to operate. Slick operators were beginning to control their frequency with crystal circuits, but that was a new technique, for later.
The code in this superb demonstration says, VVV (a lot of times, which is a standard test signal) TEST TUNED NOT TUNED (hence ‘TNT’, which means ‘tuned plate, not tuned grid’), TX DE (thanks from) KJ4RGH (his call sign) 73 (an old railroad telegraphy best wishes).