6888 Transmitter
One of my many hobbies is working with ancient vacuum tube equipment, much of which are AM radios. However, as Ben Krasnow of Applied Science says: "after spending hours getting the radio working, you'll find that it sounds terrible." This has nothing to do with the radio, but is due to the fact that AM radio these days is mostly used for political talk shows and modern music, neither of which I particularly care for when listening to an old radio.
To play whatever I want on the radio, I need an AM transmitter, one that operates in the normal AM broadcast band of about 500-1500 kHz. Surprisingly, the FCC allows this for low power transmitters (there's a similar rule for FM, which is why you can buy those little adapters that make your phone play over your car's radio), though I kind of doubt anyone would mind as long as the signal can't be detected outside your property. Besides, these days the FCC is busy making sure the cable companies can keep your internet bill high.
Anyways, after completing a solid state version of a transmitter which turned out to only travel a few millimeters, I decided my next version would be tube based. After all, tubes operate at a much higher voltage, and while I am no expert in RF, it stands to reason that a higher voltage applied into the antenna would make for a better range.
The version I decided to build is the 6888 transmitter designed by Norm Leal and others on the Antique Radio Forums. Specifically, I based my design on a schematic drawn by Jon Stanley. The original version of the 6888 transmitter uses an audio transformer to match a low impedance audio output from a computer or audio player to the high impedance input of the 6888 tube, but I did't have a transformer that fit the bill, so I used this version which uses a 6AB4 tube as an audio preamp to do the same job.
Theory
Jon has a pretty good explanation of the theory of this transmitter on his site, but for the uninitiated I will describe the workings here.
The heart of the transmitter is the 6888 tube, which was originally designed (supposedly) for tube based computers. What is special about this tube is that is has two control grids, which is an input to the tube that controls the flow of electricity from the cathode to the plate. The inputs are independent, so we can control the flow through the tube with two things. On one, we feed in a 1 kHz signal, which becomes the carrier wave our transmitter will transmit on (ie, which frequency we will tune to). On the other, we feed in the audio; this setup serves to modulate (turn on and off) the flow through the tube, in a manner proportional to the magnitude (or amplitude) of the signal. The flow through the tube is now an amplitude modulated carrier (hence, AM).
The signal comes out at the plate of the 6888, which is where we attach our antenna, and an inductor-capacitor tank circuit, which resonates with the signal. By adjusting the tank circuit to be more or less resonate, we can adjust the range of our transmitter.
The 1 kHz signal is supplied by a TTL can oscillator. Put in 5V in one side, and we get a 1 kHz square wave out the other. The 5v comes from a Zener diode that we use on the cathode of the 6888 instead of the typical resistor, to both power the oscillator and set the cathode voltage.
The power supply for the whole thing is done with a silicon bridge rectifier on the 120VAC that comes in from the wall, and is filtered with resistors and capacitors, forming a low pass filter, just like the B+ power supply filtering in most tube radios.
Construction
I happened to have a cubeish metal box, about 4 inches square, with no lid, which I think was once part of some model train equipment. With the addition of an aluminum plate lid on which to mount the tubes, this became the perfect build platform in which to construct my transmitter, although it is a bit cramped working from the bottom.
I needed a power transformer to power the tube heaters at 6 volts, but for some reason I only had 12V ones, probably ones from the inside of the common 12V wall-wart style adapters. This was no problem; I just unwound 1/2 of the secondary winding, turning it into a 6V one. Unwinding the secondary also created some extra space on the transformer spool, and because I counted the turns as I unwound it, I had enough information to wind on a ~120V winding of thinner wire, turning this into an isolated system. This is not strictly needed to make the transmitter work, but it does make it a bit safer, with neither the tubes (or more importantly, the metal chassis) connected directly to either of the power cord terminals.
I didn't have an inductor on hand, so I decided to wind an air cored one on a 3D printed form. A lot of the ARF users who built this transmitter used a toridal coil, which does have the advantage of a small size, but it seems that an air cored coil works fine. Also, for the trimmer capacitor I used a tiny mylar trimmer normally used in transistor radios, instead of the more common air gapped trimmers used in tube equipment. Again, this seems to work absolutely fine, although I did have to cover up the metal screw that holds the knob on, as I found out the hard way that this screw connects directly to B+ voltage!
Lastly, I added a power switch and neon indicator to easily turn the transmitter on and off, and tell that it is on. Most of the equipment in the shop has an indicator light, so that when we turn off the room lights to leave for the day, we can make sure that everything is turned off for the night.