1-Tube AM Broadcast Transmitter

This little transmitter came about as a result of my interest in vintage radio receivers, and the lack of anything worth listening to on the AM broadcast band (540 - 1700 kHz). With a transmitter, I could broadcast my own choice of music, and listen to it on my old radios.

There are a lot of transmitter projects to be found in Internet-land, including lots of one-tube circuits. Unfortunately, most of the one-tube circuits use pentagrid converters which are very nonlinear by design, and tend to produce very bad sounding signals. I had a few 6M11 compactron tubes that I’d collected for another project which was never completed. So, I decided to design my own circuit using one of these tubes. A 6M11 has two triode sections, and a pentode section. I decided to use one triode for the oscillator, one triode for the audio preamp/modulator, and the pentode section for the power amplifier. Screen grid modulation is used.

Here is the schematic:

Larger Version Here

How it works

The oscillator section, V1-b, is a standard Colpitts oscillator using a ceramic resonator (or a crystal). Feedback is from the cathode circuit. The output of the oscillator section is coupled to the control grid (G1) of the pentode section. Using a separate oscillator section, provides excellent isolation from factors which might otherwise cause frequency drift or frequency modulation. In testing, I was not able to detect any FM distortion in the transmitted signal.

Audio preamp section, V1-a amplifies the audio input, and the resulting signal is DC coupled to the screen grid of pentode section V1-c through the NE-2 neon lamp. The neon lamp drops the DC level at the screen by about 50 volts without attenuating any of the audio. This voltage difference provides a good quiescent operating point for the pentode while providing a higher voltage at the triode plate. The variable resistance R1, in the triode plate adjusts the triode plate and pentode screen supply together for the best operating point. Now, the obvious question: Why use a DC coupled circuit? Why not use AC coupling and bias the stages separately to get the best operating point for each stage? The first reason was simply to minimize any signal attenuation resulting from all of the extra bias components. The second reason is that I like neon lamps. The third reason turned out to be the accidental discovery that this type of DC coupling allows for carrier control. Carrier control is the principle of adjusting the average carrier level in a transmitter as the average modulation varies. At low levels of modulation, the average carrier value is kept to low values, and at high modulation levels, the average carrier level is increased. This reduces the average power consumption of the transmitter, but allows high peak power output when required. It also has the advantage of working as an audio expander, and reduces problems of under-modulation and over-modulation. Once adjusted properly, it’s very forgiving of changes in audio source levels.

The carrier control works as follows. The preamp grid and input coupling capacitor act as a clamp circuit. The V1-a grid is biased very close to 0 volts when no audio is present. With an audio input signal present, positive excursions of the input signal cause a very small grid current to flow, which charges the input blocking capacitor creating a more negative bias on the grid. Hence, the grid bias voltage follows the peak of the input signal, and positive peaks of the input signal are clamped to ground level. Therefore, as the audio input level increases, the grid bias goes more negative, and the average plate current decreases. When the plate current decreases, the plate voltage goes up, which in turn increases the voltage to the screen of V1-c, increasing the average carrier level. The 1 Megohm grid leak resistor in the grid circuit of V1-a provides a discharge path for the grid current and the charge on the capacitor. The value of this resistor and the value of input coupling capacitor determine the time constant of the carrier control. The values shown in the schematic seem to be optimum for correct carrier attack/decay time, and give the best sound quality.

Output from the plate of V1-c is coupled to the antenna through a pi matching network. I used a tapped coil, because I wasn’t familiar enough with antenna matching (especially with a random length antenna) to zero in on the best value. Also, the optimum inductance will vary depending on the operating frequency of the transmitter. The coil shown in the schematic covers the AM broadcast band fairly well, when coupled to reasonably short antennas (~10 feet long).

Sound quality from this transmitter is excellent. It is easily capable of delivering an acceptable signal even to hifi AM receivers. I tested this with a Sony ST-JX450A AM stereo, FM stereo tuner, which is capable of wideband AM reception. In wideband mode, the audio quality was better than what I could receive from commercial AM stations.

This transmitter does have a couple of shortcomings though. The first is that it requires a bit more than line level audio to drive it to full modulation. This could be overcome by using an audio input transformer to increase the drive level, or add an additional preamp stage. I use an inexpensive Radio Shack mixing console which puts out enough signal to drive the transmitter quite well.

The second shortcoming is that even with three stages, this transmitter doesn’t put out as strong a signal as you could get from a one tube circuit using a dual control pentode with suppressor modulation. However, it has enough output when properly matched to a 10 foot antenna to be heard over a distance of 100 feet, which is enough to get to any radio in my house.

Here is a view of the underside of the chassis:

Wiring is point to point, using the tube socket for mounting many of the components. This results in very short lead lengths which is good for radio frequencies, even though it doesn’t necessarily look very neat. The toroidal chokes shown in this picture were later found to be extremely lossy, and were replaced with better ones. Never use mystery ferrite! Immediately in front of the tube socket is a part of an IC socket which is used to hold the ceramic resonator, but the resonator is missing in this photo.

Here is a view showing the transmitter with the antenna matching network, which was built as a separate unit The audio source is from the MP3 player in the foreground.

In the background is an inexpensive passive signal strength meter, which I used for adjusting the matching network for maximum output. Antenna matching is a critical part of setting up any transmitter, especially something as low powered as one of these. When correctly adjusted, the amount of power to the antenna goes up dramatically, and is clearly visible on the signal strength meter. The red wire connected to the matching unit is the base end of a 10 foot wire antenna. The signal strength meter is about a foot away, and is sensing the signal with a small 4 inch built-in antenna (beyond the right side of the photo).

Here is a trapezoidal scope trace showing the modulation envelope, when at approximately 100% modulation.

Note that the curvature in the envelope indicates the presence of some distortion. This was not audible however, and I didn’t consider it significant compared to the severe type of distortion which would occur from over-modulation. When the audio level is changing, the carrier control effect is very noticeable, as the zero percent modulation point on the left hand side remains stationary, and the wide full modulation part on the right, expands further to the right in time with the audio.

Antenna Matching Update - 2015-09-07

Please refer to this page for updated information about:

Matching a Part 15 transmitter to a short antenna

It discusses the antenna pi matching network used in the above schematic, as well as a simpler and more efficient antenna matching network.

Back to:

Projects

Home

This page last updated: March 15, 2022