Back in the 70s, I remember my friend Richard G4EIE building a one-transistor transmitter for 160m. It comprised a BFY51 transistor, crystal, L-C tuned circuit, a resistor, a couple more capacitors and a carbon microphone. It put out 250mW AM or 750mW CW (using a morse key in place of the mic). It was surprisingly effective!
I’ve wanted to build this circuit ever since and surprisingly I waited 35 years!
The circuit was devised by Eric Elsey G3YUQ and was published in the Winter edition of Sprat in 1975.
It has been republished in many other places since including RSGB Amateur Radio Techniques and Practical Wireless.
The Carbon Mic took a bit of finding. Old telephone handsets (40 years+) have them in and that’s your best way of finding one. The benefit of a carbon mic is that it is essentially carbon granules between two conductors – it can take some voltage. The mic has an impedance of around 600-ohm and when you speak into it you will see the impedance change; in my tests between 300 and 1200-ohm typically.
Topband crystals are hard to find other than 1843.2kHz, which luckily is the QRP calling frequency.
The L-C circuit uses a 800pF capacitor in parallel with an iron dust core tuned inductor (quarter-inch diameter, 40T of 24swg enamelled). Coil formers with tuned slugs are hard to find these days and perhaps a better bet would be a toroid although it would have to be big enough to take a good few turns.
Rather than hunting down specific exact value of components, I built the circuit using multiples of smaller value components. The 1W 22K resistor was made up of 4 x 0.25W resistors in series and the capacitors were made up of 100pF and 200pF capacitors in parallel.
As you can see, the circuit was built on perforated board with component legs used as the ‘tracks’ and soldered underneath the board.
On my first test, I fried the coil! I’d forgotten to put a series capacitor in the output to the aerial and my dummy load made a DC short circuit! This was awful for my plastic coil former which melted and warped so the iron dust core won’t easily screw in and out of the coil any more now.
My first real tests successfully oscillated and created about 500mW with 12V input when fired into a dummy load. Testing into an aerial proved problematic at first. It was harmonic rich which gave rise to reflected power problems!
Low Pass Filter
I decided to build a low-pass filter which cleaned up the harmonics quite nicely. The LPF was simply two pi-networks, each having 1500pF caps in the legs and a T68-2 toroid with 26T of 24swg. I built this on stripboard as it provided a decent ground for the signal route through the circuit.
As my topband activity is generally at the other end of the band, I thought I would try a 2MHz ceramic resonator and pull it LF using a series variable capacitor. This was an illuminating disaster! Ceramic resonators can be pulled much further in frequency than crystals but I found that the other capacitors in the circuit including my ATU all worked to pull the frequency. My ATU not only matched the aerial but tuned the oscillator too!!!
My best test used 5 ceramic resonators in parallel (this reduced the ‘pull’ factor) but the frequency was very unstable. On CW the notes sounded like bird-song! On AM, additional capacitance in the carbon mic was varying the frequency too.
This design isn’t suitable for use with ceramic resonators as it stands!
Testing on Air
My first tests on air were with Richard G4EIE who lives about 5 miles away. 250mW AM was detectable, even with modern domestic topband QRM.However the modulation level is fairly low. Modulation reduces the emitter current in the BFY51 and on a power meter it appears to drive the power downwards.
The good news was that 750mW CW was RST 599 with no churp!
The next step will be to address
- Poor modulation level on AM
- Use of ceramic resonators as appropriate crystals are scarce