2008-04-20
PSK-31 Beacon Mark 1
There are at least two main ways to generate PSK-31 using a microcontroller. You can generate the baseband keying waveform, apply it to a DBM then amplify up the signal, or you can use a switched phase selector (perhaps a DBM or an XOR gate) and envelope modulate the final amplifier. The later method has efficiency and IMD bonuses but I decided to start with the former.
I'm using the ATtiny13V which is quite a small device, only 1K of Flash, so the program must be kept nice and small. To this end I decided not to work initially with PWM to generate the smooth baseband signal, but rather to just bit-bang a square baseband signal. Keying the DBM with this produces horrible key clicks, but it allows testing of the varicode and other major parts of the device.
Some initial problem with the DBM were fixed, giving better decoding:
The signal is very broad, but it appears to work quite well. I had to trim the F_CPU figure to get a sufficiently accurate frequency reference for good demodulation. I should probably xtal lock to final device, but it appears that the internal RC oscillator of the ATtiny13 may be good enough... Peter Marx VK2TPM suggested trimming the oscillator using the OSCCAL value. This is a good idea which I'll look into, but for now I am just being lazy.
The very much WIP software is available for your inspection.
Baseband Filtering
I'm having trouble with simply low-pass filtering the baseband signal to shape it suitable for direct DBM modulation. This seems to distort it too much and causes poor decoding. Much more maths and work to be done before I get that part working.
Thoughts
The keying envelop is 90 degrees out of phase with the phase transition step. With some clever design it might be possible to utilise this fact to produce a transmitter that has only a single baseband input signal. I've sketched a suitable circuit, but it is pretty ugly. It is likely much easier to just PWM the envelope signal. I'll have to try it.
Accidental Reception
Patrick VK2PN who lives pretty close to me, maybe 2 km, received the leakage from the prototype lash-up on my bench! I was using my free-running LC signal generator to experiment with the keyer. It is only barely stable enough on 20 metres for the task, but once warmed up my PC can hold lock for a few minutes.
This is rather disturbing actually. The signal must have been radiated by my antenna which terminates in the shack quite near my bench. I'll have to be more careful not to pollute the local area with my experimental noise!
2 comments.
Updates
- 2008-04-27: PSK-31 Beacon Mark 2
- More progress on the PSK-31 beacon.