The warble effects mentioned could certainly be caused by the wrong loop filter values and this warrants investigating. Another problem which I have observed is that if the PLL circuitry is close to the RF output circuitry, especially the loading coil, the RF can disrupt the PLL operation causing a warble effect. In my experiments, mechanical shielding solved this problem. Just mention this as something else to check when looking for PLL problems.
Here's the link to where I first reported this:
Neil
I don't know what you call them but they look like a metal box like shape. It covers transistors from getting any RFI. You could somehow shield the PLL chip from the transistor picking up RFI from it.
Cool idea let us know if you solve the issue with this one.
I am curious how you are going to mix stereo down to mono or has that been taken care of already? I know a couple ways to mix the left and right channels down to mono without altering the circuit, one would be to use a mixing board that offers the option of stereo or mono, then of course software based mp3 players like Winamp have the mono option under Options / Playback
I once tried about 20 tracks mixed down to mono using Audacity and was not entirely happy with the end result.
Anywho, keep up the great work and hope to hear more about your design.
Barry of Blue Bucket Radio 1620 AM
My findings were from a version of Polk's printed circuit board, no output stages or loading coils, just a few feet of wire on the output terminal of the board. I tried different capacitors in the loop and could reduce the noises, but I think it should be much better than it is.
The FM receiver was tuned to a harmonic of the Polk circuit, and the audio recorded on a wave editor, then the wave was shown in spectrum mode. That showed a solid line at 10 khz, and lots of random noise activity below 1 khz, with clear spectrum in between.
The loop filter is like some audio AGC filters or AM radio AVCs I've seen, so it seems like designing a good loop filter would be like designing a good AGC filter.
I just looked at 3 different designs, all using the 4046 PLL IC, with output on pin 13, and VCO input on pin 9, and the filter between the input and output. One is the Polk, the Lythall and the one used in the AMT 3000 transmitter. Polk's uses a 10 k resistor on the output, the others use 100 k and more filtering.
That other link is interesting reading radio8z, and it just goes to show that oscillators should be isolated in serious transmitters, especially free running oscillators, but really all of them should be.
The Part-15 antenna system seems like a mini Tesla coil in a lot of ways, so there could be lots of potential difference with any metal in the area. Tesla builders typically have to cage the electronics, especially with high power coils.