Here are some examples:
http://Noel.ProDigitalSoftware.com/Foru ... dHours.wma
http://Noel.ProDigitalSoftware.com/Foru ... Hours2.wma
http://Noel.ProDigitalSoftware.com/Foru ... Hours3.wma
http://Noel.ProDigitalSoftware.com/Foru ... Hours4.wma
http://Noel.ProDigitalSoftware.com/Foru ... Hours5.wma
The waterfall display doesn't really show anything too exciting, with maybe just a little dark area showing out beyond the where there's energy. I don't know if that might indicate there's some kind of overload happening or what.
It seems to happen at various signal strengths, though perhaps more (or it's more noticeable) when the signal is strong. It happens whether or not I use AM or SAM mode. AGC doesn't matter, nor does adjusting the audio width. Neither does the setting for the sample rate or decimation correct it.
I've heard it on other parts of the announcer's speech, but much less often. I don't think I've ever heard the word "hours" without it. I just caught it happening pretty badly on one of the weather announcements...
http://Noel.ProDigitalSoftware.com/Foru ... eather.wma
Any ideas? It's not a big problem, just a minor annoyance.
I'm starting to wonder if it could be a problem with the transmission, since I haven't sensed it on other stations, but their equipment is rather more expensive than mine, so I'm not quick to rush to that judgment.
Reason: No reason
I'm beginning to suspect their transmitter. According to the NIST "bible" on WWV (NIST SP 250-67) they use one model of transmitter on 2.5 and 20 MHz and another on 5, 10 and 15. The former uses linear amplification to produce 2.5 kW and the latter produces 10 kW using plate modulation, which is more efficient (an obvious consideration at the higher power level). I'm beginning to suspect some inherent nonlinearity in their plate modulation.
I can post some snapshots of my spectral displays if there is still interest.
I've implemented coherent AM in my SDR, but I've noticed that it doesn't work very well on WWV, or in fact anywhere on HF. If I put the I channel on the left speaker and the Q channel on the right, the tones and voice tend to be spread between the two channels even though my PLL seems to track the carrier just fine. If I listen to just the I channel (the normal operating mode for a coherent AM demodulator) the voice announcements seem even more distorted than in AM mode with a conventional envelope detector.
I think my code is OK. When I tune in a local AM broadcast station, I get the expected result: the audio is completely on the left channel and the right channel is just noise. I've chalked this up to HF propagation but I'm still not sure. One theory is that the PLL is tracking the sum of WWV and WWVH's carrier, which will be out of phase to both their programs. But I see the problem even when WWVH is totally inaudible.
Reason: No reason