[time-nuts] WWVB PSK testing update
paulswedb at gmail.com
Fri Jun 8 15:57:50 UTC 2012
To update the group.
Been working on this for a while. Thanks to Jfor John for comments.
Since I am in the Boston area always a bit of a challenge for wwvb and now
the psk change to the modulation.
Lots of experiments with simply doubling and then dividing by 2 approaches
to removing the psk.
Some approaches adding gain then full wave rectify, analog multipliers,
Exclusive Or gate doublers, injection locked and Ringing oscillators, PLL,
adaptive slicing etc. The signals only about 6 uv during the day here so
the 1s were stable but the 0s 17 db down were simply sort of noise. It
really hit me a week ago. The 2' loop just wasn't cutting the mustard.
Now have a 10' square loop and new preamp. That brings the daytime signal
to 60 + uv while driving 140' of coax and for the first time I can actually
see and trigger on the 0 reliably.
So its been interesting during this test that is running. By using a soft
clipper to eliminate the 17 db modulation difference. Then slicing and
converting to a cmos logic level signal; and differentiating on both sides
adding together and dividing by 2. A HP VLF117 seems to track the signal. A
Fluke 207 does not. Though interestingly what it does is consistently ramp
in 1 direction. Looks like a saw tooth.
But its not at all perfect or even close.
Lightning and noise cause the divide by 2 flip flop to shift 180 degrees
randomly. Its interesting that there is enough noise that the VLF 117s
integration time constant creates a noisely line, but does track the
correct phase as compared to a pretty clean line without PSK.
So all of this is not the answer yet and will keep digging.
More information about the time-nuts