When I first read it, back when it was first posted on the FCC site, my
immediate reaction was, why would the FCC take what was suggested seriously?
You never, ever use baud rate to measure timing.
In the digital world, you'd oversample the bits at a couple of hundred kilo
samples/second if needed, and do a cross-correlation to get much higher
temporal resolution. In fact, if you use a Barker code or a long PN code
(arguably illegal for amateurs), you can get better SNR from a sequence and get
a very nice sharp peak. If you can't use a Barker code or a PN sequence, even
a series of RYRYRY can be cross correlated.
For goodness sakes, you don't decimate it to the baud rate and *then* take a
cross-correlation -- *that* would give the poor resolution as reported. All of
us have sound cards that can sample to 48,000 samples/second or better, right?
For doing the measurements he was attempting, it does not take a higher baud
rate. It just takes good engineering practice.
By the way, the exercise may be futile anyway when there is strong multipath
(or even when one path splits into two rays). A rectangular pulse sent to the
ionosphere goes through both Doppler spreading and multi-path -- what you
receive back is no longer a nicely shape rectangular pulse. And the
propagation velocity through a dispersive media may not be the speed of light
either.
73
Chen, W7AY
_______________________________________________
RTTY mailing list
RTTY@contesting.com
http://lists.contesting.com/mailman/listinfo/rtty
|