2018-04-14 17:41:35 UTC
looking less wonderful.
I wired 1k and 100pF in series from the 1pps to the 50 ohm input. (when
I first tried it, it was just some wire twisted together.. maybe 10 pF).
Why those values? I found the 100pF caps first.
In any case, when I digitize with the RTL tuned low (5 Mhz)- I get two
spikes - one for rising, one for falling.
BUT, when I tuned the RTL up band (say, to 15 MHz), I started to get a
more "flat topped pulse" in my total power detection (rms over 1
Turns out, when the 1pps is *high*, it's also kind of noisy, and that
couples into the RTL through the larger capacitor.
See attached "betweenpulse" and "duringpulse" spectra - I happened to
run this one at 1.44 MHz sample rate so it's +/- 720 kHz.
I have no idea if 14.8 MHz is some sort of significant frequency inside
the uBlox, or if it's just grounding the wire which shunts noise that's
being picked up.
I suppose I could fool around and find a capacitance that couples just
the "edge" through, but not the noise during the "on time". or some sort
of more complex circuit.
Or, I could build some sort of cal burst generator.
But I think what I'll do is go back to the original plan of radiating a
pilot tone with some modulation to all channels.. (see pulses3.png)