Hacker News Reader

Comment by nimish | Hacker News Reader

I wonder what signal processing methods exist to deal with terrestrial noise sources instead of relying on the brute force approach.

View on HN

Not sure what you mean by the brute force approach. The problem is that the terrestrial noise sources are many orders of magnitude stronger than the faint signals that the ultra-sensitive receivers are designed to receive, and easily saturate them. The only recourse is to excise any data taken during an RFI episode. It would be extremely difficult, if not impossible, to do science in many RF bands of interest were it not for the NRQZ.

Not who posted the parent comment, but "brute force approach" = having the NRQZ I'd imagine.

I figured as much. The question about why it is needed brought to mind something I was told when I started work there, over three decades ago. It was something along the line that all the signal energy recorded by the observatory’s telescopes integrated over their lifetimes amounted to less energy than is expended by one flea making one jump. I don’t know exactly how accurate that is, but it does rather vividly illustrate the problem.

Well, for fun, that's 2.25 ergs, over 30 years that's an average of -126dBm

That's not that low for signal processing these days, and it'd be much more practically useful to develop the ability to filter out noise to that level than attempting to suppress it over a very broad area. My understanding is that it's the sensitivity of the detectors that's the real issue: they get pushed into clipping since their dynamic range is not that high. A similar thing happens with optical transceivers.

[0]: https://journals.biologists.com/jeb/article-abstract/47/1/59...

Reply on news.ycombinator.com