Hacker News Reader

Comment by leetbulb | Hacker News Reader

This is surprisingly easy to do with some cheap hardware[0]. There is plenty of information if you just search something like "NOAA RTL-SDR." Building the antenna is the most labor intensive part, but still very easily done. I haven't personally, but it's on my to-do list for this summer.

[0] https://www.rtl-sdr.com/rtl-sdr-tutorial-receiving-noaa-weat...

View on HN

I would say it's one of those ruses where it's easy to get basic results, but you can spend (tens of) thousands of dollars and years of learning and refinement to achieve anything approximating error-free results.

It's a lot of fun though. Diving into SDR really opens up a new dimension of reality to you and lays bare the magic that can happen when you are able to directly apply complex mathematical constructs to real data coming out of a wire attached to, well, nothign.

I got a little obsessed with it actually and had to step away from it for a while lol.

> I got a little obsessed with it actually and had to step away from it for a while lol.

Same here. I burnt a ton of time just scanning for voice communications and analyzing ADS-B. Once you start playing around in GNU Radio and decoding more exotic stuff, you really start going down the RF rabbit hole. It's an amazing hobby for learning about electronics.

> Once you start playing around in GNU Radio and decoding more exotic stuff

Do you have any tips or resources on how to get to a competency level where you get to decode arbitrary digital signals you find from a scan?

I've read through all the GNU Radio documentation, along with the examples on their wiki and have even written my own blocks but keep finding that opening a grc file that someone else authored, there will be a critical block I've never seen before (and likely, never see again) which hampers experimentation. As soon as I attempt to decode a real-life signal, it feels like I dove in too fast -- there's too many unknowns, since the end result is [presumably] not going to be a nice, human-readable string. However, occasionally, I find challenges like [1] which are great, since you have the expectation that the signal _can_ be decoded; it's just that I've found examples like these to be quite rare.

[1] https://www.gnuradio.org/blog/2018-02-21-gnu-radio-challenge...

Not the person you asked but once gnuradio flowgraphs started to function materially outside of basic flowgraph blocks stitched together, or worse, wrapped by some python script, I would have to tap out.

It's one thing to demodulate some manchester encoded OOK/FSK signal from some ISM-band hardware monitor vs. 64-QAM or GMSK. At some point it, to me, becomes indistinguishable from magic. Fluency in DSP is, IMHO, mandatory to be broadly effective with GNU Radio.

I'm fairly fluent in both DSP and software, and find GNU radio indecipherable.

A plain C/C++ implementation is often so much clearer than a flow graph built in python.

Copy that. By far the most informative bit of dsp code for me was the core libraries in SDRSharp.

>I would say it's one of those ruses where it's easy to get basic results, but you can spend (tens of) thousands of dollars and years of learning and refinement to achieve anything approximating error-free results.

Like any hobby?

This confused me. I swear I thought i wrote 'like any good hobby' at the end of the sentence. Totally agree!

Congrats, you made me purchase another gizmo. :)

Anyway, I'm a licensed HAM and I've played with electronics since forever. I was thinking to get a small SDR module to tinker with.

Yes, there is also a Reddit https://www.reddit.com/r/RTLSDR/

Reply on news.ycombinator.com