Hacker News new | past | comments | ask | show | jobs | submit login

It's really easy to forget how complex SDR is, and how much expertise it really requires to fully understand if you need to write receivers/processors/filters yourself. There's a TON of great stuff out there for hobbyists that just want to get some cool stuff working though. I tried my hand decoding a soil temp/humidity sensor broadcast and was quickly reminded how complicated it can be for an initiate.

I know almost nothing about actual SDR, but I've got a cheap SDR receiver/antenna inside a glass basement door that receives temp/humidity data from ~10 433Mhz transmitters in various rooms in my house + outbuildings, and a weather station that reports temp/humidity/lignthing strikes/rain amount/wind speed+direction, and lux. All that goes to an influx DB instance, and has a set of graphana dashboards built on top of it. Took me a couple evenings to get set up, and now I've got real time + historical environment data about everything I care about at my house; including high humidity alerts in rooms with dehumidifiers, freeze warnings for a crawl space, and a bunch of other stuff. It has been wildly reliable.






I think as a counter point, Michael Ossmann, the creator of HackRF One has worked really hard to make it a lot more accessible.

In one of my favorite YT videos ever he talk about how he design RF PCB. The point of it is kind of like, the field of RF and SDR is incredibly deep and complicated, but at the same time, most filters/mixers/amplifiers and everything else needed for making an SDR are ICs that one can buy from Digikey and put them together (or likely have assembled).

So, not denying how deep SDR can be, but it can also be accessible with enough effort!

https://www.youtube.com/watch?v=TnRn3Kn_aXg


If you are serious about getting into SDR, having background in DSP can take you very far. Most of the painful SDR dragons are based in stuff that was optional at university time, but is otherwise not that hard to pick up once you know it exists.

Understanding the frequency domain, why it's useful, and how to get there & back, is like 80% of the puzzle.


What do you wish you knew about dsp and frequency back then that you know mow?

Ubiquity of the concepts. I'd have paid more attention if I had known. Things like aliasing, kernels and minimum viable sample rate show up pretty much everywhere.

If you were to give advice / do a talk on more details or areas to apply those suggestions what might they be?

Sorry just trying to get more detail out of you and apply it towards some code to make it practical. :)


Getting some hands on experience with the signal processing side of SDR, with GNU Radio can help with understanding things like negative frequencies, complex signals, etc.

It's open source, and you can just play with your audio ports for starters. Later adding a $40 rtlSDR kit goes a long way. I used mine to build a VOR receiver.


As a pilot with an amateur radio license (though an inexperienced ham) this sounds really interesting. Could you elaborate more on building the VOR receiver?

The transmitter has an electronically switched antenna that virtually moves in a circle at 3600 rpm, causing FM modulation phased with direction. The carrier is also AM modulated with a reference phase signal.

I built a gnu radio flowgraph to receive both and display heading to the VOR.


Sorry I was working from memory yesterday... It's 1800 rpm / 30 Hz, not 60.

https://en.wikipedia.org/wiki/VHF_omnidirectional_range


Woah coming from a dsp / synth background that’s wild to hear.

I've played around with Gnu Radio, and even though I've been able to do some very sophisticated things with GRC, I also realize just how little I really know. Part of that is that the math makes my head spin. I'd be lost without those blocks doing a lot of the work for me.

Does anyone have any recommendations for GNU Radio "guides"? I would like one which is a bit hand-holdy but steps through the theory along with a the practical example of building it in GRC, preferably with known good captures of the intended signal so that I don't need to spend time wondering if I've botched my capture somehow.

I've just started getting into SDR - what receivers and transmitters are you using? Would love to give a similar project a go at my place!

If you want to decode transmissions from common weather sensors, be sure to check out rtl_433 [https://github.com/merbanan/rtl_433]. If you have and RTLSDR dongle, it should just work.

RTL-SDR is an inexpensive way to get started, only a receiver though. Transmitters are more complicated (legally) since a lot of spectrum is off limits without a license in most jurisdictions. They have a kit with a couple basic antennas to get you started.

https://www.rtl-sdr.com/buy-rtl-sdr-dvb-t-dongles/


The RTL-SDR dongle with GQRX is highly recommended, and go well with a NanoVNA.

https://www.tindie.com/products/hcxqsgroup/4-nanovna-v2-plus...

LimeSDR are a bit more involved, but like most SDR transceivers still need band-pass filters and power amplifiers to perform well. Works well when combined with a GPS locked 10MHz GPSDO like a BG7TBL. =3


> weather station that reports temp/humidity/lignthing strikes/rain amount/wind speed+direction, and lux

Do you get this information from radio waves? would love to know more about your setup.


Not OP but the idea is simple: You put transmitters somewhere outdoor (or wherever you want to measure) plop a battery in there and be set for a few years.

Such weather stations were really commonplace 20 years ago, I remember my grandparents having a sensor outside and a basic LCD display in the kitchen displaying the outdoor temperature and humidity. These days we want that data digitally on our phones or home assistant so you need a receiver that talks TCP/ip and runs a real os. That's where the SDR comes in to bridge the gap between primitive RF Tech and modern computing. Of course you could also put an esp8266 outdoors, which natively talks wifi, but then you lose range and your battery life goes from years to weeks.


Apparently 433mhz environmental sensors is a whole category: https://www.amazon.com/wireless-433Mhz-weather-sensors-Home-...

433MHz is attractive because it's low frequency allows it to propogare farther with less energy input than higher freqs (900MHz; 2.4GHz) would need and does not suffer from nearly as much reflection off of obstacles thanks to that longer wave.

There are several frequency ranges in the US that are unlicensed for transmission. But don't confuse unlicensed with a lack of rules governing what you are allowed to transmit; how often you can transmit and for how long. Because you can plop a 433MHz transceiver into anything, you want to be careful that you're not clogging up the local airwaves by not knowing to know the rules. Also, most smart meters (near me anyway) operat in this band sending out pulses every so often. They mesh together to relay the data towards a central collector. Thanks to that low frequency, hundreds of meters can be visible at times showing up as tiny chirps all over this area of the spectrum. Unfortunately this also means that some cheap receivers (just looking for any signal on a very specific frequency in that range, can be randomly triggered by this 'noise'. But also, because it's used by utilities, to want to make sure they don't end up having an issue with meter readings because you began running a wifi link over 433MHz.


This is a tool for receiving a lot of 433MHz stuff: https://github.com/merbanan/rtl_433

They're using weather stations that broadcast on 433MHz around their property and receive those messages with their SDR, then they pipe that into their database and display system.

Are you piping to influx db through ignition?



Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: