Guide to Software Defined Radio | Hacker News Reader

Guide to Software Defined Radio

luaradio.io

543 pointsphsilva115 comments

Even if you have no idea what you're doing, I very much recommend buying one of the RTL-SDR dongles on amazon for ~$20-30 and playing around. For the cost of a movie ticket, an afternoon messing with one will teach you a MASSIVE amount of stuff about how RF works.

I tried to set up an SDR for my journalist girlfriend, and like the rest of you hackers, reading the manual, figuring stuff out and getting something working is usually pretty easy for me -- but there is a whole lot to software-defined radio.

What were you having issues with? Usually the RTL-SDR + SDR# (sdr sharp) works without any config. Once you start to look at more expensive devices like the Lime mini things get weird though.

It was a physical device and I was setting it up to read the police scanners. I do have problems sitting too long with documentation, though, I'm sure if I more patiently worked through it I could have set it up.

Start with the spectrum analyser app, it is a nice way to see what transmissions are around. This leads to "what is that".....

Seconded. I was playing with mine last week and had it set up to read pager messages. Turns out that hospitals still use pagers heavily, and pager data is completely unencrypted.

Adding, just in case anyone wants to do this on their own: IANAL, but my understanding is that in the US it is not against the law to capture and decode these plain-text messages, but it is probably against the law to publish that information elsewhere or act on it in some way (an example from the link below was intercepting taxi service text messages in order to gain a business advantage for your own taxi business).

Source: https://www.fcc.gov/consumers/guides/interception-and-divulg... via https://www.reddit.com/r/RTLSDR/comments/4nhg7h/legalities_o...

The text messages that float about in the ether these days aren’t messages between individuals, they’re monitoring alerts (refrigeration systems, computers), and medical (doctors being paged, order to clean hospital rooms, calls for medical transport). That’s the case at least here in my medium-sized town in the US Midwest. It could be very different elsewhere.

Oh yeah, this has actually been a bit of a controversy in Vancouver recently[0][1], including speculation that the problem may be national. Local hospitals have now removed diagnosis data from the transmissions, but apparently still broadcast everything else, including "patient name, age, gender marker, their attending doctor and room number" (as of Sept 2019)[2].

Vancouver Coastal Health claims they "have no information to suggest private patient information has been used in any malicious way", which is a very disingenuous statement to make, because there's no conceivable way for them to know who has received radio transmissions or made use of such data maliciously. To be frank, I find this pretty mind-blowing and it's disappointing that even in the face of press/public attention, it's not being remedied.

[0] https://openprivacy.ca/blog/2019/09/09/open-privacy-discover...

[1] https://www.ctvnews.ca/health/pager-systems-used-in-healthca...

[2] https://openprivacy.ca/blog/2019/09/26/pager-breach-update/

Sit outside that place with a 4G connected phone and just send it up to a website called VancouverCostalData.com and I bet you someone will care.

... and go to jail for stealing patient data and break HIPAA laws. Nice thought experiment, though.

They are broadcasting patient data. It’s not more stealing than me publishing in a web site CB conversations.

Then set up a mock-up site at https://10.3.42.8/, and just show a journalist on your laptop. Ask to remain anonymous, then switch off your web server and let them know that "it has since been taken down, but anybody else could make a new one".

That's their story. They can then go on to describe the actual issues, namely the plain-text broadcast that made your proof of concept possible, and state in all honesty that it'd be ten minutes work for a bad actor to set up another, similar, site – and less to just gather (and later sell) the private medical data. Plus, it's not wildly unethical! (Might still be illegal, though.) This is drawing attention to an issue that they're well aware of, and are doing nothing about, in a way that minimises public access to these medical records – one could argue it's positively ethical.

… Though, actually, it'd probably be better to do to somebody in charge first, since perhaps only with such a mock-up would they truly understand. They might genuinely not be aware, thinking it's just some obscure issue.

HIPAA doesn't apply to individuals, it only applies to health institutions and their employees and business associates[0].

[0] https://www.hhs.gov/hipaa/for-individuals/guidance-materials...

Plus, Canada.

Can't those only receive and not send?

That’s correct.

There are some more expensive options that do but it's not recommended if you don't have a amateur radio license

If you have no idea what you’re doing, then you should not be transmitting. There’s plenty of good clean fun to be had on the receive side, though.

Receiving radio is perfectly fine and you aren't at risk of interfering with anyone.

Transmiting is very "dangerous" if you don't know what you're doing, and likely illegal unless you're licensed (you're only allowed to broadcast on certain frequencies below certain power limits without a license, but you can still cause trouble).

Tl;Dr Please do not experiment with transmitting if you don't know what you're doing.

pedantic but noteworthy: amateur radio licensees are not allowed to broadcast, that is transmit with no intention to communicate two way

Even more pedantic: There is an exception to the broadcasting rule.

>Communications directly related to the immediate safety of human life or the protection of property may be provided by amateur stations to broadcasters for dissemination to the public where no other means of communication is reasonably available before or at the time of the event.

For some values of "those" yes, for others No. Some are receivers that can only receive but you can also get transmitters or transceivers that will do both. Most of the cheap dongles are receive only. Disclaimer: Your local laws regarding the legality of transmitting (or even receiving) on a particular frequency at any given power may vary.

For various reasons (such as, it can cause interference), I do not want to transmit, and would rather have one that cannot transmit or that has a hardware switch to disable transmitting when you do not want to transmit.

yeah! I had picked one up a while ago to play with it and had some fun and stashed it away for a while. Well last week i was BBQing and the receiver end of my little remote temp monitor stopped working. Found rtl_433 that already is setup to read the output from the sender portion and in like 10 minutes had it graphed on grafana with alerts for temps!

Can you show this off? It sounds like something I need to have in my life. What TSDB did you use? What did your conversion from rtl_433 to a metric look like?

It was a bit hacky, i just had rtl_433 output the values as json, piped them into a mqtt topic and used node-red to inject the values in influxdb

nothing fancy

If you set up a HackRF in Chicago, you can hear police dispatch. It's pretty nifty.

Sir, you just received the most news.ycombinator.comish comment award of the year.

Any recommendation for one to buy?

Just hop on Amazon and search for rtl-sdr dongle. AFAIK they're all roughly the same since they're based off the same hardware; any with decent reviews should be fine. I think mine was a Nooelec brand or something; comes with the dongle and an telescoping antenna.

I recommend getting one from here: https://www.rtl-sdr.com/buy-rtl-sdr-dvb-t-dongles/

Great hardware with driver/software compatibility

My advice would be the same. These units are mostly built on the same parts, the big difference is build quality (metal dongle vs plastic) and what antenna comes bundled (antenna quality/size matters a lot for anything beyond just tinkering).

It's more than just build-quality. The actual RF design is very different. The RTL-SDR V3 dongle was purpose built to be an SDR device. Most of the generic plastic dongles were designed to watch broadcast television.

+1 on this one: well built, optimized for small signal reception (TV dongles are not) therefore lower noise etc. bias tee for active antennas, metal case. Still at a good price. I have two of them and never regretted the purchase. One day I'l probably get a HackRF or something similar, although I'm still good with the rtl-sdr.com dongles. If they made a small one with enclosed upconverter to be put in my laptop bag to be carried when I take a trip in the mountains I'd get one immediately. I've tried one of those no-name Chinese little boxes with HF and V-U inputs but performance is inferior compared to the above dongle.

Instead of a HackRF, I would suggest a PlutoSDR.

Are there any cheap upconverters that will make use of its insane bandwidth? I'd like being able to receive down to the HF but the Pluto doesn't go below 325MHz.

I recently got a couple of the NooElec NESDR SMArt v4 dongles, and I've been quite happy with them. I also have a few no-name dongles, with one of them that has been running rtl_433 for years now to graph outdoor and basement temperature and humidity from off-the-shelf sensor units.

The NESDRs are much more accurate than the no-names and don't drift. I was flabbergasted when I plugged one in, tuned it to a nearby NOAA weather transmitter, fired up gqrx, and saw that it was spot-on without having to adjust the frequency error. Also, they have SMA connectors instead of the dodgy barrel connectors on the no-names.

I either had a terrible antenna, interference or reception at my apartment and so my dongle gathered dust for a few years until I camr across an Android SDR Touch app. Using a USB-OTG I could easily listen in on the go for cheap.

Agreed! I picked one up and learned about ADSB protocol and usage, and now have a live view of the air traffic in my vicinity. Sure, I could use one of the websites that tracks that info, but there's something about viewing it on your own equipment.

AIUI, those websites are only hackers like you crowdsourcing that info anyway.

I believe that’s correct. FlightRadar will actually give you a free pro membership if you run a receiver in your house (basically a Raspberry Pi, SDR dongle, and an antenna).

Fun fact: Most of those sites filter out a subset of air traffic, notably military and anyone that pays to have their flight data suppressed (typically celebrities or other high profile people). If you run your own equipment you’ll see everything within range of your antenna.

The one notable exception is https://www.adsbexchange.com/, which actually allows you to just show just military aircraft or just aircraft that have been tagged as interesting (typically planes owned by high profile people like Bill Gates, large corporations, or news and police helicopters).

It’s kind of fun to zoom their map out to the continental US and put on the military filter. At any given moment there are a surprising number of military aircraft over the US.

Hackaday (the other hacker news?) just posted a link to this ready-to-go Raspbian distro pre-packed with SDR software for the Raspberry Pi:

https://pisdr.luigifreitas.me

In what country is a movie ticket $20? That seems unusually expensive to me.

This is the price of a movie ticket in the United States, at least in California.

Source: lived in LA and the Bay Area.

3D movies cost €15 and IMAX tickets cost €18 in Berlin

I lost complete interest in SDR due to the fact that antenna's are very important. I am ok with the tools etc but the fact that I cannot do anything with the antennas and the interference put me off.

I mean... you could go buy or build an antenna appropriate for the signal you want to monitor? If you've got the space for it, a wideband discone like this one [1] out on the balcony and up on your roof paired with a wideband LNA from Mini-Circuits will give you acceptable reception from low VHF all the way up through S-band. Or, if you are looking for a specific frequency, build yourself a dipole of the correct length (it's really not hard).

You can't cheat physics, you do in fact need the right antenna. I don't know what else to tell you.

[1] https://www.diamondantenna.net/d3000n.html

Man, this seems like such a great resource; wish I had found it when I first got into SDR. I wrote a blog post about controlling ceiling fans using SDR (https://blog.hmac.io/2019/10/25/making-dumb-fans-smart-using...), but I glossed over my initial struggles just getting up to speed with this stuff. All the concepts of SDR are straightforward and fairly intuitive, but it's the software stack and actually using the tools that's hard. The whole field is niche enough that you end up stubbing your toe with every step you take in that world.

Googling around and trying to figure out where to even begin comes up with so many fragmented, unhelpful pieces of information. You either end up being pointed at Gnu Radio, which amounts to an incomprehensible behemoth for a newbie, or you find the numerous lighter weight pieces of software which aren't very clear on what exactly they're good for and are often unmaintained.

Luckily my first project was rather simple; ceiling fan remotes don't exactly use the most advanced protocols. Once I found CubicSDR and fiddled around with it enough I was able to dump the radio signals to an audio file and just tease the rest out in Audacity. My blog post mostly covers the nightmare that was the TX side of things.

Plug for the “universal radio hacker” open source tool. It made it absolutely childs play to reverse engineer a wireless thermometer signal which i now feed into a dynamo db table and present as a graph via alexa. My kamado joe is at the bottom of the garden and when i’m smoking or bbq’ing i can bring up the meat and the grill temps on any screen in the house. Fantastic.

There was a nice paper on it at woot18: https://www.usenix.org/node/220563

Rachel Kroll (wrote The Bozo Hour making fun of incompetent sysadmins and management) created a web portal for an SDR in C. I found her blog posts in the project very interesting.

> http://scanner.rachelbythebay.com/

This write up is a great introduction to what can be done and is nearly word for word what I have been through so far with SDR and amateur radio.

Another super interesting thing I have done with my RTL-SDR is track NOAA and Meteor weather satellites and decode the images in realtime as they fly by overhead

That’s amazing. Did that require any custom code?

Thanks for posting, I didn't know about LuaRadio.

After playing with some examples I noticed the weird behaviour of the gnuplot windows that get focused and refreshed (including decorations) at every sample so that beside wasting cpu power for unneeded drawing, it takes away focus from the shell where the program was invoked, making impossible for example to press ctrl+c to stop the program. I was forced at every test to ctrl+alt+f1 to a full screen console to kill the process.

There's a nice Lua GUI project called TekUI which includes a basic graphing control that could be imported and extended to do the same without the need of an external dependency like gnuplot.

  http://tekui.neoscientists.org/screenshots.html (look at about half of the page)

  http://tekui.neoscientists.org/releases/

It's highly portable, a few years ago I managed to compile and use it on a A10 ARM processor board (Hackberry if memory serves).

Thanks, I'll take a look! New plotting sinks are easy to add and I really would like to migrate to something more stable. Gnuplot has been finicky with many people's setups, and it's been a real sore point. They seem to run best with the wxt output on Linux.

Thanks what a good link to SDR, the link feels like a Christmas present. SDRPlay is a good SDR receiver a step up from RTL-SDR for exploring the radio spectrum. Plus to exploring software defined radios there is a good community built around ham that one can join.

I got a SDRplay some time ago from a co-worker. I haven't used it yet because of the behemoth that is SDR, but with this source I feel I can finally do something with it.

I want to broadcast (unidirectional) UDP packets over the air on 2.4 GHz or 5 GHz at 3+ Mbit/s at maximum (legal) power. Is there a good way to do this with any current SDR hardware? Or is my best bet to ab(use) wifi cards for this purpose?

What does "broadcast UDP packets" mean? They'll need some sort of physical encoding; if wifi, why not use a wifi card? If not wifi...what are you trying to do by broadcasting UDP packets?

Broadcast meaning unidirectional, with support multiple receivers, and no association or acknowledgements required.

Correct me if I'm wrong, but I don't believe there is a "broadcast mode" for wifi.

So I'm asking if there is some SDR software/hardware combo that works well for this purpose, or if re-purposing wifi is the best option.

UDP requires some physical layer, which 802.11 provides. There are others, but in any case, you can't just "broadcast UDP" over the radio by itself.

Yes, of course. Any suggestions on SDR-based hardware/software options?

We broadcast UDP over various satellite transponders. We have also done this over terrestrial links. We've done UDP file broadcasting over DVB-S/DVB-T/LoRa/FSK/etc with forward error correction in the client/receiver. Feel free to contact me. Email in my profile.

That was the question, how to implement those layers. Why complain about terminology, what does it contribute? Can you recommend specifics?

The question was too underspecified to give a good answer. It's the equivalent to asking "can I send data over the internet"? The answer to the question as-asked is "yes", but the details will be substantially different depending on what you're actually trying to do.

The question was precisely much more specific than "data over internet".

They indicated that they had UDP packets, which is the specification of the application layer, it means that they have an application that already works with UDP and they just want to send them over the radio, bypassing WiFi protocols/cards etc, from an SDR to an SDR directly. It means that whatever solution is, if it can work with UDP, if the underlying layer translates data into UDP, it will work with their application. That's a very precise specification.

> from an SDR to an SDR directly

That wasn't actually stated. Maybe OP has a bunch of laptops with WiFi cards and no access point and wants to broadcast data to them with the laptops listening in monitor mode. Or maybe OP has a bunch of laptops connected to an access point and wants to broadcast to them using IP multicast. Or maybe OP really does have a bunch of software-defined radios and wants to directly encode UDP packets into IQ samples. Don't you think those would be three very different solutions?

You definitely can broadcast packets via wifi. I mean, that's essentially what you're doing when talking to an AP. All the other wifi antennas/chips in the area can see the data you're sending. You'd have to do some hacking on the receive side to get the data out, but that should be pretty straightforward.

To receive, just set your interface to monitor mode, set the right channel, dump and filter. tcpdump works fine, though you probably want to read and process the data with a custom application.

You can use SDR to transmit any I/Q stream you want. I/Q is just a stream of two numbers for the strength of two phase-shifted carrier waves. How you encode your data (UDP packets) in the I/Q stream is up to you. SDRs like the LimeSDR Mini support sample rates of up to 30 MS/s, therefore it should be perfectly possible to achieve high throughput rates. On the receiver side you will also get an I/Q stream. The problem though is that it is not going to be the same of the transmitted I/Q stream. There might be a lot of noise, the carrier may be slightly shifted, the phase is not the same. Therefore you have to use some modulation to encode and decode the data. It is really up to you what kind of modulation to use. Some traditional ones are OOK (on-off keying), and some of the more complicated ones are APSK. LimeSDR Mini comes with example Python code to transmit raw I/Q. RTL-SDR should be enough for a passive receiver, but it is limited to 3.2 MS/s.

Wifi is indeed the easiest route, you will need a card supporting injection mode:

https://befinitiv.wordpress.com/wifibroadcast-analog-like-tr...

When I was messing around with ADS-B stuff for work I found this website to be very helpful with lots of ideas/links for projects: https://www.rtl-sdr.com/

I have some other questions:

1. Can Usenet (and perhaps also Unusenet) be used with amateur radio?

2. Can received AM/FM broadcasts be used as a input for SoX?

3. Is time synchronization possible with software defined radio?

4. Can it be used with a RF cable (such as to connect to a VCR)?

For 1.) Yes, with caveats. No encryption would be allowed. Typical data transfer rates are abysmal so it wouldn't be great for transferring files. And it might get interesting legally if you do transfer a file with music. I believe that would be broadcasting music which isn't allowed.

2.) Haven't tried this specifically, but I'd say almost certainly yes. Gnuraduo is all about creating audio processing pipelines.

3.) Yes, almost definitely, though there might be some technical aspects to it depending on what you're trying to accomplish.

4.) Yes, but that would likely require extra unnecessary steps.

The caveats you mention would probably be irrelevant with text newsgroups. However, what I am asking is if there are specific protocols for transferring Usenet articles with amateur radio, and how those protocols work.

One project I want to do with SDR is relay + amplify my car smart key so that I can open my car from 50 feet away (as opposed to needing to be right next to it)

Is this possible?

Technically: yes

The tricky part is having your key-fob close enough to a transceiver that can then rebroadcast it's signal. Which would no doubt run afoul of broadcast regulations.

Devices like the RTL-SDR work on as receive-only. I'd suggest checking out your local Amateur-Radio club to learn the laws, get a license, and have fun with projects like that.

Key fobs are at 315, an ism band, so 0.5w is fine for everyone. Some are actually at 433, mostly European stuff. Technically not an ism band in the US but if the fcc doesn't go after VW i doubt they will come after you. Unless you are stealing cars with it.

I'll add that there are no SDRs from ettus, lime, or hackrf that put out even .25w without a separate amp. So he won't break the limit by accident.

YardStick One: software-controlled antenna port power (max 50 mA at 3.3 V)

Is this enough to broadcast a signal stronger than my car key? If not, what do I need to do to get a stronger antenna/more power?

0.055A * 3.3V ~ 0.180w. That is way louder than your key fob.

What would my key fob approximately be around amp/volt/wattage wise?

So recor and look at your car door signal. See how long the pulse train is for one click. Lets call that time N samples. Have a running sum of the last N samples' absolute value. When you see that sum peak above some threshold, switch to TX and play back those N samples plus some on either end for good measure. Newer cars with UWB wont work like this. They do an exchange and measure distance.

I actually don't want to do the "one click" to unlock. I want to do "is the key near the vehicle".

Ah, that is slightly harder. You have to hear the car's challenge and then have a decent antenna pointed at the fob, repeat the challenge, then quickly switch to rx to hear the response, them replay the response. Even with the antenna pointed at the fob you should be loud enough for the car because it is close.

You can even start the car this way. The car will keep running even out of range of the fob for safety reasons.

PSA: Don't rob from or steal cars. That is wrong. Don't be a bad guy. They might have worked hard for what they have and be really nice people.

what frequency does the car transponder challenge the key and what frequency does the key respond at?

is it a rolling code challenge like unlock/lock?

Rolling yes, all on same freq. 315 mostly. Some 433

Do you have an email/IRC I can reach you at?

Also some new ones use ultra wide band. A super short pulse that can tell how far the fob is based on how long it took to respond to the challenge.

Are there any places selling amplifiers? It seems to be a closely guarded secret (because unlicensed operators must be protected from themselves, I guess). I found some but they wouldn't have been suitable for any testing I wanted to do with the LimeSDR I have.

> Devices like the RTL-SDR work on as receive-only.

How would I accomplish this with two YardStick Ones (that do transmit)?

Most sdr don't have a ton of power. But probably. An sdr is Overkill though. You just need an amp and some antennas. This is how those attacks that let people open and start your car when your key fob is on your nightstand work.

Check out “HackRF.

Yea, some people do this to break into cars

PLEASE PLEASE don't do this unless you are licensed by the car thievery union.

Unions aren't libertarian enough - can one get a freelancer's license? Asking for a friend.....

Edit: /sarcasm. PLEASE don't do this.

Yeah it's not a collective either. Maybe it's a bureau?

Unions 2.0 on a social network near you.

go to alibaba and look for range extenders/RFID repeaters designed for car thieves.

Oh great Xmas gift for myself, better receiver :-)

Reply on news.ycombinator.com