close-circle
Close

Fundamentals of RF HackChat Transcript

A event log for Fundamentals of RF Hack Chat

How do bits get turned into radio waves and back to bits? We'll be discussing RF modulation, theory, and other basics in this chat!

ShaynaShayna 11/03/2017 at 19:051 Comment

Sophi Kravitz : HI everyone, we're going to get started

Sophi Kravitz : @Nick, welcome to the chat!

Nick : Hello everyone!

Sophi Kravitz : Can you introduce yourself and tell us a little about the kinds of things you like to work on?

Sophi Kravitz : @Nick will be discussing the Fundamentals of RF.

Sophi Kravitz : If you're new to the chat, welcome!

Sophi Kravitz : We add questions and discussion topics to this sheet: https://docs.google.com/spreadsheets/d/14DApYIwcEGZl9mGEpwNx7DsVLW19e6A74DpYT5C6mO8/edit#gid=0

Nick : Sure! I'm Nick, information security engineer by day, with a background in network engineering and Linux systems administration. I've also been an amateur radio operator since 1999 (recently upgraded to an extra-class license) and have been playing with electronics as a kid. I like to sit back and watch the waves resonate in my antennas. At home I fiddle with radios, SDR, electronics, RC aircraft, and whatever other projects I can find to start but not finish ;)

Sophi Kravitz : Does anyone have a really basic question?

Sophi Kravitz : (because nothing on that sheet looks basic)

Chris : What is RF?

davedarko : my book question is basic

Robert Marosi : Do you use Smith charts in your work, at all?

Michael Harpe : Can you describe how the signal flow in a DSP receiver is implemented? I understand that there's an A/D converter and then we do an FFT but how do I actually do anything useful with that raw data

Nick : RF is Radio Frequency. It's an electromagnetic wave you get when you wiggle electrons back and forth. That wave will propagate, and stimulate electrons in resonant conductors. The rate at which you wiggle the electrons is the frequency.

Nick : @davedarko I started out with ham radio, there are study guides for the technician-class license. I highly recommend starting out with that, it gives you a good opportunity to learn more about RF and electronics basics. Beyond that, check out Crystal Set to Sideband, a course that takes you through constructing the most basic radio receiver to a single-sideband transceiver: http://www.wa0itp.com/crystalsetsssb.html

Nick : @maclean.at basic project for kids, check out Crystal Set to Sideband as well, the earlier projects there will be a good place to start. Building a crystal receiver is easy, and kinda neat to pull radio signals out of the air and hear them (especially cool as it doesn't need any sort of power supply!)

Nick : I'm going to pop some notes in the question spreadsheet so I remember what I answered :)

davedarko : cool :)

snp.ididnotsignupforthis : Are there any brands of Crystal Receiver sets you recommend? I'm homeschooling an freshman who is very into electronics and tech.

maclean.at : Thanks @Nick

Soul_Est : @snp.ididnotsignupforthis Please post your question in the document here: https://docs.google.com/spreadsheets/d/14DApYIwcEGZl9mGEpwNx7DsVLW19e6A74DpYT5C6mO8/edit#gid=0

Nick : @snp.ididnotsignupforthis No specific sets/kits that I know of. There are a few on Amazon though they don't seem to have good reviews, not sure why.

Sophi Kravitz : @Nick what is I/Q?

Chris : @Sophi Kravitz oh that's a good one

Nick : Pax has a question about I/Q data...I/Q is in-phase and quadrature, which is another way of representing a periodic signal like an RF signal. It breaks it apart into real and complex elements, and makes it easier to work with mathematically.

Frank Buss : when I tried to build an AM receiver as a kid from a kit, it didn't work, can be difficult with no help, could lead to bad reviews

Nick : A couple ways to visualize it...the in-phase signal is the original signal, and the quadrature signal is phase-shifted 90 degrees.

Kevin : For getting started you can also check out books and projects from Doug DeMaw who used to be a regular contributor to QST magazine in the US.

Wilfredo Velázquez-Rodríguez : I don't know if this question is too low-level, but one thing I've never understood about RF (and I suppose signal transmission, period) is how data is encoded in those transmissions. Looking at an analog system, how does a receiver for say, a radio, interpret the signals and turn them into audio? The information is not encoded in the frequency, because frequency is used to filter (tune), and it's not encoded in the magnitude, since that would be volume, so, how is it done? I very well may be wrong with some of my assumptions here.

I'm going to post this question in the document, as well. I think it's related to row 6.

NickIf you take a sine wave and project it onto an X/Y plane where in-phase (real) is X and quadrature (complex) is Y, as time progresses you'd see a point moving in a circle around that plane.

Sophi Kravitz : So I/Q is a data representation?

Sri Gogineni : How does RF apply to technologies like Wi-Fi? Is there a book you recommend for wireless/RF fundamentals?

davedarko : @Wilfredo Velázquez-Rodríguez watch this :)

Nick : Yes, you can take an RF signal and pass it through an I/Q decoder (which just spits out the original and 90-degree phase-shifted signal) and then sample those components to do work on. You can also go the opposite direction, generate in-phase and quadrature components, combine them, and then you have your RF signal.

KBP : Stupid question: does higher the frequency give higher bandwidth?

Nick : Mathematically it makes things easier to work with. It also gives you some flexibility in SDR architecture. With an SDR, you have to sample the signal with an analog-to-digital converter to turn the analog waves into bits. Now, if you sample just the original signal, Nyquist says you need to sample at twice the frequency, otherwise you'll get aliasing effects. However, if you're sampling I and Q separately, you can sample each AT the frequency of the signal. So you have a design decision between one faster ADC, or two slower ADCs.

Kevin : @KBP: Frequency and bandwidth are not directly related. When you use higher frequencies you are often allowed to use signals requiring more bandwidth.

Nick : Further reading/watching on I/Q:

http://whiteboard.ping.se/SDR/IQ

(I REALLY like w2aew's videos)

KBP : Thanks.

Nick : @Henrik Enggaard Hansen does that answer your question about how to sample signals as well?

KBP : How flexible can a SDR be? What limits them in terms of frequency capability?

Wilfredo Velázquez-Rodríguez : @davedarko That is an excellent link. Thank you

davedarko : sure :)

Wilfredo Velázquez-Rodríguez : Thank you for your responses as well, Nick.

Nick : @kbp @kevin they are tangentially related...it is a lot easier to design a system with 40MHz bandwidth when you're operating at 2.4GHz than it is when you're operating at 30MHz. The higher you go, the smaller the same bandwidth difference is in relation to the frequency itself...hm...not sure I'm explaining that well.

Nick : @Charlie Lindahl afraid I don't have any experience with LoRa, or the service providers that offer LoRa data transit.

Nick : @programagor Can't help you with distributed element circuit design, but this will help you with routing microwave traces: https://www.microwaves101.com/encyclopedias/mitered-bends

Kevin : @Nick, Right. I wasn't going to get in to it in too much detail at the moment. For any given frequency you need to check with the local (Federal) government organization governing radio to see what bandwidth is allowed at various frequencies.

Sophi Kravitz : Just paged @w2aew

Nick : @kevin yep

KBP : thanks to both. kevin and nick

Kevin : Another suggestion for anyone wanting to get in to RF (ie. design/build circuits), start with lower frequencies (such as 30MHz or lower). RF design gets in to "black magic" when you start going up in to UHF frequencies and higher.

SG : hi @Nick , how does RF apply to technologies like Wi-Fi / bluetooth / LoRA? Like how does one go from radio / ham radio to these higher frequencies that are "black magic"?

Nick : @themartinm For getting started, I recommend KiCAD (check out @Chris Gammell 's Getting To Blinky series, https://contextualelectronics.com/courses/getting-to-blinky/ ) @oshpark is great for prototype boards. As with all things hardware, compiling atoms takes a lot longer than compiling bits ;) So unfortunately there is time in board re-spin. You can also look into Manhattan style circuit design for some prototyping work on lower-frequency circuits (https://hackaday.com/tag/manhattan-style/) Also Mike Ossmann's talk on RF circuit design

Nick : For hams in the group, take a look at the BiTX40 radio. It's a 40 meter SSB rig, mostly surface mount (comes assembled already except for a few connectors and controls), but it's laid out in a way to let you see each individual block of the radio and get a better grasp of the components in each block.

davedarko : how low can you go in frequencies?

davedarko : (sorry for throwing in, should be a quick one though)

Nick : @davedarko DC to Daylight :) (meaning anywhere from DC, up to light wavelengths)

davedarko : in Hertz :D ?

Chris : Heh that's a cool way to put it.

Kevin : @davedarko In North America amateur radio operators are now being allowed to use some frequencies below 500kHz. There are aircraft navigational beacons operating down as low as around 100kHz or so.

Nick : you need a REALLY REALLY REALLY BIG antenna though :D

davedarko : nice :D

Kevin : :)

themartinm : thanks for the tips @Nick !

Frank Buss : at about 8 Hz you can receive the Schumann Earth Resonance

Chris : What is that?

Kevin : Could be fun trying to do R/C control of a submerged model submarine if you have to use ELF frequencies. :)

Frank Buss : https://en.wikipedia.org/wiki/Schumann_resonances

Jose Marcelino : How do you choose which digital modulation to use? Why isn't everything using QAM or BPSK. Is it because of complexity of the receivers/transmitters?

Chris : ah thank you, sorry im in class googling isn't really an option

Nick : @Foalyy For bandwidth, if your center frequency is 100MHz and your bandwidth is 10MHz, ideally yes, you'd be using 95-105MHz. Depending on how well designed and filtered the transmitter is though, you might have harmonics or other spurious emissions elsewhere in the band, and you'll probably leak over the edges of that chunk a bit. It's not like a discrete square block, think of it more like the cross-section of a speed bump (or road hump depending on where you're from).

Foalyy : some kind of bell curve I guess? Thanks for the explanation :)

Nick : Michal's question about ILS...it looks like it has a set of VHF transmitters modulated at different frequencies for position, and a set of UHF transmitters in the same arrangement for glide slope. I would guess the ILS receiver is looking for a point where the beat frequency between those two modulations is maximized, indicating you're right between them. I'm not too sure though. http://instrument.landingsystem.com/ is what I just skimmed.

Foalyy : quick follow-up : how does this RF bandwith relates to the usual "bandwith" use of the word (in kbps)? Does it only depends on the modulation? Is there a theoretical limit?

Nick : @Foalyy Aha, now we get into information theory! (yay Claude Shannon!)

Kevin : :)

Nick : (btw, there was just an interview with the author of a biography about Claude Shannon on the Embedded.FM podcast by @Elecia White and Christopher White

Nick : ok hang on, let me get my notes for this one

Foalyy : maybe that wasn't such "quick" follow up question then!

Nick : Hehehe. Indeed. It's not, but it IS fascinating.

Nick : Going to list my references first: https://en.wikipedia.org/wiki/Channel_capacity https://en.wikipedia.org/wiki/Shannon%E2%80%93Hartley_theorem

Kevin : @Foalyy The short answer is that modulation does affect the bandwidth for a signal. The how it affects it for a given modulation would not be a short answer.

Nick : There is a theoretical maximum throughput capacity of a given data channel depending on the RF bandwidth and signal/noise ratio. This is roughly represented by Capacity ~= 0.332 * Bandwidth (in Hz) * 10log(10)Signal/Noise

Nick : So if we have a signal with 20MHz bandwidth, at -70dB, with a noise floor of -90dB, we have a theoretical maximum of 86mbit/sec

Nick : If we improve our S/N ratio from 20dB to 50dB, same bandwidth, we'd get up to 112mbit/sec

Nick : those logarithms will get you every time

Nick : If we double the RF bandwidth though, we can double the throughput. 50dB SNR, 40MHz bandwidth, 225mbit/sec

Nick : (and yes I had these written down ahead of time)

Nick : As far as modulation itself goes, yes, different modulation schemes will have different throughput. They'll also depend on better/worse SNR, maybe utilizing forward error correction to overcome poor signal conditions, etc. There are a lot of variables that go into throughput of a signal, unfortunately I've not studied them all to give more detail there.

Foalyy : really interesting to see how noise impacts bitrate, and that capacity is proportional to the RF bandwith. Thanks a lot Nick!

Sophi Kravitz : Question from Shlomo: How would you choose an antenna for transmitting through a person? wearables etc

Nick : ok, @Sri Gogineni So, wifi is still RF, just up in the microwave band. It's attenuated much more easily by materials, reflected more easily, etc. What sort of detail are you looking for as far as getting into wifi and other 2.4/5GHz stuff?

Frank Buss : isn't it dangerous to transmit through a person?

Sophi Kravitz : there seem to be several products (medical) that transmit through people

Sophi Kravitz : cameras etc

Ramon Schepers : depends on time and power transmitted

Nick : @Shlomo Zippel I'm not sure. Humans are hard, RF-wise. We're big squishy bags of mostly water, and water is not RF's friend. It attenuates stuff quite a bit. You'd want to get some human analog material in a roughly human shape into an anechoic chamber and test different antennas for radiation patterns. I'm sure there are modeling software packages, but I'm not familiar with what they are, as I'm guessing they're all well outside my price range :)

Sri : @nick if its the same fundamentals RF. How does range and bandwidth increase as you go from radio -> wi-fi -> 4G LTE

Frank Buss : I always wear my tinfoil hat when I use my mobile phone :-)

Nick : @Frank Buss depends on power, time, frequency, and where on the person it is. Different parts of the body absorb RF in different ways. Note that RF radiation is NOT ionizing radiation (such as x-rays, gamma particles, etc) and does not knock atoms off your DNA. The effects of RF radiation on the body are purely a heating effect. Oh, and maybe you'll hear radio stations in your teeth if you have dissimilar metallic tooth fillings touching each other. :)

Sri : I think I understand some different strategies for modulation (how to package/transmit data) and encoding strategies, but is there a underlying principle besides add more channels / bandwidth and find a frequency that doesnt interfere with materials?

Shlomo Zippel : @Nick thanks! Yeah, trying to balance 2.4ghz which has all of the FCC modules and plenty of antennas to choose from -- with the fact that 2.4Ghz is terrible for going through squishy water :)

Kevin : @Frank Buss, @Ramon Schepers It depends on duration, power level, and frequency. There is a lot of debate ongoing about the safety of the use of cell phones being held near the head.

Shlomo Zippel : *FCC certified modules

Nick : @Shlomo Zippel See if you can find modules in the 900MHz ISM band, it's got better penetration characteristics.

Frank Buss : isn't a hair dryer much more powerful regarding induced energy?

Nick : Yep. Harder to modulate for sending signals though :)

Nick : @Sri Gogineni so wifi vs LTE...LTE is lower in frequency (700/800/900/1900/2100MHz I think) and higher power output than wifi, but you have a problem of lots more people using it. Wifi is Carrier Sense Multiple Access....which basically means all stations are trying to talk at the same time and stepping on each other.

Nick : LTE does Code Division Multiple Access which lets a bunch of stations talk on the same frequency, and does some fancy math to pull each individual station out of the mess. I'm not familiar enough with how it works to explain.

Sri : @nick i think im with you.. so people developed different strategies for 'talking over each other' like TDMA, CDMA, FDMA

Nick : @David For rtl-sdr stuff, I like doing aircraft monitoring (shameless plug for a podcast I was on about ADS-B receive, http://hamradio360.com/index.php/2017/06/13/ham-radio-360-ads-b-with-nick-kk6lhr/ ) Also check out reddit.com/r/rtlsdr There's a lot of getting-started stuff in the sidebar.

Mariano : @Nick I think LTE = Magic so complex but works..

Sophi Kravitz : We're going to wrap up in 5 minutes- please feel free to stay and chat some more afterwards! If you're just arriving, @Shayna will be posting a transcript soon

Nick : @Sri Gogineni yep. TDMA had an interesting side effect...remember the days when you'd hear the "BZZ BZZ BZZZZ BZZ BZZ BZZZZZ" in your computer speakers when your phone was nearby?

Sri : Yea, you still heard that a lot when travellng around some countries :)

Nick : TDMA worked by splitting up a channel into multiple time slots. You would get to talk one out of every, say, 32 time slots. But the act of turning the radio on and off that quickly created some spurious emissions which were picked up by audio amplifier circuits.

Sri : I thought it was your phone checking in for messages and such and nearby speakers happened to resonate

Nick : @Mariano It's just math :) Granted, it is math way over my head though.

Sri : So the rapid on-off of the radio creates EM waves that get picked up by nearby speakers?

Nick : Yep.

Sri : Gauss all over again

Frank Buss : not the speaker, the amplifier, and only the badly designed ones :-)

Frank Buss : for the speaker you would need much more power

Sri : @Frank Buss added digging into speakers and amplifiers on my todo list :)

Nick : Ok! I've gotta run, need to get back to work. Hopefully I got everyone's questions answered! Hit me up on Twitter if you like, I'm @explodinglemur :)

Discussions

Neil K. Sheridan wrote 11/03/2017 at 23:02 point

I only came across LoRaWAN last week, since it was mentioned, this is stack https://www.thethingsnetwork.org/wiki/LoRaWAN/Home

  Are you sure? yes | no