Software Defined Radio is fundamentally a different way of looking at radio spectrum

Discussion in 'Amateur Radio News' started by VK6FLAB, May 4, 2019.

ad: L-HROutlet
ad: l-rl
ad: abrind-2
ad: Left-2
ad: HRDLLC-2
ad: L-MFJ
ad: MessiPaoloni-1
ad: L-Geochron
ad: Left-3
  1. VK6FLAB

    VK6FLAB Ham Member QRZ Page

    Foundations of Amateur Radio

    Software Defined Radio is fundamentally a different way of looking at radio spectrum

    We think of radio as operating on a specific frequency. We select an antenna resonant on a single band. We configure the radio for that same band and then turn the dial or the VFO, or Variable Frequency Oscillator to a particular frequency within that band.

    All of our language is geared towards this concept of tuning, of picking out, selecting one special tuned, resonant frequency and listening to it.

    I've said this before, but that's not actually what's happening.

    Your radio is receiving all RF frequencies, all of them, all at the same time, all the time. Your antenna is better at hearing some frequencies than others, but that doesn't stop it from hearing everything at once. Your radio is getting all that RF information at the antenna connector. After that, every step along the way is removing unwanted information, first it removes all the bands you're not listening to, then the VFO selects which part of what remains to let through to the decoder and the result finally arrives at the loudspeaker.

    Ultimately, all your radio lets you play with is what's left over. Say about 3 kHz bandwidth. Using traditional radio, if you want to listen to two repeaters, you either need to switch back and forth quickly, or you need two receivers.

    Now without going into how precisely, imagine an SDR with a bandwidth of 3 MHz, one thousand times larger than your traditional radio. Before you think I'm being fanciful, a $25 gadget can do this. This means that you could process most if not all of the 2m amateur band and then pick out which bits you'd like to decode. You could decode all the local FM repeaters, an overflying satellite, the International Space Station SSTV, a beacon, Morse, Packet, RTTY and simplex contacts, WSPR, APRS, EME, whatever is happening on 2m, all at the same time.

    Let me say that again. All of the 2m band, all at the same time.

    The point is that all this information is there, all the time. We can opt to decode or ignore the information. In a traditional radio, you can only decode one signal at a time, but on an SDR, you can extract as much or as little as your computer can handle. Some SDR language talks about using multiple receivers, but a better description is multiple decoders.

    This means that software defined radio is fundamentally a different way of looking at radio spectrum. Instead of filtering out everything we don't want to decode, we select which decoder to apply to which part of the spectrum.

    With an SDR you could represent the 2m band as a 3 MHz slice of spectrum as a series of measurements. There is no loss if you reuse the numbers, so if you process the same data multiple times, you have no loss of signal, no deterioration, no extra noise.

    All we do is feed the same data into each decoder, pick out the bit we want to decode and have at it.

    There is a misconception that you need serious computing power to do this. That's not strictly accurate. A $5 Raspberry Pi single board computer is more than powerful enough to do this. You can argue that this is serious computing power, compared to what we used to land on the moon it is, compared to your mobile phone, it isn't.

    I fully intend to go into the maths behind this, but it's not scary, despite what you might think or have been taught. My week has been about the maths and it's become clear to me that there are lots of explanations around, each trying harder than the next to scare you away.

    If you feel the need to run screaming for the hills when you hear the words Nyquist, Shannon and Fourier, then get it out of your system and come back when you're ready.

    I'd like to mention that I've been working on how to explain this over much of the week, I've lost count of the number of drafts I've written, but it keeps coming back to the words that are almost as old as I am: My god, it's full of stars.

    No doubt you might be convinced that I've lost my marbles and that I'm going well outside the Foundations of Amateur Radio, but I have to confess, this is what radio is today, and I'm thrilled to be here learning more about how this all works. Hopefully you are just as thrilled.

    I'm Onno VK6FLAB

    TL;DR This is the transcript of the weekly 'Foundations of Amateur Radio' podcast - for other episodes, see
    K0UO likes this.
  2. AC1GX

    AC1GX Ham Member QRZ Page

    I may be wrong, but I think of it as mere measurements of electric field at a high sampling rate. We know any real world waveform can be represented as the sum of a bunch of sine waves. With SDR we look at the entire wave form, and use math to tease out the component frequencies we are interested in.
    With SDR, we can do this at our leisure. We can record our sampled waveform. And we can go back and decode any of the various transmissions which were strong enough to affect our recording.
    This is my understanding anyway. I'm not clever enough to actually do it myself.
  3. KA0HCP

    KA0HCP XML Subscriber QRZ Page

    Basic disagreement. An SDR radio does NOT imply that it has a spectrum display. Many radios today are SDR of one flavor or another, but do not have any spectrum display.
  4. WQ4G

    WQ4G Ham Member QRZ Page

    Not really a different way of looking at radio spectrum... It's a different way of looking at radios.

    Nothing about the spectrum has changed. It is there just as it always has...

    But, the radios (hardware) we use to access the spectrum has changed (is changing). Software and computer chips are taking the place of master oscillators, mixers, and tuned circuits. And, it is computer hardware and software, not spectrum, that makes it possible to do those things which were not possible before. Being able to receive and listen to (to use) 100 different frequencies all at the same time is a result of the hardware and software advancements that have been made (in the radio). It is not the result of any changes in the spectrum, or the way in which we measure the spectrum, or how we view the spectrum.

    Dan KI4AX
    KA0HCP and KP4SX like this.
  5. KQ9J

    KQ9J Premium Subscriber QRZ Page

    The vast capabilities of modern software defined radio cannot be overstated.

    But aren't you really doing the same thing either way? You can use your analog rig to filter out what you DON'T want to process, or you can use your computer to select what you DO want to process.
    Seems to be sort of the same end result :)
    AC0GT, K4AGO and WQ4G like this.
  6. SM0AOM

    SM0AOM Ham Member QRZ Page

    There are not "fundamental differences".

    The underlying principles behind SDR are simply the fact that there is a time domain dual to every action taken in the frequency domain.

    An SDR does indeed "look" at the whole spectrum from DC up to the Nyquist frequency at once, but uses operations in the time domain to select and demodulate the part(s) of this which are of interest.

    Features like a spectrum displays and very steep bandpass filtering are easier to realise using SDR technologies than by analogue, but are by no means exclusive.

    However, fundamental properties such as strong signal handling and spurious responses are valid also in the SDR realisations.

    KA0HCP and WQ4G like this.
  7. IK6UGD

    IK6UGD Ham Member QRZ Page

    The SDR will also be the future in our radios. I'm using the hackrf which has a bandwidth of 20 mhz and I can see graphically and listen to everything that happens in HF.
    It is able to receive and transmit with 100mW from 3 mhz to 6Ghz in all modes and to the bandwidth you want.
    inside there is not a coil, not a software-only trimmer.
    73 ik6ugd Claudio
    G7VJH and AC0GT like this.
  8. IK6UGD

    IK6UGD Ham Member QRZ Page

  9. W1YW

    W1YW Ham Member QRZ Page

    Yes, that. FLAB is lucky Ulrich didn't see this....there are so many wrong statements here (in the OP), its hard to know where to begin.

    Congrats to Ulrich Rohde, N1UL for his recent CAS IEEE Award for inventing SDR.
    WQ4G and K4AGO like this.
  10. WB6CXC

    WB6CXC Ham Member QRZ Page

    I think you two are being overly critical. Sampling the broadband RF signal and doing the processing digitally in the time domain is indeed significantly different than the "analog VFO/mixer/filter/demodulator" design of traditional radios. You might argue with the term "fundamental" but that's a nit at best. And of course you *could* build a broadband analog receiver and use multiple VFO/(etc) stages to demodulate multiple signals in that band,or drive a panadaptor but so what?
    And yes, the front-end (and elsewhere) issues of noise and linearity remain an issue for SDR as well as traditional radios, but that doesn't negate anything the OP said.

    W1YW, could you please point some one of these wrong statements? I'm seeing simplifications and analogies, but no real mistakes.
    AC0GT likes this.

Share This Page