Is anyone playing with Software defined radio?

[Photonic]

I'm just wondering. I've got a little TV dongle coming soon and I plan on seeing what I can do with it.

[lvictorlucas]

I'm just wondering. I've got a little TV dongle coming soon and I plan on seeing what I can do with it.

I have designed many during my working years. I don't know anything about your TV dongle but I have seen what appears to be a similar item at WalMart for about $10.00 that I assumed just contained the internet addresses of live TV stations. If you decide to do a software radio one of the first decisions is to design or buy a front end that outputs RF, IF or baseband. The RF front end will contain a low noise amp, filter, AGC, and down converter. IF contains amps, filters, AGC and final down conversion to baseband. The baseband will contain more amps, filters, AGC and a/d converter. I know this sounds like a lot of stuff you can skip, and I did design a 2.4 GHz direct down conversion system for 802.11b, but if you want to learn it is the way to go. If your signal is fairly narrow band (audio bandwidth) you could sample at IF but to start I would suggest sampling at baseband. That way you have a lot more time to process each sample. To start assume you know the modulation parameters i.e. AM broadcast radio. This way you don't have to build a transmitter, have a good strong signal with a piece of wire and don't have to figure out the modulation that is usually done with FFTs and a lot of higher order non-stationary non-Gaussian statists. Demodulation, carrier and symbol tracking is next and your done. To start I would also suggest assuming a continuous signal that is above the noise floor and skip as mush acquisition stuff as possible. Acquisition is the hard part of almost all radios and assuming a continuous signal with known parameters, no transec or encryption, and above the noise floor makes life easy to start. Most of the systems I did were some type of digital modulation QPSK, QAM, OFDM, or more complicated ones. I would also suggest you do the coding in MatLab. It will save you a lot of time over C or Fortran. Free knockoffs are available on line. I did not use Simulink as I usually did my own custom non-linear signal processing, but Simulink is great for undergraduate courses where you don't want to get into the details. I usually go to the Wednesday evening get to gathers if you want to discuss your ideas. If you just want to send data a system like 802.11b is very easy and you would learn a lot doing an entire system. -Vic

[dkoop]

I am just getting interested in SDR and that is one of the reasons I searched out a Spokane Maker Space. Here is a couple of leads that I am following:

http://www.kickstarter.com/projects/moss...f=category

Interesting, but if you are on this site, you probably want something that you can make. That Kickstarter is for an open source SDR - all of the schematics, designs, and code can be found here:

https://github.com/mossmann/hackrf

Hope this helps!