Naive software defined radio?

The most straightforward SDR is an ADC with an antenna plugged into it.
But how to make one in practice?

What i want is a receiver for a range of 1Mhz and below, that would produce essentially a recording of all EM waves in this range, kind of like a microphone picks up all the air oscillations to make a sound waveform.
Surprisingly enough, i was unable to google up anything for such a direct approach to an SDR, or anything remotely simple.
That suggests that it's either terribly difficult, or quite useless.

As i understand it, after an antenna there should be a filter for 50Hz and similar loud stuff, then some sort of amplifier to get the signal into detectable range for ADC, and a 2 MSPS or higher (twice the target frequency) ADC.
A byte stream from the ADC would then go into an FPGA/CPLD to be either recorded or processed.

The question is - can it be done, and was it done before?
If yes, are there ready-made schematics, or at least any guidelines on which components to choose and what problems to expect?

Nice idea - it's been done, very successfully!

**broken link removed**

http://www.arrl.org/files/file/QEX_Next_Issue/Jan-Feb_2011/QEX_1_11_Ahlstrom.pdf

http://openhpsdr.org/mercury.php

...and several others. Search for Direct Sampling SDR receivers.

Here are a couple of commercial ones:

http://www.winradio.com/home/g33ddc.htm - this has a 50MHz-wide real-time spectrum analyser!

**broken link removed** - this one plugs into your network. I want one!

An amazing thing to do with these is FFT the whole range and see the full RF spectrum. That's an interesting thing to do when playing with antenna systems too... you can really see the changes you are making.

I've been thinking such thoughts myself recently too. The principles are simple, although I'm not good enough to code the FPGA for that without a lot of research. A 1MHz wide receiver should be fairly easy on the hardware side.

Finding out that a £15 DVB-T receiver dongle (with a modified driver) can be used as an SDR got me started on this track.

FoxyRick - these were being demonstrated and their functions explained at a convention in Basingstoke last week, just for my interest, is that where you found out about them? I have one here.

Brian.

Thank you for replies.
However, these links point to the same kind of overly-complex all-inclusive solutions for 30-120 Mhz upper range, with downsampling, and corresponding rare parts to handle all that.
While i'm looking for a simple, bare thing - antenna, filter, amplifier, ADC, nothing after.

In particular, i'm interested in which ADC to pick - there are literally thousands of them that apparently fit the base parameters.

I'm no expert in this but I think you might have problems with such a simple design. In essence all you do is connect an antenna to a bandpass filter, say 1KHz to 1MHz so you eliminate most power line borne interference. Then digitize at a suitably high rate and run an FFT on the samples to recover the energy spectrum.

The difference in the SDR examples shown is they derive I & Q samples by mixing quadrature oscillator signals with the incoming radio signal. So they are zero-IF receivers with I and Q outputs. Your simple digitizer has no oscillator so extracting I and Q samples would be very difficult. It might also be difficult to isolate a single station using only software. As I said, I'm not an expert so prhaps others may be able to advise better than me.

Brian.

@Brian,

No, although I wish I had been there, it would have been interesting.

I first discovered SDR (as something worth playing with, I already knew of the concept) when I came across the idea of modifying a TV dongle for a PC and using it as a general SDR. For £15 I suddenly had a remarkable little receiver on my PC that would work from 60MHz to 1.5GHz! That got me interested in making a better one and I just researched it from there. There is some fascinating kit available now; some of it so fascinating that they won't even tell you what it does without a letter from your mum (or MI5).

@Artlav,

I've not come across anything much simpler that uses direct sampling, other than VLF techniques of feeding the aerial straight into the sound card on a PC. The basic techniques are very similar though, and easier from a hardware point of view. The software/DSP side is the same.

I suppose that, given the effort of making such a limited receiver, one might as well add a little extra effort and make a much more flexible one.

As the the ADC, you are right, there are a lot to choose from, and a good many of them would be perfectly acceptable at the sort of speeds you are talking about. By the way, you need something after the ADC (I know you know that) and at those speeds, it could be a PC. high-speed USB 2.0, firewire, or ethernet would get your samples in at full speed. That's the sort of thing I have ended up thinking about rather than implementing everything in an FPGA. It might even be possible to use some of the existing open-source software then.

In principle, you can digitize such a wideband signal and then, by digital signal processing, filter your signal of interest. The digital filtering can be as selective as you want provided that adequate arithmetic (i.e. enough number of bits in fixed or floating point) is used.
But a practical limitation of that approach is dynamic range. Your signal of interest may be weak (compared with the power of the whole spectrum you are digitizing) and there is not only quantization noise, by spurious components produced (in the ADC) by strong signals out of your band that can mask your signal.
Regards

Z

Take a look at this video for more information:

www.batc.tv then click on "Film archive", "BATC Convention 2012", select that category. then select lecture 6 "Using DVB sticks for SDR" and click "select stream". It's 16:9 widescreen so I suggest clicking the "full screen" icon under the picture or it will look squashed.

This is one of the Basingstoke lectures, they are all very interesting. I'm the guy on the second row from the back in the audience, on the left next to my better half.

Brian.