On 01/29/2011 12:37 PM, Phil Behnke wrote: > Thanks for the tips! I'm definitely going to investigate hooking the > ADC directly to the FX2. Looks like there are some inexpensive dev > boards for the FX2 on ebay, although I have no idea how good they are. > I figure I will try to develop the receiver to always grab 1MHz worth > of bandwidth, and then give the user finer filtering abilities by > using digital filters on the PC side. > That's probably a reasonable approach. The FX2 is actually a fairly-capable chip. It's used in the USRP1 from Ettus Research, which can "pump" upto 16Msps(complex, 8-bit) over USB-2.0, so 1Msps should be more than doable.
You have to make certain that your I and Q lines are low-pass filtered, fairly stiffly, prior to sampling. If you're sampling at 1Msps, the I and Q lines need to be low-pass filtered to 500KHz. There's a nice passive-filter designer on-line at: http://www.wa4dsy.net/filter/filterdesign.html I've used it for other radio astronomy projects in the past Other things to keep in mind: o you'll need enough low-noise gain ahead of the down-converter to make up for the *terrible* noise figure that's typical of mixers and ADCs. At HF, in radio astronomy, you'll probably need about 50dB. o You'll need a good bandpass filter at RF. Again, the above-mentioned site should help here. A good approach is to use a amp-filter/amp-filter/amp-filter topology, which gives you distributed gain and filtering, and makes the individual filter stages manageable. If you designed your bandpass RF filter for an Fc of 20Mhz, and bandwidth of 5Mhz or so, it'll improve your usable dynamic range, and prevent driving your gain stages into compression, due to "other muck" that your antenna will inevitably "see". -- Principal Investigator Shirleys Bay Radio Astronomy Consortium http://www.sbrac.org _______________________________________________ Discuss-gnuradio mailing list [email protected] http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
