On 09/01/2010 06:47 PM, Sam K. wrote:
I’m not sure what you’re asking for, but I’ll take a stab at guessing
I’ll describe what the USRP2 does on the receiver side:
o Complex samples the input (which is complex, that is I + Q,
baseband, typically) at 100Msps
o Filters and decimates that complex sample stream down to whatever
sample rate/bandwidth you wish to appear across the
1GiGe interface (and hence into your Gnu Radio application).
Your Gnu Radio application then does whatever it does with
than complex-sampled stream. That stream is a time-series with
fixed and uniform timing, so any discrete-time-series “math”
you want to do on that stream will “work”.
o Both the sample clock and synthesizer clocks can be synchronized to
an external source, via the 10MHz SMA inputs, or via
the so-called “mimo bus”.
o The FPGA “assists” in the programming of the various daughter-cards
programmable elements, like the PLL synthesizers and
variable-gain elements in the gain chain.
The transmit side is the logical reverse, with the D/A sampling at
200Msps, and the FPGA interpolating your application data stream
as appropriate, and presenting it as a complex sampled baseband
stream to the D/A. From there, it is presented to complex mixers
on the daughtercard to produce the desired final RF signal.
I don’t know if this even comes close to answering your question.