[Discuss-gnuradio] Sanity check - BPSK/FSK and IQ sample stream

[Bahn, William L Ctr USAFA/DFCS]

I'm finally getting back to working with the USRP and would like a sanity
check on if I am thinking about something correctly.

Let's say that I want to generate a BPSK signal at a data rate of, say, 1000
bits per second. Is all I have to do is generate an antipodal (+/-V where V
is some constant amplitude) baseband signal at 1000 bps and send that as the
I data (or the Q) and leave the other one all zeros?

Does that result in a BPSK waveform at the carrier frequency? If not, what
does it result in?

If I wanted to do 4PSK, would I just use +/-V on the I and +/-V on the Q as
my four data values thus doubling my data rate for the same sampling rate.
In theory, if I can support 4MSa/s (I/Q samples) across the USB link then I
could get 8Mb/s transmitted. Does this sound right (ignoring issues
associated with receiving and demodulating it)?

What about if I want to do FSK? For example, let's say that I want to use
the frequencies 450MHz +/- 1MHz (so 449MHz and 451MHz). Is there a way to
tune the USRP to a center frequency of 450MHz and send it a corresponding IQ
sample stream? If so, what would that sample stream need to look like? They
way I am presently visualizing it, I would need to tune the USRP to, say,
448MHz and then generate 1MHz and 2MHz waveforms placed on the I (or Q)
leaving the other zero. I suppose I could tune to 449MHz and use a DC value
and a 1 MHz waveform, but that seems like it would result in a pretty good
degree of asymmetry if there weren't very many samples in the bit period.
I'm also concerned about asymmetry in the other case, as well, although it
seems like it would be better. 
 

TIA - Bill

smime.p7s
Description: S/MIME cryptographic signature