USRP complex multiplication question

Dear All,
I have a question regarding the first complex multiplication process
with
the IF sinusoid in the USRP. Which of the scheme mentioned below is
actually
happening in the USRP?

Let’s say that two ADCs provide the sampled received signal to the next
stage of processing. Let’s call the sampled stream of ADC_0 as A and
sampled
stream of ADC_1 as B.

SCHEME-1

The NCO produces two sinusoids that are in quadrature phase with each
other.
These are sin(2pift) and cos(2pift). Let’s call the
sin(2pift)stream as C and
cos(2
pift) stream as D. The multiplier after the ADC stage does the
following.

  1. Multiplies the cos(2pifc*t) stream, C, with the sampled data
    stream of
    ADC_0, which is the stream named A and the product is AC.
  2. Multiplies the sin(2pifc*t) stream with the sampled data stream of
    ADC_1, which is the stream named B and the product is BD.

Then sends the resultant product from these two steps as a complex
sample,
( AC + jBD )

SCHEME-2

The NCO produces two sinusoids that are in quadrature phase with each
other.
These are sin(2pift) and cos(2pift). The multiplier after the ADC
stage does the following.

The multiplier after ADC stage then multiplies the ADC outputs with the
NCO
outputs as
( A + jB ) * ( C + jD) = (AC - BD) + j(AD + BC)
and sends the above output as a complex sample.

Which one is true?

Thanks

On Wed, 2011-05-25 at 12:14 -0500, John A. wrote:

Dear All,

I have a question regarding the first complex multiplication process
with the IF sinusoid in the USRP. Which of the scheme mentioned below
is actually happening in the USRP?

I should also clarify – USRPs are zero-IF quadrature receivers
(excepting the TVRX daughterboards). The daughterboards (excepting the
TVRX series) have quadrature outputs and send complex baseband signals
to the two ADCs.

–n

On Wed, 2011-05-25 at 12:14 -0500, John A. wrote:

following.

the NCO outputs as
( A + jB ) * ( C + jD) = (AC - BD) + j(AD + BC)
and sends the above output as a complex sample.

Which one is true?

SCHEME-3

The NCO feeds a (continuously increasing) phase to a CORDIC, which
rotates the incoming samples by that phase value.

–n

SCHEME-3

The NCO feeds a (continuously increasing) phase to a CORDIC, which
rotates the incoming samples by that phase value.

–n

I’m trying to understand the implementation of the CORDIC in the USRP.
Does the CORDIC rotate both I and Q samples by the same phase value? I
would think that the Q samples would have to be rotated by the phase
value offset by 90 degrees for the quadrature mixing.

Thanks,

Rob