I am using an AD9854 Quadrature DDS and upconverting its output to between 700MHz and 1500MHz via a high quality quadrature-modulator ($1250 worth of quad-mod!). The quad-mod, with correct inputs, suppresses the lower sideband and the carrier in its output very well.
However, I was seeing the following in the quad-mod's RF o/p spectrum:
- strong baseband leakage from the DDS
- no carrier suppression
- no lower sideband suppression
I have greatly improved carrier suppression by ensuring that the I and Q DAC current-output DC-levels and signal-amplitudes are exactly balanced via the DDS's OSK function. This also somewhat improved lower-sideband suppression.
However I believe that the I and Q signals arriving at the quad-mod's inputs have some phase-skew causing the baseband leakage and the remainder of the incomplete lower-sideband suppression. Even if I and Q are precisely 90-degrees apart coming out of the DDS, they may not stay that way as they travel through the reconstruction filter and along transmission lines etc.
I used X-Y mode on my scope to obtain a Lissajous figure displaying the phase-relationship between I and Q and I got a very elongated oval shape (a perfect circle is the result of exactly 90-degrees phase difference). I used 1MHz out of the DDS to minimise high-frequency effects on the scope-probes although I will later need to check phase at max DDS output as well (100MHz).
On the AD9854 how can I trim the phase-difference between I and Q?