Category: Hardware
Product Number: AD9175
I am testing a JESD204B-based RF signal chain and observing an unexpected spectrum at the output.
My setup is as follows:
- MPSoC generates a complex sine/cosine waveform using a DDS Compiler IP (1 MHz baseband, sysref = same jesd204b tx_coreclk) .
- The I/Q data is transmitted via JESD204B (Mode 0, 12 Gbps) to the AD9175.
- The AD9175 operates at a 300 MSPS input data rate (with interpolation x16 from a 4.8 GHz DAC clock).
- The internal NCO is used to upconvert the signal to around 1 GHz.
- The RF output is measured using a spectrum analyzer.
Expected result:
- A clean single-sideband tone around 1 GHz ± 1 MHz.
Observed result:
- The spectrum is spread over a wider bandwidth with a “hill-shaped” profile instead of a sharp tone.
- There is significant spectral spreading and possible image components.
- The signal is not a clean CW tone.
What I have already verified:
- DDS output (I/Q) is correct.
- JESD204B link is stable (no errors, ILAS successful).
- SYSREF is configured as one-shot.
- tvalid/tready handshake is stable (tready always asserted).
- I/Q swapping does not change the result.
- Clock sources appear stable.
Question:
What could cause this kind of spectral spreading when using complex I/Q input with the AD9175 NCO?
Could this be related to incorrect complex datapath configuration, I/Q mapping, or mixer/NCO settings inside the AD9175?
I don't really know why this result came out.
Any guidance on what to check in the AD9175 configuration would be appreciated.
