AD-FMCDAQ2 / AD9680 ADC Capture Instability at 1 GSPS

I am currently experiencing an issue with the AD-FMCDAQ2-EBZ when operating the ADC at 1 GSPS.

The board uses the AD9680, a dual-channel, 14-bit, 1.0 GSPS JESD204B ADC, together with the AD9523-1 clock generator. According to the ADI documentation, the AD-FMCDAQ2-EBZ is designed to support 1 GSPS operation, and the reference design uses a 1000 MHz clock for the ADC. However, in my setup, the ADC output becomes inconsistent when operating at 1 GSPS. During repeated captures, I sometimes obtain mostly noise instead of the expected signal. In other captures, the signal is present but is significantly noisier than expected, and the waveform varies considerably from one capture to another.

I have attached three captures showing the ADC output at 1 GSPS:

As can be seen, the captured waveform is not consistent between captures.

For comparison, I performed the same test at 500 MSPS, where the captured signal is significantly more stable and repeatable:

Since the AD9680 and AD-FMCDAQ2-EBZ are specified for 1 GSPS operation, I would like to understand what could cause this difference between 500 MSPS and 1 GSPS.

In particular, I would appreciate advice on the following:

1. Clocking: Are there any specific AD9523-1 configuration requirements or clock-quality considerations that are particularly important for stable AD9680 operation at 1 GSPS?

2. JESD204B: Could the higher ADC sampling rate expose a JESD204B link configuration, lane-rate, or FPGA transceiver timing issue that would not appear at 500 MSPS?

3. ADC configuration: Are there specific AD9680 settings that should be checked when switching between 500 MSPS and 1 GSPS?

4. Reference design: Are there any recommended ADI reference-design settings or register configurations that should be used as a baseline for 1 GSPS operation?

5. Analog input/clocking: Could the instability be related to the input signal level, clock amplitude/quality, or other analog considerations that become more critical at 1 GSPS?

6. Known issues: Are there any known issues with the AD-FMCDAQ2-EBZ or AD9680 that could cause intermittent/noisy captures specifically at 1 GSPS?

The fact that the same setup produces a stable waveform at 500 MSPS but an inconsistent waveform at 1 GSPS makes me suspect that there may be a configuration, clocking, JESD204B, or FPGA-interface issue rather than simply an increase in ADC noise.

Any suggestions for troubleshooting or recommended measurements/configurations would be greatly appreciated.

Thank you.