Tool/software:
Hi Teams,
1. Use chip: ADC12DJ5200RF
2. Configuration mode: Dual channel 5GSPS
3. 204 interface configuration: Use JESD204B 8B/10B, JMODE3
4. Analog signal input: 5G/2.5G/1.25G/0.625G/0.3125Gbps, 5 levels in total
5. Phenomenon description: The customer uses this ADC for data acquisition after coherent receiver (ICR). When the input signal is at the highest level of 5Gbps, it is in QPSK debugging format, and both the I and Q channels are at 2.5Gbps, thus meeting the 2x sampling relationship.
The low-speed fourth gear is in BPSK modulation format, and the customer has not changed the ADC's working mode, fixed to use dual channel 5GSPS mode.
For the initialization configuration of ADC, customers can refer to section 7.3 of the IC datasheet.
Now it has been found that the sensitivity of the received signal varies greatly from the theoretical value of the algorithm (about 3-7dB), and the difference widens as the input signal rate decreases. For example, at 5G speed, there is a 3dB difference from the theoretical value of the algorithm, and at 0.625G, there is a 6dB time difference.
The sensitivity testing method is to add an attenuator in the link and increase the attenuation until the bit error rate exceeds the FEC error correction limit.
The customer used a high-speed oscilloscope to collect signals from the ICR backend and ADC frontend (positions P1 and P2 in the figure), and performed algorithm processing on the data in Matlab. It was found that there was also some degradation here, but it did not cause such a large difference in sensitivity.
The data collected by the ADC received by the FPGA was exported to an ILA file through VIVADO and processed by an algorithm in MATLAB. It was found that the difference was 3-7dB. The data collected by the ADC here introduced significant degradation to the signal quality.
6. Ask questions
The customer only referred to section 7.3 to configure and use the ADC. Is there anything else that can be improved? It can be helpful for their sensitivity issues.
Thank you
BR
