Hello,
I'm using the Abaco FMC120 which uses an ADS54J60 ADC. When I provide an input at the center frequency, the amplitude varies wildly. Other input signals work just fine.
ADS54J60 is being clocked at 1GHz, operating in second Nyquist, which puts center frequency at 750Mhz. I also see the same issue in first Nyquist when supplying an input at 250MHz.
First Test:
- Signal Generator configured to produce 750MHz at 50% duty cycle (500ms/1000ms), 0dBm
- Spectrum Analyzer configured for Zero Span at 750MHz, sweep time of 5ms. Averaging is enabled.
- What I see: After the pulse falling edge, the signal does not immediately fall to the noise floor, but instead decays over a few milliseconds.
Second Test:
- Signal Generator configured to produce 750MHz at 100% duty cycle, -15dBm
- Spectrum Analyzer configured for Zero Span at 750MHz, sweep time of 1000ms. Averaging disabled.
- What I see: Instead of a smooth line, I see a lot of ripple in amplitude.
Third Test:
- Repeat of the first test, but at 751MHz.
- What I see: After the pulse falling edge, the signal drops immediately to the noise floor
Fourth Test:
- Repeat of the second test, but at 751MHz.
- What I see: A very smooth line in amplitude.
This behavior seems to only occur within +/- 50kHz of the center frequency. The attached plots were created by wiring the ADC to a DAC output in an FPGA with minimal processing.
I found a post on this forum with details on how to make changes to the DC Offset Correction (Freeze and Unfreeze), Interleave Engine (Freeze, Unfreeze, Bypass, and Read/Write Coefficients). When I froze the DC offset correction, I saw the behavior improve.
Any suggestions on what I should do to address this issue? Pulsed CW at center frequency is a common use for my application. What are the tradeoffs of freezing the DC offset correction? What kind of performance can I expect? Thanks!
