We have an ADS54J60 on an Abaco FMC120. We recorded Gaussian broadband noise (400 MHz bandwidth) with the 16 bit ADC from the ADS54J60 with a sampling rate of 1 GS/s. We then calculated the histogram by counting the occurrences of each of the 2**16 quantization levels. The result is shown below. The expected outcome is a Gaussian distribution. The measured outcome is that every fourth bin is about 2 to 3 times more probable than the adjacent 3 bins. DC offset was frozen for this measurement.
Histogram:
Zoom into the middle region:
Looking at the DC offset idle channel histogram when the DC offset correction is disabled, one of the peaks is much higher than the three others and shows a similarly strange pattern. According to the datasheet of the ADS54j60 the four peaks should be similar in nature (Figure 146). I'm not sure if this is related to the problem at hand.
ADC0 to ADC3 all behave the same.
Is this behavior expected? What is the reason for the behavior? Is this something which can be improved? It generates quite some problems for our application, possibly rendering the use of the ADS54J60 impossible.
Thank you very much for your help!
