With a new lot of produced PCBs we see a failure with the ADC122S021, which previous were working OK (same PCB, same SW).
As can be seen on attached file, an extra tracking pulse is required, which gives an offset in the reading of data. However, if the temperature of the ADC122S021 is increased (not much with a heatgun) the failure dissapears. Also adding some capacity to the SCLK seems to have a positive effect. I have studied the timing and cannot find any obvious reason. Is any reason/solution known for this problem? Thanks.