I'm using the DDR Batch wizard to analyse a DDR4 bus. I have a failing tDQSS measurement which is failing by a substantial amount compared to the other DQS in the system.
Whilst there are a couple of borderline results here, the most troubling appears to be DDR_A_DQS3 which has a massive launch time of 1192pS compared to the nominal of 250pS-28pS (ish). When I examine the waveforms by clicking the link, it almost looks as though the DDR wizard is measuring at the wrong point:
Looking at the launch measurement point in the top waveform, it measures the DDR clk after 6 pulses. However, looking at the DQS waveform on the bottom, the launch measurement is taken after 5 pulses. The DDR batch simulator was set to ignore the first 5 initial transitions.
Looking at one of the other tDQSS measurements which passed, the CK and DQS are both measured after the same number of pulses. Do you know what might cause this behaviour and how to fix it?