Generally we use extraction to calculate the coupling C and from that we calculate the worst case offset (in mV). For digital circuits it is the timing window in which the aggression can happen that is tricky to calculate, not the C or the offset. You need very precise waveform propagation across the gates to capture the effect.

Dear Tajinder,

welcome,

Your question is a conceptual one and occurs when you have coupled transmission line. You put a stimuli on one line and see its interference on the nearby intended line. Say you have a stimulation waveform s(t) and the signal coupled to the receiving line is r(t). One normally has to purposely specify s(t) such that in case of digital signal s(t) will be in the form of rectangular pulses of duration equal to the bit time Tb. In case of analog signal s(t) will be purposely sinusoidal wave with specific frequency or frequencies.

The receiving line will react with a signal r(t) in case of digital circuits so long r(t) < the threshold value of the logic value the receiver digital circuit at its output will correctly interpret the logic value say one, otherwise it will coin falsely logic one. Therefore one has to keep interference signal level always within the threshold value of the intended logic.

In case of analog circuits r(t) will be sinusoidal and seen as interference I to the original signal Sor in the receiving line.

In this case it is the Sor/N+I which is the quantity to be evaluated. It is so that the allowed interference is set by the allowed signal to noise plus interference ratio.

Best wishes

THANKS DR ABDELHALIM ZEKRY.

Your answer was really helpful in understanding the basic concept of crosstalk-induced noise for coupled interconnect lines. please make it clear whether r(t) indicates the peak of the noise pulses that is received on the victim lines and positive peaks is to compared for logic 1 threshold and negative for those of logic 0.

Yes Tajinder,

When you apply a narrow pulse with a width equal it will couple deferentially to the victim transmission line. As the pulse width decreases it will transfer more power to the victim line and its positive amplitude grows such that it can be interpreted as logic one by the logic circuits terminating the victim line. According to this concept you can understand all the types of interference.

mathematically r(t)= K dS/dt , where K is the coupling coefficient which is normally smaller than one. You can study the interference by the simulator as i described in my post. Or you can study it experimentally.

Best wishes