I think it is better to run the driver into an optimum test-load (optimum-channel) with no-loss over very wide-band frequency-range. Then extract the power-spectral density of the signal and then try to find out if we shopped the power-spectral density @ freq 'fc' then we will be able to construct back the signal meeting the receiver specs. The idea depend on assuming an optimum-channel with S21=1 and with linear phase with frequency, then applying low-pass filter to the signal power spectral density and then convert back the signal to TD and find out whether that signal can be detected by the receiver or not if we use such low-pass filter of 'fc'. Does it make sense? Hany Fahmy (Nvidia Corporation). ----- Original Message ----- From: si-list-bounce@xxxxxxxxxxxxx <si-list-bounce@xxxxxxxxxxxxx> To: Lakshmi N. Sundararajan - PTU <lakshmi.s@xxxxxxxxxxx>; si-list@xxxxxxxxxxxxx <si-list@xxxxxxxxxxxxx> Sent: Sat Mar 06 13:20:54 2010 Subject: [SI-LIST] Re: frquency limit of a channel > Suppose assume I have a high speed serial link at 6Gbps. The > nominal rise time of the signals on this channel is 150ps. > Given this rise time, the bandwidth required to transmit this > signal is 0.35/tr = 2.33Ghz. At 6 Gbps, the period of a single bit is 166 ps. Within that period, the single must rise and fall. With a 150 ps edge rate, the rise and fall is 300 ps, which exceeds that of the 166 ps bit period. The math does not compute. As a general rule, 0.35/tr should always exceed (data rate)/2. The exception is if you're dealing with non-traditional signal protocols such as PAM, QAM, QPSK, etc... where symbol rate must be considered. > So, to study this channel behavior, is it correct to only look at > s-param frequency output till say 3Ghz. > Can any higher frequency data points on this s-param be > ignored and still correctly model the channel behavior? Ignoring the apparent discrepancy from above, it is useful to look beyond 3 GHz. There is no general consensus in industry as to the max frequency of consideration. 0.35/tr is a common, so is (data rate)/2. My experience leads me to believe that either is inadequate in most cases. Generally, it's best to consider at least 1/tr or even 1.5/tr [that is, 3X or 5X of the edge rate]. If the passive channel is very clean, the most modeling will be sufficient with 0.35/tr. However, "very clean" cannot be determined unless you look beyond 0.35/tr. Resonances are prominent in high-speed channels, and it is all to common to have a dramatic drop in S21 at 0.4/tr to 0.5/tr that will certainly affect the eye opening. Pat Zabinski Mayo Clinic ------------------------------------------------------------------ To unsubscribe from si-list: si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field or to administer your membership from a web page, go to: //www.freelists.org/webpage/si-list For help: si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field List technical documents are available at: http://www.si-list.net List archives are viewable at: //www.freelists.org/archives/si-list Old (prior to June 6, 2001) list archives are viewable at: http://www.qsl.net/wb6tpu ----------------------------------------------------------------------------------- This email message is for the sole use of the intended recipient(s) and may contain confidential information. Any unauthorized review, use, disclosure or distribution is prohibited. If you are not the intended recipient, please contact the sender by reply email and destroy all copies of the original message. ----------------------------------------------------------------------------------- ------------------------------------------------------------------ To unsubscribe from si-list: si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field or to administer your membership from a web page, go to: //www.freelists.org/webpage/si-list For help: si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field List technical documents are available at: http://www.si-list.net List archives are viewable at: //www.freelists.org/archives/si-list Old (prior to June 6, 2001) list archives are viewable at: http://www.qsl.net/wb6tpu