AUI cables needed

Pete> Richard wrote: Pete> It's some prime-number-fraction (17 IIRC but it's been a while) Pete> of the wavelength of the data bits (and it's too late tonight Pete> to work it out[1]) The idea is to ensure that the interference Pete> effects introduced by the discontinuities of the taps are not Pete> additive. Essentially, yes. That is the purpose. It's not actually a round number multiple, or round fraction multiple, of the wavelength. Using the minimum velocity factor given in the 802.3 and DIX specs (0.77), 2.5 meters translates to 0.10823 wavelengths, or 1/9.24 wavelengths. That makes sense, it's not a round number or 1/round number, so you don't get constructive interference even if you have transceivers separated by several marks. Then again, it's all somewhat approximate because the spacing is fixed while the velocity factor only has a minimum spec. But you don't get in trouble (with a round number) until you have a VF of 0.83, which is pretty high for small coax so the designers may have figured that practical cable would be 0.77 or above, but below 0.83. paul