hsplab@ecsvax.UUCP (03/15/84)
Terminating resisters on disk drive cables provide better wave (e.g., signal)
transmission characteristics. This is probably of little importance in an
IBM PC where the disk cable is less than a foot, but in some units where the
disk drives are connected with up to ten feet of cable, the signal rapidly
deteriorates. If the terminating impedance is wrong (e.g., infinite if un-
terminated), the signals may reflect back and forth on the signal line
similar to an undamped shock absorber. Thus the term standing wave. Some
senior electrical enginneering courses discuss this in logic design; most
do not. I found a valuable reference in a book by Matick, Transmission Lines
for Digital and Communications Networks (1969) McGraw-Hill. The theory has
not changed even though the book is old. Usually signals traversing on a
cable requires consideration as a transmission line when the traversal time
in the cable begins to approximate the timing of the logic gate. Most TTL
gates have propagation times of 10-15 ns. Propagation in wire is approx.
1 ns/foot. Thus the ten foot disk drive cable should be treated as a
transmission line and may require termination, depending on the equipment
design.
David Chou
University of NC, Chapel Hill
...!decvax!mcnc!ecsvax!hsplabwmb@sun.uucp (Mitch Bradley) (03/17/84)
> Usually signals traversing on a cable requires consideration as a > transmission line when the traversal time in the cable begins to > approximate the timing of the logic gate. Most TTL gates have > propagation times of 10-15 ns. Actually, the propagation time of the gate doesn't matter as much as the rise/fall times of the output. Rise and fall times depend on the logic family and the loading. TTL of the "F" variety (Fairchild Advanced Schottky TTL) can have fall times on the order of 1 nsec. Obviously, "F" is not recommended for driving long unterminated lines! Mitch Bradley Sun Microsystems, Inc.