KLUDGE@mit-mc.arpa (03/22/83)
From: "James J. Frimmel,Jr." <KLUDGE@mit-mc.arpa> You can almost always beat the RS-232 spec's in various ways, since EIA was trying to define something which would work with most existing (at the time) lines/hardware. Their problem was that EIA could not limit the conductor size/shielding/braid/etc., so they set limitations which would work on just about any type of line in a noisy environment. I have seen 3,000 foot connections through a factory work (300 BPS only), and I've seen 56kbps at 10 feet (shielded coax). For specific situations, the thing to do is to try unmodified RS-232 over normal lines (up to 1,000 feet, separate shields if farther) and see if it works. If you have problems, then fix as needed: 1. Shielded cable, larger conductors. 2. add balanced line drivers/receivers (RS-422) for data lines (clocks if used) 3. in bad cases, use "MODEM eliminator" type short-haul MODEMS The real problem with RS-232 is that design engineers (who should know better) who design serial interfaces never read the EIA spec's. I've seen everything: mis-labeled/misplaced signal lines, confusion of DTE/DCE, non-standard signals on the connector (Cromemco System 3, TRS-80 Model III, AARRRGGGGHHHH!!) which can blow up modems, computers, terminals, people, etc. Standards don't do us any good unless we read them and apply them! flamin' with gas, Kap'n Kludge