>> Strictly speaking, the RS422 voltage levels should be more in line >> with typical RS485 signals (differential on top of a positive common >> mode voltage), but the Mac doesn't seem to adhere to this convention. >> Most models seem to use asymmetrical signals about 0v, like RS232. > That's not consistent with what you just described. Can you explain > more clearly how the Mac does not "adhere"? I suspect you mean it > will "tolerate" RS-232 rather than "use" RS-232 levels. No. The Macintosh serial port hardware was designed as RS422 but certain details were chosen to provide easy and complete EIA/TIA-232 (aka RS232C) compatibility. See Inside Macintosh, volume III, pages 22 to 25. I quote: Macintosh serial ports conform to the EIA standard RS422, which differs from the more common RS232C standard. While RS232C modulates a signal with respect to a common ground ("single-ended" tranmission), RS422 modulates two signals against eash other ("differential" tranmission). The RS232C receiver senses whether the received signal is sufficiently negative with respect to ground to be a logic "1", whereas the RS422 receiver simply senses which line is more negative than the other. This makes RS422 more immune to noise and interference, and more versatile over longer distances. If you ground the positive side of each RS422 receiver and leave unconnected the positive side of each transmitter, you've converted to EIA standard RS423, which can be used to communicate with most RS232C devices over distances up to fifty feet or so. > I'd use the resistor, 220 ohms, in the RS-422 input just in case. Apple makes no mention of a resistor in the adapter cable. I've never used such a resistor in over 15 years of Macintosh network consulting and support. But then I've connected Macs to networking hardware, not PIC programmers. > The RS422-to-RS232 conversion works because all standard RS232 chips > have a threshold set at 1.5 volts not 0V as the standard implies. Since > the RS422/RS485 driver is a simple buffer and outputs a guaranteed 0.5 > to 2.5V swing, this works fine. The Macintosh, using the cable described, meets both the transmitter and receiver specs of EIA/TIA-232. It doesn't depend on the non-standard input thresholds obtained using recevier chips with floating threshold control pins. In the original Macintosh system, Apple's designers used AMD 26LS30 driver and 26LS32 receiver. The driver's power supplies were +5V (Vcc) and -5V (Vee) [volume III, page 26, figure 4. Diagram of Serial Ports]. So the RS422 TxD- output pin was very close to the letter of the EIA/TIA-232 voltages at the transmitter pin and certainly met the voltage requirements of EIA/TIA-232 at the receiver pin with any reasonable cable. Succeeding Macintosh models have, to the best of my knowledge, followed this serial port hardware model. Lee Jones ------------------------------------------------------------------- Jones Computer Communications lee@frumble.claremont.edu 509 Black Hills Dr, Claremont, CA 91711 voice: 909-621-9008 -------------------------------------------------------------------