karn (12/11/82)
Regarding the proposed packet radio standard for Oregon: The Amsat/Amrad AX.25 standard, since it uses HDLC at the link layer, allows the use of the HDLC link controller chips such as the Intel 8273 and WD 1933. Those chips do bit stuffing for 8-bit transparency, CRC-16 calculation and checking, and 8-bit address recognition. Since these chips do so much of the work, we didn't feel it worthwhile to sacrifice using them just so we could have error correcting codes; over short propagation delays, its just easier to retransmit a bad packet than to try to reconstruct it from the error codes. AX.25 allows for specified routing of a packet, by having a series of fixed-length callsign fields. The last callsign in the string has the lsb set (this is compatible with X.25 address extension). Routing from one station to another is up to the choice of the relaying station; it could indeed fire off duplicate copies over two links if desired if the links were considered unreliable. Now that we have a protocol with has been agreed to by a larger percentage of the amateur packet radio community than ever before, the area which has the greatest potential for improvement and standardization is modulation techniques. Clearly, using Bell 202 (FSK) modems on FM channels is inefficient. In fact, the link margins available on the dedicated Phase 3B packet channel won't support it at any reasonable bit rate. We are looking at two promising methods, minimum shift keying (MSK) and binary DPSK. The latter will be used for the Phase 3B engineering telemetry channel. I'd like to get in contact with anyone experimenting with such modulation techniques to share experiences. Phil Karn, KA9Q