Path: utzoo!attcan!uunet!hayes!tnixon From: tnixon@hayes.uucp (Toby Nixon) Newsgroups: comp.dcom.modems Subject: Re: 19200 bidirectional extension to V.32 bis. Message-ID: <2104@hayes.uucp> Date: 15 Oct 90 14:47:07 GMT References: <1990Oct7.221627.12252@larouch.uucp> <69427@lll-winken.LLNL.GOV> <1895@hayes.uucp> <69515@lll-winken.LLNL.GOV> Organization: Hayes Microcomputer Products, Norcross, GA Lines: 62 In article <69515@lll-winken.LLNL.GOV>, casey@gauss.llnl.gov (Casey Leedom) writes: > Jeeze, and I thought we were having a hard time now finding serial > interfaces that could go fast enough. I have a feeling we're going to > have to go with EIA-422. (Or is that EIA-423? What's the one with two > wires for transmit and two for receive and uses differential voltage > swings to determine 0/1? If I remember right it can go up to several > hundred Kbps ...) EIA-422 is the "balanced" interface (two separate wires for each signal), EIA-423 is the "unbalanced" interface (common ground, bit still uses lower voltage levels than 232 so it can go faster). The international standards the are compatible are, respectively, V.11 and V.10. Often, the "balanced"-type circuits are used for signals that change at the bit rate (data, clocks, and sometimes RTS and CTS), while "unbalanced"-type circuits are used for signals that don't change as frequently (DSR, DCD, DTR, RI). Previously, the EIA had defined only a 39-pin connector (RS-449) for use with these signals. But within the past couple of years, a new standard, EIA-530, has defined a 25-pin interface that appropriately mixes 422 and 423-type signals for a fully-functional interface. Some chip manufacturers are producing driver and receiver chips that are switchable between 232 and 530, so that manufacturers (like Hayes and Telebit) can produce modems that could support either interface. EIA-530 can support up to 1 megabit/s on cables the length typically used for modems, and rates of up to 100kbit/s over a thousand feet. The USA has proposed an EIA-530-compatible standard to both CCITT and ISO. > But seriously, there are some serial devices which don't even go up as > high as 38.4Kbps and I personally haven't seen higher rates used on > EIA-232. Even V.32bis/V.42bis is going to feel choked with 38.4Kbps > under some conditions. I sure CCITT is considering standardizing a modem > interface at the same time ... I've run EIA-232 on very short (one-meter) cables at 115,200bps without errors. Nevertheless, it is better to use the "proper" interface. People will definitely want to go to intelligent buffered serial interface cards (like Hayes ESP) when data rates get up this high. One thing I should point out is that the primary force behind pushing data rates beyond 14,400bps is NOT the demand from the asynchronous PC-type users. It is the large mainframe and networking environment, where communications is primarily synchronous and cannot benefit from the current techniques used for data compression. While those of us transferring uncompressed interactive async data streams can get up to 38,400bps today with V.32/V.42bis and up to 57,600 with V.32bis/V.42bis, these synchronous users are still stuck back at 9600 and 14,400. They're looking for something in between these rates and going to Switched Digital or Fractional T1. -- Toby ----------------------------------------------------------------------------- Toby Nixon, Principal Engineer Fax: +1-404-441-1213 AT&T: !tnixon Hayes Microcomputer Products Inc. Voice: +1-404-449-8791 CIS: 70271,404 Norcross, Georgia, USA BBS: +1-404-446-6336 MCI: TNIXON UUCP: ...!uunet!hayes!tnixon Internet: hayes!tnixon@uunet.uu.net