HardWare I/O Modems Signal Standards

Standards determine whether a given modem can successfully connect with any other modem in the world. Bell 103 Uses a kind of modulation called Frequency Shift Keying, or KSK, in which a specific tone frequency signifies a digital one and another a digital zero. Each change in the modem's signal thus carries one bit of digital information. Consequently, the Bell 103 standard is the only one in which the baud rate (the rate at which signals change) is equal to the data rate of 300 bps (Bits Per Second). Bell 212A is the next logical step and the second modem standard to find wide application in the United States. It achieves a data-transfer rate of 1,200 bps by using phase modulation of a carrier signal. The phase of a fixed tone, called a carrier wave, is shifted by any of four phase angles. Under Bell 212A, the carrier wave can change phase up to 600 times per second, or 600 baud. The four phase states are sufficient to represnet any of four two-bit patterns. Thus each baud can carry two bits of information, raising the actual throughput to twice the baud rate of 1,200 bps. While widely used in America, many foreign countries prohibit the use of Bell 212A, prefering instead the similar international standard, V.22. NOTE: Why do rockwell controllers, in Bell 212A mode require a CCITT V.22 BIS answer tone? Cause their controller code is buggy! However, they've shipped a few million of them, so it's a "standard" now. LAP-B stands for Link Access Procedure, Balanced, an error-correction protocol designed for X.25 packet-switched services like Telebit and Tumnet. Some modem makers adapted it to their dial-up products before the V.42 standard was agreed upon. For example, the Hayes Smartmodem 9,600, form Hayes Mirocomputer Products, includes LAP-B error-control capabilities. LAP-M is an acronym for Link Access Procedure for Modems and is the error-correction protocol used by the CCITT V.42 standard. MNP Microcom Networking Protocol is an entire hierarchy of standards, starting with MNP Class 1, a no longer used error correction protocol, and running to MNP Class 10, designed to induce the highest data transfer performance from poor connections, especially those found in cellular phone systems. MNP classes 2 though 4 deal with error control and are in the public domain; classes 5 though 10 are licensed by Microcomm MNP-2 is designed to work with any modem capable of full-duplex communications. It works by confirming each byte as it is sent - by having the receiving modem echo back each character. MNP-3 improves on MNP-2 by wirking synchronously instead of asyncronously. As a result, no start and stop bits are required for each byte, trimming the data-transfer overhead by 25 percent or more. MNP-4 is an error correcting protocal that also yields some data compression. It incorporates two innovations. The first, Adaptive Packet Assembly, allows the modem to package data in blocks or packets sent and error-checked as a unit, The second, Data Phase Optimization, eliminates repetitive control bits from the data traveling across the connection to streamline trasnmissions. Together these techniques can increase the throughput of a modem by 120 percent at a given bit rate. MNP-5 is purely a data compression protocol that squeezes some types of data into a form that transmits faster. MNP-5 can compress up to a factor of two, effectively doubling the speed of data transmissions. On files that have been already compressed, however, MNP-5 may actually increase the transmission time. MNP-6 is designed to help modems get the most out of telephone connections, independent of data compression. Using a technique called Universal Link Negotiation, modems can start communicating at a low speed, then, after evaluating the capabilities of the teliphone line and each modem, switch to a higher speed. MNP-7 is a more efficient data compression algorithm (Huffman encodeing) than MNP-5, premitting increases in data throughput as much as threefold on some data. MNP-8 was never released. MNP-9 is desinged to reduce the transmission overhead required by some common modem operations. The acknowledgments of data packets is streamlined by combining each acknowledgement with the next data packet instead of sending a separate confirmation byte. While some error-correction schemes require all information transmitted after an error to be resent, an MNP-9 modem only requires the incorrect data to be sent again. MNP-10 is a set of Adverse Channel Enhancements that help medems work better with poor connections, compensating for line noise, echo problems, and limited bandwidth. Modems with MNP-10 will make multiple attempts to set up a transmission link, optimize the size of data packets for a connections, and adjust to the highest rate possible. V.22 is the CCITT equivalent of the Bell 212A standard, a transfer rate of 1,200 bps at 600 baud. It uses the same form of modulation as Bell 212A but is not compatible with the bell standard, because it uses a different protocol to set up the connection. Some modems support both standards and allow you to switch between them. V.22bis was the first true world standard, adopted in both the United States and Europe. It allows a transfer rate of 2,400 bps at 600 baud by using a technique called trellis modulation that mixes two simple kinds of modulation; quadrature and amplitude. Each baud has 16 states, enough to code any pattern of four bits. Each state is distinguished both ny its phase relationship to the unaltered carrier and its amplitude (or strength) in relation to the carrier. There are four distinct phases and four distinct amplitudes under V.22bis V.32 is an international high speed standard that permits data-transfer rates of 4,800 and 9,600 bps. At 4,800 bps, it uses quadrature amplitude modulation simiar to Bell 212A, but at 2,400 baud rather than 212A's 600 baud. At 9,600 bps, it uses trellis modulation similar to V.22 bis's 600 baud) and with a greater range of phases and amplitudes. Note that while most Group III FAX machines and modems operate at 9,600 bps, a FAX modem with 9,600 bps capability isn't necessarily compatible with the V.32 standard. V.32bis extends the V.32 standard to 14,400 bps, while allowing intermediary speeds of 7,200 and 12,000 bps in addition to the 4,800 and 9,600 bps speeds of V.32. Note that all of these speeds are multiples of a basic 2,400 baud rate. The additional operating speeds available to V.32bis are generated using different ranges of phases and amplitudes in the modulation. At 14,400 bps, there are 128 potentially different phase/amplitude states for each baud under V.32bis, enough to encode seven data bits in each baud. Because there are so many phases and amplitude differences squeezed together, a small change in the characteristics of a telephone line might mimic such a change and cause transmission errors. Consequently, error detection and correction become increasingly important as transmission speed goes up. V.42 is a world wide error correction standard designed to help make V.32, V.32bis and other modem communications more reliable. V.42 incorporates MNP-4 as an "alternative" protocol. That is, V.42 modems can communicate with MNP-4 modems, though a connection between the two won't use the more sophisicated V.42 error correction protocol. At the beginning of each call, as the connection is being negotiated between modems, a V.42 modem will determine whether MNBP-4 or full V.42 error correction can be used by the other modem, MNP-4 being the second choice. V.42bis is a data compression protocol endorsed by the CCITT. Different from and incompatible with MNP-5 and MNP-7, V.42bis is more efficient than either. On some forms of data, it can yeild compression factors up to four, potentially quadrupling the speed of modem transmissions. (With PCs, the effective saximum communication rate is limited by the serial port itself to no more that 38,400 bps.) Unlike MNP-5, V.42 never slows transmission of "incompressible" data. Worst case operation is the same speed as would be achieved without compression.