XMODEM CRC Error Checking Protocols XMODEM CRC ERROR CHECKING PROTOCOLSData corruption caused by noise is a common problem when transferringfiles from one computer to another over telephone lines. A major sourceof noise for data transmission is impulse noise. These pulses or spikeson the line typically have a duration of 10 msec. To the human ear,such pulses sound like little clicks. To a 1200 baud modem, they arethe death knell for 12 bits of information.To combat the problem of data corruption during file transmission, filetransfer programs which use error checking protocols have been devised.Typically, a program on the sending end examines a block of data beingsent and adds various error checking characters to the data according tothe error checking protocol. A similar program on the receiving endexamines the received data, determines what the error checkingcharacters should be, and compares the expected values with the receivedvalues.
If the two agree, the block of data is accepted as valid. Ifthe two don't agree, the block of data is deemed corrupted andretransmission of the data is requested.The XMODEM file transfer program currently available on the GRC-VAXcluster supports two types of standard XMODEM error checking protocols.These are checksum (also known as longitudinal parity), and cyclicredundancy code (also known as polynomial code). The type of errorchecking used is determined by the receiving program when it first tellsthe sending program that it is ready to receive a file. If thereceiving program transmits a NAK character (ASCII 21), it is requestingchecksum error checking.
If the receiving program transmits a 'C'(ASCII 67), it is requesting cyclic redundancy code error checking.In order to understand the two different error checking protocols, it isfirst necessary to understand how XMODEM transfers files. XMODEMtransfers files in blocks consisting of 128 data characters each. Addedto each block are block identification characters and error checkingcharacters. The total number of characters in a block total 132 ifchecksum error checking is used, and 133 if CRC error checking is used.The actual placement of characters within a block is as follows:1 SOH (start of header, ASCII 1)2 block number3 complement of block number4 DATA (1st character).131 DATA (128th character)132 error checking byte (Checksum and CRC mode)133 error checking byte (CRC mode only)The determination of the error checking bytes depends upon which errorchecking protocol is being used. In checksum mode, the ascii values ofeach of the 128 data characters are added up and the low byte of theresulting sum is used as the error checking character. This method willdetect most errors, but it is possible that two or more of the datacharacters may be corrupted in such a way that their checksum remainsthe same. For example, the characters 'BB' will add up to the samevalue if they are inadvertently changed to 'AC'.
Because of this, amore sophisticated method of error checking is usually preferred.In CRC mode, the 128 data characters are treated as a continuous streamof 1028 bits. Two error checking bytes (16 more bits) are appended tothe end of this stream, and the resulting set of 1040 bits is thought ofas one very large binary number. The last 16 error checking bits arechosen in such a way that the resulting binary number will be evenlydivisible by a previously agreed upon binary constant (sometimes calleda generating polynomial).
Srw2024 Ing Code Using Xmodem Crc Windows 10
If any one of the 1040 bits of informationbecome corrupted during transmission, the large binary number theyrepresent will no longer be evenly divisible by the previously agreedupon binary constant.By carefully choosing the binary constant, it is possible to drasticallyminimize the number of errors which go undetected. XMODEM uses a binaryconstant value of 0100001 (equivalent to a generatingpolynomial of X^16 + X^12 + X^5 + 1), which is a standard valuerecommended for use by CCITT (Comite Consultatif Internationale deTelegraphique et Telephonique). By using this value it is possible todetect all single and double bit errors, all errors involving an oddnumber of bits, all burst errors of length 16 or less, 99.997% of 17-biterror bursts, and 99.998% of 18-bit and longer bursts.The mathematical theories of polynomial arithmetic lie behind theoverwhelming success of this type of error checking. Although thecalculation required to compute the last 16 error checking bits may seemcomplicated, it turns out that there is a simple method using shiftregisters which makes the method easily applicable and highly attractivefor computer use.-David Deley (1984)BIBLIOGRAPHY:Belonis, J.James II. XMODEM version 5.53 (FORTRAN source code) DECUStape VAX-96, June 1984.Housley, Trevor.
Crc Ccitt Xmodem
Data Communications and Teleprocessing Systems.Prentice-Hall, 1979.Peterson, W.W., and D.T. Brown.: 'Cyclic Codes for Error Detection'.Proc. 228-235, Jan. 1961.Puzman, Josef., and Radoslav Porizek. Communication Control inComputer Networks.
John Wiley & Sons, 1980.Tanenbaum, Andrew S. Computer Networks. Prentice-Hall, 1981.