CompuServe Thread

No. I repeat, you are off base on this issue. Darn it, I was AFRAID you were going to make me dig up my docs on this one. Excerpted from Electronic Design, January 24, 1985. Used without permission. "How RLL 2,7 upstages MFM and FM" "The read/write heads of disk drives… detect data by sensing the presence or absence of transitions between [magnetic] fields. The disk recording density is limited by how closely the flux reversals can be placed." "The maximum BIT density… is a… different matter. Depending on the method used to encode the data, the same number of flux reversals can be made to represent fewer or more bits." "RLL data encoding improves on the double-density encoding method of MFM and even more on the single-density encoding method of FM, because it represents a given bit stream using fewer flux reversals… it can record up to 60% more bits than MFM and transfer data 50% faster." "FM… needs the most flux reversals… it defines each bit cell by writing a flux reversal at the cell's leading edge. This leading flux reversal is… a clock bit. If the value of the cell is 1, a second flux reversal occurs midway in it. For a 0, no further flux reversals occur until the leading edge of the next bit cell. Therefore FM requires flux reversals for all clock bits and all data 1s." "MFM retains flux reversals for 1s but removes all clock bits except those between consecutive logic 0s. The MFM bit cell… is defined simply by a phase-locked oscillator synchronized with the data bit stream." "RLL code has nothing that can be defined as either a clock bit or a data bit. Rather, it turns data patterns into code patterns. Any binary data string can be broken up into combinations of just seven data patterns (Table A). Each of these is matched to an RLL code pattern that ends with either two or three 0s and starts with no or up to four 0s. Thus any combination of code patterns will have a minimum run of two 0s and a maximum run of seven 0s — hence the name RLL 2,7." "All three encoding methods are ultimately constrained by the nature of the recording process. The fields that establish the reversals cannot be too far apart, lest they become too weak for the phase-locked oscillator to lock onto. Too close, and they cancel one another out." "RLL 2,7 offers the best tradeoff… between flux reversal density and bit density." "TABLE A: Binary Pattern RLL Pattern 11 1000 10 0100 000 100100 001 001000 010 000100 0110 00100100 0111 00001000 "For example, a hexadecimal 7B looks like 0111 1011 in binary and 00001000 0100 1000 in RLL 2,7." (Woof. I need a page scanner!) I hope this explains RLL, and the differences between it and MFM, to your satisfaction. If not, please don't hesitate to inquire further.