DP8473A

On Sat, 29 Dec 2012, Christian Corti wrote: Two different things, although they both involve "index" Oversimplification (for those not familiar, or accustomed to disk formats (Atari, Commode, Apple ][), where the index of the disk is not used): There is a small hole in the disk. On most formats (other than the above), there is a small hole through the medai and a light and photocell. When light shines through the hole, a pulse is sent back to the disk controller, called an "index pulse". It is a hard mark of the start of the track. In addition, a few drives, including some Teac 55's use that to determine whether the disk is spinning. The stock Option Board software relies on the index pulse. NEC chose to resync their PLL when it sees that pulse. That means that the NEC chip is "blind" or "deaf", and can not see ANYTHING for a little while after the pulse, "but is fully enabled in plenty of time for the start of data." WD does NOT do that. Some disk formats, that use WD controllers, in order to squeeze the maximum capacity on the disk end up starting their first sector during that time. Therefore, there can be some difficulties reading such disks (such as Cromemco, TRS80, etc.) The index hole and its resulting pulse are used for identifying the start of the track for formatting, but are generally not needed for normal successful reads and writes. Therefore, many of the problems reading such disks can be circumvented by covering the index hole. Except for some drives, AND an unsuccessful read may generate the wrong error code. PLEASE don't design controllers that "go blind" after index. "Index Address Mark" is a completely different issue! The IBM/WD disk formatting pattern was originally called "IBM Soft Sector Disk Format", but when the 5150 came into existence, too many people thought that "IBM Format" meant PC-DOS. Without going into the full blather about splice gaps, resynchronization, etc., . . . Each sector has a special byte (with non-standard timing pulses) ("ID address mark"/IDAM) to mark the beginning of the sector header, and a "Data Address Mark"/DAM (a different byte with non-standard timing pulses) to mark the start of the data. In addition, there is an INDEX address mark/IAM (a different byte with non-standard timing pulse) to mark the start of the track. Since a read attempt might start anywhere in the track, to avoid excessive latency, most systems do NOT wait to see index. They will start reading as soon as they see what appears to be data, and wait for the desired sector header to come around. Accordingly, if everything is going smoothly, index is never really needed other than to know where to start a track when formatting. If that wasn't more than you want to know about it, or if you are offended by the oversimplifications, we could go through the reasons and structures of every part of the IBM/WD soft-sector disk format. -- Grumpy Ol' Fred cisin at xenosoft.com