Bits:30009717
Spring til navigation
Spring til søgning
Download Link
Metadata
BitStore.Metadata_version: 1.0 BitStore.Access: public BitStore.Filename: Comet32_harddisk.BIN BitStore.Size: 60116480 BitStore.Format: BINARY BitStore.Ident: 30009717:1 BitStore.Digest: sha256:26887b7e050a8801faf86df7cb3915adb41852a8465a5ef4bbf57da364a38288 BitStore.Last_edit: 20260202 phk DDHF.Keyword: COMPANY/ICL/COMET/DISK ARTIFACTS DDHF.Genstand: Genstand:11000898 DDHF.QR: 50005879 Media.Summary: Comet 32 harddisk image Media.Geometry: 117415s 512b Media.Type: ST506 Disk Media.Model: Vertex V170 Media.Serial: 44000 Media.Description: In the Comet 32 this disk is attached to a Shugart 1610 "ICU" SCSI-to-ST506 controller, which evidently hides the bad sectors found during formatting, from the host computer. The first physical sector on the drive is hidden from the host, and contains the drive geometry, a table of bad sector offsets, and other information we have not figured out. The only documentation we have for the controller, is for a variant named "1610-3", and it contains no mention of this bad sector handling, and requires that the host to configure the ST506 geometry for the physical drive, so we conclude the Comet 32 uses a later and more advanced version of the controller. With help from David Gesswein, we figured how the bad sectors found at formatting time are marked and handled, and the bad sectors thus found matches the "ERROR MAP" sticker on the disk drive perfectly one to one. Like the documented variant, this latter controller probably supports assigning alternate tracks as with SCSI command 0x0E, but we found no evidence of that, nor any other dynamic reassignments, on the drive. We have chosen to archive the drive contents, as the SCSI host computer would have seen it, and corresponding to the content we would have gotten, if we had dumped the disk through the SCSI interface of the Shugart controller. We do not know if the controller supports SCSI command 0x25 (READ CAPACITY), conveys the size of the drive to the host in some other way, or if the host is supposed to "just know" that, for instance from a disk-label of some kind. The first hidden sector contains several bytes which might or might not record the number of sectors reported to the host, but we have chosen to include in this image all sectors we could read from the drive, likely including, sectors the controller reserved for dynamic bad sector handling. When examined with the AutoArchaeologist, this image gives rise to no inconsistencies or other indications of trouble. *END*