> On Oct 26, 2025, at 8:21 PM, David Gesswein via cctalk
> <[email protected]> wrote:
>
> On Sat, Oct 25, 2025 at 07:30:58PM -0400, Bill Gunshannon wrote:
>>
>>
>> As long as we're talking about it, is there anywhere that
>> images for it are being archived?
>>
> I've linked to a few images people have sent me or I've made but I don't
> know of any large collection.
> https://www.pdp8online.com/mfm/status.shtml
>
>> And, can one build an image on an emulator (like SIMH or Ersatz-11)
>> and use it on a real system or will I need to build all my images
>> on real hardware?
>>
> For many machines yes. For many DEC controllers what the OS sees isn't
> exactly the same as whats on the disk. Some space is used by controller and
> some remap bad sectors but don't use the sector headers for that. Someone
> would need to take the time to figure out what the controller is doing and
> write code to manupilate the images.
>
> I know RQDX3 remaps. Not sure about other controllers.
MSCP controllers do "bad block replacement", producing an error free logical
address space from a physical space that has bad blocks in it. In some early
controllers, like the UDA50, BBR required help from the OS driver. Also, MSCP
introduced logical addressing not tied to the physical geometry of the device,
enabling such features as variable sector counts on inner vs. outer tracks.
PRO disk controllers don't do any remapping, though the drivers do. For the
hard disk that amount simply to skipping the first sector, which is reserved
for the powerup self test. For the floppy drive, the driver software
implements the interleaving, track skew, and oddball cylinder numbering
(matching what the RQDXn controllers do in the controller firmware).
Some earlier devices, like the RM80, have spare sectors (one per track on the
RM80). I think that was handled by flagging a bad sector with a header bit
saying it was replaced, at which point the driver would use the spare instead.
Most older DEC disks were simply a collection of physical sectors with bad
sectors handled in the OS. The RK05 nominally had 3 spare cylinders but at
least in RSTS those were never used. Bad sectors were typically marked as "in
use" in the file system, so they would not be allocated. That information
would be initialized from the DEC Std. 166 bad block table, if present on the
particular device, and in any case bad blocks found by the pattern test pass of
the disk initialization utility. Blocks going bad while in use tended to be a
bit more problematic; RSTS/E had ways to handle this but they are quite
primitive.
paul
