On Wed, 20 Mar 2002, John Wilson wrote:

> >From: Johnny Billquist <bqt@update.uu.se>
> 
> >> Potential nitpick:  I don't know about bad144 the *program*, but the
> >> Massbus-style last-track bad block file is also used on the RK06/RK07,
> >> and (sort of) on the RL01/RL02 (minus the 18-bit copies).
> >
> >18-bit copies?
> >Ah, perhaps you're talking about if the packs are used for -18(36) bit
> >systems. Hmmm, I think you really need to reformat the packs to use them
> >on the other system, and thus need to rewrite the bad block list.
> 
> I've never seen the actual DEC spec for this, just the stuff in the RK07
> manual, so I may be wrong but my understanding is:  the last track (with
> the bad block file on it) is always formatted in 16-bit mode even when
> used on an 18-(36-)bit system.  There are two copies (actually, many copies
> of each, but two *sets*) of the bad block file, one of which shows which
> sectors are bad when the pack is formatted in 16-bit mode, the other shows
> where the bad spots land when you format the pack in 18-bit mode.  The
> RL01/RL02 has no 18-bit mode so they omit all the 18-bit stuff in the bad
> block file.

Okay. I haven't read the STD 144 myself either, just the description in
the RSX manuals. And here is what it says (The chapter on BAD):

6.4.1.2 BAD and Last-Track Devices - BAD records bad block
informtaion differently on last-track devices than on non-last-track
devices. Last-track devices include the RK06/07, RL01/02, RP07 and
the RM02/03/05/80. The last track is divided into two areas, the
Manufacturer's Detected Bad Sector File (MDBSF) and the Software
Detected Bad Sector File (SDBSF). The MDBSF is created when the
manufacturer formats the pack. This operation also sets bits in any
header that is marked bad in the MDBSF and sets the SDBSF to be empty.
When you run BAD, entries are made in the SDBSF. BAD storage allows
126(10) entries of bad block addresses. The information contained in
the two last-track files is combined to form [0,0]BADBLK.SYS when you
issue the MCE INI command.


Note that the bad block list (also described in the bad144 manual page) is
not used on RP04/05/06.

But you might be right about 18-bit systems using a 16-bit format of the
last track. I seem to have some vague memory agreeing with that. It's been
a long time since I played with a 36-bitter though...

> The RK06/RK07 *does* support 18-bit mode but, does anyone know of a system
> which actually used it?  I don't think the 10s or 20s ever did (DM:s are
> really too small for them anyway), but I thought I read somewhere that the
> PDP-15 had a Unibus (option?), could that be it?

PDP-15 definitely used Unibus, but I would think that the PDP-15 had
ceased production before the RK06 came. Not sure, though.

Also, the RH11 can be used on both 16 and 18-bit unibus, IIRC.

	Johnny

Johnny Billquist                  || "I'm on a bus
                                  ||  on a psychedelic trip
email: bqt@update.uu.se           ||  Reading murder books
pdp is alive!                     ||  tryin' to stay hip" - B. Idol