Some nice progress with hardware and Ultrix in the PUPS archive

[wkt@henry.cs.adfa.oz.au (Warren Toomey)]

In article by Michael Sokolov:
>    As I was thinking about how else can I get correctly written Ultrix
> tapes, the following idea sneaked into my mind. The PUPS archive already
> contains PDP-11 Ultrix. How about VAX Ultrix? With Warren's permission, I
> would be more than happy to upload my VAX Ultrix tape images (I have them
> sitting as files on my DOS disk). Once that is done, would someone
> volunteer to download them, write them to two TK50 cartridges, and send
> them to me? TIA for any help.

We have to be careful with 3rd party UNIXes. We would have to get approval
from DEC (or whoever they are) to allow us to add their product into the
archive. I did this with V7M and the later Ultrix-11s before DEC got bought
out.

	Warren

Received: (from major at localhost)
	by minnie.cs.adfa.edu.au (8.9.1/8.9.1) id BAA04900
	for pups-liszt; Sun, 6 Dec 1998 01:36:15 +1100 (EST)
Received: from timaxp.trailing-edge.com (trailing-edge.wdn.com [198.232.144.27])
	by minnie.cs.adfa.edu.au (8.9.1/8.9.1) with SMTP id BAA04895
	for <PUPS at MINNIE.CS.ADFA.OZ.AU>; Sun, 6 Dec 1998 01:36:03 +1100 (EST)
Received: by timaxp.trailing-edge.com; Sat, 5 Dec 1998 9:33:16 -0500
Date: Sat, 5 Dec 1998 9:33:16 -0500
From: Tim Shoppa <SHOPPA@trailing-edge.com>
To: msokolov at harrier.Uznet.NET
CC: PUPS at MINNIE.CS.ADFA.OZ.AU
Message-Id: <981205093316.2de002c3 at trailing-edge.com>
Subject: Re: System Industries MSCP disk controller problem
Sender: owner-pups at minnie.cs.adfa.edu.au
Precedence: bulk

>   I have just tried configuring the board as you have suggested and it
>works! Thanks! You have saved the Project!

"It has to be simple or I can't do it :-)".  In this case (like just
about all), it's a matter of getting rid of unknowns.
   
>   When I tried starting WOMBAT, it also worked. However, there are tons of
>configuration parameters you can set there.

Oh yeah - the Webster controllers are wonderfully configurable.  It's
incredibly useful, as one physical disk can be partitioned up many ways
into "virtual" MSCP units - especially useful for OS's which don't deal
well with large units (or when it's desirable to have many different 
versions of many different OS's on the same disk.)

>Do you have a full manual for this controller? Would you mind sending me a
>xerox copy?

Sure, for the cost of copying and postage.  Or I'd trade you for some
surplus of yours.
   
>   What about VAXen? Can you still call 1-800-DIGITAL with a _BIG_ credit
>card in hand and order a VAX?

Yep - high end 3100 and 4000 units are still available.

(It looks like someone has already pointed you towards the Emulex QD21
and QT13 docs available around the net.)   
   
>   And while we are at it, I have a question about 9-track tape transports
>and controllers in general. I often hear about something called Pertec.
>What is it? Is it some kind of standard interface that nearly all tape
>transports use?

Nearly all non-SCSI, non-proprietary interface tape drives will have
either (or both) Pertec formatted and Pertec unformatted interfaces.

Pertec unformatted drives have three cables; one carries read data, 
another carries write data, and the third carries control signals.  It's
a rather low-level interface - the controller tells the drive to go
forward, then reads or writes data, with the controller responsible
for all the timing.  The controller gets to do the nitty-gritty work
of repositioning and spacing forward and backward over tape marks.

Pertec formatted drives have two 50-pin cables.  It's a higher-level
interface, where the "formatter" does a lot of the nitty-gritty work, and
the controller in the backplane just has to spew out data bytes (or
take them in).

Many times you'll see a Pertec unformatted drive with a formatter bolted
onto it with two 50-pin cables coming out.  (In most cases, a formatter
could control multiple physical drives.)  Other times you'll find
"embedded formatter" drives, where there is no 3-cable unformatted
interface lurking inside.

> It is what this QT13 controller is for? (It has 2 50-lead
>connectors.)

Yep, Pertec formatted.  The QT13 is nice because it'll emulate
either MU: or MS:-type devices.

> I remember at CWRU there was a very neat-looking tape
>transport called Cipher. It looked EXACTLY like DEC TS05, suggesting that
>it's what the TS05 really is.

Yep, a TS05 is a rebadged Cipher F880 (with some slightly-different 
firmware).

> That one also had a two-50-lead-cables
>interface as far as I could tell (it was disconnected). It also appeared to
>be dual-ported.

It probably had 4 50-pin edge connectors, so that multiple drives could
be chained on the same Pertec-formatted-bus.  (There are terminators at
each end of the bus, much like SCSI.)  There's provisions for at least
4 formatters per bus, with each formatter potentially running multiple
drives.

> Does this mean that DEC also used this Pertec interface?

Yep.  Actually, the DEC TS05/TSV05/TU80 controllers are rebadged Dilog boards
(again, with slightly different firmware - for instance a DEC TU80 controller
will only work with TU80 drives or their CDC equivalent, the Keystone.)

>Then why are there different DEC controllers for different DEC 9-track tape
>transports (TSV05 for TS05, KLESI for TU81+, etc.)? If they were all Pertec
>one controller would fit all, wouldn't it? Or is that some DEC 9-track tape
>hardware uses Pertec and other DEC hardware doesn't?

At the guts of most DEC tape drives, you will often find either a Pertec
formatted or unformatted interface.  TU80's and TS(V)05's are simple
Pertec formatted interfaces.  Other times they convert to some other
interface (Massbus, LESI, etc.) before the cables come out to the "real
world".
   
>   Moving on to the next and last unidentified flying board. This one is a
>total mystery. The board is labeled "SIGMA INFORMATION SYSTEMS, INC." and
>"ASSY NO 400135 REV A". There are NO microprocessors, ROMs, or any other
>LSI chips on the board, only SSI/MSI chips, resistors, and capacitors. It's
>a dual-height board. It has 3 8-switch packs. There is one 40-lead shrouded
>header connector on the board, and a straight 40-lead ribbon cable connects
>it to a bulkhead which has a female 37-lead D-sub connector on the outside.
>Any ideas on what in the world is this?

Any chance it's a simple parallel I/O?

Could also be the bus interface for a Sigma and/or DSD MFM drive controller
(if so, it probably emulates RL02's).  The assembly number sounds vaguely
DSD-like.

-- 
 Tim Shoppa                        Email: shoppa at trailing-edge.com
 Trailing Edge Technology          WWW:   http://www.trailing-edge.com/
 7328 Bradley Blvd		   Voice: 301-767-5917
 Bethesda, MD, USA 20817           Fax:   301-767-5927

Received: (from major at localhost)
	by minnie.cs.adfa.edu.au (8.9.1/8.9.1) id BAA04942
	for pups-liszt; Sun, 6 Dec 1998 01:44:43 +1100 (EST)
Received: from timaxp.trailing-edge.com (trailing-edge.wdn.com [198.232.144.27])
	by minnie.cs.adfa.edu.au (8.9.1/8.9.1) with SMTP id BAA04937
	for <PUPS at MINNIE.CS.ADFA.OZ.AU>; Sun, 6 Dec 1998 01:44:33 +1100 (EST)
Received: by timaxp.trailing-edge.com for PUPS at MINNIE.CS.ADFA.OZ.AU;
          Sat, 5 Dec 1998 9:42:42 -0500
Date: Sat, 5 Dec 1998 9:42:42 -0500
From: Tim Shoppa <SHOPPA@trailing-edge.com>
To: PUPS at MINNIE.CS.ADFA.OZ.AU
Message-Id: <981205094242.2de002c3 at trailing-edge.com>
Subject: Re: Some nice progress with hardware and Ultrix in the PUPS archive
Sender: owner-pups at minnie.cs.adfa.edu.au
Precedence: bulk

>   As I was thinking about how else can I get correctly written Ultrix
>tapes, the following idea sneaked into my mind. The PUPS archive already
>contains PDP-11 Ultrix. How about VAX Ultrix?

As Warren pointed out, there might be a problem with putting such
tapes on the PUPS archive.  I'd be glad to run off TK50's from images
for you, though I think your earlier idea, about installing from
the miniroot image that's commonly put on 4.2- and 4.3BSD derived
distributions, is a *far* better idea as it avoids using a TK50
tape drive at all.

It's not that tape copies are bad ideas - it's just that TK50's
are so slow.  If you were coming from 9-track or DLT or something
fast, that wouldn't be so bad.

If you can get just about any OS running on your
Q-bus machine, under any CPU - i.e. NetBSD, RT-11, 2.11BSD, RSX,
whatever you might have - then you can just write the miniroot
straight to a "scratch" disk.  You can also write the root
dump and tar savesets straight to another scratch disk, in "raw"
format, if you desire.

-- 
 Tim Shoppa                        Email: shoppa at trailing-edge.com
 Trailing Edge Technology          WWW:   http://www.trailing-edge.com/
 7328 Bradley Blvd                 Voice: 301-767-5917
 Bethesda, MD, USA 20817           Fax:   301-767-5927