>The description was cut/paste directly from the "Field Guide".
>Maybe you could go through the latest version and point out
>other errors?
I have to sheepishly admit that I think the TC13/QT13 error
in the Field Guide was probably a typo I made myself 5-6
years ago!
Tim.
>I have one -AB that has a socket for the FPJ11 and one -AB that doesn't
>have a socket installed. Are these the same board?
I would gusess so (assuming that it really is a -AB revision and
some handles haven't been swapped around.)
> Is installing a
>socket on the board without one difficult (being it's multilayered)?
Not so bad if installed pin-at-a-time.
>I have a couple of 1MB M7458-AH (MSV11-RA) PMI boards. Although the
>Micronotes refer only to MSV11-J boards, I assume that what it has to
>say applies to the MSV11-R also?
Amazing! Someone actually read the micronotes! About 3/4 of the
Q-bus questions asked here are answered in great technical detail
by the micronotes, yet it seems like most folks still ask without
bothering to consult them...
Yes, the MSV11-R's are just as PMI as the -J's.
>I also have a 4MB non-PMI board. What sort of performance loss do I
>get by using the non-PMI memory?
It depends on the I/O vs CPU memory load. It's a visible performance
gain, but not incredible - I'd say 10-30% for most things.
> My goal is to run 2.11BSD, so more
>memory may be better than faster memory.
1 Mbyte is sufficient for 2.11BSD, and 2 Mbytes is more than
enough.
Tim.
>This is Q-Bus magtape controller:
>TC13: Emulex Pertec-interface tape drive controller. Switchable TU81
>TMSCP (MU:) or TS-11 (MS:) emulation.
Nope, a TC13 is MS:-only. I think you're confusing it with the QT13,
which is switchable between TMSCP and TS11.
Tim.
On Oct 18, 12:14, David C. Jenner wrote:
> Subject: Re: Need Info on DEC 11/84 Board, M8190
> CLASSICCMP(a)trailing-edge.com wrote:
> >
>
> > A M8190-AB is a 15 MHz (11/73) CPU with "warm floating point", but
> > *will* accept the FPJ11 as an option.
>
> I have one -AB that has a socket for the FPJ11 and one -AB that doesn't
> have a socket installed. Are these the same board? Is installing a
> socket on the board without one difficult (being it's multilayered)?
Normally I'd say that meant the one without the socket wasn't likely to
work with the FPJ11 even if you fitted it (they omitted the socket for a
reason) but if they have same etch revision, maybe that's not the case.
> I have a couple of 1MB M7458-AH (MSV11-RA) PMI boards. Although the
> Micronotes refer only to MSV11-J boards, I assume that what it has to
> say applies to the MSV11-R also?
Yes, put them above the processor and they use PMI; put them below and they
run as ordinary QBus memory.
> I also have a 4MB non-PMI board. What sort of performance loss do I
> get by using the non-PMI memory? My goal is to run 2.11BSD, so more
> memory may be better than faster memory. (Of course, I can always try
> to find more, faster memory!)
There's quite a difference on memory-intensive tasks -- several tens of %.
But the good news is that 2.11 wil run fine in 2MB. Whether more is
better than faster will depend on how much memory you need, and whether you
often use enough to make it do a lot of swapping to disk.
--
Pete Peter Turnbull
Dept. of Computer Science
University of York
At 15:04 18-10-98 -0400, you wrote:
<snip>
>TC1310201-FSH is on the big sticker.
I recognize it. That's a TC13 (or TC131?) tape controller. It provides
control for Pertec-interface drives. As I recall (someone check me on
this!) it emulates TS11, but can also handle 6250BPI density drives.
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Bruce Lane, Owner and head honcho, Blue Feather Technologies
(Hamateur: WD6EOS) (E-mail: kyrrin(a)jps.net) (Web:
http://table.jps.net/~kyrrin)
SysOp: The Dragon's Cave (Fido 1:343/272, 253-639-9905)
"Our science can only describe an object, event, or living thing in our own
human terms. It cannot, in any way, define any of them..."
On Oct 17, 23:50, William Donzelli wrote:
> Subject: Unibus Emulex board - ID, please...
> I am sorting out my DEC cards, figuring out which ones will be included
on
> my upcoming "junk sale" list. I came across a Unibus card made by Emulex,
> and I can not figure out its model number!
>
> It has two 50 pin connectors on the end, along with one red LED and a 4
> place DIP switch. Near the middle of the board are two 10 place DIP
> switches. Notable chips include a 2901, six 24 pin ROMs, a 40 pin DIP
with
> a sticker over it, and lots of glue.
It could be any of several things, for example it could be a serial line
multiplexer, or a Pertec tape drive controler. What's written on the
sticker on the 40-pin chip? The first two letters and two digits of the
number that's usually on those chips would tell us what it is. Are there
any stickers left on the ROMs? Anything silkscreened on the board, or
written in the etch?
--
Pete Peter Turnbull
Dept. of Computer Science
University of York
>I'm in Ottawa, and on my obligatory trip to the computer salvage place,
>picked up an M8190 Unibus 11/84 card.
How do you know that you've got an 11/84 "capable" 11/84? There
were many revisions:
A M8190 without any suffixes is a 15 MHz (11/73) CPU with an
unusable socket for the FPJ11 floating point processor.
A M8190-AB is a 15 MHz (11/73) CPU with "warm floating point", but
*will* accept the FPJ11 as an option.
A M8190-AC is still a 15 MHz (11/73) CPU, with a FPJ11 factory installed.
A M8190-AD is a 18 MHz (11/83 or 11/84) CPU with "warm floating point",
taking the FPJ11 as an option.
A M8190-AE is a 18 MHz (11/83 or 11/84) CPU with the FPJ11 factory
installed.
Even the above suffixes don't actually pinpoint the exact revision of
the CPU with respect to FPJ11 bugs, but they serve as a starting point.
To further complicate matters, within digital the CPU is identified
differently based on whether it has PMI memory installed or not.
(And, of course, it matters whether the PMI comes before or after
the CPU in the backplane!) So, for example, you may find a 18 MHz
CPU with non-PMI memory called a 11/73, but with PMI memory it's called
a 11/83.
An 11/84 is a 18MHz KDJ11-B with PMI memory and a KTJ11-B Unibus adapter
tacked on. I'm pretty sure the M8190 ROM's are the same whether
it's to be used as a 11/84 or an 11/83, and that the Unibus boot
ROM functionality is added in the KTJ11-B, which somehow coordinates
these issues with the CPU. (It's explained in the KTJ11-B processor
handbook, but I never read through it all without falling asleep.)
To a large extent, all this confusion is caused by the desire for
folks to say "this machine is a 11/x3" where x is 7 or 8. If you
really insist on the official classifications, you have to read
three of the Micronotes:
uNote 25: FPJ11, KDJ11-A compatibility
uNote 30: PMI on KDJ11-B and MSV11-J
uNote 39: KDJ11-A and -B differences.
The above micronotes - and many more - available from:
http://sunsite.unc.edu/pub/academic/computer-science/history/pdp-11/hardwar…
>1. Any other support boards needed, or will this board run standalone,
>presuming available RAM?
If, indeed, you have the 11/84 capable version, you'll also need the
KDJ11-B and the 11/84 backplane, along with PMI memory, to have a
real 11/84 CPU.
>4. Is anybody else using this card successfully?
I've got several in use as "11/73B"'s.
--
Tim Shoppa Email: shoppa(a)trailing-edge.com
Trailing Edge Technology Voice: 301-767-5917
7328 Bradley Blvd Fax: 301-767-5927
Bethesda, MD, USA 20817
At 06:06 AM 10/18/98 EDT, you wrote:
>Hi:
>The bus is Intel Multibus. Look in Intel documentation for Multibus or
>80/10A.
>Paxton
>
>
Hi Paxton,
Thanks, I will look.
-Dave
>>To further complicate matters, within digital the CPU is identified
>>differently based on whether it has PMI memory installed or not.
>>(And, of course, it matters whether the PMI comes before or after
>>the CPU in the backplane!) So, for example, you may find a 18 MHz
>>CPU with non-PMI memory called a 11/73, but with PMI memory it's called
>>a 11/83.
>In fact... you could find the system with PMI memory, but installed
>in the wrong place, and it will be identifed - by software - as an
>11/73B. If you put the same memory in the right place, it will be
>identified as an 11/83.
And if you put the FPP in your pocket instead of the CPU board, it
won't be identified by the software either :-). Your point is good
though - you've got to put PMI memory in the right slot or you don't
get the advantages!
>>If, indeed, you have the 11/84 capable version, you'll also need the
>>KDJ11-B and the 11/84 backplane, along with PMI memory, to have a
>>real 11/84 CPU.
>But the board with Qbus memory in the qbus of an 11/84 system box
>should work...
There are lots of never-officially-supported combinations which
do work.
>I've also taken one of the KDJ11-B 18mhz boards, removed the 18mhz
>clock and replaced it with a 20mhz clock... the boot ROM correctly
>identifies the clock speed, and the system runs rock solid...
Maybe you mean MHz? :-). Sorry, ads where sellers claim
spectacular milliHertz performance are one of my pet peeves. (Along
with specs calling for compatibility with the ASC-2 character set
and construction plans calling for DB-9 connectors!) Though it would
be an interesting exercise to construct a Pentium II-type computer
based on relays just so that it does top out around 300 milliHertz!
One thing I never understood about metric prefixes is how come
many of them are misused or, even more mysteriously, not used. For
example, everybody around the world uses the term "metric ton" when
the perfectly acceptable (and SI-preferred) term "megagram" is
exactly equivalent (and to my ears sounds better!) And why say
"ten thousand metric tonnes" when "10 gigagrams" is available?
Tim "but let's keep the Metric Buttload" Shoppa. (shoppa(a)trailing-edge.com)