VT103 (was Re: Reviving DEC RX01 power supply)
Jerome H. Fine
jhfinedp3k at compsys.to
Mon Sep 10 14:16:04 CDT 2007
>Ethan Dicks wrote:
>I have upgraded backplanes in the past. I found the easiest way to do
>it was to use recycled DEC backplane power strips - bits of thin metal
>that have a small (pin-sized hole) alternating with a larger hole that
>misses the next pin. I recovered some from a PDP-8/L that arrived to
>me with broken backplane. DEC used to sell it as a separate item for
>customers who were working with foundation modules and such. I don't
>know where I'd get it now.
Jerome Fine replies:
Since I had never even heard of a DEC power strip, using insulated wire wrap
seemed to be the best solution. The solder points stick out a bit. So
the wire about 1/8" was about right. Perhaps it was also prudent to
insulation back another 1/4" before soldering and then pushing the
forward again after making the connection and cooling off.
>>... However, an RQDX1,2
>>was out of the question since they require 6 Amps for
>>the 5V. An RQDX3 is probably OK when used with a quad
>>CPU like the M8189 or the M8190 (since there are no
>>boot ROMs on any RQDXn controller). I guess that you
>>could type in a boot program (if you don't do it too often)
>>using hardware ODT - it takes about 5 minutes.
>I was already presuming a SCSI interface.
I realize, but most everyone else reading this might not have a SCSI
host adapter for the Qbus. I also have a SCSI host adapter for the
Qbus and for a time it was the most costly Qbus board that I ever
>>Either a dual or a quad KDJ11 can be used. Almost EVERY
>>quad KDJ11 is really the same as a PDP-11/83 except for
>>the 18 MHz crystal. And since you can't use PMI memory
>>without further modification of the first 2 slots to add
>>ABCD like the first 3 slots on a BA23, it seems that you
>>are stuck with an ordinary PDP-11/73.
>True enough. I was more thinking of the advanced processors (11/93?)
>for any onboard memory, not for PMI.
>From my point of view, the PDP-11/93 is vastly over rated and even more so
over priced. On the other hand, for the VT103, the PDP-11/93 actually
makes a lot of sense when you are looking at only 4 quad backplane slots
and have an important project that needs 5 quad slots with a normal CPU
and separate memory. But a BA23 box is not that much bigger and heavier
if more than 4 quad slots are essential.
As for the speed of the PDP-11/93, when I did buy one to sell to a
customer over 10 years ago, I did a few benchmarks. For pure
CPU speed, the PDP-11/93 is about 10% faster than the PDP-11/83
which agrees with the crystal of 20 MHz vs 10 MHz, respectively.
However, the PDP-11/83 is only about 35% faster than the quad
PDP-11/73 (M8190) which has a 15 MHz crystal and uses normal
memory. Add PMI memory to the PDP-11/73 (almost all quad
PDP-11/73 CPUs are able to handle this) and the speed difference
is only about 15% - which means that the PMI memory speeds up
a PDP-11/73 by about 20%. At one point, PMI memory demand
was so low that it sold for less than normal Qbus memory. Shows
what people really know.
>>Most VT103 systems came with a paddle card that was intended
>>to be used with a DLV11-J and a dual PDP-11 Qbus card. The
>>paddle card connected the terminal and keyboard to the 4th
>>channel on the DLV11-J that was then strapped as consol
>>and the normal 25 pin serial connector on the back could
>>then be connected to a second serial channel on the DLV11-J
>>(and probably sent to a serial line printer).
>I do not have this card, but do know about it, and have been looking
>for one, casually, for 20 years.
I have one which I would gladly trade if someone was truly interested in
RT-11 software. Eventually I will need to abandon my collection, so ...
>>Since both replaceable media drives can also
>>be attached to a PC running E11...
>We work in different worlds... I'd be doing any off-line prep work
>with Simh and Linux.
Since E11 can also run on Linux, I should consider moving.
For some of my work, speed is important - sieving for prime numbers.
E11 is MUCH, MUCH faster than Simh.
But maybe someone would be interested in a faster drive interface
for Simh. The HD: drive under E11 would probably be ideal for
emulation by Simh and that would mean the the RT-11 device driver
would work under Simh as well. The interface is almost trivial.
Does anyone want the details?
>>Can I answer any other questions?
>Since I'm not likely to be attempting this for a year or two (after
>getting back from another winter at Pole), I'll probably ask questions
>as they come up. For now, though, I can easily put together a KDF11
>system, so I'll probably start there. I was just interested in what I
>might need to put together a SCSI-based KDJ11 system with 4MB.
Just add a hard drive and a DLV11-J with the M8186. With
a quad M8189, you already have 2 serial ports. Probably
best in that case to use the BA23 cabinet kit (if you have
one) for the serial ports. Using the paddle board with
an M8189 might be a problem since I seem to remember that
the ports used with the paddle board did not use the baud
rate selection on the DLV11-J but rather the baud rate was
selected by the VT103.
How many 4 MB Qbus memory cards do you have? Do you
have any PMI Qbus memory cards? Christlen made a 4 MB PMI
board, but by the time I became interested, it was not
possible to justify the expense.
If you attempted to send a reply and the original e-mail
address has been discontinued due a high volume of junk
e-mail, then the semi-permanent e-mail address can be
obtained by replacing the four characters preceding the
'at' with the four digits of the current year.
More information about the cctalk