Hi,
Pls, have a look at this description of Signetics 25000 series 9046 XN
component: http://pichotjm.free.fr/humr/2.html
Can you explain me, what is Vff? Is it a good idea to decrease this voltage
in order to increase reliability?
But what will happen with our 50 Hz european supply?
I need help...
(please, answer with simple english)
JM Pichot
Not only did Cricklewood have them, but they arrived Saturday morning.
That effected a partial fix.
The timing generator is a string of four bit shift registers clocked
with 20Mhz.
Various combinations of their outputs are used to set and reset
bistables made out of cross coupled 7440's.
As the problem I'm fixing is you can't load memory, which is a one shot
operation, its slow old work.
A spare 8330 would confirm if that's where the problem is and I expect
would speed up the repair on the current one.
I'll continue with fixing it and keep an eye open for a spare. Getting
one from the US would be an option.
I also return all systems to working order. Apart from a couple of
VT420's with ticking SMPSU's everything I have restored runs. (Three
VAX's, Two alphas, a Rainbow 100+, DECMateII, Pro350, Pro380, a DS3100,
a VS3100 and assorted LA printers and VT terminals)
Rod
-----Original Message-----
From: cctech-bounces at classiccmp.org
[mailto:cctech-bounces at classiccmp.org] On Behalf Of Tony Duell
Sent: 28 March 2008 22:52
To: cctalk at classiccmp.org
Subject: Re: Looking for a M8330 or a couple of SN74194's
>
> Hi Tony
> I agree that the concept of originality is important in what we
do.
> The system I have is original in the sense that all the parts are from
> the right era. However the current configuration is not that as stated
> on the factory label. You could also say that any item with a
If this is the machine I think it is, I know not only the chap you got
it from, but also the owner before that. IIRC, it originally had a TU56
(single drive versiun, and no I don't mean a TU55) which is currently on
my desk, hooked up to my PDP8/e. And the RX8/e is not original.
If you look at one of the CPU logic boards, you'll see one of the chips
(I forget what) has been replaced. I did that over 10 years ago to cure
a problem with the top 4 data lines in some operation or other. The TU56
was 'payment' for that repair...
> date of later than the manufacturing date and not the result of a
> repair or upgrade is not original.
True, and some museums, alas, take it that way, and refuse to instal
replacement parts to keep their machines going, even if said parts are
contemporarty with the machine. Personally, I like all my classics to
run.
However, I also like to use as few spares as possible. If your clock
generator was missing, or physically cracked in half, or... then I would
agree you need an M8330 board. But if you can get your old one to run
again, even by using more modern parts (I would have no problems fitting
LS chips if they would work electrically, for example), then I'd rather
keep the spare board for somebody who really did need it.
>
> Thanks for the tip re Cricklewood I will contact them.
They may only have LS or something like that. As I said, I needed some
7495s, a rather oscure shift register chip that I couldn't find
anywhere.
They had 74LS95, which worked fine in my HP9820. OK, not original, but
at least I can use the machine again.
-tony
Johnny writes:
> The PDP-11 CPU is a 16-bit machine, so all code will always just deal
> with 16-bit values.
>
> The PDP-11 can have different buses, with different behaviours, but
> basically we can divide them into three groups.
> 16-bit bus
> 18-bit bus
> 22-bit bus
>
> Machines with a simple 16-bit bus is almost nonexistant, but there was
> some early Q-bus CPUs with a pure 16-bit address range, I seem to remember.
>
> 18-bit means all Unibus machines, as well as some Q-bus machines.
>
> 22-bit means most Q-bus machines.
Obviously written from a 11M+ user's perspective. Ignore all those puny Unibus and Q-bus machines without MMU, will you? But they still run RT-11 just fine!
Lots of Unibus and Q-bus and not-anything-bus-11's all do quite fine with 16 bit address spaces.
While the Unibus electrically has 18 address bits from the get-go, the Unibus 11's without memory management (or without the memory management option) don't use the last two. Even "later" Unibus CPU's like the 11/04 still came without MMU (and it was not an option, either!)
Obviously "early" and "later" are taken to be Lasnerian.
Tim.
I don't suppose you still have the old M8330 do you?
Rod
-----Original Message-----
From: cctech-bounces at classiccmp.org
[mailto:cctech-bounces at classiccmp.org] On Behalf Of Charles H Dickman
Sent: 02 April 2008 02:01
To: On-Topic and Off-Topic Posts
Subject: Re: Looking for a M8330 or a couple of SN74194's
Tony Duell wrote:
> DEC scheamtics have the odd error on them (I've heard it claimed said
> errors are deliberate, either to stop the machine being copied, or as
> copyright traps on the diagrams themselves). They're also somewhat
> unconventionally drawn (flip-flops may appear to have 4 outputs!), but
> I've got used to them. I must say that if I need to re-draw a
> sechematic, I use pen and paper, not a CAD system...
>
A year and a half ago I was playing with a PDP-8/e and trying to get an
EAE working with it. I ended up needing a later version M8330. My
original boards were rev 1 and the EAE didn't work until about rev 3.
The drawings I found online were WAY different than the boards I had.
For the EAE, it looked like there were some serious timing issues, so
the later board revisions were quite significant.
So, if your boards don't match the drawing rev, don't expect there to be
a match.
-chuck
Hi
Having worked for DEC for a number of years I'm used to the odd logic
symbols.
They do explain them in the manual. I don't use a CAD system but
something much more useful.
Visio. Define a DEC logic symbol once, add connection points in the
right places, drop as many copies as you need, join connection points
and print on A0 paper. Modify drawing as much as you like and print a
new copy each time.
Iv'e moved the boards back a bit and as soon as I get a few mins I'll
see if I can spot why it won't write to memory.
Rod
-----Original Message-----
From: cctech-bounces at classiccmp.org
[mailto:cctech-bounces at classiccmp.org] On Behalf Of Tony Duell
Sent: 01 April 2008 20:31
To: cctalk at classiccmp.org
Subject: Re: Looking for a M8330 or a couple of SN74194's
>
> Hi
> Its not test gear that's the problem its access to the board. My Tek
> 465 will trigger on single pulses with no problem.
> However whilst it's recommended that the the first CPU board (the
> 8330) be in slot 2 (The front panel is in No1) its not mandatory. So I
> can move the boards back a bit and get to the front of the 8330.
All Omnius slots are the same, and in an 8/e backplane, AFAIK all slots
are plain Omnibus.
The only time the top connectors are used on the M8330 is if you have
the EAE board set installed, I think. If you do, take it out for the
moment (you don't need it when trying to sort out a timing generator
problem).
You can then put the M8330 in whatever slot you want.
An extender board is useful, but by no means essential :-).
Another trick that's worked for me in the past if I can't re-arrange the
boards and don't have an extneder is to solder a hnadful of wires to
useful testpoints on the board and put it back in the cardcage. Then
probee the free ends of the wires (make sure they don't short to
anything!). Normally 2 or 3 iterations of doing that will get me to the
faulty chip.
>
> Manual I have, (very good) Schematics also (Awful) so bad I may even
> redraw at least the 8330. I have an A0 HP plotter so its easy to do
> drawings.
DEC scheamtics have the odd error on them (I've heard it claimed said
errors are deliberate, either to stop the machine being copied, or as
copyright traps on the diagrams themselves). They're also somewhat
unconventionally drawn (flip-flops may appear to have 4 outputs!), but
I've got used to them. I must say that if I need to re-draw a
sechematic, I use pen and paper, not a CAD system...
-tony
pichotjm <pichotjm at free.fr> wrote:
>
>Pls, have a look at this description of Signetics 25000 series 9046 XN
>component: http://pichotjm.free.fr/humr/2.html
It is not always necessary to go to the expense of buying a 25000 series chip.
I found some WOM in my Dec Alphaserver 1000A recently. I'm not interested in
this bleeding-edge stuff (besides its too new to be on-topic here) so I have now
replaced it with some ordinary memory. I wonder if I could get a good price for
the WOM on ebay?
>Can you explain me, what is Vff? Is it a good idea to decrease this voltage
>in order to increase reliability?
6.3Vac is a common voltage for valve (tube) filaments. Given that the 25000 series
is implemented in VLSI (Vacuum large scale integration), I suspect Vff is the supply
to the filaments for the integrated valves on the WOM chip.
Decreasing Vff should reduce power dissipation and improve the lifetime of the chip.
However, it may also result in poorer noise immunity.
Either that or Vff is the reference voltage used by the flip-flops and decreasing it
will unbalance the chip by creating an excess of flops over flips.
>But what will happen with our 50 Hz european supply?
Typically, these things are designed for the American market with no thought given to
widely used standards in other parts of the world. I suspect we will have to use
motor-generator sets to power this type of circuit. Solid state frequency converters
will not be suitable for the critical power requirements here.
>I need help...
>
>(please, answer with simple english)
>
>JM Pichot
>
Regards,
Peter.
Dwight asked:
> Pardon me for asking but are you sure the chip can provide
> enough drive? Most floppies expect to terminal with about
> 150 Ohms. That is more than an unbuffered output of most
> chips are designed for.
> Just asking?
Good point--but the datasheet states right in the opening remarks
that it can be connected directly to the disk drive with its own
internal 48 ma buffered output and schmitt inputs. The datahseet's
easy to find on the web--several places have it.
Here, for example:
http://www.tranzistoare.ro/datasheets2/36/366021_1.pdf
I like the DP8473--unlike the Intel 82077AA and later chips, it still
knows how to do FM correctly--and will even write 128 byte sectored
MFM, which is something that Intel was ever able to do. Don Maslin
and I worked through a pile of FDC chips before we found this one.
I think there was a Future Domain card (1680?) that brought all 4
floppy drives out to 2 headers--and I certainly could have added one
to do that myself for 4 drives, but 3 on one cable was the easiest.
Cheers,
Chuck
Hi Everybody
I was reading some doco last night (specifically the DEC RL11/01 Disk
Subsystem Training Handout), and I am confused.
(That is nothing new...)
In the "Subsystem commands section", the addresses specified for DAR and
CSR are 774404, and 774400 respectively, but in the back of the
document, under the 'toggle in programs' section, the DAR and CSR are
shown as being at 174404 and 174400 respectively.
Hmmm. I see three possibilities:
1. I have (re) discovered a documentation error - I wonder if DEC, no,
HP, no Compaq, would be happy for me to submit an amendment :-)
2. The hardware throws away the upper bits so both references are
correct. Hmmm Possibly not :-)
3. There is some other stuff happening that I missed...
If anybody could shed some light on this I would appreciate it. The
background to the problem is that I can not convince my shiny,
wonderful, (and noisy) DL01 to perform the seek test - Me thinks there
is a problem....
Doug
Oh I forgot.. The 420's ..
I have not done anything because
a) It only happened last week
b) I cant find any drawings.
Rod
-----Original Message-----
From: cctech-bounces at classiccmp.org
[mailto:cctech-bounces at classiccmp.org] On Behalf Of Tony Duell
Sent: 31 March 2008 22:52
To: cctalk at classiccmp.org
Subject: Re: Looking for a M8330 or a couple of SN74194's
>
>
> Not only did Cricklewood have them, but they arrived Saturday morning.
I'm even luckier in that I live near enough to Cricklewood Electronics
that I can pop over there and pick the chips up the same day ;-)
> That effected a partial fix.
> The timing generator is a string of four bit shift registers clocked
> with 20Mhz.
OK.
> Various combinations of their outputs are used to set and reset
> bistables made out of cross coupled 7440's.
> As the problem I'm fixing is you can't load memory, which is a one
> shot operation, its slow old work.
I don;t know what test gear you have, but I find a simple logic probe to
be more use than a 'scope for this sort of problem. The former can
easily detect narrow pulses that are difficult, if not impossible, to
see on a non-storage 'scope. Of course an 'Advanded Logic Probe' as HP
called the LogicDart is even nicer, but...
Do you have the maintenance manual? Not the printset (schematics), but
the book which is a low-level circuit description? It's well worth
reading!.
I would have thought it wouldn't be too hard to keep on hitting the DEP
key and seeing what, if any Omnius signals are generated. And then
totrace back the missing one(s).
[...]
> I also return all systems to working order. Apart from a couple of
> VT420's with ticking SMPSU's everything I have restored runs. (Three
Any reason you';ve not fixed those? Ticking SMPSUs are often quite
simple to fix, maybe just dried-up capacitors.
-tony
> Date: Tue, 1 Apr 2008 10:47:25 +0200 (CEST)
> From: Christian Corti
> The 8473 directly supports four drives, so you can do the "trick" a second
> time for a fourth drive. And even some motherboard Super I/O chips have
> support for four drives, the outputs are simply not used.
My needs and the number of spare conductors on the cable
simultaneously gave out with 3 drives. Somewhere in my collection
I've got a controller (may not even be SCSI; could be ESDI or
MFM/RLL) that had a jumperable option to drive 4 floppies from a
single cable by using the drive select jumpers and a common motor
control line. The two downside issues are that your drives have to
have 4 drive selects (or you had to be clever with your cable twists)
and when one motor comes on, they all do. Better to use two headers
and two cables, IMOHO.
Cheers,
Chuck
