Price of core memory
Daniel House
"lwin" <lwi...@bbs.cpcn.com> wrote in message news:9f3vii$f...@netaxs.com...
>
> Is there a _large_ group of people out there interested in _early_
computer
> history and collecting old computer stuff? (By early, I mean someone
> who knows that computers existed long before the PC came out.)
>
Not sure how "large", but there is definitely a group of dedicated computing
enthusiasts interested in pre-PC days. I collect IBM "iron" and embed
microcontrollers to create visual simulations, as in the following pages.
http://home.nc.rr.com/deh/GALLERY/GALLERY.HTM
If anyone knows the whereabouts of obtainable 360/370 paraphernalia, control
panels especially, I'd appreciate the leads.
I've had lots of people ask if they are for sale ... Sorry, they are not.
The larger ones take hundreds of hours to put together and it's basically a
labor of love. That said, I do have a few spare 3803 microcontroller based
creations in hand made red oak frames that I'd be happy to swap in trade for
something I don't have (360/370 related only).
Michael Roach
In article <9f3vii$f...@netaxs.com>, lwin <lwi...@bbs.cpcn.com> wrote:
>Is there a _large_ group of people out there interested in _early_ computer
>history and collecting old computer stuff? (By early, I mean someone
>who knows that computers existed long before the PC came out.)
Of course. Hey, people still keep steam engines running. Why do they do
it? Because they like it that way 8^)
One thing I remember is the tricks used to make programs small and fast,
because storage was small and computers were slow. I remember using
instructions for data on the PDP-8 because the low order 8 bits were
some character I needed to use (usually a CR or LF.)
Small, fast programs aren't being written for the modern machines but
now I'm just ranting..because I like it that way!
--
No problem is so large it can't be fit in somewhere.
Michael L. Umbricht
It is a spare for a Honeywell 316 that I am restoring.
See http://conan.ids.net/~mikeu/h316/ for info on the H316
And see http://shrimp.osfn.org/~mikeu/images/ for jpg's of core.
-mikeu
win wrote:
>
> Is there a _large_ group of people out there interested in _early_ computer
> history and collecting old computer stuff? (By early, I mean someone
> who knows that computers existed long before the PC came out.)
>
> Comments anyone?
CBFalconer
>
> x-no-archive: yes
>
> I've noticed on ebay several sales for computer magnetic core
> memory. I was surprised to see it to be fairly expensive and
> a number of people bidding on each offered item.
>
> As one interested in computer history, I wouldn't mind owning a nice
> chunk of core; but frankly not so much at the prices being bid. (I did
> get one sheet from the Boston Computer Museum).
>
> I am surprised that so many people are interested in such old artifacts.
> (Other old artifacts seem to sell well, too.)
>
> Is there a _large_ group of people out there interested in _early_ computer
> history and collecting old computer stuff? (By early, I mean someone
> who knows that computers existed long before the PC came out.)
>
> was interested in the very latest and greatest stuff, Internet, etc.,
> but no interest at all in any old stuff prior to the PC.
>
> Comments anyone?
Back around '63 we built an instrument (a cheap nuclear pulse
height analyzer) around a 1 bit by N core. The point was to
eliminate wires. It had X, Y, and a sense wire, with no inhibit
line, and the primary object was to increment a count. Reading a
bit automatically zeroed it, so if the bit had been set the next
operation was to advance address and repeat. If the bit had been
zero the next operation was to set it and exit.
This was Frank Petree's idea. He and I later built one of the
first desk top computers after an abortive effort to expand the
concept into a GP computer. The end result used a 2k x 4
conventional core memory.
In those days we built logic primarily from diodes, resistors, and
transistors and the count of little bitty parts was important.
--
Chuck F (cbfal...@my-deja.com) (cbfal...@XXXXworldnet.att.net)
http://www.qwikpages.com/backstreets/cbfalconer :=(down for now)
(Remove "NOSPAM." from reply address. my-deja works unmodified)
mailto:u...@ftc.gov (for spambots to harvest)
Douglas W. Jones,201H MLH,3193350740,3193382879
by never...@panics.com.invalid (Michael Roach):
>
> One thing I remember is the tricks used to make programs small and fast,
> because storage was small and computers were slow.
Try programming on embedded microcontrollers such as the Microchip PIC
family. What was once a relay-rack sized PDP-8 is now an 18 or 28 pin
DIP package, but that includes USART, parallel port, 8K of code ROM and
422 bytes of data RAM. Many of those old tricks are still very valuable!
Doug Jones
jo...@cs.uiowa.edu
Douglas W. Jones,201H MLH,3193350740,3193382879
>
> Is there a _large_ group of people out there interested in _early_ computer
> history and collecting old computer stuff? (By early, I mean someone
> who knows that computers existed long before the PC came out.)
See the newsgroup alt.sys.pdp8 for a modestly organized group of folks who
collect and maintain DEC PDP-8 computers, many from the 1960's remain in
operating condition, with 4 to 32K 12-bit words of core. I have 4 machines,
2 with 8K and 2 with 16K.
Those of us who try to collect working machines don't approve of those who
break up the machines to make trophies people can hang on their office
walls. Much of the commerce on Ebay is in such trophies!
Doug Jones
jo...@cs.uiowa.edu
Michael Roach
In article <9f5q1m$b8s$1...@flood.weeg.uiowa.edu>,
Douglas W. Jones,201H MLH,3193350740,3193382879 <jo...@cs.uiowa.edu> wrote:
>From article <9f3vii$f...@netaxs.com>, by lwi...@bbs.cpcn.com (lwin):
>>
>> Is there a _large_ group of people out there interested in _early_ computer
>> history and collecting old computer stuff? (By early, I mean someone
>> who knows that computers existed long before the PC came out.)
>
>See the newsgroup alt.sys.pdp8 for a modestly organized group of folks who
>collect and maintain DEC PDP-8 computers, many from the 1960's remain in
>operating condition, with 4 to 32K 12-bit words of core. I have 4 machines,
>2 with 8K and 2 with 16K.
I've been wondering...where do you get new (or usable) dectapes? Are
the originals still usable? How old is the oldest original that's still
usable? Is it true you can wind vcr tape onto a spool and have it work?
>Those of us who try to collect working machines don't approve of those who
>break up the machines to make trophies people can hang on their office
>walls. Much of the commerce on Ebay is in such trophies!
Has anyone resurrected a trophy and gotten it to work in one of the old
machines?
--
Don't steal; thou'lt never thus compete successfully in business.
Cheat.
-- Ambrose Bierce
Brian {Hamilton Kelly}
mi...@shrimp.osfn.org "Michael L. Umbricht" writes:
> I picked up a Litton 4KW x 16 core plane on eBay rather cheap.
Just so that we can get a feel for prices, how much is "rather cheap"?
(I recall purchasing an 8KiWord "extension memory" for an Elliott 920B in
1968. This worked out at almost exactly one pound per (18-bit) word.
With inflation making 1968 prices about 14x nowadays, then your bit of
bits ought to have cost around USD75,000.)
BTW, how do you get a 16-bit word into ONE core plane? IME, one bit ==
one plane.
--
Brian {Hamilton Kelly} b...@dsl.co.uk
"We have gone from a world of concentrated knowledge and wisdom to one of
distributed ignorance. And we know and understand less while being incr-
easingly capable." Prof. Peter Cochrane, formerly of BT Labs
Al Kossow
Kelly}) wrote:
> In article <3B15E552...@osfn.org>
> mi...@shrimp.osfn.org "Michael L. Umbricht" writes:
>
> > I picked up a Litton 4KW x 16 core plane on eBay rather cheap.
>
> Just so that we can get a feel for prices, how much is "rather cheap"?
>
eBay #1234635532 may 6 $15.27 Varian core assy
Unmolested, as far as I can tell. Haven't tried it in my 620L-100
yet, though.
DEC core sets (esp for 8E's) go for a LOT more. When someone was
recently scrapping a large lot of DEC boards, I bought all of the
11/40 32k core board sets that he had, with the intent of using
them as spares for the front end processor of a KL10.
Michael Roach
In article <9f63lv$q...@netaxs.com>, lwin <lwi...@bbs.cpcn.com> wrote:
>x-no-archive: yes
>
>> Of course. Hey, people still keep steam engines running. Why do they do
>> it? Because they like it that way 8^)
>
>steam engine and an electronic computer.
>
>A running computer just sits there. In contrast, a running steam
>engine, to its fans, has lots of neat sounds, sensations, and visuals.
A running computer has switches and blinkenlights.
You can put a cup of coffee in an old computer to warm it up and not
worry about it spilling 8^)
>(And I think there's a lot more railroad buffs out there willing to
>do the incredible amount necessary to rehab and run a steam engine
>than there are computer buffs willing to do likewise to old hardware.)
>
>I must admit I'd love to see a working S/360, just for nostalgia's
>sake. But I also must admit I'm NOT sure I'd be so quick as to run
>down and write a program, keypunch up the cards, wait in line at
>the card reader, wait for the job to be run, and wait for the printout
>only to discover I left out a single comma and have to do it all
>over again.
I have a hunch that many of the old computers running today are personal
computers so you won't have to wait in line.
--
Loose bits sink chips.
Douglas W. Jones,201H MLH,3193350740,3193382879
by never...@panics.com.invalid (Michael Roach):
>
> the originals still usable? How old is the oldest original that's still
> usable? Is it true you can wind vcr tape onto a spool and have it work?
Real DECtape is made using a sandwich of 2 mylar layers with the oxide
between. As a result, the oxide can't flake off, so the stuff has an
extremely long shelf life. The fully redundant recording format also
helps (each bit is stored in duplicate on two different non-adjacent
tracks of the tape, so you can punch a quarter-inch hole anywhere in the
tape with no data loss).
Would 3/4 inch VCR tape work? Probably, but it might have different
wear characteristics on the tape head than legitimate DECtape.
I have a few cubic feet of DECtapes that date from the 1960's (they came
with one of my PDP-8/L systems), and although I haven't restored my
drives to working order, I've sent some of those tapes to others, and
they check out as good tapes.
> Has anyone resurrected a trophy and gotten it to work in one of the old
> machines?
I've rebuilt one 4K core module that came to me broken. I had to replace
3 of the transformers on the sense wires, and since I couldn't get
replacements, I had to manufacture them myself from small ferrite cores
and 28 gauge wire.
Doug Jones
jo...@cs.uiowa.edu
Heinz W. Wiggeshoff
> x-no-archive: yes
>
>> Of course. Hey, people still keep steam engines running. Why do they do
>> it? Because they like it that way 8^)
>
> But aesthectically, there's a big difference between a running
> steam engine and an electronic computer.
>
the rolling power blackouts.
...
> sake. But I also must admit I'm NOT sure I'd be so quick as to run
> down and write a program, keypunch up the cards, wait in line at
> the card reader, wait for the job to be run, and wait for the printout
> only to discover I left out a single comma and have to do it all
> over again.
Option 1) You write the program onto an approved coding form, such as
the IBM SYSTEM/360 ASSEMBLER CODING FORM GX09-0010-04,
which I also used for PL/I (and 24 line by 80 col. screen
layouts). The result is sent for keypunching and verifying.
You may get a red face when the keypuncher finds logic errors.
2) You find a glass tube and compose on CRJE, being sure to
save your work every 5 minutes and _never_ overflow the lib!
3) In more recent times, say 1973, you use WYLBUR on either a
2741 (slow, noisy) or a tube.
CBFalconer
>
... snip ...
>
> I must admit I'd love to see a working S/360, just for nostalgia's
> sake. But I also must admit I'm NOT sure I'd be so quick as to run
> down and write a program, keypunch up the cards, wait in line at
> the card reader, wait for the job to be run, and wait for the printout
> only to discover I left out a single comma and have to do it all
> over again.
Something over 15 years ago Charlie Gingrich, in the Yale
Accellerator Labs, mentioned to me he had a 360 in some back room,
that he had snaffled when the Yale Computer Center was about to
junk it. He even offered it to me.
Unfortunately Charlie died since then. His boss was DA Bromley,
Physics Chairman. Maybe you can get one through him :-) There
are incredible things in the warrens beneath Yale.
Al Kossow
>
> Unfortunately Charlie died since then. His boss was DA Bromley,
> Physics Chairman. Maybe you can get one through him :-) There
> are incredible things in the warrens beneath Yale.
Someone asked me a few months ago what would be a good system to
have running on display at the CMHC. I told him a 360/50, since
I thought it was old enough to be historically interesting, and
small enough that you might have a chance to keep one running,
and it had an interesting operators console :-)
Unfortunately, I wonder if there is a complete one still to be found,
and if you did find it, could you find any mid/late 60's software
for it.
gla...@glass2.lexington.ibm.com
Well, in 1957, IBM was selling it for under US$1/bit, so if you're paying
more than that, after adjusting for inflation, then you're probably paying
too much. :*)
738 One Megabit Magnetic Core Storage
32768 words by 36 bits
First IBM memory product using transistors
April 1957
$1,040,000 sale price (later reduced to $940,000 - under $1.00 per bit!)
Dave
P.S. Standard Disclaimer: I work for them, but I don't speak for them.
gla...@glass2.lexington.ibm.com
I think almost all of us in here hate the trophy hunters. But, sometimes
that auction place can be a good place to pick up spare parts for old
machines. I know we all hate to see a perfectly good machine torn apart,
but sometimes it's necessary to do this in order to supply parts so that
a dozen other machines can continue to live.
Dave
P.S. Standard Disclaimer: I work for them, but I don't speak for them.
P.P.S. The core plane on my wall was already destroyed beyond repair
by the time it came to me, most likely by the CE who replace it in the
S/360 it came from since it had failed and was being replaced (e.g.,
It's not a working core plane that was scavenged to make a trophy!).
It's one thing to make a trophy out of a failed and unrepairable part,
and quite another to make a trophy out of a perfectly good antique computer
part.
Douglas W. Jones,201H MLH,3193350740,3193382879
by gla...@glass2.lexington.ibm.com:
>>(I recall purchasing an 8KiWord "extension memory" for an Elliott 920B in
>>1968. This worked out at almost exactly one pound per (18-bit) word.
I have a very early 1960's vintage core plane (never used, with factory
baggies still taped over each bundle of wires) with the original wrapping
plastic. It's 8K bits, and the price on the wrapper is 8865.02
>>BTW, how do you get a 16-bit word into ONE core plane? IME, one bit ==
>>one plane.
Depends on the era. The PDP-8/E core memories I have have either 4K or
8K 12-bit words in one physical plane, arranged as 12 postage-stamp sized
patches that share addressing lines, where each patch has its own sense
line. That is, each patch is one logical plane, but all are "woven" into
a single physical plane.
I also have a 32K by 16bit core module that I haven't opened up (it may
be in working condition) from the late 1970's. I suspect it is of similar
construction, but with the physical plane carefully folded to to make it
fit in the requisite area. The 8K by 12-bit plane is done without folds.
Doug Jones
jo...@cs.uiowa.edu
Jay Maynard
>Unfortunately, I wonder if there is a complete one still to be found,
>and if you did find it, could you find any mid/late 60's software
>for it.
The 360/50 would run MVT quite well...and that's still pretty easy to lay
hands on, thanks to Rick Fochtman and Sam Knutson. You can get it from
http://www.cbttape.org .
Al Kossow
<B0AFC96EB953D5DE.30A72E38...@lp.airnews.net>,
jmay...@conmicro.cx wrote:
> On Thu, 31 May 2001 13:12:56 -0700, Al Kossow <a...@spies.com> wrote:
> >Unfortunately, I wonder if there is a complete one still to be found,
> >and if you did find it, could you find any mid/late 60's software
> >for it.
>
> The 360/50 would run MVT quite well
It could, and you wouldn't see all the warts of the early versions
(historically accurate software to match the vintage of the hw).
It is as if you were to run an original Macintosh with late 80's
versions of the Mac OS. Although the basic hardware was the same,
the software had years of evolution/features.
That's the problem with preservation of software. No one bothers
to save the old versions, unless they have dusty decks that wouldn't
work after a software upgrade.
Joe Morris
>Someone asked me a few months ago what would be a good system to
>have running on display at the CMHC. I told him a 360/50, since
>I thought it was old enough to be historically interesting, and
>small enough that you might have a chance to keep one running,
>and it had an interesting operators console :-)
...especially the DC OFF lever switch on the upper left. I recall
being in the IBM Data Center in Atlanta (my shop was doing prep
work for the soon-to-be-delivered 360/65) and watched my local
IBM sales rep casually lean his hand against the console while
we were talking. He got tired of leaning against the console
and straightened up -- and his hand dropped along the front of
the console, hitting the (unguarded) DC OFF switch. Sadly, pressing
the lever *down* activated the DC OFF function.
We didn't have anything running on that machine when he hit the
switch, so it wouldn't have been so much of a problem -- except that
the /50 was CTCA-coupled to an adjacent 360/40 that was running
production SBC jobs. (There was *nothing* to warn us that the
two machines were linked.) The data center manager threw a
tantrum in front of all the customers on the floor; I don't know if
there was any link to it but the next time we went to Atlanta
for work there the center had a new manager.
Joe Morris
Don Chiasson
> From article <9f61jc$ohs$1...@ausnews.austin.ibm.com>,
> by gla...@glass2.lexington.ibm.com:
>
> >>(I recall purchasing an 8KiWord "extension memory" for an Elliott 920B
in
> >>1968. This worked out at almost exactly one pound per (18-bit) word.
>
> I have a very early 1960's vintage core plane (never used, with factory
> baggies still taped over each bundle of wires) with the original wrapping
> plastic. It's 8K bits, and the price on the wrapper is 8865.02
In the early 70's, I bought 1.5 microsecond core memory for a
Honeywell H316 and it cost $4k for 4k words, 8K bytes. The computer could
hold a max of 32k words.
At that price, 1MB would have cost $500k. As a guess, each 4k unit
was about 3"x3"x12" or 1/16 of a cubic foot. Thus 1MB would have occupied
about eight cubic feet, plus power supplies, cooling fans and cabling.
Don
e-mail: it's not not, it's hot.
Heinz W. Wiggeshoff
MFT yes. But MVT, unless severely chopped, I don't think so.
Jay Maynard
wrote:
>Jay Maynard (jmay...@thebrain.conmicro.cx) writes:
>> The 360/50 would run MVT quite well..
> MFT yes. But MVT, unless severely chopped, I don't think so.
There are a bunch of references to running MVT on a 360/40 with the storage
key protect feature...so a /50 should be enough, as long as you don't try
pushing the envelope, I would think...but then again, I wasn't around in
those days,so I'm only guessing.
Heinz W. Wiggeshoff
Well, I was at Transport 1970-72. I can't recall the core size, but
512K would be most likely on the /50 running MFT. During the day,
a 128K partition was humungous. About the same time, I took a project
out to a service bureau that ran MVT on an /85 with a honking 2 Mb of
core. Lots of room for HASP, several job initiators etc. CPU usage
was closely metered there and charged at $1200/hour (Cdn).
Charlie Gibbs
(Heinz W. Wiggeshoff) writes:
> Option 1) You write the program onto an approved coding form, such as
> the IBM SYSTEM/360 ASSEMBLER CODING FORM GX09-0010-04,
> which I also used for PL/I (and 24 line by 80 col. screen
> layouts). The result is sent for keypunching and verifying.
> You may get a red face when the keypuncher finds logic
> errors.
Verifying? Program cards were never considered worthy of
verifying. I found it faster and easier to sneak onto a
keypunch and do it myself, rather than get writer's cramp
using those damned coding forms - even for RPG.
--
cgi...@sky.bus.com (Charlie Gibbs)
Remove the first period after the "at" sign to reply.
Charlie Gibbs
(Michael Roach) writes:
>You can put a cup of coffee in an old computer to warm it up and not
>worry about it spilling 8^)
An operator made the mistake of putting his coffee cup on the card
punch. When the punch started up its lid vibrated enough to make
the cup walk right off the edge... <splat>
jmfb...@aol.com
never...@panics.com.invalid (Michael Roach) wrote:
>
>In article <9f5q1m$b8s$1...@flood.weeg.uiowa.edu>,
>Douglas W. Jones,201H MLH,3193350740,3193382879 <jo...@cs.uiowa.edu>
wrote:
>>From article <9f3vii$f...@netaxs.com>, by lwi...@bbs.cpcn.com (lwin):
>I've been wondering...where do you get new (or usable) dectapes?
Towards the end of TOPS-10, SDC ran out of DECtapes, and couldn't
order them. A panic button was pushed. Coincidently, the Software
Support group was getting disbanded causing them to dump the
contents of their offices into dumpsters. We were sifting
through dumpsters to retrieve all the DECtapes. I get sick
just thinking off everything that was thrown away that week.
>Are
>the originals still usable?
Yup.
>How old is the oldest original that's still
>usable?
<shrug> I've got one laying around here that I wrote stuff
on in 1972.
>Is it true you can wind vcr tape onto a spool and have it work?
I have no idea. The feel of the tape is different. And isn't
the DECtape wider?
>
<snip>
/BAH
Subtract a hundred and four for e-mail.
David Horvath, CCP
>
>x-no-archive: yes
>
>I've noticed on ebay several sales for computer magnetic core
>memory. I was surprised to see it to be fairly expensive and
>a number of people bidding on each offered item.
>
>As one interested in computer history, I wouldn't mind owning a nice
>chunk of core; but frankly not so much at the prices being bid. (I did
>get one sheet from the Boston Computer Museum).
There is something about core that chip-memory just doesn't have -- the ability
to see the bits (not their status though). It is also non-volitile. I
remember playing around with an IBM System/1130 that had 16 Kb of 3.6 uSec
core. You could stop the system, cycle the power (even leaving it off
overnight) and restart it in the morning from the point it stopped!
Personally, I have a couple of DEC PDP/11 core memory boards (sorry, not for
sale) that had been replaced with MOS memory. One of these days I'll get around
to getting them mounted and framed for my wall.
- David
--
David B. Horvath, CCP dhor...@nosuch.cobs.com
Consultant, International Lecturer, Adjunct Professor
Author of "UNIX for the Mainframer" and other books.
*** remove "nosuch." when replying ***
Joe Morris
Um...that's stretching it a bit. The largest 360/40 was the H model (with
the huge *256K BYTES* of memory). It might be possible to shoehorn an MVT
system into the machine, but it would be tight, and the largest region
would probably be unusable for any application that could make use of
the features that MVT offered above and beyond what was in MFT.
Performance would have been a bear as well, since the /40 had a 2-byte
bus with a 2 usec memory cycle.
My PPOE did run MVT with HASP in a 512 MB system, shoehorning the
OS, HASP, and all related baggage into 256 KB, but only by stripping
out everything possible. (I even wrote a transient processor for HASP
although it was abandoned when we finally were able to reduce the
footprint to 256KB without it.) We ran with it for about nine months
until our budget allowed us to buy a 1 MB box of 8 usec/dword LCS.
(Even with this we were tight for fast memory. The folk at ORNL and I
eventually wrote local mods to IPLTXT to allow us to load parts of
the nucleus into slow memory.)
Joe Morris
Joe Morris
>In article <9f64bn$51f$2...@news.panix.com> never...@panics.com.invalid
>(Michael Roach) writes:
>>You can put a cup of coffee in an old computer to warm it up and not
>>worry about it spilling 8^)
>An operator made the mistake of putting his coffee cup on the card
>punch. When the punch started up its lid vibrated enough to make
>the cup walk right off the edge... <splat>
...and very few operators failed to learn by the hard way that one does
not put *anything* on the top of a 1403-N1 printer. Each newcomer
had to experience the "sudden sinking feeling" for him/herself.
(The 1403-N1 printer had a motor-driven cover that opened automatically
whenever the printer needed attention, such as paper-out.)
Joe Morris
CBFalconer
>
... snip ...
>
> There is something about core that chip-memory just doesn't have -- the ability
> to see the bits (not their status though). It is also non-volitile. I
> remember playing around with an IBM System/1130 that had 16 Kb of 3.6 uSec
> core. You could stop the system, cycle the power (even leaving it off
> overnight) and restart it in the morning from the point it stopped!
Provided you took some care in the drivers and power down
sequencing, not to mention power up. One of the selling points of
my 1965 era desktop machine was to pull the plug during a demo,
make suitable selling incantations, re-insert it, and carry on.
Jeff Jonas
>>> history and collecting old computer stuff?
>>Those of us who try to collect working machines don't approve of those who
>>break up the machines to make trophies people can hang on their office
>>walls. Much of the commerce on Ebay is in such trophies!
I want to agree but where are these folks when I had large things that
needed a new home? I had to pull apart things before they were discarded
since space is a rare thing even in colleges.
I'd love to have saved the 2: IBM system 1130s I used in college,
for nothing gives the feel of an I/O system than seeing the lights
for the dedicated registers and being able to halt and single step the
entire system with REAL switches and lights, not this virtual-debugger stuff!
But I did not have a house or the space for them either :-(
At least I gave the core planes to fellow EEs who appreciated them.
I have my core planes on display when possible
with a huge button "I am rotten to the core" and a little sign
"Jeff's core museum, entrance: 2 bits" (yet nobody ever left a quarter).
I wish I had taken photos of the IBM Kingston NY facility's
"Historical Society" display: they showed how the huge trunk-sized
core racks were made, tested and repaired!
Even the IBM archives don't have that
(but they did send me a copy of the IBM songbook!)
--
Jeffrey Jonas
jeffj@panix(dot)com
The original Dr. JCL and Mr .hide
Michael Roach
In article <465.551T51...@sky.bus.com>,
Charlie Gibbs <cgi...@sky.bus.com> wrote:
>In article <9f64bn$51f$2...@news.panix.com> never...@panics.com.invalid
>(Michael Roach) writes:
>
>>You can put a cup of coffee in an old computer to warm it up and not
>>worry about it spilling 8^)
>
>An operator made the mistake of putting his coffee cup on the card
>punch. When the punch started up its lid vibrated enough to make
>the cup walk right off the edge... <splat>
I once created a similar configuration with a washing machine and an
open bottle of bleach 8^o
OK so you have to put the cup in a rack that has no moving parts.
--
Ducharme's Axiom:
If you view your problem closely enough you will recognize
yourself as part of the problem.
Al Kossow
> where are these folks when I had large things that
> needed a new home?
How long ago did you have to dispose of this stuff?
The net has helped a LOT to get the word out that people are
interested in preserving old hardware.
It's unfortunate that you had to scrap out the 1130's, since
there are a couple of groups now trying to either get one or
restore the ones that they have. The software and docs are
even harder to find, since it is easier to throw out.
Another example would be the core assembly in TCMHC's 1620.
It isn't working, and apparently they weren't able to find
another one, since they had to build a semiconductor replacement
to get the machine into a running state.
I'm sure there's one hanging on someone's wall now that could
have been used.
gla...@glass2.lexington.ibm.com
>jmay...@thebrain.conmicro.cx (Jay Maynard) writes:
>
>>On 1 Jun 2001 00:26:09 GMT, Heinz W. Wiggeshoff <ab...@FreeNet.Carleton.CA>
>>wrote:
>>>Jay Maynard (jmay...@thebrain.conmicro.cx) writes:
>>>> The 360/50 would run MVT quite well..
>>> MFT yes. But MVT, unless severely chopped, I don't think so.
>
>>There are a bunch of references to running MVT on a 360/40 with the storage
>>key protect feature...so a /50 should be enough, as long as you don't try
>>pushing the envelope, I would think...but then again, I wasn't around in
>>those days,so I'm only guessing.
>
>Um...that's stretching it a bit. The largest 360/40 was the H model (with
>the huge *256K BYTES* of memory). It might be possible to shoehorn an MVT
>system into the machine, but it would be tight, and the largest region
>would probably be unusable for any application that could make use of
>the features that MVT offered above and beyond what was in MFT.
>
>Performance would have been a bear as well, since the /40 had a 2-byte
>bus with a 2 usec memory cycle.
I show the /40 having a 2-byte bus, but with a 2.5 usec memory cycle time
(at least, if I'm reading the chart correctly). See the chart I've added
below.
>
>My PPOE did run MVT with HASP in a 512 MB system, shoehorning the
|
Err, I hope you meant KB!
>OS, HASP, and all related baggage into 256 KB, but only by stripping
>out everything possible. (I even wrote a transient processor for HASP
>although it was abandoned when we finally were able to reduce the
>footprint to 256KB without it.) We ran with it for about nine months
>until our budget allowed us to buy a 1 MB box of 8 usec/dword LCS.
>
>(Even with this we were tight for fast memory. The folk at ORNL and I
>eventually wrote local mods to IPLTXT to allow us to load parts of
>the nucleus into slow memory.)
>
>Joe Morris
Here's a summary from a table in appendix D of the book 'Programming
the IBM 360' by Clarence B. Germain (1967):
360 MODELS
# Core Size 25 30 40 50 65 75 85 91-95 44 67 20
B 4K......X...........................................X.
C 8K......X...X.......................................X.
D 16K......X...X...X...................................X.
E 32K......X...X...X...........................X.........
F 64K..........X...X...X.......................X.........
G 128K..............X...X...X...................X.........
H 256K..............X...X...X...X...............X..X......
I 512K..................X...X...X...X..............X......
J 1024K..................X...X...X...X.....X........X......
K 2048K..............................X.....X........X......
L 4096K..............................X.....X...............
2361 1024K..................X...X...X.........................
2361 2048K..................X...X...X.........................
CD 24K....................................................
DE 48K....................................................
Cycle time(uS) .9 1.5 2.5 2.0 .75 .75 1.04 .75-.125 1.0 .75 3.6
per _ bytes 2 1 2 4 8 8 16 8 4 8 1
This agrees with the information contained in my copy of
'IBM Field Engineering Education Student Self-Study Course'
SR23-3062-7 (November 1973).
gla...@glass2.lexington.ibm.com
>x-no-archive: yes
>
>> Of course. Hey, people still keep steam engines running. Why do they do
>> it? Because they like it that way 8^)
>
>But aesthectically, there's a big difference between a running
>steam engine and an electronic computer.
>
>engine, to its fans, has lots of neat sounds, sensations, and visuals.
>
>(And I think there's a lot more railroad buffs out there willing to
>do the incredible amount necessary to rehab and run a steam engine
>than there are computer buffs willing to do likewise to old hardware.)
>
>I must admit I'd love to see a working S/360, just for nostalgia's
>sake. But I also must admit I'm NOT sure I'd be so quick as to run
>down and write a program, keypunch up the cards, wait in line at
>the card reader, wait for the job to be run, and wait for the printout
>only to discover I left out a single comma and have to do it all
>over again.
Well, I don't know about an S/360, but some of us are still running S/370
machines:
CP Q CPLEVEL
VM/ESA Release 1, Service Level 111 - 370 Feature
Generated at 10/13/93 09:11:15 LCL
IPL at 05/30/01 16:50:10 EST
Ok, so it's out of date, and completely unsupported, it still runs.
Nick Spalding
> "David Horvath, CCP" wrote:
> >
> ... snip ...
> >
> > There is something about core that chip-memory just doesn't have -- the ability
> > to see the bits (not their status though). It is also non-volitile. I
> > remember playing around with an IBM System/1130 that had 16 Kb of 3.6 uSec
> > core. You could stop the system, cycle the power (even leaving it off
> > overnight) and restart it in the morning from the point it stopped!
>
> Provided you took some care in the drivers and power down
> sequencing, not to mention power up. One of the selling points of
> my 1965 era desktop machine was to pull the plug during a demo,
> make suitable selling incantations, re-insert it, and carry on.
I remember visiting an installation where the machine (I can't
remember what its name was but it was about the size and shape of a
Nova 2) had an excellent power fail recovery mechanism. It had a
teletype as its input device and it ran a Fortran system that had to
be loaded off paper tape via the 75cps reader which took about half an
hour. If you switched it off this process had to be gone through so
the on-off switch was taped in the on position and to shut down they
pulled the plug out of the wall socket. Next day, just plug it in
again and off it went.
--
Nick Spalding
Jim Saum
>jmay...@thebrain.conmicro.cx (Jay Maynard) writes:
>
>>On 1 Jun 2001 00:26:09 GMT, Heinz W. Wiggeshoff <ab...@FreeNet.Carleton.CA>
>>wrote:
>>>Jay Maynard (jmay...@thebrain.conmicro.cx) writes:
>>>> The 360/50 would run MVT quite well..
>>> MFT yes. But MVT, unless severely chopped, I don't think so.
>
>>There are a bunch of references to running MVT on a 360/40 with the storage
>>key protect feature...so a /50 should be enough, as long as you don't try
>>pushing the envelope, I would think...but then again, I wasn't around in
>>those days,so I'm only guessing.
>
>Um...that's stretching it a bit. The largest 360/40 was the H model (with
>the huge *256K BYTES* of memory). It might be possible to shoehorn an MVT
>system into the machine, but it would be tight, and the largest region
>would probably be unusable for any application that could make use of
>the features that MVT offered above and beyond what was in MFT.
IIRC, the official minimum supported memory sizes in the late OS/360
releases were 64K for PCP, 128K for MFT, 256K for MVT, and (after
release 20) 384K for MVT with TSO. However, these were too small to
get anything real done. I think they were just enough to IPL the
system, process operator commands, and run trivial-size jobs.
(Although a colleague of mine claimed he had been able to do useful
work in the late 1960s on a 64K 360/30 under PCP.)
- Jim Saum
Steve O'Hara-Smith
"Don Chiasson" <don_ch...@notmail.com> wrote:
DC> At that price, 1MB would have cost $500k. As a guess, each 4k unit
DC> was about 3"x3"x12" or 1/16 of a cubic foot. Thus 1MB would have occupied
DC> about eight cubic feet, plus power supplies, cooling fans and cabling.
That would seem about right, I recall seeing a 1Mb core in ~1977 at
Cambridge just before it was ripped out and replaced by 4Mb of semiconductor
RAM, something around 18" square and most of the way to the ceiling IIRC.
--
Directable Mirrors - A Better Way To Focus The Sun
gle...@encompasserve.org
>>
> ... snip ...
>>
>> There is something about core that chip-memory just doesn't have -- the ability
>> to see the bits (not their status though). It is also non-volitile. I
>> remember playing around with an IBM System/1130 that had 16 Kb of 3.6 uSec
>> core. You could stop the system, cycle the power (even leaving it off
>> overnight) and restart it in the morning from the point it stopped!
>
My first computing job was as an operator/technician of some
dedicated computer driven typesetting machines, circa 1976. We had two
of them, each with a whopping 32 Kb of core memory. Boot and load
of the monitor program was done via paper tape. After loading, they
could be powered on and off,only needing a reload when the code
went wild or diags cleared things out.
One day, the reader on one of the croaked it, and we absolutely
had to get it on the air. I thought about it a while, and loaded
its code into the other system - then I powered them down, transferred
the four memory cards to the machine with the dead reader, and powered
it back up. The memory contents were still valid, and it was back
on the air.
Lee K. Gleason N5ZMR
Control-G Consultants
lgleason at houston dot rr dot com
gla...@glass2.lexington.ibm.com
>
>In article <9f63lv$q...@netaxs.com>, lwin <lwi...@bbs.cpcn.com> wrote:
>>x-no-archive: yes
>>
>>> Of course. Hey, people still keep steam engines running. Why do they do
>>> it? Because they like it that way 8^)
>>
>>But aesthectically, there's a big difference between a running
>>steam engine and an electronic computer.
>>
>>A running computer just sits there. In contrast, a running steam
>>engine, to its fans, has lots of neat sounds, sensations, and visuals.
>
>
>You can put a cup of coffee in an old computer to warm it up and not
>worry about it spilling 8^)
Err, spilling the cup of coffee or the computer? After all, some old
computers were water cooled, and some of the old DASD and printers used
hydraulic actuators. :*)
>>(And I think there's a lot more railroad buffs out there willing to
>>do the incredible amount necessary to rehab and run a steam engine
>>than there are computer buffs willing to do likewise to old hardware.)
>>
>>I must admit I'd love to see a working S/360, just for nostalgia's
>>sake. But I also must admit I'm NOT sure I'd be so quick as to run
>>down and write a program, keypunch up the cards, wait in line at
>>the card reader, wait for the job to be run, and wait for the printout
>>only to discover I left out a single comma and have to do it all
>>over again.
>
>I have a hunch that many of the old computers running today are personal
>computers so you won't have to wait in line.
>--
>Loose bits sink chips.
gla...@glass2.lexington.ibm.com
>In article <3B169A5E...@my-deja.com>, cbfal...@worldnet.att.net wrote:
>
>>
>> Unfortunately Charlie died since then. His boss was DA Bromley,
>> Physics Chairman. Maybe you can get one through him :-) There
>> are incredible things in the warrens beneath Yale.
>
>Someone asked me a few months ago what would be a good system to
>have running on display at the CMHC. I told him a 360/50, since
>I thought it was old enough to be historically interesting, and
>small enough that you might have a chance to keep one running,
>and it had an interesting operators console :-)
>
>and if you did find it, could you find any mid/late 60's software
>for it.
Somewhere, I have a standalone COFFEE application demonstration program[1]
that will IPL from a S/360 or S/370 card reader. Of course, it needs a
3215 or similar console device for I/O.
[1] Some people might be tempted to call this a game program, but we all
know that you're not supposed to run games on mainframes. :*)
GerardS
|> Al Kossow wrote:
|>Unfortunately, I wonder if there is a complete one still to be found,
|>and if you did find it, could you find any mid/late 60's software
|>for it.
|
| The 360/50 would run MVT quite well...and that's still pretty easy to
lay
| hands on, thanks to Rick Fochtman and Sam Knutson. You can get it from
| http://www.cbttape.org .
... providing it had at least 256K. It could run on a 128K machine (as
we
tried to do that, and that was with MVT 13 or 14 or so, and it ran like
a
dog (no room for HASP either), and we ended up with (as I recall) about
a
80K region, and that's with no readers or writers running. We had to
revert to MFT to get a 98K partition, plus a reader and writer.
Gerard S.
Charlie Gibbs
gle...@encompasserve.org (gleason) writes:
> One day, the reader on one of the croaked it, and we absolutely
>had to get it on the air. I thought about it a while, and loaded
>its code into the other system - then I powered them down, transferred
>the four memory cards to the machine with the dead reader, and powered
>it back up. The memory contents were still valid, and it was back
>on the air.
Cute.
Once upon a time a friend of mine built IMSAI 8080s and used them as
serial port concentrators for a B1700. (We each wound up with a full
IMSAI's worth of spare parts. :-) The boxes out in the field had no
peripherals aside from the serial ports. To get new updates to them
he'd attach a 6-volt lantern battery to the backup power pads on a
spare memory board (an IMSAI RAM-4A board, 4K worth of 2102 chips),
drop it into the development box, and copy the code onto the board.
Then he'd power down the development box, pull the memory board,
and drive out to the remote site with it, with the battery keeping
the chips alive. At the remote site he'd do the reverse.
CBFalconer
Back about the same era (mid 70's) my hardware also had 4k 2102
boards (not IMSAI etc, I considered that hardware an abortion).
Somebody produced a giant 16k static chip, the firm was named
something like EMI (but not EMI) on 24 pin dips, and I built a 16k
board around them. It worked fine, after I learned a few things
not to do.
The interesting thing about the chips, as I found by accident, was
that they held their data with power off. Most of the time I
could do a CRC over a memory area, kill the power, go back on
later, and the CRC over the same area was unchanged. I have no
idea what was going on in them. I even told the firm what I had
found, since I thought they were on the track to an electrically
reprogrammable PROM, and that would have been a big deal then.
Everything died off with nary a whimper.
Don Quixote
[Wonderful core]
> One day, the reader on one of the croaked it, and we absolutely
>had to get it on the air. I thought about it a while, and loaded
>its code into the other system - then I powered them down, transferred
>the four memory cards to the machine with the dead reader, and powered
>it back up. The memory contents were still valid, and it was back
>on the air.
AOL.
Did this with GTE equipment (IS/1000), three 16K boards as I recall. My boxes
were side by side though.
--
'After I started drinking yesterday I didn't do anything else stupid.
That would seem to speak for itself.'
Adam, in the Monastery.
Jim Stephens
Wiggeshoff) wrote:
The 360/50 at University of Missouri, Rolla ran with 512K of main
memory, and 1mb of LCS. Originally it ran MVT, Waterloo Fortran,
and about 30 terminals running CPS.
The CPS system was taken down and the 2741's were all allocated
to be remote TSO terminals for the UMC 370/148's system.
The LCS was then free to have two wonderful 512mb initiators, or
if you requested it 1mb.
You had to request that the Waterloo system be taken down, and
about 1 time / hour 1mb jobs would be run.
One heavy user of the large initiator was a researcher doing Xray
crystalography.
Also the system ran Hasp all the time
DASD was 9 2314's and Tape was 2 - 2401 (9 track).
Power was by motor generator out in back of the computer center,
the system did not take power directly off the mains w/o going thru
a converter.
jmfb...@aol.com
Ptui. Games were really good for load tests, swap tests, and
comm tests. However, there can be too much of a good thing.
/BAH
Subtract a hundred and four for e-mail.
Chris Baird
> those who break up the machines to make trophies people can hang on
> their office walls.
In my experience, most of these trophy hunters are pretender wannabes;
who coincidentally now all own Leathermans, Dilbert t-shirts, and Nerf
weapons because they're such status symbols in alt.sysadmin.recovery.
One example ripped a Pyramid Mini apart so he could use the mainboards
inside as wallpaper for the equipment room of an Internet Cafe. :p
--
Chris,,
"And programmers really didn't drink beer before Linus Torvalds made
it hip with that 'virtual beer' crap."
Joe Morris
> jcmo...@jmorris-pc.MITRE.ORG (Joe Morris) writes:
>>My PPOE did run MVT with HASP in a 512 MB system, shoehorning the
> |
> Err, I hope you meant KB!
Errr...yes. In fact, I was having fun with a similar mistake (but
in the opposite direction) on the part of one of our group leaders
yesterday when we were talking about hard disk size in new systems:
he was suggesting that we standardize on 30 MB disks.
> 360 MODELS
> # Core Size 25 30 40 50 65 75 85 91-95 44 67 20
> B 4K......X...........................................X.
> C 8K......X...X.......................................X.
[etc]
There were also some hybrids, especially in the CPU boxes that used
external memory. Our /65, for example, eventually got a third
quadrant of memory for a total of 768 KB (yes, *KB*. I got it right
this time), causing its model number to change from 65 I (512 KB)
to 65 IH (768 KB). And for the newer readers here, I should note
that this upgrade required that the entire CPU be reclocked because
of the wire lengths to the outside memory box...so it wasn't just
a matter of dropping another DIMM into a socket.
Joe Morris
Joe Morris
>IIRC, the official minimum supported memory sizes in the late OS/360
>releases were 64K for PCP, 128K for MFT, 256K for MVT, and (after
>release 20) 384K for MVT with TSO. However, these were too small to
>get anything real done. I think they were just enough to IPL the
>system, process operator commands, and run trivial-size jobs.
>(Although a colleague of mine claimed he had been able to do useful
>work in the late 1960s on a 64K 360/30 under PCP.)
My not-at-all-infallible memory says that in our /40 I had built a PCP
system with an 18K nucleus, so in a 64K machine that would still have left
46K for the user program -- easily enough for the E-level language
translators and still far larger than all but maybe a few 14xx systems.
The MFT nucleus was only 28K (up from 26K when we upgraded to version 14),
meaning that in 10K we got extra features like multiple partitions [1],
asynchronous console support and error recovery, and the like.
[1] the word "partition" here refers to *memory* allocation, not
*disk* allocation as the used by Microsoft.
Joe Morris
Anne & Lynn Wheeler
> The MFT nucleus was only 28K (up from 26K when we upgraded to version 14),
> meaning that in 10K we got extra features like multiple partitions [1],
> asynchronous console support and error recovery, and the like.
part of my share '68 presentation on MFT 14 and CP/67 performance
optimization for a 768k 360/67.
http://www.garlic.com/~lynn/94.html#18
The kernel was expanded to 82k bytes with various resident options. The
job scheduler was defined as 100k bytes (aka it needed 100k region in order
to start a process/job step).
HASP was gen'ed at 118k bytes using 1/3rd 2314 tracking buffering.
Effectively 82kbyte kernel plus resident HASP 118kbytes resulting in
200kbyte fixed storage requirements.
That left 768kbytes-200kbytes ... 568kbytes for a few regions.
in something similar to the above work ... a recent thread on job step
processing overhead ("estimate jcl overhead")
http://www.garlic.com/~lynn/2001e.html#60
http://www.garlic.com/~lynn/2001e.html#61
http://www.garlic.com/~lynn/2001e.html#68
--
Anne & Lynn Wheeler | ly...@garlic.com - http://www.garlic.com/~lynn/
Jim Saum
>The MFT nucleus was only 28K (up from 26K when we upgraded to version 14),
>meaning that in 10K we got extra features like multiple partitions [1],
>asynchronous console support and error recovery, and the like.
I assume that was the original MFT. I ran a Rel 19.6 MFT-II nucleus
that was 40K.
- Jim Saum
GerardS
|> jcmo...@mitre.org wrote:
|
| >The MFT nucleus was only 28K (up from 26K when we upgraded to version
14),
| >meaning that in 10K we got extra features like multiple partitions
[1],
| >asynchronous console support and error recovery, and the like.
|
| I assume that was the original MFT. I ran a Rel 19.6 MFT-II nucleus
| that was 40K.
I remember our MFT nuc (at NDSU) was 18.something K. I don't remember
the Rel, though, this was in 1969 and I think the Rel was the early
teens.
I believe we started with Rel 9 or 10 (in 1967), replacing an IBM 1620.
Gerard S.
jchausler
Michael Roach wrote:
> In article <9f5q1m$b8s$1...@flood.weeg.uiowa.edu>,
> Douglas W. Jones,201H MLH,3193350740,3193382879 <jo...@cs.uiowa.edu> wrote:
>
> I've been wondering...where do you get new (or usable) dectapes? Are
> the originals still usable? How old is the oldest original that's still
> usable? Is it true you can wind vcr tape onto a spool and have it work?
I just saw a dual drive DECtape unit for sale for $15 at a flea market
yesterday. Unfortunately it had been outside for a while and would
require serious effort to make functional. I was tempted but having
too much junk awaiting resurrection already, passed it by. And although
I have a number of 30 year old DECtapes, I'm not sure I'd want to
risk running them on anything but a tested fully functioning drive. I'm
a little fuzzy on this, but I when magnetic tape is manufactured, I
understand the magnetic particles are "aligned" on the tape depending
on intended use although I've never read any details on this so I may
be dead wrong. Supposedly the alignment on standard audio tape (and
"standard" computer tape) is along the length of the tape whereas
home (helical) video tape is diagonally across the tape and old time
"quad" as was used for broadcast video for many years was directly
across the tape. This is essentially always in the direction the head
moves across the tape. However, it is my understanding that the
specs for DECtape were so loose that almost anything of the right
width would work. DECtape is wider than home video tape. I'm not
home at the moment and so cannot measure it but I think its around
3/4 inch wide. IIRC the original "professional" helical video cartridges,
Sony Umatic were of this width.
> >Those of us who try to collect working machines don't approve of those who
> >break up the machines to make trophies people can hang on their office
> >walls. Much of the commerce on Ebay is in such trophies!
Unfortunately, sometimes all one can afford to do is save the "trophies". I
have saved a number of front panels where the rest of the machine was
either out and out junked or parted out. Had I not saved the front panel,
it too would have gone in the trash. On a couple of occasions, the machine
was already in the dumpster when I got to it.
Chris
AN GETTO$;DUMP;RUN,ALGOL,TAPE
$$
jchausler
Charlie Gibbs wrote:
> In article <9f67am$t4e$1...@freenet9.carleton.ca> ab...@FreeNet.Carleton.CA
> (Heinz W. Wiggeshoff) writes:
>
> > Option 1) You write the program onto an approved coding form, such as
> > the IBM SYSTEM/360 ASSEMBLER CODING FORM GX09-0010-04,
> > which I also used for PL/I (and 24 line by 80 col. screen
> > layouts). The result is sent for keypunching and verifying.
> > You may get a red face when the keypuncher finds logic
> > errors.
>
> Verifying? Program cards were never considered worthy of
> verifying. I found it faster and easier to sneak onto a
> keypunch and do it myself, rather than get writer's cramp
> using those damned coding forms - even for RPG.
Briefly in my first "real job" in the early 70's I was forced to use
coding forms and have "lower paid" keypunchers do the
punching for me. This resulted in so many keypunching errors
that I complained to "management" (hoping to be able to
do the punching myself during which I would sometimes find
logic errors as the act of keypunching for me acted like another
pass at a desk check). Management's solution was to require
verifying. Unfortunately, not only would the verify step sometimes
not find an error, it would mark as incorrect places that were
correct in the original effort. I sometimes wonder how accidental
this was as the keypunchers did not like punching or verifying
program decks.......Fortunately, "management" eventually came
to its senses.
Keith F. Lynch
> A running computer just sits there. In contrast, a running steam
> engine, to its fans, has lots of neat sounds, sensations, and visuals.
So do many early computer peripherals. For instance an ASR-33
teletype, which, with its noise, vibration, and smell of hot oil,
makes 110 baud seem fast.
--
Keith F. Lynch - k...@keithlynch.net - http://keithlynch.net/
I always welcome replies to my e-mail, postings, and web pages, but
unsolicited bulk e-mail sent to thousands of randomly collected
addresses is not acceptable, and I do complain to the spammer's ISP.
Al Kossow
<jcha...@earthlink.net> wrote:
> I just saw a dual drive DECtape unit for sale for $15 at a flea market
> yesterday. Unfortunately it had been outside for a while and would
> require serious effort to make functional.
There are several people that read this group that would be
willing to make that effort, esp if it had G888 read amps in
it.
jmfb...@aol.com
jchausler <jcha...@earthlink.net> wrote:
>
>
>Charlie Gibbs wrote:
>
>> In article <9f67am$t4e$1...@freenet9.carleton.ca> ab...@FreeNet.Carleton.CA
>> (Heinz W. Wiggeshoff) writes:
>>
>> > Option 1) You write the program onto an approved coding form, such as
>> > the IBM SYSTEM/360 ASSEMBLER CODING FORM GX09-0010-04,
>> > which I also used for PL/I (and 24 line by 80 col. screen
>> > layouts). The result is sent for keypunching and
verifying.
>> > You may get a red face when the keypuncher finds logic
>> > errors.
>>
>> Verifying? Program cards were never considered worthy of
>> verifying. I found it faster and easier to sneak onto a
>> keypunch and do it myself, rather than get writer's cramp
>> using those damned coding forms - even for RPG.
>
>Briefly in my first "real job" in the early 70's I was forced to use
>coding forms and have "lower paid" keypunchers do the
>punching for me. This resulted in so many keypunching errors
>that I complained to "management" (hoping to be able to
>do the punching myself during which I would sometimes find
>logic errors as the act of keypunching for me acted like another
>pass at a desk check).
<grin> You should just admit that you wanted to play with
the keypunch yourself. Coding on papers isn't any fun if
you can't play with the equipment while doing it.
> ... Management's solution was to require
>verifying. Unfortunately, not only would the verify step sometimes
>not find an error, it would mark as incorrect places that were
>correct in the original effort.
It sounds like your shop had no coding standards. For instance
when tabs or spaces were used. How comments were entered. How
the programmer wrote zeroes, letter Zs, numbers, etc.
Or, the keypunch group were applying data interpretation standards
to code which does not work at all on cards.
Also I bet that shop had a different person verifying than
the one who keyed it in. That was another failsafe mechanism
for data, but I don't see how it could be successful with
code.
> I sometimes wonder how accidental
>this was as the keypunchers did not like punching or verifying
>program decks.......Fortunately, "management" eventually came
>to its senses.
Did you ever ask them why they didn't like it? Programmers
could be a real pain in the keester because they didn't know
a thing about data entry processes.
Joe Morris
Absolutely. MFT-II (introduced at OS/360 release 16) added all sorts of
goodies that had been available only under MVT, but made the nucleus
so large that our 128 KB machine had no hopes of ever running it. We
had to wait until the $$ were available to buy the /65 in late 1969.
Joe Morris
Joe Morris
>I remember our MFT nuc (at NDSU) was 18.something K. I don't remember
>the Rel, though, this was in 1969 and I think the Rel was the early
>teens.
>I believe we started with Rel 9 or 10 (in 1967), replacing an IBM 1620.
OS/360 release 12 arrived (IIRC) in early 1967, and 13 showed up later
that year. I can place it since my shop got its first /360 (the /40 that
I've referred to in other postings) in 3Q67, and our initial tests were
run under release 12. Fortunately, release 13 came out before we went
into production; one urgent problem I recall was that the R12 COBOL-F
library blew up whenever it issued a WTOR (request for operator text
input).
*Why* I recall that bug from 34 years ago I don't know...
Joe Morris
Joe Morris
>Briefly in my first "real job" in the early 70's I was forced to use
>coding forms and have "lower paid" keypunchers do the
>punching for me. This resulted in so many keypunching errors
>that I complained to "management" (hoping to be able to
>do the punching myself during which I would sometimes find
>logic errors as the act of keypunching for me acted like another
>pass at a desk check).
For a long time much of my programming technique was to write the code
on a legal pad with lots of lines shoehorned into odd places; I would
then sit down and keypunch it, composing the running commentary as
I went. Having to come up with the narrative as I keypunched it forced
me to think about what was actually going on, and often led to improved
(or at least corrected) design.
Joe Morris
Joe Morris
>So do many early computer peripherals. For instance an ASR-33
>teletype, which, with its noise, vibration, and smell of hot oil,
>makes 110 baud seem fast.
Well, it *is* fast if you've just graduated to an ASR-33 from a model 28.
(Assuming that the /28 hasn't been regeared to run at 100 wpm.)
Joe Morris
jchausler
Joe Morris wrote:
Yes, Exactly! I would make up the comments as I was keypunching.
jchausler
jmfb...@aol.com wrote:
> In article <3B1A695D...@earthlink.net>,
> jchausler <jcha...@earthlink.net> wrote:
> >Briefly in my first "real job" in the early 70's I was forced to use
> >coding forms and have "lower paid" keypunchers do the
> >punching for me. This resulted in so many keypunching errors
> >that I complained to "management" (hoping to be able to
> >do the punching myself during which I would sometimes find
> >logic errors as the act of keypunching for me acted like another
> >pass at a desk check).
>
> <grin> You should just admit that you wanted to play with
> the keypunch yourself. Coding on papers isn't any fun if
> you can't play with the equipment while doing it.
There was that ;-) but I can touch type reasonable well but my
handwriting has been compared to a doctor's, I cannot even read
it after a couple of days and filling in coding forms was out and
out painful to me.
> > ... Management's solution was to require
> >verifying. Unfortunately, not only would the verify step sometimes
> >not find an error, it would mark as incorrect places that were
> >correct in the original effort.
>
> It sounds like your shop had no coding standards. For instance
> when tabs or spaces were used. How comments were entered. How
> the programmer wrote zeroes, letter Zs, numbers, etc.
We had standards. I believe the "problem" was that it just made more
work for the keypunchers so they couldn't sit around and chat and
drink coffee as much. Further this was assembly code mixed with
numeric and text comments and was unlike the highly structured stuff
they usually did. I also don't think they or there management thought
it was important. This was the "data processing" department doing the
keypunching. I was in "systems engineering" and our coding was for
real time control customer systems, i.e. no benefit (as far as they
could see despite that this was the "product" we were selling and thus
eventually what paid their wages) to their department.
> Or, the keypunch group were applying data interpretation standards
> to code which does not work at all on cards.
Again, it was a different style of data on the cards to what they were
familiar with.
> Also I bet that shop had a different person verifying than
> the one who keyed it in. That was another failsafe mechanism
> for data, but I don't see how it could be successful with
> code.
Yes, usually but again they didn't seem to care. That attitude
was "take what you get".
>
>
> > I sometimes wonder how accidental
> >this was as the keypunchers did not like punching or verifying
> >program decks.......Fortunately, "management" eventually came
> >to its senses.
>
> Did you ever ask them why they didn't like it? Programmers
> could be a real pain in the keester because they didn't know
> a thing about data entry processes.
"We" asked their management. Again there was a lack of desire
to perform the service across the board.
jchausler
"Keith F. Lynch" wrote:
> lwin <lwi...@bbs.cpcn.com> wrote:
> > A running computer just sits there. In contrast, a running steam
> > engine, to its fans, has lots of neat sounds, sensations, and visuals.
>
> So do many early computer peripherals. For instance an ASR-33
> teletype, which, with its noise, vibration, and smell of hot oil,
> makes 110 baud seem fast.
Yes, and I can still go down into my utility room and turn on one
of my 33's and experience it :-) I've always said the smell of hot
teletype oil at 3 AM in the morning (when you've been up all
night working on some project) is a unique experience.
Chris
AN GETTO$;DUMP;RUN,ALGOL,TAPE
$$
The above .sig was frequently typed on model 33's and 35's
back in the mid 60's in my neck of the woods.
Julian Thomas
>For a long time much of my programming technique was to write the code on
>a legal pad with lots of lines shoehorned into odd places; I would then
>sit down and keypunch it, composing the running commentary as I went.
>Having to come up with the narrative as I keypunched it forced me to
>think about what was actually going on, and often led to improved (or at
>least corrected) design.
I did the same, altho the handwritten section may have been less organized
than yours.
--
Julian Thomas: jt . jt-mj @ net http://jt-mj.net
remove letter a for email (or switch . and @)
In the beautiful Finger Lakes Wine Country of New York State!
Boardmember of POSSI.org - Phoenix OS/2 Society, Inc
http://www.possi.org
-- --
Never ascribe to malice that which is adequately explained by
incompetence. - Napoleon Bonaparte
Charlie Gibbs
(Joe Morris) writes:
That sounds exactly like my "coding forms", except that I usually
used the back of old program listings - or failing that, scrap
printer paper. Not only did I fill in comments and put together
all of those inserted bits written anywhere with arrows running
around the page to where they belonged, but I wouldn't bother
writing out "boilerplate" code at all. It's as if I was using
a mental preprocessor, with output feeding directly to the keypunch.
It saved *immense* amounts of time.
Of course, it was often hard to get time on a keypunch, which led to
my doing a lot of keypunching after hours. The alternative - sneaking
into the keypunch area when the girls were off for lunch or coffee -
was potentially dangerous if I interrupted one of their jobs and
cleared their programs and auto-dup fields.
Finally we got a little manual punch. This one was even more
primitive than that IBM 001 or whatever it was - it had a single
die which you slid up and down to the desired row after positioning
it to the desired column. Slow and painful, but it sometimes saved
the day when I desperately needed a fix but couldn't wheedle my way
onto a keypunch (and they'd have been too busy to look at a coding
form anyway). I stuck a little tag on it which read: "Programmers
have priority on this punch!"
jmfb...@aol.com
jchausler <jcha...@earthlink.net> wrote:
>
>
>jmfb...@aol.com wrote:
>
>> In article <3B1A695D...@earthlink.net>,
>> jchausler <jcha...@earthlink.net> wrote:
>> >Briefly in my first "real job" in the early 70's I was forced to use
>> >coding forms and have "lower paid" keypunchers do the
>> >punching for me. This resulted in so many keypunching errors
>> >that I complained to "management" (hoping to be able to
>> >do the punching myself during which I would sometimes find
>> >logic errors as the act of keypunching for me acted like another
>> >pass at a desk check).
>>
>> <grin> You should just admit that you wanted to play with
>> the keypunch yourself. Coding on papers isn't any fun if
>> you can't play with the equipment while doing it.
>
>There was that ;-) but I can touch type reasonable well but my
>handwriting has been compared to a doctor's, I cannot even read
>it after a couple of days and filling in coding forms was out and
>out painful to me.
Ah! Us Tape Prep girls were able to read anybody's hand writing.
;-)
>
>> > ... Management's solution was to require
>> >verifying. Unfortunately, not only would the verify step sometimes
>> >not find an error, it would mark as incorrect places that were
>> >correct in the original effort.
>>
>> It sounds like your shop had no coding standards. For instance
>> when tabs or spaces were used. How comments were entered. How
>> the programmer wrote zeroes, letter Zs, numbers, etc.
>
>We had standards. I believe the "problem" was that it just made more
>work for the keypunchers so they couldn't sit around and chat and
>drink coffee as much. Further this was assembly code mixed with
>numeric and text comments and was unlike the highly structured stuff
>they usually did. I also don't think they or there management thought
>it was important. This was the "data processing" department doing the
>keypunching.
Our company had two groups: one for the programmers and one for
pure data entry. We were different buildings and AFAICT, never
mixed...well...except for that one time some programmer got a bright
idea to use cards and I was the card expert...I was the only person
who knew about cards.
> I was in "systems engineering" and our coding was for
>real time control customer systems, i.e. no benefit (as far as they
>could see despite that this was the "product" we were selling and thus
>eventually what paid their wages) to their department.
I can see why mixing the two would be a bad idea.
>
>> Or, the keypunch group were applying data interpretation standards
>> to code which does not work at all on cards.
>
>Again, it was a different style of data on the cards to what they were
>familiar with.
Yup. Got that. In my very young years in this biz, I did both
types of keypunching.
<snip>
>> > I sometimes wonder how accidental
>> >this was as the keypunchers did not like punching or verifying
>> >program decks.......Fortunately, "management" eventually came
>> >to its senses.
>>
>> Did you ever ask them why they didn't like it? Programmers
>> could be a real pain in the keester because they didn't know
>> a thing about data entry processes.
>
>"We" asked their management.
Big mistake. You should have queried a worker with great
humility ;-). That's how I always gathered enough information
to design processes.
<snip>
Joe Morris
>That sounds exactly like my "coding forms", except that I usually
>used the back of old program listings - or failing that, scrap
>printer paper. Not only did I fill in comments and put together
>all of those inserted bits written anywhere with arrows running
>around the page to where they belonged, but I wouldn't bother
>writing out "boilerplate" code at all. It's as if I was using
>a mental preprocessor, with output feeding directly to the keypunch.
>It saved *immense* amounts of time.
Ah, those wonderful little arrows that snaked their way around
the listing when you realize that you can move THIS block of
code over ------> THERE and save a few bytes of memory and/or
reduce path lengths by a few instructions. I can argue that
the inability to do this with online editors is one of the reasons
that so much garbage gets passed off as professional "programming"
in today's market.
>Of course, it was often hard to get time on a keypunch, which led to
>my doing a lot of keypunching after hours. The alternative - sneaking
>into the keypunch area when the girls were off for lunch or coffee -
>was potentially dangerous if I interrupted one of their jobs and
>cleared their programs and auto-dup fields.
At my shop there were always 026 (later 029) punches in the programming
offices. Oddly enough from today's viewpoint where most professionals
expect relatively quiet surroundings, none of us had any problems with
the noise from those keypunches even though one of them was in an office
I shared with three other programmers. (Now, when somebody brought in
a Univac VIP it was quickly banished to the back of the office, far
away from anyone's desk. That [censored] thing was LOUD!)
Joe Morris
David Gesswein
> by never...@panics.com.invalid (Michael Roach):
>>
>> I've been wondering...where do you get new (or usable) dectapes? Are
>> the originals still usable? How old is the oldest original that's still
>> usable?
>
DECtapes for archiving written on PDP-8 or PDP-9/15. The oldest labels had
dates in the early 70's. I have had reasonable sucesss, 2/3 of the tapes
I could get an error free read and many others I could read if I used the
other transport or had a few bad blocks. Some of the problem looks like head
alignment since some of batches that look like they came from the same
place will read better on my left transport, other batches read better on
the right transport.
In article <9f63ss$iis$1...@flood.weeg.uiowa.edu>,
jo...@cs.uiowa.edu (Douglas W. Jones,201H MLH,3193350740,3193382879) writes:
> Real DECtape is made using a sandwich of 2 mylar layers with the oxide
> between. As a result, the oxide can't flake off, so the stuff has an
> extremely long shelf life. The fully redundant recording format also
> helps (each bit is stored in duplicate on two different non-adjacent
> tracks of the tape, so you can punch a quarter-inch hole anywhere in the
> tape with no data loss).
>
Do you know more about this? Visually it just looks like a oxide coating
on a Mylar backing, http://foldoc.doc.ic.ac.uk/foldoc/foldoc.cgi?LINCtape
seems to say it was a very thin coating so it might not be visible. I also
was seeing what would safely clean a tape to see what would help tapes
with bad blocks and found alcohol will remove the oxide coating so the
protective layer isn't very thick.
Some DECtape/tu56 information and pictures
http://www.pdp8.net/tu56/tu56.shtml
David Gesswein
http://www.pdp8.net/ -- Run an old computer with blinkenlights
Charlie Gibbs
(Joe Morris) writes:
>"Charlie Gibbs" <cgi...@sky.bus.com> writes:
>
>>Of course, it was often hard to get time on a keypunch, which led to
>>my doing a lot of keypunching after hours. The alternative - sneaking
>>into the keypunch area when the girls were off for lunch or coffee -
>>was potentially dangerous if I interrupted one of their jobs and
>>cleared their programs and auto-dup fields.
>
>At my shop there were always 026 (later 029) punches in the programming
>offices. Oddly enough from today's viewpoint where most professionals
>expect relatively quiet surroundings, none of us had any problems with
>the noise from those keypunches even though one of them was in an
>office I shared with three other programmers. (Now, when somebody
>brought in a Univac VIP it was quickly banished to the back of the
>office, far away from anyone's desk. That [censored] thing was LOUD!)
I worked in Univac shops so the majority of machines were VIPs.
However, I heard that they were prevalent even in many IBM shops -
they were a lot more productive than 029s. In fact, an IBM CE
(who was working on the one or two 029s that seemed to still be
present everywhere) once told me that the VIP caught IBM with
their pants down, and that as a result they rushed out the 129.
In his opinion (which I respected since he had to repair them)
the result was half-baked.
But you're right, a roomful of VIPs going full-bore was loud -
the roar of the basic 1701's card feed mechanism was augmented
by the machine-gun rattle of 1710's print unit. (Univac produced
a 1610, which was a lot quieter because they de-rated it to half
speed, but it never caught on because the 1710's speed was its
major selling point). The follow-on model, the 1800, had much
better soundproofing - but by then cards were on the way out,
and it was too late for them to sell in significant numbers.
Their 1900 CADE (Computer-Assisted Data Entry) system and
other card-to-tape systems had their brief day in the sun,
then everything went online and dedicated data-entry hardware
died a natural death.
Steve O'Hara-Smith
jcmo...@jmorris-pc.MITRE.ORG (Joe Morris) wrote:
JM> "Charlie Gibbs" <cgi...@sky.bus.com> writes:
JM> Ah, those wonderful little arrows that snaked their way around
JM> the listing when you realize that you can move THIS block of
JM> code over ------> THERE and save a few bytes of memory and/or
JM> reduce path lengths by a few instructions. I can argue that
JM> the inability to do this with online editors is one of the reasons
But that's why vi has 26 named buffers for you to play with. All you
have to do is keep track of them in your head :)
--
Directable Mirrors - A Better Way To Focus The Sun
Roger Johnstone
Horvath, CCP) wrote:
> There is something about core that chip-memory just doesn't have -- the
ability
> to see the bits (not their status though).
IC memory is visible, just a bit smaller. All you need is an EPROM with a
window...and a microscope.
--
Roger Johnstone, Invercargill, New Zealand
Apple II - Future Cop:LAPD - Warcraft II
http://homepage.mac.com/rojaws
______________________________________________________________________
"Never underestimate the bandwidth of a station wagon full of tapes hurtling
down the highway."
Andrew S. Tanenbaum, Computer_Networks, Second Edition, p. 57
Douglas W. Jones,201H MLH,3193350740,3193382879
Horvath, CCP) wrote:
> There is something about core that chip-memory just doesn't have -- the
> ability to see the bits (not their status though).
Late model core memory (post 1970) uses such tightly packed cores of such
a small size that you really can't see the bits without a magnifying
glass. To the unaided eye, each patch of core memory is just a satin
black area, surrounded by a border of tightly packed hair-thin copper
wires. A number of these patches are usually packed together in a single
physical core plane, and they're usually laminated to a plastic substrate
to prevent mechanical damage.
Doug Jones
jo...@cs.uiowa.edu
CBFalconer
Does anybody remember 'rope' memory. The core equivalent of a
masked PROM.
--
Chuck F (cbfal...@my-deja.com) (cbfal...@XXXXworldnet.att.net)
http://www.qwikpages.com/backstreets/cbfalconer :=(down for now)
(Remove "NOSPAM." from reply address. my-deja works unmodified)
mailto:u...@ftc.gov (for spambots to harvest)
Al Kossow
> Does anybody remember 'rope' memory. The core equivalent of a
> masked PROM.
Depending on how it was implemented, it could be changed in the field,
so it was more like a fuse link PROM than a mask-programmed ROM.
The original PDP-14 industrial controller used this, and you
encoded your ladder logic by rewiring the cores. The later designs
of the 14 used PDP-8 core and was downloaded through a serial link
from a purpose built controller make from an 8E.
The microstore in a PDP-9 is based on this technology too.
Douglas W. Jones,201H MLH,3193350740,3193382879
by a...@spies.com (Al Kossow):
>> Does anybody remember 'rope' memory. The core equivalent of a
>> masked PROM.
> The original PDP-14 industrial controller used this, ...
> The microstore in a PDP-9 is based on this technology too.
The mid-level microstore for the HP9100 computing calculator used this.
The low-level microcode was stored in a diode matrix ROM. The low-level
microcode was used to implement the interpreter for the mid-level
microinstruction set, and this, in turn, interpreted keypresses and user
programs for the calculator. This was the first RPN programmable
calculator, new in 1968. Fully transistorized, except for the CRT!
Doug Jones
jo...@cs.uiowa.edu
Jim Saum
>Does anybody remember 'rope' memory. The core equivalent of a
>masked PROM.
The Apollo Guidance Computer used on both the Command and Lunar
Modules during flights to the moon had a rope ROM. The flight software
development group at MIT had for each mission a person called a "rope
mother", i.e., what we today would call a build or release
coordinator.
- Jim Saum
gla...@glass2.lexington.ibm.com
Well, that was true in the early days of integrated circuit storage
chips, but the dimensions have gotten so small now that I question
whether the wavelength of visible light is small enough to resolve
the individual bits. Chip makers have been using X-ray lithography,
electron-beam lithography, and other exotic processes to shrink the
dimensions of the chips because the wavelength of ordinary light is
too long to resolve the images. If I'm remembering correctly, the
wavelength of green light is about 5350 Angstroms, which is about .5
micrometers, and some chip lands have been below that.
I remember some of the middle generation EPROMs acting as a
diffraction grating when viewed under visible light. The early
EPROMs didn't exhibit this effect (well, at least, not as strongly).
I also remember pealing the gold foil off of SRAM chips over 15
years ago and looking at the chips (Some had the void (partially?)
filled with thermal conduction compound, but others were just
void.)[1]. Plus, I have fuzzy recollections that some epoxy chips,
if hit precisely enough with a hammer, would fracture the epoxy and
result in a viewable (but non-functional) chip.
[1] This is not an endorsement to do this; I'm merely relating what
has been done in the past. Some of the thermal conduction compounds
may be extremely toxic or have other adverse environmental or
biological effects. This should only be attempted by trained
professionals in the appropriate labratory environment.
I've thought about trying the foil pealing with some of the more
modern chips, such as 80386, 80486, or even Pentium chips [2]. The
geek in me thinks that this would make a wonderful jewel for a tie
pin (especially if it was covered with a clear epoxy for protective
purposes). Of course, I haven't worn a tie for over 10 years, so
it's real low on the priority list. :-/
[2] See the warning for footnote [1]. Don't do this at home. If
you do do this at home and kill yourself, don't come running to me.
Dave
P.S. Standard Disclaimer: I work for them, but I don't speak for them.
P.P.S. "Never underestimate the bandwidth of a 747 with a cargo-hold full
of tapes." - Anonymous
Prof. Richard E. Hawkins
>[2] See the warning for footnote [1]. Don't do this at home. If
>you do do this at home and kill yourself, don't come running to me.
You may, however, haunt him . . .
:)
hawk
--
Prof. Richard E. Hawkins, Esq. /"\ ASCII ribbon campaign
doc...@psu.edu Smeal 178 (814) 375-4700 \ / against HTML mail
These opinions will not be those of X and postings
Penn State until it pays my retainer. / \
CBFalconer
Sounds something like the organization of my desktop machine of
1964. I described it in terms of instructions executing
milli-instruction executing micro-instructions executing
nano-instructions executing pico-instructions.
The pico level was a 1 bit serial adder. The nano level
implemented a one digit excess 3 decimal adder. The micro level
implemented 10 digit 9's complement floating point arithmetic.
The milli level implemented algebraic statements. The instruction
level was described by a key press. More or less.
I ran into the rope concept there. Didn't use it.
Jim Stewart
There was a 256 word Omnibus option for the PDP 8 that used this
technology as well.
Eric Smith
> The original PDP-14 industrial controller used this, and you
> The microstore in a PDP-9 is based on this technology too.
And there was a core rope memory RTM module for the PDP-16.
The Apollo Guidance Computer used core rope memory.
The HP 9100A and B calculators used a core rope memory for their
control sequencer, 64 words of 29 bits IIRC. The actual microcode
was 512x64 of inductively coupled printed-circuit board memory, which
achieved the then-amazing density of over 1000 bits per square inch.
AFAIK, core rope memory was invented by IBM for the control store of
some of the first System/360 models. IBM called it Transformer
Read-Only Store (TROS).
Other early 360 models used capacitive schemes. The 360/30 control
store was exceptionally clever; it used mylar 80-column punched cards,
with row metalization on one side and columns on the other. IIRC they
used 64 columns and stored 12 microwords per card, one per row. The
control store was programmed by running the virgin cards through an
ordinary card punch. Of course, the punch needed to support column
(or row) binary mode.
Ben Franchuk
> The mid-level microstore for the HP9100 computing calculator used this.
> The low-level microcode was stored in a diode matrix ROM. The low-level
> microcode was used to implement the interpreter for the mid-level
> microinstruction set, and this, in turn, interpreted keypresses and user
> programs for the calculator. This was the first RPN programmable
> calculator, new in 1968. Fully transistorized, except for the CRT!
>
> Doug Jones
> jo...@cs.uiowa.edu
Ain't modern technology wonderful. 1930 1945 1959 1968 1973 ...
It all was state of the ART when it came out.
Ben.
Tony Duell
: From article <aek-060601...@il0502a-dhcp38.apple.com>,
: by a...@spies.com (Al Kossow):
: > In article <3B1E4FA2...@my-deja.com>, cbfal...@worldnet.att.net wrote:
: >> Does anybody remember 'rope' memory. The core equivalent of a
: >> masked PROM.
:
: > The original PDP-14 industrial controller used this, ...
: > The microstore in a PDP-9 is based on this technology too.
: The mid-level microstore for the HP9100 computing calculator used this.
: The low-level microcode was stored in a diode matrix ROM. The low-level
I have to disagee with that.....
In the 9100, the microcode is a core-on-a-rope memory. Physically it's on
the leftmost board of the set of plug-in boards in the machine. It is the
lowest level of programming in the machine, and is something like 64
locations of 29 bits each.
The main program store (next level up) is a PCB ROM. Not a diode matrix.
There is a _seriously_ multi-layer board in the middle of the bottom of
the machine (16 layes or so). It forms a ROM by inductive coupling
between the tracks on the various layers. There are 512 address tracks
(at least in the 9100B, which is what I have) driven by transistors on the
end of the board and 64 sense loops which couple (or not) to the address
tracks. These are then connected to sense amplifiers alongside the ROM board.
Tbe big board above the ROM that's covered in diodes is not really a ROM
in the conventional sense. I call it the 'gating board'. It's AND and OR
gates (only, no inverting functions, and IIRC, no connections to the
power line) that go between the flip-flops in the machine. They provide
functions like incrmenting and decrementing of the registers, shifts,
memory address decoding, etc.
The other boards in the machine are :
2 boards each with 20 JK flip-flops (the boards with red handles). The
lack of inverters in the logic is not a major problem since each
flip-flop provides a true and an inverted output.
2 boards for the user read/write core memory. The exact split of the
electronics between these 2 boards depends on the version. The -B has one
board with the core address drivers on it, and the other board contains
the core plane and the data sense/inhibit circuity. The address drivers
have to be matched to the core plane, so these boards are tied together
by double-width handles on the -B. In the -A, I beleive that 1 board
contains the address drivers and core plane, the other the data
circuitry. These boards don't have to be matched, and so they have
separate handles.
There are 2 boards down the sides of the machine, which connect the ROM
assembly to the gating board. The one on the left contains the microcode
branch logic and the error flip-flop. The one on the right is part of the
instruction decodeer for the main ROM.
The main ROM assembly is 4 boards soldered together. The one in the
middle (as I have said) is the ROM itself. The ones down the sides are
the sense amplifiers. The one across the back is the address decoder.
: microcode was used to implement the interpreter for the mid-level
: microinstruction set, and this, in turn, interpreted keypresses and user
: programs for the calculator. This was the first RPN programmable
: calculator, new in 1968. Fully transistorized, except for the CRT!
Contrary to popular belief there are 8 ICs in a 9100B. They are on the
card reader PCB, and are simple op-amps. 2 per read channel (amplifier
and comparator IIRC).
The card format is quite simple. There are 4 tracks in each direction on
the card. 3 of them are data tracks, the other one is a clock track. On
one edge of a pulse on the clock track, the odd-numbered bits are
read/written to the data tracks. On the other transition, the
even-numbered bits are read/written. The main memory is 6 bits wide, user
program codes are 6 bits (there are 64 possible program instructions) and
so on.
-tony
David Powell
in alt.folklore.computers wrote:
>> From article <9f5v3j$2u9$1...@news.panix.com>,
>> by never...@panics.com.invalid (Michael Roach):
>>>
>>> I've been wondering...where do you get new (or usable) dectapes? Are
>>> the originals still usable? How old is the oldest original that's still
>>> usable?
>>
>Its hard to say when they were last written but I have read at least 50
>DECtapes for archiving written on PDP-8 or PDP-9/15. The oldest labels had
>dates in the early 70's. I have had reasonable sucesss, 2/3 of the tapes
>I could get an error free read and many others I could read if I used the
>other transport or had a few bad blocks. Some of the problem looks like head
>alignment since some of batches that look like they came from the same
>place will read better on my left transport, other batches read better on
>the right transport.
>
Yes, same problem on my TU56 . It's been years since I set up the
timing and skew on a DECtape drive, and I don't intend to fix it
unless the drive fails. If I did, then it would simply alter the
readable sub-set of my tapes. Annoyingly, I have some 30+ year old
DEC source files that I cannot read. They are on "certified" DECtapes,
and probably have only been read once, to produce working copies.
Now, if I'd have archived those instead, there's a fighting chance I
could read them.
When DECtapes were "state of the art" we had two TU56s, and the people
down the corridor had one, and a line printer. We used to walk a tape
down the corridor to get our printing done. It worked well, until the
units drifted, and a tape was only readable on the drive that
formatted it. So we called in DEC FS to set all of them right. Big
mistake. Some of our tapes were unreadable on all drives. The next
time we had just one drive done, copied everything to tapes formatted
on that drive, then did the other units.
Wise after the event, I think the trick for DECtapes was to buy only
"certified" tapes, never reformat them, and get the drive fixed if it
drifted. Then you should be able to read any tape on any drive. It's
too late now.
Regards,
David P.
jchausler
Charlie Gibbs wrote:
> Of course, it was often hard to get time on a keypunch, which led to
> my doing a lot of keypunching after hours. The alternative - sneaking
> into the keypunch area when the girls were off for lunch or coffee -
> was potentially dangerous if I interrupted one of their jobs and
> cleared their programs and auto-dup fields.
Oh yes! We had only one 029 in our department and so when it was in
use would go over to data processing and beg to use one of their's. It
seemed to me that if they planned to use it anytime that day, it was
unavailable and if I casually wandered in at lunch time and started
using one of their punches, lost track of the time and didn't leave before
lunch ended, I would catch hell ;-)
David Gesswein
David Powell <ddotp...@netnospamscapeonline.co.uk> writes:
> On Tue, 05 Jun 2001 12:43:54 GMT, d...@munin.sabir.com (David Gesswein)
> in alt.folklore.computers wrote:
>
> Yes, same problem on my TU56 . It's been years since I set up the
> timing and skew on a DECtape drive, and I don't intend to fix it
> unless the drive fails. If I did, then it would simply alter the
> readable sub-set of my tapes. Annoyingly, I have some 30+ year old
> DEC source files that I cannot read. They are on "certified" DECtapes,
> and probably have only been read once, to produce working copies.
> Now, if I'd have archived those instead, there's a fighting chance I
> could read them.
>
I have tweaked the head position a couple times for a tape I really wanted
to read. Some I have been able to read that way and others not. That was
the transport that was flakier when I got it, the other I am leaving alone
as the reference. The irritating thing is that the checksum method used
isn't robust enough to detect double bit errors on same track, "good" read
can still have bad blocks. I normally read each tape twice when archiving,
especially if I see retries.
I can give a shot on reading them on my drive if you wish.
David Gesswein
http://www.pdp8.net/ -- Run an old computer with blinkenlights.
Have any PDP-8 stuff you're willing to part with?
Dennis Ritchie
David Powell wrote:
...
> Yes, same problem on my TU56 . It's been years since I set up the
> timing and skew on a DECtape drive, and I don't intend to fix it
> unless the drive fails. If I did, then it would simply alter the
> readable sub-set of my tapes.
> ... It worked well, until the
> units drifted, and a tape was only readable on the drive that
> formatted it. So we called in DEC FS to set all of them right. Big
> mistake. Some of our tapes were unreadable on all drives. The next
> time we had just one drive done, copied everything to tapes formatted
> on that drive, then did the other units.
>
> Wise after the event, I think the trick for DECtapes was to buy only
> "certified" tapes, never reformat them, and get the drive fixed if it
> drifted. Then you should be able to read any tape on any drive. It's
> too late now.
In the early 70s, when we used DECtapes heavily for backup, we had
the same situation several times. All the drives would drift in
alignment, so we couldn't read older things. We learned that getting
the FE to align all drives was a bad idea, because then we could not
read newer things. So we learned to leave at least one drive untouched,
so as to be able to copy tapes using the untouched drive to
a newly aligned one.
But it also began to become clear that the alignment tapes used by the FEs
varied nearly as much as the tapes written by our own drives. There
was no golden standard.
Dennis
Don Quixote
[Tapes, alignment, and such]
Dennis Ritchie might have said:
>But it also began to become clear that the alignment tapes used by the FEs
>varied nearly as much as the tapes written by our own drives. There
>was no golden standard.
As a once and perhaps future FE let me reassure you that what was issued to
many was old crap, and some FE's considered them stock for levelling the couch
first, and reference second.
They were right, but for the wrong reasons.
My knowledge of this begins with disk packs, but I'm sure someone will detail
the poor quality of rocks they were issued, and I will read that story with
glee.
Eric Smith
> In the early 70s, when we used DECtapes heavily for backup, we had
> the same situation several times. All the drives would drift in
> alignment, so we couldn't read older things. We learned that getting
> the FE to align all drives was a bad idea, because then we could not
> read newer things. So we learned to leave at least one drive untouched,
> so as to be able to copy tapes using the untouched drive to
> a newly aligned one.
What's especially aggravating is the alignment instructions in the
TU56 Maintenance Manual, which basically says "send it to the factory."
It would have been nice if they had documented the actual procedure, and
the materials and tools necessary to do it, even if they said "don't do
this in the field."
Joe Morris
>In the early 70s, when we used DECtapes heavily for backup, we had
>the same situation several times. All the drives would drift in
>alignment, so we couldn't read older things. We learned that getting
>the FE to align all drives was a bad idea, because then we could not
>read newer things. So we learned to leave at least one drive untouched,
>so as to be able to copy tapes using the untouched drive to
>a newly aligned one.
>But it also began to become clear that the alignment tapes used by the FEs
>varied nearly as much as the tapes written by our own drives. There
>was no golden standard.
Cue the ancient saying, "There's nothing new under the sun."
The problem of tape alignment isn't unique to the computer world; it's
also well-known (far too well-known) in audio systems as well. Many
years ago I worked as a TV broadcast engineer; one of the routine
maintenance tasks (no, I can't recall the frequency) was to check the
alignment of the heads on all of our audio tape decks. The master
alignment tape was carefully segregated from all other tapes to prevent
any accidental damage to it; that tape was used to aligh every deck.
I recall at least once we got in a tape that had been recorded on a
*very* badly misaligned deck; we wound up deliberately misaligning
one of the decks (tuning by ear, aka "earballing") to make it sound
right, then copied it to a good tape and realigned the hacked deck
back into spec.
The station was barely surviving at the time, but it did have a certified
alignment tape from Ampex. Other stations weren't so picky and were
known to use normally-recorded tapes to "earball" decks into something
vaguely resembling standard alignment.
Joe Morris
Alexandre Pechtchanski
>Dennis Ritchie <d...@bell-labs.com> writes:
>
>>In the early 70s, when we used DECtapes heavily for backup, we had
>>the same situation several times. All the drives would drift in
>>alignment, so we couldn't read older things. We learned that getting
>>the FE to align all drives was a bad idea, because then we could not
>>read newer things. So we learned to leave at least one drive untouched,
>>so as to be able to copy tapes using the untouched drive to
>>a newly aligned one.
>
>>But it also began to become clear that the alignment tapes used by the FEs
>>varied nearly as much as the tapes written by our own drives. There
>>was no golden standard.
>
>Cue the ancient saying, "There's nothing new under the sun."
>
>The problem of tape alignment isn't unique to the computer world; it's
>also well-known (far too well-known) in audio systems as well. [...]
I remember that all tape recorders I ever owned or saw (all home user) had
alignment screws on the heads. Of course, for home recorders there wasn't such
luxury as alignment tapes - we did it by ear ;-)
--
[ When replying, remove *'s from address ]
Alexandre Pechtchanski, Systems Manager, RUH, NY
Steve O'Hara-Smith
Alexandre Pechtchanski <alex*@*rockvax.rockefeller.edu> wrote:
AP> I remember that all tape recorders I ever owned or saw (all home user) had
AP> alignment screws on the heads. Of course, for home recorders there wasn't such
AP> luxury as alignment tapes - we did it by ear ;-)
You were supposed to make one when the deck was new :)
Steve O'Hara-Smith
gla...@glass2.lexington.ibm.com wrote:
> void.)[1]. Plus, I have fuzzy recollections that some epoxy chips,
> if hit precisely enough with a hammer, would fracture the epoxy and
> result in a viewable (but non-functional) chip.
Ceramic pack chips could be split along the seam to reveal a fully
functioning chip (with about a 70% success rate after practice). I left
one running in a TTL clock (my chip abuse testbed for some time) for a
couple of months before it died.
Tony Duell
: What's especially aggravating is the alignment instructions in the
: TU56 Maintenance Manual, which basically says "send it to the factory."
: It would have been nice if they had documented the actual procedure, and
: the materials and tools necessary to do it, even if they said "don't do
: this in the field."
Yes, 'service' manuals like that drive me mad as they're totally useless.
Similalry 'service' manuals that are little more than board-swapper
guides with no information on how to repair the boards are useless too.
Maybe the average FE couldn't (economically) repair individual PCBs, or
do TU56 alignments, but the situation has rather changed since those days.
A serious hardware-biased hobbyist is going to have access to tools and
equipment _way_ beyond the stuff that FEs worked with. Said hobbyist also
probably has more knowledge and experience of electronic repairs. And he
certainly has a semi-infinite amount of time to spend fixing the thing
(in my case, I _enjoy_ fiddling with old computer hardware. I enjoy
staring at the scope or logic analyser, OK...). And of course, there's
now often no factory that you can send the thing back to to have it
realigned.
In serveral cases I've had to work out repair procedure for machines
where the manufacturers manuals are totally inadequate.... And in general
I've managed to do it, at least well enough to get the machines working
again. Alignment, though, is more of a problem -- you can align something
well enough to work, but you might well have interchangeability problems
with other drives. You can read your own tapes/disks but you can't use
them on other drives.
I've got a few service manuals that while they start out telling you how
to swap boards, they then continue with 'The faulty PCB should generally
be returned to the <foo> factory for repair. Should this be impossible,
the following information will assist in finding the fault'. It then
contains full schematics, waveform charts, and so on.
-tony
Brian Inglis
wrote:
>In <9f63lv$q...@netaxs.com>, lwi...@bbs.cpcn.com (lwin) writes:
>>x-no-archive: yes
>>
>>> Of course. Hey, people still keep steam engines running. Why do they do
>>> it? Because they like it that way 8^)
>>
>>But aesthectically, there's a big difference between a running
>>steam engine and an electronic computer.
>>
>>A running computer just sits there. In contrast, a running steam
>>engine, to its fans, has lots of neat sounds, sensations, and visuals.
>>
>>(And I think there's a lot more railroad buffs out there willing to
>>do the incredible amount necessary to rehab and run a steam engine
>>than there are computer buffs willing to do likewise to old hardware.)
>>
>>I must admit I'd love to see a working S/360, just for nostalgia's
>>sake. But I also must admit I'm NOT sure I'd be so quick as to run
>>down and write a program, keypunch up the cards, wait in line at
>>the card reader, wait for the job to be run, and wait for the printout
>>only to discover I left out a single comma and have to do it all
>>over again.
>
>Well, I don't know about an S/360, but some of us are still running S/370
>machines:
>
> CP Q CPLEVEL
> VM/ESA Release 1, Service Level 111 - 370 Feature
> Generated at 10/13/93 09:11:15 LCL
> IPL at 05/30/01 16:50:10 EST
>
>Ok, so it's out of date, and completely unsupported, it still runs.
Did you mean to do a CP Q CPUID or whatever shows the model?
Thanks. Take care, Brian Inglis Calgary, Alberta, Canada
--
Brian....@CSi.com (Brian dot Inglis at SystematicSw dot ab dot ca)
fake address use address above to reply
tos...@aol.com ab...@aol.com ab...@yahoo.com ab...@hotmail.com ab...@msn.com ab...@sprint.com ab...@earthlink.com ab...@cadvision.com ab...@ibsystems.com u...@ftc.gov
spam traps
Maury Markowitz
> In the early 70s, when we used DECtapes heavily for backup, we had
> the same situation several times. All the drives would drift in
> alignment, so we couldn't read older things. We learned that getting
> the FE to align all drives was a bad idea, because then we could not
> read newer things. So we learned to leave at least one drive untouched,
> so as to be able to copy tapes using the untouched drive to
> a newly aligned one.
Oddly enough, I had the exact same problem on my Atari 400. I couldn't
afford a floppy-disk, so for $100 I got a tape - geez. Mine seemed to be out
of alignment compared to the rest of the world, but I found that by pulling
off the lable on the front I could access the head-alignment screw. Four to
the left and I could access other people's tapes, four to the right and mine
worked. I was too lazy to ever copy my tapes over to the new "format",
because it would have had to be done one program at a time and I typically
had 15 to 20.
I later learned that using two cassette (audio) machines worked just as
well, but by that time I had moved on.
Maury
gla...@glass2.lexington.ibm.com
Well, no, and I was hoping that no one was going to ask, since by doing that,
I give away the secret[1]:
CP Q CPUID
CPUID = FF64072074700000
[1] Ok, for those not in the know, the 7470 CPU model corresponds to a P/370
card, which is a S/370 on a card that plugs into a PS/2 and gives you a
full and complete S/370 (with a very minor exception or two, such as 4K storage
keys). Of course, it's been obsolete for quite some time, having been
replaced by the P/390 cards (full ESA/390 on a card), although I've heard a
rumor that those may have been withdrawn from marketing.