Looking For NCR 315 Mainframe Pics
RFCOMMSYS
asking NCR was a dead end. This computer series has great nostalgic value to
me. Any links would be most appreciated.
Bill Turlock
Worked on a 315 RMC at U of Nebr at Omaha in the early 70's
Sorry, no photos, but "fond" recollections of the CRAM
(Card Random Access Memory), the most devilish device *ever*
devised for computer I/O. If you're not familiar with it,
I'll reminisce...
LMK
Bill (Double Drop!!!) Turlock
Chuck Sterling
Only one I ever saw was at the Miamisburg tech school. Noisy when it
operated, like a Boeing 707 noisy. Never heard the result of a double
card drop but was told the whole damn building knew when it happened.
Chuck Sterling
Brian Boutel
> "Bill Turlock" <"Bill Turlock"@sonnic.net> wrote in message
> news:3FF4EF5E...@sonnic.net...
> Really, really loud.
>
I worked for NCR in the early 60s. We got a 315 around December 62. Not
the RMC version, which came later. Yes, it was LOUD - so much so that I
suspect it would not be acceptible to OSH types now.
--brian
--
Brian Boutel
Wellington New Zealand
Note the NOSPAM
L0nD0t.$t0we11
Did the big NCR units use [electric motor driven] hydraulics for
moving the cards around and driving the belts that took them to
and from the read head cylinder? I can't remember if the
RCA Spectra version called the Mass Storage Unit or "moose" was
based on NCR or CDC mechanisms. Dimly recall it had the old
epoxy black dot transistors, which usually meant CDC ancestry.
Most customers kept them in a separate room, soundproofed.
--
Fan of the dumbest team in America.
jchausler
Bill Turlock wrote:
The CRAM was the center of NCR's pavilion at the 1964
New York Worlds Fair. That's the only place I've ever seen
one. Now if you want to talk about the RCA RACE unit, also
a card memory system and rumor has it designed by the same
guy who designed the CRAM, I can go for hours. Still have
a RACE card in my stash of old computer stuff. Yes, tell me
more about the CRAM......
Chris
AN GETTO$;DUMP;RUN,ALGOL,TAPE
$$
RFCOMMSYS
>
>Did the big NCR units use [electric motor driven] hydraulics for
> moving the cards around and driving the belts that took them to
> and from the read head cylinder? I can't remember if the
> RCA Spectra version called the Mass Storage Unit or "moose" was
> based on NCR or CDC mechanisms. Dimly recall it had the old
> epoxy black dot transistors, which usually meant CDC ancestry.
>
> Most customers kept them in a separate room, soundproofed.
>
I may be wrong, but I thought the NCR CRAM card was sucked down the chute,
wrapped around the read/write cylinder, and sent back up to the card rods, by
vacuum action.
I never saw a 315 CRAM unit in action, but I did see some NCR Century Series
CRAM units in action, and they were quite loud. I believe it was the evacuating
air (which caused the vacuum) that was so noisy.
Jack Russell
I think the NCR sales guys were bemused by the fact that anyone outside NCR would use the wretched things. I suspect my current partial deafness is due to standing in a small room with 10 crams going full bore!
Jack Russell
Guy wrote:
"Bill Turlock" <"Bill Turlock"@sonnic.net> wrote in message news:3FF4EF5E...@sonnic.net...
RFCOMMSYS wrote:
I remember seeing one in action at a computer show in the '60s. It was loud. Really, really loud.Bill (Double Drop!!!) Turlock
Bill Turlock
Curiously, I've googled for it before without getting all
this intelligence:
http://en.wikipedia.org/wiki/CRAM
http://content.techweb.com/encyclopedia/defineterm?term=CRAM&exact=1
http://www.aboutlegacycoding.com/default.htm?AURL=%2FArchives%2Fv3%2Fv30606%2Easp
http://www.presshere.com/html/pw8012.htm
Here it is! The very 315 I'm talking about! [Apparently my
recollection, below, of the apearance of the CRAM, is
faulty]:
http://www.unoalumni.org/About_Us/Flashback/Archive/92/index.asp
http://www.aboutlegacycoding.com/default.htm?AURL=%2FArchives%2FV5%2FV50205%2Easp
[the bit about using oil for the carrier for the iron dust
doesn't match my recollection--for the 729's, IBM supplied
some kind of nasty volatile solvent to clean the heads with,
and the label proclaimed it was also "tape developer
medium"]
Anyway--
I started at UNO in the spring of 1970, hired on the basis
of my experience as a crack operator on the U1108 RTOS at
GWC. The UNO computing center had just signed a contract
w/Univac for an 1106 (de-clocked 1108). About that time IBM
started throwing their political weight around, got the
contract annulled and made them get a M40, (I think). They
had both the 315 and the M40 operating while I was there.
Anyway, the pics in the first URL are most likely from one
of the early models of the the CRAM. Ours wasn't so
space-agey looking, and _not_ pastel! It had a glass front
door all the way to the floor, and was a bit taller.
IIRC, the card, after being selected, fell by gravity down a
channel until it met the spinning drum, where a vacuum held
it to the drum for reading and writing. To release, I think
fingers picked it off the drum and inertia shot it back up
the return channel (or mabe it was the same channel, I'm not
sure) and a solenoid-operated plate smacked it back onto the
pack, suspended by the eight " lazy-'D' " rods which were
the selection mechanism.
As mentioned earlier in the thread, if some of the address
notches in the cards became flexed due to wear, it was
possible for more than one card to drop simultaneously. The
channel/drum gap wasn't quite wide enough for two cards to
pass, but since they both couldn't get to the gate precisely
at the same instant, one was offset from the other,
resulting in a wedge-jam just ahead of the drum, with one
card usually halfway around the drum, being chewed up as a
result.
It made a characteristic whine, audible and recognizable
throughout the machine room. Many's the time I'd be sitting
in the ops mangager's office shooting the breeze when we'd
hear the siren-like sound of the jam. Someone would shout
"double-drop!!" and we'd all rush out to the CRAM.
The first guy there would fling the door open, drop to his
knees, and without any hesitation (if you flinched you'd get
your hand chewed up) slam the heel of his hand against the
outside edge of the spinning drum to bring it to an abrupt
stop.
Surprisingly, we were able to save the majority of the cards
involved in a D.D. if we were fast enough. They were quite
rugged, despite being not too much thicker than mag tape,
maybe a bit thinner than a punch card. Most of the time,
IIRC, we'd have to open the back door of the cabinet to put
the drive belt back on the pulley.
Somewhere I've got an address template for the cards showing
which parts of the notches to trim to hard code the card
number. I scanned it once, if I can find it I'll post a URL
to it.
Fun remembrance!
Bill Turlock
Heinz W. Wiggeshoff
>
...
> knees, and without any hesitation (if you flinched you'd get
> your hand chewed up) slam the heel of his hand against the
> outside edge of the spinning drum to bring it to an abrupt
> stop.
This strikes me as superbly shitty engineering! Or perhaps
the unit predates any formal workplace safety standards for
First World countries. Every clothes washer and drier I've
ever used has the drum come to a quick halt when the lid is
opened. It shouldn't have been much of a stretch for the
design team to add such a facility before the first unit was
shipped. But then again, much like the PHBs contemplating
the Ford Pinto's gas tank design, NCR may have figured that
it's cheaper to pay off injury and damage claims. Besides,
machine room operators who lose a hand can probably be re-
trained.
Peter Flass
> I may be wrong, but I thought the NCR CRAM card was sucked down the chute,
> wrapped around the read/write cylinder, and sent back up to the card rods, by
> vacuum action.
>
> I never saw a 315 CRAM unit in action, but I did see some NCR Century Series
> CRAM units in action, and they were quite loud. I believe it was the evacuating
> air (which caused the vacuum) that was so noisy.
I never saw one in action either, but ISTR a Datamation article on CRAM.
If I haven't dropped too many memory bits, I believe the cards were
edge-notched and picked from the stack with rods. There was some
discussion here recently on CRAM, that I can't recall just now and
Google can't seem to find.
Peter Flass
>
> http://www.presshere.com/html/pw8012.htm
>
Good description of this most folklore-ish piece of equipment.
L0nD0t.$t0we11
From memory circa 1969... The RCA moose had several cards located
in bins that ran along the fairly lengthy sides of the machine.
An individual card was picked by a set of mechanical rods, in a
process that would embarass Rube Goldberg, and scootched slightly
into a set of very rapidly moving belts that slammed it the length
of the unit in mumble mumble milliseconds [100 if memory serves, but
memory isn't doing that well this morning] where it got wrapped
around the reading cylinder, held there by vacuum until done
reading, then slammed back into those belts to be stuffed into
a bin lengthwise. The vacuum made some noise, but nowhere near
as bad as the vacuum in those air capstan tape decks. The hydraulic
pump was almost as noisy as the one on the AN/FSS-7 radar. The
motor driving the pump was the only thing in the room with more
horsepower than the ones in the card punches.
L0nD0t.$t0we11
> Curiously, I've googled for it before without getting all
> this intelligence:
>
> http://en.wikipedia.org/wiki/CRAM
>
>
> Here it is! The very 315 I'm talking about! [Apparently my
> recollection, below, of the apearance of the CRAM, is
> faulty]:
>
> http://www.unoalumni.org/About_Us/Flashback/Archive/92/index.asp
That thing is dinky. The big RCA moose was more on the order
of 20-30 feet long. If can find an online image will post,
which will probably be easier if I go dig out the RCA reference
manuals that had the model number of the beast.
Bill Turlock
Yabbut--
That was 30+ years ago. The emphasis then was on getting the
job done, not about whining about it. We did it by choice,
there's a certain macho attitude about it--we felt like
heros! (Besides, nobody ever got seriously hurt, the worst
was a minor case of 'rug burn'.)
There's another thread here about 1401 memory size, where I
talk about the 4k one at GWC, then the Giant 8k one at CICV.
I _actually_ got a Medal for 'doing my job' there. That's
how rare that attitude is.
Bill
Heinz W. Wiggeshoff
>
> I never saw one in action either, but ISTR a Datamation article on CRAM.
> If I haven't dropped too many memory bits, I believe the cards were
> edge-notched and picked from the stack with rods.
That must be the cards pictured in DEK's TAoCP, vol 3. B-)
Jack Russell
Of ten they did not fall which is why an operator would then riffle the size of the pack. The card then fell down and chute and was picked up on a drum. I think it was held there by air pressure. When you had finished with the card the air pressure was removed and the card flew up the exit chute and back into the pack. I cannot remember but think it was assisted in ts journey by air pressure. It did not matter where it went in the pack (the end I suspect).
The worst thing was when several cards fell at once (I guess the hooks wore) and they all arrived at the drum together which then tore them up out of spite!
Jack Russell
L0nD0t.$t0we11
> Each cram card was encoded by the "hooks" at the top, my memory says 128
> combinations but I may be wrong. so card zero only had one hook
> (probably two actually) and card 127 had all but two, You could select
> any card by turning the appropriate rods so the one card was free to fall.
> Of ten they did not fall which is why an operator would then riffle the
> size of the pack. The card then fell down and chute and was picked up on
> a drum. I think it was held there by air pressure. When you had finished
> with the card the air pressure was removed and the card flew up the exit
> chute and back into the pack. I cannot remember but think it was
> assisted in ts journey by air pressure. It did not matter where it went
> in the pack (the end I suspect).
> The worst thing was when several cards fell at once (I guess the hooks
> wore) and they all arrived at the drum together which then tore them up
> out of spite!
>
> Jack Russell
That gravity feed appears to be a big difference between the
RCA Moose and the CRAM. The picking of the card by the
notches and such was probably pretty close... but on the
RCA the card had to travel a goodly distance horizontally
to get to the read area from the bins. And was slammed
along with a set of hydraulically driven belts. Plus on
the RCA, the bins were stuck along one side [optionally
both sides] of the machine and the read area was on the
right hand end.
Don't remember any double feed issues, but when the card
was being slammed back toward a bin, those belts could
easily generate enough force to simply rip all the addressing
hooks off in milliseconds.
Someone else here had mentioned the RCA unit, haven't seen
them come back in.... only one I ever saw was in Lincoln
Nebraska at the old Bankers Life.
Jack Russell
Jack Russell (and his minkeys).
.
jchausler
"L0nD0t.$t0we11" wrote:
I never heard the RCA RACE called the "moose", but from
what you're describing, we're talking about the same unit.
At CIT in the 60's there was a pair of RACE drives using
an RCA 501 as their controller. The 501 was interfaced
to the pair of Bendix G-20's (re branded CDC) which
served as the primary computer facility at the school through
most of the 60's. As a part time computer operator on
weekends, I had to open the drives and clean the read
drum and heads a couple times a shift. The cards ( I still
have one) were notched and there were 128 possibilities.
If the drive got a card which did not match its selection,
it stuck it out a slot. The operator could then examine
the card for damage. The typical problem was that one
of the fingers between the notches had been bent or
broken. Sometimes it could be straightened out. There
was another slot into which you could then stick the
card and it would check it again. From my experience
with bent fingers, this worked successfully most of the
time. Our drives only had the bins along one side. I
was unaware that one could get them along both sides
and my memory tells me that the drive cabinet would not
have been wide enough to do this. Maybe a different
model. I vaguely recall that the original drives were
replaced with a newer unit about the time I came there.
One had to be sure to go over the the 501 and shut
down the program before opening the windowed access
to the read station otherwise it might try and select a
card while your hands were in the way. Not well
interlocked. I never saw any double feed issues
either. The RACE units provided for the majority
of user file storage for this system. My sig below is
a command to this file system which would be at the
end of a program (either card or TTY RJE, the $$
only being necessary on the TTY's as the next job's
job card would do the same thing from cards) which
would terminate the batch edit and save the results to
the file system before calling out the language compiler
to CLG the source you had just edited. I never saw
a RACE other than at CIT (CMU). As I mentioned
elsewhere, it was rumored that the designer of the CRAM
and the RACE were the same guy. I don't know if this
is true or who that may be. From my experience the
unit was reasonably reliable. Every so often the 501
would get confused and need to be rebooted from the
one tape drive it had. Someone also had written a
music program for the 501 which as usual played through
an AM radio but in addition used the console printer for
percussion. I heard this program run once. I don't recall
the selection :-)
jchausler
"L0nD0t.$t0we11" wrote:
> That thing is dinky. The big RCA moose was more on the order
> of 20-30 feet long. If can find an online image will post,
> which will probably be easier if I go dig out the RCA reference
> manuals that had the model number of the beast.
I would like to see this. Maybe the Moose and the RCA
RACE are different units although the rest of your description
matches well. The RACE drives each were only about 10 to
12 feed long, a couple to three feet wide and about 5 feet high.
jchausler
Bill Turlock wrote:
True. An IBM CE taught me to change 2314 packs by
turning the drive off, immediately ripping the sliding drawer
open and putting the palms of both hands on the top platter
to rapidly slow it down. (limited by how hot your hands
got in the process). Hmmm......
As I stated in the other thread, the RCA RACE drives
were not well interlocked either and one had to be
careful.
L0nD0t.$t0we11
> I never heard the RCA RACE called the "moose", but from
> what you're describing, we're talking about the same unit.
Perhaps a variant of the same beastie. The official term
in the Spectra/70 era was MSU or mass storage unit.
However, the innards of this beast were not Spectra
generation either as discrete transistor logic or
newer, it had the old round topped black transistors
with a Spectra Channel interface that was extremely
obviously hacked crudely into an earlier generation
technology.
> At CIT in the 60's there was a pair of RACE drives using
> an RCA 501 as their controller. The 501 was interfaced
> to the pair of Bendix G-20's (re branded CDC) which
> served as the primary computer facility at the school through
> most of the 60's. As a part time computer operator on
> weekends, I had to open the drives and clean the read
> drum and heads a couple times a shift. The cards ( I still
> have one) were notched and there were 128 possibilities.
> If the drive got a card which did not match its selection,
> it stuck it out a slot. The operator could then examine
> the card for damage. The typical problem was that one
> of the fingers between the notches had been bent or
> broken. Sometimes it could be straightened out. There
> was another slot into which you could then stick the
> card and it would check it again. From my experience
> with bent fingers, this worked successfully most of the
> time. Our drives only had the bins along one side. I
> was unaware that one could get them along both sides
> and my memory tells me that the drive cabinet would not
> have been wide enough to do this. Maybe a different
> model.
I first encountered the beast in 1969, well into the
era of the Spectra 70/45 to which it was attached.
The internals didn't look much like a 501 or even
301 electronics styles.
The one unit I did encounter had bins on both sides,
was mumble mumble 20-30 feet in length. But being
so much later, it probably had been goldberg'd even
further by the RCA sustaining group.
Mainly my memory of the Spectra MSU was that the system
catalog would break down every now and then and it would
go into serial search looking for the particular
insurance record. Recovery was to offline the pig and
recover from the system console, as it could take a
loonnnnng time for it to read each card in each bin on
both sides looking for the specific record. A software
bug the Cherry Hill guys swore up and down never happened
on pure disk systems, and not even on the smaller
single sided MSU. I left RCA before ever hearing of
any good fix.
PS. 501. Is that the one that had the RCA tape drive
where it actually *weighed* the tape in the columns for
servo balance? I remember guffawing at the drive the
first time I saw it [at RCA Records in Indianapolis] but
don't recall which of the antiquated RCA processor
systems used it.
Joe Morris
>True. An IBM CE taught me to change 2314 packs by
>turning the drive off, immediately ripping the sliding drawer
>open and putting the palms of both hands on the top platter
>to rapidly slow it down. (limited by how hot your hands
>got in the process). Hmmm......
H'mmm indeed. The 2314 drawers on all the systems I worked with
had a toothed wheel attached to the spindle; whenever the drawer
was opened a pawl dropped onto the wheel to lock it so that the cover
could be screwed onto or off of the pack.
If you opened the drawer while the disk was still spinning at a moderate
rate the pawl didn't have time to drop into the teeth as they spun past,
so you got a click-click-click noise. Once the disk slowed down to the
point that the pawl could drop, however, it did so and the spindle was
brought to an instantaneous stop -- with the obvious opportunity for the
rotational energy to be expended by bending various parts of the drive
and/or deforming the bearings.
Of course, if the CE has locked off the pawl (and even with it unlocked
the inertia of the disk usually allowed you to attach and remove the
covers) then the above doesn't apply -- but the box is out of spec,
and you've got a possible safety problem.
Joe Morris
jchausler
"L0nD0t.$t0we11" wrote:
> Roughly 1/6/04 08:14, jchausler's monkeys randomly typed:
>
> > I never heard the RCA RACE called the "moose", but from
> > what you're describing, we're talking about the same unit.
>
> Perhaps a variant of the same beastie. The official term
> in the Spectra/70 era was MSU or mass storage unit.
>
> However, the innards of this beast were not Spectra
> generation either as discrete transistor logic or
> newer, it had the old round topped black transistors
> with a Spectra Channel interface that was extremely
> obviously hacked crudely into an earlier generation
> technology.
This looks like a later device than the ones I was
familiar with.
> I first encountered the beast in 1969, well into the
> era of the Spectra 70/45 to which it was attached.
> The internals didn't look much like a 501 or even
> 301 electronics styles.
The ones I worked with were removed from service
in 1968 IIRC. Although, again, I vaguely recall that
the drive units had been changed out once with a newer
model just as I got there in 1966, I believe the original
installation was around 1964 + or -.
> The one unit I did encounter had bins on both sides,
> was mumble mumble 20-30 feet in length. But being
> so much later, it probably had been goldberg'd even
> further by the RCA sustaining group.
This is much larger than the units I worked with. The
name of the units were definitely RCA RACE (Random
Access Card ???, or something like that). As I said I
have a "RACE CARD" unused in its envelope. It says
RCA on the envelope but not RACE. It, however, says
that "This envelope contains one magnetic card for
replacement in the 3488". I assume 3488 is the model
number of the drive. The card is 4 1/2 inches wide by
16 inches long and reasonably heavy plastic. The notches
on my card say it is address 119 (out of 128). Again,
the cards are quite sturdy. I remember one of the Philco
CE's (who maintained the system, both the RACE and its
RCA 501 controller and the Bendix G20's to which it
was interfaced) telling me he used the used cards for
hinges for his dog door :-) I found this "new" card
in its envelope ("Do not open this envelope until the
machine has been set properly to accept the enclosed
card") under the false floor where the RACE had been
installed, several years after the units had been removed.
> PS. 501. Is that the one that had the RCA tape drive
> where it actually *weighed* the tape in the columns for
> servo balance? I remember guffawing at the drive the
> first time I saw it [at RCA Records in Indianapolis] but
> don't recall which of the antiquated RCA processor
> systems used it.
No, this single tape drive was an RCA unit but IIRC
it had the typical vacuum columns.
jchausler
Joe Morris wrote:
> jchausler <jcha...@earthlink.net> writes:
>
> >True. An IBM CE taught me to change 2314 packs by
> >turning the drive off, immediately ripping the sliding drawer
> >open and putting the palms of both hands on the top platter
> >to rapidly slow it down. (limited by how hot your hands
> >got in the process). Hmmm......
>
> H'mmm indeed. The 2314 drawers on all the systems I worked with
> had a toothed wheel attached to the spindle; whenever the drawer
> was opened a pawl dropped onto the wheel to lock it so that the cover
> could be screwed onto or off of the pack.
Yes
> If you opened the drawer while the disk was still spinning at a moderate
> rate the pawl didn't have time to drop into the teeth as they spun past,
> so you got a click-click-click noise. Once the disk slowed down to the
> point that the pawl could drop, however, it did so and the spindle was
> brought to an instantaneous stop -- with the obvious opportunity for the
> rotational energy to be expended by bending various parts of the drive
> and/or deforming the bearings.
Yes but in addition to the clicking noise, it did help slow the disk down
but I don't recall it coming to an "instantaneous stop". Of course, the
palms of my hands were helping too and maybe when it got slow enough
for the pawl to drop, the friction of my palms was enough to essentially
stop the disk instead of expending the rotational energy against the
pawl. I did this a lot at that site. It was almost SOP. This was a
"production shop" with an emphasis on "production". A lot of the
software running was from the tape days and a typical job would
mount and unmout quite a number of drives (although nothing like
the "real" tape jobs, or is that "reel", which could keep several
operators working almost constantly).
> Of course, if the CE has locked off the pawl (and even with it unlocked
> the inertia of the disk usually allowed you to attach and remove the
> covers) then the above doesn't apply -- but the box is out of spec,
> and you've got a possible safety problem.
It definitely made the rather loud clicking noise. Now, on the other
hand, had I done this at CIT/CMU with their 2314's I would likely
have been thrown off the top floor of the computer center......
Heinz W. Wiggeshoff
>
...
> hand, had I done this at CIT/CMU with their 2314's I would likely
> have been thrown off the top floor of the computer center......
.... no doubt landing in the outdoor fountain used to cool the
the machine room boxes, (a la James Bond, Inspector Clueless etc.).
Legend has it that such a fountain at Ottawa's Carleton University
a bottle of dishwashing detergent was added at the start of each
school year.
B-)
Jeff Jonas
>> knees, and without any hesitation (if you flinched you'd get
>> your hand chewed up) slam the heel of his hand against the
>> outside edge of the spinning drum to bring it to an abrupt stop.
> This strikes me as superbly shitty engineering! Or perhaps
> the unit predates any formal workplace safety standards ...
It's surprising how many parts you can REMOVE from a device
and still have it running.
Safety interlocks tended to have overrides,
some temporary "as needed", some totally defeated.
A retired professor I know is now an expert on safety issues,
testifying about cases of industrial accidents that were avoidable
had the safety interlocks remained operational.
On the bright side, we have the technology and wisdom how to make
things safe with infra-red "curtains", dead-man switches, etc.
I remember seeing a huge hydraulic paper-cutter that required the operator
to press deeply recessed switches with BOTH hands,
standing far away from the cutter.
comp.risks is full of stories to the contrary, though :-(
I remember the day I learned how to make the safety-override switch
pop up from the 1442 card reader-punch (I still have that part as a souvenior).
It had spinning and whurring parts all around,
kinda like the atomic bomb from Goldfinger!
Similarly, I have the switches from the Telefile disk drives
(2: 14 inch 11 platter diskpacks on drawers that slid out).
Opening the cabinet released the button safety interlock switch,
but it was not just a momentary push putton:
it had a third position.
It pulled out further to the "on/enable" position
once the metal guard was unscrewed and moved aside.
Of course, none of that was clearly labelled "override",
that was at least SOME attempt of safety thru obscurity!