1 Dec
2017
1 Dec
'17
10:12 a.m.
https://www.ebay.com/itm/263005049078
EBay listing for a "Soviet Magnetic Ferrite Core Memory Board". It looks
like 20 something gigantic cores and a lot of diodes. I am guessing it is
some kind of ROM, but it doesn't look like a rope memory. And maybe the
cores are not cores at all, but some sort of inductor. I've not seen this
before.
1 Dec
1 Dec
6:38 p.m.
On 12/01/2017 07:12 AM, Tony Aiuto via cctalk wrote:
https://www.ebay.com/itm/263005049078
EBay listing for a "Soviet Magnetic Ferrite Core Memory Board". It looks
like 20 something gigantic cores and a lot of diodes. I am guessing it is
some kind of ROM, but it doesn't look like a rope memory. And maybe the
cores are not cores at all, but some sort of inductor. I've not seen this
before.
Probably a driver board for core memory.
--Chuck
11:05 p.m.
On 2017-Dec-01, at 7:12 AM, Tony Aiuto via cctalk wrote:
https://www.ebay.com/itm/263005049078
EBay listing for a "Soviet Magnetic Ferrite Core Memory Board". It looks
like 20 something gigantic cores and a lot of diodes. I am guessing it is
some kind of ROM, but it doesn't look like a rope memory. And maybe the
cores are not cores at all, but some sort of inductor. I've not seen this
before.
That's very funny.
It looks to be a core rope memory that hasn't been programmed.
Other organisations might be possible, but it looks like a pulse-transformer type of
core-rope,
where the cores are just for ordinary induction, not switching/memory cores.
- the matrix of black what-look-to-be diodes would be data-wire isolation diodes
- the little brown 'stools' are wire routing posts
- you can see the mulit-turn sense windings (bluish) already present on the cores
- above the cores are the sense amplifiers or 1st stage thereof
- there is one wire through all the cores, perhaps a test wire for core and sense amp
response
Each data-wire would start at one of the solder pins in the pin matrix on the left, weave
through the cores to encode the data,
turn back 180, then 90 degrees around one of the stools to drop down and terminate at the
solder pin by an isolation diode.
There would be another board for decoding the address to 1-of-x and 1-of-y.
I didn't count precisely but it looks like it would be 256 words of 20 bits.
That might be a date code of 6847 on a cap (or is it 6B47?), so perhaps earlier than the
listing-stated 1981.
Actually, it kind of hints at it in the description: "With out Firmware ROM wire
(empty slots)"
2 Dec
2 Dec
11:44 a.m.
On Fri, Dec 1, 2017 at 11:05 PM, Brent Hilpert via cctalk <
cctalk at classiccmp.org> wrote:
On 2017-Dec-01, at 7:12 AM, Tony Aiuto via cctalk
wrote:
I think that is the most likely case.
https://www.ebay.com/itm/263005049078
EBay listing for a "Soviet Magnetic Ferrite Core Memory Board". It looks
like 20 something gigantic cores and a lot of diodes. I am guessing it is
some kind of ROM, but it doesn't look like a rope memory. And maybe the
cores are not cores at all, but some sort of inductor. I've not seen this
before.
That's very funny.
It looks to be a core rope memory that hasn't been programmed.
Other organisations might be possible, but it looks like a
pulse-transformer type of core-rope,
where the cores are just for ordinary induction, not switching/memory
cores.
- the matrix of black what-look-to-be diodes would be data-wire
isolation diodes
- the little brown 'stools' are wire routing posts
- you can see the mulit-turn sense windings (bluish) already
present on the cores
- above the cores are the sense amplifiers or 1st stage thereof
- there is one wire through all the cores, perhaps a test wire for
core and sense amp response
Each data-wire would start at one of the solder pins in the pin matrix on
the left, weave through the cores to encode the data,
turn back 180, then 90 degrees around one of the stools to drop down and
terminate at the solder pin by an isolation diode.
There would be another board for decoding the address to 1-of-x and 1-of-y.
I didn't count precisely but it looks like it would be 256 words of 20
bits.
That might be a date code of 6847 on a cap (or is it 6B47?), so perhaps
earlier than the listing-stated 1981.
Actually, it kind of hints at it in the description: "With out Firmware
ROM wire (empty slots)"
Ah, you read the description. I just looked at the title and saw "with the
firmware". My addled brain made the leap to a external firmware, which made
no sense. "firmware ROM wire" would be a clear case for rope memory.
On Fri, Dec 1, 2017 at 11:32 PM, Charles Anthony <charles.unix.pro at gmail.com
wrote:
The last picture has "???-5". Some googling takes us to
https://bg.wikipedia.org/wiki/%D0%94%D0%97%D0%A3
"DZU is a factory in Stara Zagora , a major producer of magnetic disk
storage devices (hard drives and floppy disks) during the rise of computer
production in Bulgaria in the 1970s and 1980s, century. Today it is part of
VIDEOTON Holding ZRt., Hungary [1] ."
The article says it was a disk drive factory, but maybe...
-- Charles
Given the cleanliness of the board and other things the seller is offering,
my guess now is that this NOS from the DZU plant.
Thanks, everyone.
5:03 p.m.
My understanding is that DZU made disk storage subsystems including both drives and
control units and that many if not all into the 1980s were copies of IBM disk storage
products.
The recollection is that IBM 2314 era SCU used a form of TROS where words were strips of
film that punched a hole a one or a zero at each transformer (core) location to route the
current around or through a core. The 2841 may have had the same technology. This could
be done relatively easily in the field.
At Memorex they used wire rope as TROS for the same function as I think did other PCMs.
This was smaller and allowed faster clocks than film strips but was difficult but not
impossible to field upgrade. Next generation SCUs went to writable control store's
loadable from FDD's.
So this might be a 2314 era TROS so its date could be early 70s given first PCM 2314's
SCUs didn't ship until the early 1970s so I have a hard time thinking the Russian
system beat US capitalism :-)
But if they copied the 2841 and it used TROS then it could be late 60s
Tom
-----Original Message-----
From: Tony Aiuto [mailto:tony.aiuto at gmail.com]
Sent: Saturday, December 02, 2017 8:45 AM
To: Brent Hilpert; General Discussion: On-Topic and Off-Topic Posts
Subject: Re: Can anyone identify what this board is/does?
On Fri, Dec 1, 2017 at 11:05 PM, Brent Hilpert via cctalk < cctalk at
classiccmp.org> wrote:
On 2017-Dec-01, at 7:12 AM, Tony Aiuto via cctalk
wrote:
I think that is the most likely case.
https://www.ebay.com/itm/263005049078
EBay listing for a "Soviet Magnetic Ferrite Core Memory Board". It
looks like 20 something gigantic cores and a lot of diodes. I am
guessing it is some kind of ROM, but it doesn't look like a rope
memory. And maybe the cores are not cores at all, but some sort of
inductor. I've not seen this before.
That's very funny.
It looks to be a core rope memory that hasn't been programmed.
Other organisations might be possible, but it looks like a
pulse-transformer type of core-rope, where the cores are just for
ordinary induction, not switching/memory cores.
- the matrix of black what-look-to-be diodes would be
data-wire isolation diodes
- the little brown 'stools' are wire routing posts
- you can see the mulit-turn sense windings (bluish) already
present on the cores
- above the cores are the sense amplifiers or 1st stage
thereof
- there is one wire through all the cores, perhaps a test wire
for core and sense amp response
Each data-wire would start at one of the solder pins in the pin matrix
on the left, weave through the cores to encode the data, turn back
180, then 90 degrees around one of the stools to drop down and
terminate at the solder pin by an isolation diode.
There would be another board for decoding the address to 1-of-x and 1-of-y.
I didn't count precisely but it looks like it would be 256 words of 20
bits.
That might be a date code of 6847 on a cap (or is it 6B47?), so
perhaps earlier than the listing-stated 1981.
Actually, it kind of hints at it in the description: "With out
Firmware ROM wire (empty slots)"
Ah, you read the description. I just looked at the title and saw "with the
firmware". My addled brain made the leap to a external firmware, which made no sense.
"firmware ROM wire" would be a clear case for rope memory.
On Fri, Dec 1, 2017 at 11:32 PM, Charles Anthony <charles.unix.pro at gmail.com
wrote:
The last picture has "???-5". Some googling takes us to
https://bg.wikipedia.org/wiki/%D0%94%D0%97%D0%A3
"DZU is a factory in Stara Zagora , a major producer of magnetic disk
storage devices (hard drives and floppy disks) during the rise of
computer production in Bulgaria in the 1970s and 1980s, century. Today
it is part of VIDEOTON Holding ZRt., Hungary [1] ."
The article says it was a disk drive factory, but maybe...
-- Charles
Given the cleanliness of the board and other things the seller is offering, my guess now
is that this NOS from the DZU plant.
Thanks, everyone.
1 Dec
1 Dec
11:32 p.m.
On Fri, Dec 1, 2017 at 7:12 AM, Tony Aiuto via cctalk <cctalk at classiccmp.org
wrote:
https://www.ebay.com/itm/263005049078
EBay listing for a "Soviet Magnetic Ferrite Core Memory Board". It looks
like 20 something gigantic cores and a lot of diodes. I am guessing it is
some kind of ROM, but it doesn't look like a rope memory. And maybe the
cores are not cores at all, but some sort of inductor. I've not seen this
before.
The last picture has "???-5". Some googling takes us to
https://bg.wikipedia.org/wiki/%D0%94%D0%97%D0%A3
"DZU is a factory in Stara Zagora , a major producer of magnetic disk
storage devices (hard drives and floppy disks) during the rise of computer
production in Bulgaria in the 1970s and 1980s, century. Today it is part of
VIDEOTON Holding ZRt., Hungary [1] ."
The article says it was a disk drive factory, but maybe...
-- Charles
2 Dec
2 Dec
11:17 a.m.
This is a wire rope ROM, but it is unpopulated, meaning that there are no wires strung
through the cores to encode anything.
The little pegs located below the cores are for routing the wire.
The diodes are all for address selection, and the circuitry above the cores is the sense
amplifiers to take the small current induced in the "secondary" winding around
the core and amplify it to a level usable by the rest of the logic. It is a 20-bit wide
word. I didn't take the time to try to figure out how many words that there are, but
I'd suspect something on the order of 256 to 512 words could be stored.
To program it, thin magnet wire would be soldered to pads at the outputs of the address
decoding (one pad for each "word"), and then threaded through (or not) each of
the 20 cores to encode 1's and 0's, and then mass terminated on another pad.
The threading operation would be tedious. Generally there would be a special needle-like
tool that would be used to thread the wire for each word.
Definitely an interesting piece.
-Rick
---
Rick Bensene, The Old Calculator Museum
http://oldcalculatormuseum.com
-----Original Message-----
From: cctalk [mailto:cctalk-bounces at classiccmp.org] On Behalf Of Tony Aiuto via cctalk
Sent: Friday, December 01, 2017 7:12 AM
To: General Discussion: On-Topic and Off-Topic Posts
Subject: Can anyone identify what this board is/does?
https://www.ebay.com/itm/263005049078
EBay listing for a "Soviet Magnetic Ferrite Core Memory Board". It looks like 20
something gigantic cores and a lot of diodes. I am guessing it is some kind of ROM, but it
doesn't look like a rope memory. And maybe the cores are not cores at all, but some
sort of inductor. I've not seen this before.
2424
days inactive
2425
days old
6 comments
6 participants
participants (6)
-
bhilpert＠shaw.ca -
cclist＠sydex.com -
charles.unix.pro＠gmail.com -
rickb＠bensene.com -
t.gardner＠computer.org -
tony.aiuto＠gmail.com