20 Oct
2019
20 Oct
'19
7:35 a.m.
I was just listening to a video on the Voyager space craft. It used an interesting type of
memory, called magnetic wire memory. There is only a little bit of information of it on
the web. It is clever in that has a non-destructive read. I just wondered if any one else
was familiar with this type of memory.
Dwight
20 Oct
20 Oct
7:43 a.m.
search for NDRO and thin film memory
Univac used it in their commercial and military computers
On 10/20/19 7:35 AM, dwight via cctalk wrote:
I was just listening to a video on the Voyager space
craft. It used an interesting type of memory, called magnetic wire memory. There is only a
little bit of information of it on the web. It is clever in that has a non-destructive
read. I just wondered if any one else was familiar with this type of memory.
Dwight
8 a.m.
On Sun, Oct 20, 2019 at 07:43:34AM -0700, Al Kossow via cctalk wrote:
search for NDRO and thin film memory
Univac used it in their commercial and military computers
On 10/20/19 7:35 AM, dwight via cctalk wrote:
> I was just listening to a video on the Voyager space craft. It used an interesting
type of memory, called magnetic wire memory. There is only a little bit of information of
it on the web. It is clever in that has a non-destructive read. I just wondered if any one
else was familiar with this type of memory.
> Dwight
>
Nice article here:
http://www.technikum29.de/en/devices/plated-wire-storage
Never saw it myself only saw core. I don't think DEC ever used plated wire.
>
- Diane
--
- db at FreeBSD.org db at db.net http://www.db.net/~db
8:36 a.m.
https://ntrs.nasa.gov/search.jsp?R=19740018598
from their web page, "SCI" built the Voyager
memory, I've not found details in the on-line docs
https://ntrs.nasa.gov/search.jsp?R=19810001583
On 10/20/19 7:43 AM, Al Kossow via cctalk wrote:
search for NDRO and thin film memory
Univac used it in their commercial and military computers
On 10/20/19 7:35 AM, dwight via cctalk wrote:
I was just listening to a video on the Voyager
space craft. It used an interesting type of memory, called magnetic wire memory. There is
only a little bit of information of it on the web. It is clever in that has a
non-destructive read. I just wondered if any one else was familiar with this type of
memory.
Dwight
8:57 a.m.
https://vipclubmn.org/Articles/PlatedWire.pdf
https://vipclubmn.org/Articles/PlatedWireAddendum.pdf
and
https://vipclubmn.org/Articles/Wired_Up.pdf
The 9300 used it, Donzelli says it wasn't very reliable
http://s3data.computerhistory.org/brochures/sperryrand.univac9000.1967.1026…
9:04 a.m.
and NCR rod memory
https://pdfs.semanticscholar.org/60db/392e7b39d0b1697ead9deb033a4ae488dbe7.…
9:05 a.m.
https://www.computerhistory.org/collections/catalog/102620884
On 10/20/19 9:04 AM, Al Kossow via cctalk wrote:
and NCR rod memory
https://pdfs.semanticscholar.org/60db/392e7b39d0b1697ead9deb033a4ae488dbe7.…
11:35 a.m.
It can't be all that bad if it has been flying in space for all these year without
crashing the processors. The CMOS processor has lost several CMOS memory bits but the main
control processor is still chugging away on its original wire memory.
There are a lot of factors in the entire system.
Dwight
________________________________
From: cctalk <cctalk-bounces at classiccmp.org> on behalf of William Donzelli via
cctalk <cctalk at classiccmp.org>
Sent: Monday, October 21, 2019 8:13 AM
To: Al Kossow via cctalk <cctalk at classiccmp.org>
Subject: Re: plated wire memory
The 9300 used it, Donzelli says it wasn't very
reliable
And I do not remember where I heard that. It may have been from a
Univac old-timer.
--
Will
20 Oct
20 Oct
8:46 a.m.
On 10/20/19 7:43 AM, Al Kossow via cctalk wrote:
search for NDRO and thin film memory
Univac used it in their commercial and military computers
Apparently not for long. The 1107 was a thin-film memory machine, but
the 1108 et seq. was not, IIRC. At least I don't recall any mention in
the documentation, whereas it was a big thing with the 1107.
My guess is that it was difficult to produce and expensive.
--Chuck
7:45 a.m.
I remember an IBM engineer talking about this at our ham radio club. The
wire was coiled inside a drum and pulses were sent down the wire.? The
'read head' was? a magnetic pickup at the other end of the coil - and
access time was however long it took the pulse to arrive at the other
end.? Therefore storage capacity was inversely proportional to data
quantity, however at that time I was working with 660kB Univac FH330
drums for swapping and the 2-ton Fastrand for 164kB of long-term
storage, so it has to be taken in context!
Although the read was actually non-destructive, the pulse had to be
regenerated to go around agaiun.
Is that maybe what you are thinking of?
cheers,
Nigel
On 20/10/2019 10:35, dwight via cctalk wrote:
I was just listening to a video on the Voyager space
craft. It used an interesting type of memory, called magnetic wire memory. There is only a
little bit of information of it on the web. It is clever in that has a non-destructive
read. I just wondered if any one else was familiar with this type of memory.
Dwight
--
Nigel Johnson
MSc., MIEEE
VE3ID/G4AJQ/VA3MCU
Amateur Radio, the origin of the open-source concept!
You can reach me by voice on Skype: TILBURY2591
If time travel ever will be possible, it already is. Ask me again yesterday
This e-mail is not and cannot, by its nature, be confidential. En route from me to you, it
will pass across the public Internet, easily readable by any number of system
administrators along the way.
Nigel Johnson <nw.johnson at ieee.org>
Please consider the environment when deciding if you really need to print this message
7:58 a.m.
On Sun, Oct 20, 2019 at 10:45:27AM -0400, Nigel Johnson via cctalk wrote:
I remember an IBM engineer talking about this at our
ham radio club. The
de VA3DB
wire was coiled inside a drum and pulses were sent
down the wire.? The
'read head' was? a magnetic pickup at the other end of the coil - and
This sounds like delay line memory.
Lot's of info on this on web.
access time was however long it took the pulse to
arrive at the other
end.? Therefore storage capacity was inversely proportional to data
quantity, however at that time I was working with 660kB Univac FH330
drums for swapping and the 2-ton Fastrand for 164kB of long-term
storage, so it has to be taken in context!
Although the read was actually non-destructive, the pulse had to be
regenerated to go around agaiun.
Is that maybe what you are thinking of?
cheers,
Nigel
On 20/10/2019 10:35, dwight via cctalk wrote:
--
- db at FreeBSD.org db at db.net http://www.db.net/~db
I was just listening to a video on the Voyager
space craft. It used an interesting type of memory, called magnetic wire memory. There is
only a little bit of information of it on the web. It is clever in that has a
non-destructive read. I just wondered if any one else was familiar with this type of
memory.
Dwight
--
Nigel Johnson
MSc., MIEEE
VE3ID/G4AJQ/VA3MCU
Amateur Radio, the origin of the open-source concept!
You can reach me by voice on Skype: TILBURY2591
If time travel ever will be possible, it already is. Ask me again yesterday
This e-mail is not and cannot, by its nature, be confidential. En route from me to you,
it will pass across the public Internet, easily readable by any number of system
administrators along the way.
Nigel Johnson <nw.johnson at ieee.org>
Please consider the environment when deciding if you really need to print this message
7:59 a.m.
On Sun, Oct 20, 2019, 09:45 Nigel Johnson via cctalk <cctalk at classiccmp.org wrote:
I remember an IBM engineer talking about this at our
ham radio club. The
wire was coiled inside a drum and pulses were sent down the wire. The
'read head' was a magnetic pickup at the other end of the coil - and
access time was however long it took the pulse to arrive at the other
end.
You are describing delay line memory. Plated wire memory is more akin to
core memory.
https://en.wikipedia.org/wiki/Plated_wire_memory
Kyle
10:22 a.m.
On 10/20/2019 09:45 AM, Nigel Johnson via cctalk wrote:
I remember an IBM engineer talking about this at our
ham
radio club. The wire was coiled inside a drum and pulses
were sent down the wire. The 'read head' was a magnetic
pickup at the other end of the coil - and access time was
however long it took the pulse to arrive at the other
end. Therefore storage capacity was inversely
proportional to data quantity, however at that time I was
working with 660kB Univac FH330 drums for swapping and the
2-ton Fastrand for 164kB of long-term storage, so it has
to be taken in context!
No, that is acoustic delay line memory, and is a serial
access type of data storage All data is lost if the
equipment is powered down. Plated wire memory is a
random-access type of memory using principles similar to
core memory, except the magnetic material is a magnetic film
plated onto the copper wires. There are a few other forms
of NDRO such as Biax that use cores with two holes in them,
one for the sense/inhibit wire and one for the select wires.
Jon
1:50 p.m.
It is funny that the most common memory used today is a DRO type memory. The read destroys
much of the charge on a DRAM cell, requiring a write back of the data.
Dwight
________________________________
From: cctalk <cctalk-bounces at classiccmp.org> on behalf of Jon Elson via cctalk
<cctalk at classiccmp.org>
Sent: Sunday, October 20, 2019 10:22 AM
To: Nigel Johnson <nw.johnson at ieee.org>; General at ezwind.net <General at
ezwind.net>; Discussion@ <ezwind.net:On-Topic and Off-Topic Posts cctalk at
classiccmp.org>
Subject: Re: plated wire memory
On 10/20/2019 09:45 AM, Nigel Johnson via cctalk wrote:
I remember an IBM engineer talking about this at our
ham
radio club. The wire was coiled inside a drum and pulses
were sent down the wire. The 'read head' was a magnetic
pickup at the other end of the coil - and access time was
however long it took the pulse to arrive at the other
end. Therefore storage capacity was inversely
proportional to data quantity, however at that time I was
working with 660kB Univac FH330 drums for swapping and the
2-ton Fastrand for 164kB of long-term storage, so it has
to be taken in context!
No, that is acoustic delay line memory, and is a serial
access type of data storage All data is lost if the
equipment is powered down. Plated wire memory is a
random-access type of memory using principles similar to
core memory, except the magnetic material is a magnetic film
plated onto the copper wires. There are a few other forms
of NDRO such as Biax that use cores with two holes in them,
one for the sense/inhibit wire and one for the select wires.
Jon
2:23 p.m.
On 10/20/19 1:50 PM, dwight via cctalk wrote:
It is funny that the most common memory used today is
a DRO type memory. The read destroys much of the charge on a DRAM cell, requiring a write
back of the data.
Dwight
That's true today, but probably not in the near future. Persistent
memory is getting faster and cheaper.
--Chuck
21 Oct
21 Oct
8:08 a.m.
The plated wire sounds like really fussy stuff. The layer of permalloy had to be thin but
not to thin. I suspect that was to optimize the internal field compared to the outside
word select signal.
At first, I couldn't understand why they'd need to use beryllium copper but it
seems the wires needed to be free standing in the array. Pure copper would sage and
touch.
Clearly an interesting process. If it had not been for solid state, I suspect it would
have been refined to a more manufacturable process. I suspect tuning the layer thickness
could have been done while in the plating bath instead of waiting until it was outside.
Because of the size of things, I suspect it could have been done by vapor deposit on glass
substrates and miniaturized. It depended on the fact that the field had a preferred
direction and couldn't be easily rotated. Such things can be done on a flat surface
with current day processing.
Dwight
________________________________
From: cctalk <cctalk-bounces at classiccmp.org> on behalf of Chuck Guzis via cctalk
<cctalk at classiccmp.org>
Sent: Sunday, October 20, 2019 2:23 PM
To: dwight via cctalk <cctalk at classiccmp.org>
Subject: Re: plated wire memory
On 10/20/19 1:50 PM, dwight via cctalk wrote:
It is funny that the most common memory used today is
a DRO type memory. The read destroys much of the charge on a DRAM cell, requiring a write
back of the data.
Dwight
That's true today, but probably not in the near future. Persistent
memory is getting faster and cheaper.
--Chuck
20 Oct
20 Oct
8:28 a.m.
I don't know much about plated wire store, but I do know it was used in the
Manchester University MU5 computer which pioneered heuristic pipelining.
There is some info here:-
http://homepages.inf.ed.ac.uk/cgi/rni/comp-arch.pl?Ibuff/mu5-ibu.html,Ibuff/
mu5-ibu-f.html,Ibuff/menu-mu5.html
https://books.google.co.uk/books?id=YD5dDwAAQBAJ&pg=PA52&lpg=PA52&a…
22+plated+wire+store&source=bl&ots=4vqufyFe81&sig=ACfU3U07-MqiT-58mc16Yjs7C1
eFm_m4UA&hl=en&sa=X&ved=2ahUKEwiVhdb5kavlAhW5TxUIHbJ9Cz4Q6AEwAXoECAkQAQ#v=on
epage&q=%22mu5%22%20plated%20wire%20store&f=true
Dave
-----Original Message-----
From: cctalk <cctalk-bounces at classiccmp.org> On Behalf Of dwight via
cctalk
Sent: 20 October 2019 15:36
To: General Discussion: On-Topic and Off-Topic Posts
<cctalk at
classiccmp.org>
Subject: plated wire memory
I was just listening to a video on the Voyager space craft. It used an
interesting
type of memory, called magnetic wire memory. There is
only a little bit of
information of it on the web. It is clever in that has a non-destructive
read. I
just wondered if any one else was familiar with this
type of memory.
Dwight
9:34 a.m.
On October 20, 2019 at 9:35 AM dwight via cctalk
<cctalk at classiccmp.org> wrote:
I was just listening to a video on the Voyager space craft. It used an interesting type
of memory, called magnetic wire memory. There is only a little bit of information of it on
the web. It is clever in that has a non-destructive read. I just wondered if any one else
was familiar with this type of memory.Dwight
There is some interesting information, but not a lot of technical details, in the
Computers in Spaceflight book from NASA
https://ntrs.nasa.gov/search.jsp?R=19880069935
Will
"The names of global variables should start with // " -- https://isocpp.org
1670
days inactive
1671
days old
18 comments
10 participants
participants (10)
-
aek＠bitsavers.org -
cclist＠sydex.com -
dave.g4ugm＠gmail.com -
db＠db.net -
dkelvey＠hotmail.com -
elson＠pico-systems.com -
kylevowen＠gmail.com -
nw.johnson＠ieee.org -
wdonzelli＠gmail.com -
wrcooke＠wrcooke.net