Changeset 473 in t29-www


Ignore:
Timestamp:
Nov 15, 2013, 9:21:42 AM (6 years ago)
Author:
heribert
Message:

Kleine Ergänzungen

Files:
3 edited

Legend:

Unmodified
Added
Removed
  • de/geraete/magnetdrahtspeicher.php

    r284 r473  
    3535
    3636<p style="clear:both;">Im Prinzip war das eine geniale Idee, kein Fädeln der Ringkerne, kurze Zykluszeit, preiswert und maschinell herzustellen...
    37 <br/>Doch man ahnt es schon: Der Aufbau war so empfindlich, dass schon bald viele Probleme auftraten; ein Horror für jede Firma, die diesen Speicher verwendete. Bei ganz alten Speichern konnte man die Drähte noch einzeln auswechseln, später wurde alles verklebt und eine Reparatur ist unmöglich. Anfang der 70er Jahre kamen glücklicherweise die ersten Halbleiterspeicher auf den Markt, so dass die meisten Rechner mit Magnetdrahtspeicher auf Halbleiterspeicher mit INTEL-Chip´s umgerüstet wurden, so auch unsere UNIVAC 9400.</p>
     37<br/>Doch man ahnt es schon: Der Aufbau war so empfindlich, dass schon bald viele Probleme auftraten; ein Horror für jede Firma, die diesen Speicher verwendete. Bei ganz alten Speichern konnte man die Drähte noch einzeln auswechseln, später wurde alles verklebt und eine Reparatur ist unmöglich. Anfang der 70er Jahre kamen glücklicherweise die ersten Halbleiterspeicher auf den Markt, so dass die meisten Rechner mit Magnetdrahtspeicher auf Halbleiterspeicher mit INTEL-Chip´s umgerüstet wurden, so auch unsere UNIVAC 9400.<br>
     38Dagegen läuft unsere UNIVAC 9200 noch mit dem originalen Drahtspeicher, vermutlich der letzte noch funktionsfähige Magnetdrahtspeicher der Welt!</p>
  • de/rechnertechnik/speichermedien.php

    r373 r473  
    66        require "../../lib/technikum29.php";
    77?>
    8 <h2>Speichermedien</h2>
     8<h2>Historische Speichermedien</h2>
    99
    1010<p>
  • en/devices/plated-wire-storage.php

    r296 r473  
    1515<p>In the early 1970s, magnetical core memory was quite expensive and slow. That is, the cycle time (time needed to read and rewrite the information) was quite long, because all information had to be rewritten after they had been read. Semiconductor memory was not ready for series production for a long time yet, furthermore nobody even thought about non-volatile semiconductor memory at that time. So there was a need for a new consistent storage medium.
    1616  <br/>As a solution, the plated wire was invented, just before the first semiconductor memories were ready for beeing mass-produced, supposed to be a replacement for the magnetical core memories. The plated wire consists of magnetic wire which replaces the cores (see picture at the right).</p>
    17 <p>The installations from the UNIVAC 9000 series (like our <a href="/en/devices/univac9400/univac_9300.shtm">UNIVAC 9300</a>) were equipped with plated wire storages. We could offer pictures of our storage units, but the assembly of this type of memory is quite opaque. Therefore we publish self-drawn diagrams to illustrate the plated wire characteristics.</p>
     17<p>The installations from the UNIVAC 9000 series (like our <a href="/en/devices/univac9400/univac_9300.shtm">UNIVAC 9300</a>) were equipped with plated wire storages. We could offer pictures of our storage units, but the assembly of this type of memory is quite opaque. Therefore we publish self-drawn diagrams to illustrate the plated wire characteristics.
     18</p>
    1819
    1920<p>From the physical point of view the plated wire storage is a thin-film storage medium. The information carrier is a 1 micrometer thick permalloy solenoid film (constisting of 81% nickel, 19% iron) that  sheats a beryllium copper wire with 0.13mm in diameter.<br/>Therefore you could save exactly 4 words of three bits each on the pictured detail.</p>
     
    4243  In principle this was an ingenius idea: No need to thread the cores, fast cycle time, very cheap to produce...
    4344  <br/>But there is a rub; the construction was very sensitive and error-prone. While the wires could be replaced in the former storage systems, typical plated wire storages were completely stuck together so they could not be repared later on. This was a disaster for the companies which made use of this storage.
    44   <br/>Fortunately, the first semiconductor memories were mass-produced in the early 1970s, so most of the computers featuring plated wire storage could be upgraded to semiconductor memories made by INTEL.</p>
     45  <br/>Fortunately, the first semiconductor memories were mass-produced in the early 1970s, so most of the computers featuring plated wire storage could be upgraded to semiconductor memories made by INTEL.<br>
     46  Other hand our UNIVAC 9200 is still running with the original wire-memory, probably the last remaining functional magnetic wire store in the world!</p>
    4547
Note: See TracChangeset for help on using the changeset viewer.
© 2008 - 2013 technikum29 • Sven Köppel • Some rights reserved
Powered by Trac
Expect where otherwise noted, content on this site is licensed under a Creative Commons 3.0 License