Full resolution (JPEG) - On this page / på denna sida - Some New Components for Memory Applications, by Lars Arosenius
<< prev. page << föreg. sida << >> nästa sida >> next page >>
Below is the raw OCR text
from the above scanned image.
Do you see an error? Proofread the page now!
Här nedan syns maskintolkade texten från faksimilbilden ovan.
Ser du något fel? Korrekturläs sidan nu!
This page has never been proofread. / Denna sida har aldrig korrekturlästs.
these, the x-wire, can inhibit the coupling of the
flux from the y-wire to the signal wire. Since this
y-wire can carry currents which are not limited
in amplitude, very fast switching may be the result.
The main characteristics of the magnetic family
elements is found in table l13.
Tab. 1. Comparison between different magnetic
materials. S = small, M = medium, L = large.
Speed [IS Current mA read inhibit Output mV Size Cost
Ferrite Core . 1.0 850 425 100 M M
Ferrite Plate 16X16 ... 1.8 320 160 40 S S
Twistor ..... . 1.0 2 000 100 10 L S
Film ........ ..<0.1 400 400 5 M L
Low-temperature elements
The use of superconductive phenomena for memory
and logical elements has been reported previously18.
Therefore just a few remarks on some recent
developments will be given here. The main limitation of
the superconductive components has been their low
switching speed. To achieve high speed it is
necessary to produce thin superconductive films, which
necessitates evaporation either in extremly high
vacuum30 (< 10~12 mm Hg) or with very high ion
densities. Thin films of normally superconductive
material evaporated in e.g. 10~7 mm Hg do not
become superconductive at low temperatures,
probably due to impurities in the deposited film.
The problems of thin superconductive films are
not fully understood at the moment, but at the IBM
Besearch Laboratories films of good quality have
been produced, and it is hoped that better
fabrication techniques will produce thin superconductive
films with reproducible characteristics.
The search for new superconducting materials
gives alloys with higher and higher transition
temperatures. Critical temperatures of more than 30° K
have been found’1, and it may be hoped that some
day a material will be produced which is
superconductive at the temperatures for liquid air or, why
not, at room temperature.
The trapped flux memory
Trapped flux memories22’23 have been built and
tested. Some characteristics for an 8 x 8 bits array are23
Memory cycle 1 j.is
Switching time <10 ns
Signal to noise ratio ^>10
Drive current requirement 300 mA
Storage capacity 100 bits/inch2
If the production problems can be mastered the
superconductive memories and components seem to
be the most promising elements yet proposed for
large scale computer systems. The sandwich
structure of the components is very suitable for
automatic fabrication. A special problem is encountered
in the cooling. The most common cooling medium
is liquid helium, which is now readily available.
For whole systems in a bath of helium the losses
are expected to be small. For superconductive
memories connected to computers operating with con-
ventional components (which is not an impossible
thought because of the very good characteristics of
e.g. the trapped flux memory) the boiling of helium
during communication between the memory and the
rest of the computer can be rather excessive, so
for instance it is anticipated, that a 4 000 word 1 (.is
memory will use Vi liter of helium per hour.
It must be observed, however, that the problem of
cooling can be solved by the discovery of new
superconductive materials with more favorable critical
temperatures.
Final remarks
During a trip 1958 in the USA the author got a
strong impression that the low-temperature
components potentially have almost all the properties
needed for the very large and very fast computers of
the future.
They combine speed, resonable power requirements
and simple circuitry with low production cost and
with a supposed very good reliability. The magnetic
elements are still very widely used. The
introduction of new structures has increased the capabilities
of the magnetic elements considerably.
References
1. Best R: Memory Units in the Lincoln I.Y-2 Computer. Proc.
Western Joint Comp. Conference 1957.
2. Bergman C-I: Minnen i elektroniska sift er maskiner. Tekn.
Tidskrift 88 (1958) p. 307.
3. McMahon R E: Impulse Switching of Ferrites. Digest of
Technical Papers, Conference on Solid State Physics (CSSP), Philadelphia
1959.
4. Sylvan T P: Two Terminal Solid State Switches. Electronics
32 (1959) p. 62.
5. Shockley W: The Four Layer Diode. Beckman Instruments
GMBH, Munich 1959.
6. Tancrell R H: Thin Ferromagnetic Films as Function Table
Elements. M. Sc Thesis in E. E., Mass. Inst, of Technology,
Cambridge 1958.
7. Olssön C D, Pohm A V: Flux Reversal in Thin Films of 82 "U
Ni, 18 V» Fe. Journ. of Appl. Physics 29 (1958) no 3.
8. Callen II B: Rotational Remagnetization of Thin Films. Digest
of Technical Papers, CSSP, Philadelphia 1959.
Ö. Papian W N: Process Control for Thin Film Xacuum
Deposition. Memorandum 6M-5877 Lincoln Laboratory, Lexington, Mass.,
195S.
10. Gianola U F: Nondestructive Memory Employing a Domain
Oriented Steel Wire. Journ. of Appl. Physics 29 (1958) no 5 p. 849.
11. Bajciimann J A: Ferrite Apertured Plate for Random Access
Memory. Proc. IRE 45 (1957) p. 325.
12. Warren C S: Ferrite Apertured Plate Memories. Digest of
Technical Papers, CSSP, Philadelphia 1959.
13. Looney D: Recent Advances in Magnetic Devices for Computers.
To be published.
14. Bobeck A H: A New Storage Element Suitable for Large-Sized
Memory Arrays — The Twistor. Bell System Technical J. XXXVI
(1957) p. 1319.
15. Preston K, Simkins Q \V: Twistor Buffer Store. Digest of
Technical Papers, CSSP, Philadelphia 1959.
16. Bajciimann J A, Lo R W: The Transfluxor. Proc. IRE 45 (1956)
p. 1081.
17. Crane II: .1 lligh-Specd Logic System Using Magnetic Elements
and Connecting Wire Only. Proc. IRE 47 (1959) p. 63.
18. Gianola U: AIEE Conference on Nonlinear Magnetics. August
8, 1958.
19. Törnquist J-R: Lågtemperaturelektronik. Teknisk Tidskrift 88
(1958) p. 1041.
20. High Speed Computer-System Research. Quarterly Progress
Reports 1—3, Computer Comp. and Systems Laboratory, MIT,
Cambridge 1958.
21. Bell Telephone Laboratories. Private Communication.
22. Kraus C J: Operation of a Low-Temperature Memory Element.
Digest of Technical Papers, CSSP in Philadelphia 1959.
23. Kraus C J: Pro’s and Con’s on a Superconducting Memory.
Digest of Technical Papers, CSSP in Philadelphia 1959.
ELTEKNIK 1959 1 91
<< prev. page << föreg. sida << >> nästa sida >> next page >>
