

In many 'professional"
computers the use of Socketed Integrated Circuits is avoided due to the lowered
reliability of the Socketed I.C.'s. This is caused by a predisposition towards
contact generated noise, signal loss in the contact means, and a slower rise
time on waveforms. The latter is a consequence of the higher connector
resistance in the signal lead together with the internal capacity of the I.C.
The result of increased rise times is an occasional system failure, as certain
signal paths are not enabled by other timing signals, within the
"windows" necessary to keep the equipment operating.
If the resistance in the I.C. Socket
to I.C. pin contact means could be kept at a low level or lowered to that of a
new I.C. Socket, then the rise time could be kept within the specifications
necessary to maintain the critical timing of the signals.
An old and unreliable S-100 type 64k
memory board was set up on a connector extender in a Z- 80 based computer, and
measurements were made of the rise time of the chip refresh signals using a
very high impedance low capacity probe. The measurements were made:
The measurements were made before and after the application of Stabilant to the I.C. contacts.
The
procedure was duplicated on 10 of the memory chips.
Upon using the stabilant
there was an average reduction in rise time of 40% with one contact having a
reduction of 70% in its signal rise time. The board, which had hitherto been
unreliable, now functioned properly once all I.C.'s were treated.
The tests demonstrate that in the
case of Dynamic Memory I.C.'s, the reduction in rise time could well be the
difference in the I.C. performing acceptably, and failing.
If manufacturer did not require
soldered-in-I.C.'s to ensure reliability, the cost of repairing boards because
of bad I.C.'s could be reduced. Lower cost I.C.'s might be employed in
manufacturing in the knowledge that a marginal I.C. could be replaced. This
could cut the cost of parts.
With very high speed memory
applications or where very high rise times are encountered, it must be
remembered that the capacitance between adjacent contacts is increased by the
application of Stabilants on the socket insulation.
In most applications any effect of this added capacitance will be offset by the
reduced resistance of the contact; however in very high speed applications
using chip carriers with very small spacings between
contacts it may be necessary to remove the IC's - applying the Stabilants only to the metal contacts themselves. This can
be accomplished by using a piece of thin felt (saturated with Stabilant 22a) as the applicator. Remember, only a very
thin film of Stabilant is needed in this type of
application.
Revision
4
NATO/CAGE Supply Code
38948
The 15 mL size has NATO
Part # 5999-21-900-6937
The Stabilants are patented in Canada - 1987; US Patent number 4696832. World-wide patents
pending. Because the patents cover contacts treated with the material, a
Point-of-sale License is granted with each purchase of the material.
NOTICE: This data has been
supplied for information purposes only. While to our knowledge it is accurate,
users should determine the suitability of the material for their application by
running their own tests. Neither D.W. Electrochemicals Ltd.,
their distributors or their dealers assume any responsibility or
liability for damages to equipment and/or consequent damages, howsoever caused,
based on the use of this information.
Stabilant, Stabilant
and product type variations thereof, are Trade Marks of D.W. Electrochemicals
Ltd. Stabilant, Stabilant
22, and .the term "Contact Enhancer" is © 1983 Wright Electroacoustics and is used with permission.
©
Copyright 1987,'88.'89.'90. 91 - D.W. Electrochemicals Ltd.