I found a relevant paper on IEEE Xplore: "Electrical performance of
aluminium/copper bolted joints".
http://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=04643511
After an initial read, the points I noted are:
- Surface preparation is very important.
- Petroleum jelly is useful.
- Avoid load-spreading washers -- small-area high-pressure is better
than large-area low-pressure.
- A brass interface washer between the Cu and Al can be beneficial.
Note that the paper discusses bus-bar joints using a steel bolt. The
situation where the bolt sits in an Al thread is not covered.
For those who do not have access to IEEE Xplore, below are some extracts.
-
Manfred
*Electrical performance of aluminium/copper bolted joints*
R.L. Jackson, B.Sc. {Eng.), C.Eng., F.I.E.E.
Indexing terms: Power transmission and distribution, Cables and
overhead lines, Bolted joints, Aluminium I copper contacts
Abstract: The 1350 aluminium/copper bolted joint is fundamentally less
secure than the all-aluminium
connection. Differential expansion shearing force at the interface is
considered to be a significant degrading influence. As a consequence,
substantial clamping pressures are needed in addition to having regard
to surface preparation and bolting torques. Although inert greases are
used at the interface to restrict corrosion outdoors, they do not
improve joint stability. A hard plating, such as nickel, on the copper
surface is advantageous, but softer deposits (e.g. thick electrotin or
silver) cause rapid deterioration and are best avoided. An interface
washer is described which will improve the stability of the bimetal
joint and eliminate the need for interface cleaning.
Introduction
Although there has been a considerable increase in the use of
aluminium for busbars and cables, much of the terminal plant
(i.e. transformers, circuit breakers etc.) retains copper
conductors. As a result there are many instances where
aluminium cables have to be connected up to copper tails, thus
creating a dissimilar metal contact. It is well known [1,2] that
special attention has to be paid to the preparation and
assembly of aluminium surfaces to obtain stable low-resistance
aluminium/aluminium joints, whereas copper/copper joints are
much less sensitive in this regard. It is natural, therefore, to
assume the aluminium/copper interface lies between these two
extremes in terms of need for care during assembly, or at most
requiring no more diligence than the all-aluminium
connection. However, experience in service has not alway
supported this view and aluminium/copper joints, particularly
those operating at high temperature, have, on occasion, been a
source of trouble. Indeed many manufacturers of terminals for
aluminium cables consider it desirable to offer a special transition
lug in which the aluminium/copper interface (hereafter
referred to as Al/Cu) is friction- or cold-welded, and the
installer is simply left with making a copper/copper joint. ...
4 Degradation of aluminium/copper joints
It is evident that because the degradation of this joint is more
severe and rapid than the all-aluminium connection, despite
good surface preparation, then one or more of the mechanisms
proposed for contact degradation [1] is enhanced and/or there
is another degrading influence which occurs with the dissimilar
metal joint. Of the existing mechanisms, we need to consider
if direct oxidation can cause erosion of contact spots. Of the
remaining factors, both the creep relaxation and the differential
expansion which occurs with steel-bolt clamping are less
with the bimetallic joint and, as a result, the residual clamping
load will be higher. By implication, the resistance to
external mechanical influences will also be improved so that
there is a strong inference of another factor affecting performance.
This has already been suggested by Bond [3], where
evidence is presented to show that interfacial shear forces
caused by differences in thermal expansion of aluminium and
copper cause significant degradation.
4.1 Direct oxidation
Whereas the natural oxidation of aluminium is independent of
temperature up to 450°C, the growth of copper-oxide films
is very much affected by temperature [4]. ...
Provided, however, we limit our attention to
contact spot temperatures of around 100°C or less, the oxide
film on copper will grow to only about 15-20nm thickness
which is too thin to significantly erode contact spots between
well prepared surfaces. ... For the normal conditions,
however, the evidence suggests that copper contacts operating
at or below 100°C are not susceptible to progressive oxidation
attack since the film formed is self-limiting in much the same
way as aluminium. On this basis it seems unlikely that the
Al/Cu joint is any more affected by direct erosion of the contact
spots than the all-aluminium or all-copper joint.
4.2 Interfacial shear forces
The thermal expansion coefficient of aluminium is 1.36 times
that of copper. Thus, when an assembly such as that shown in
Fig. 3 is clamped at ambient temperature and then heated by
the passage of current, there will be a tendency for the
aluminium to expand relative to the copper causing movement
at the interface (as shown by the arrows). If the frictionresisting
force at the interface is sufficiently high this movement
will be slight. Movement causing a significant number
of contact spot ruptures will thus depend on the degree of
clamping, the temperature rise and the contact spot adhesion.
... there is the evidence for degradation of both
dry and greased aluminium/copper contacts at high cycling
temperature and low clamping load. The latter is an important
proviso as interfacial movement, altering the geometry of spots
in contact, is necessary to sustain the degradation. The use of
inert grease in the aluminium/copper joint is seen as helpful
in an outdoor environment where galvanic corrosion can
occur, but it does nothing to reduce the inherent sensitivity of
this interface to low clamping load and high operating temperature.
This is in contrast to the effect of grease in the all-aluminium
connection [1].
5 The role of clamping pressure
We have seen that, despite good surface-preparation
techniques, the 1350-grade aluminium/copper joint is less
stable than the all-aluminium connection. ...
it was found that it was important to ensure adequate
clamping load and, apart from a miniumum-pressure requirement,
the only effect of the pressure was in influencing the
creep relaxation of the assembly. ... What was proposed in the
paper of Reference 1 was the use of thick large-diameter
washers to ensure that excessive pressures on the bulk material
were avoided. ... The large-area low-pressure sample exhibits
more instability than the small-area high-pressure sample. This
result suggests that clamping pressure is important when
dissimilar metal joints are involved. ... There is little doubt however that
this bimetallic interface is potentially more unstable than the
autogenous joint, and this has led the writer to consider
further ideas for improvement — Section 7.
6 Aluminium in contact with plated copper
... It can be concluded that platings such as tin and silver on a
substrate of copper are not an aid to good electrical joints
when in contact with 1350 aluminium busbar. The initial
resistances are variable owing to the inability of the soft
platings to crack the aluminium oxide, and the poor
mechanical properties of the resulting contacts cause
interfacial rupture due to differential thermal expansion. In
contrast, the nickel-plated copper is a significant improvement
on the plain copper surface.
7 Improving the performance of the Al/Cu joint
It has been shown that plating a copper surface with nickel is a
way of improving the performance of Al/Cu joints. This, allied
to diligent surface preparation of the aluminium and adequate
clamping load and pressure on the interface, should ensure
low-resistance, stable contact between these metals. An
alternative approach is to consider the use of a suitable
transition material inserted between the mating faces, and
which is electrically compatible to both the parent metals. ...
Since various reports [9, 10] have shown that brass is more
stable than copper when in contact with aluminium, it is conceivable
that a washer or plate made of brass could be used as
a transition material. ... A 60/40 leaded-brass busbar (BS
2974, Czl21) was used, bolted up to 1350-grade aluminium,
both surfaces being well abraded prior to assembly. The results
of current cycling these joints are illustrated in Table 3, which
lists two typical results from each batch tested. Results on the
Al/Cu joint are included for comparison. There is clearly an
improvement in performance and further experiments of the
type shown in Fig. 4 show none of the erratic resistance
changes ascribed to differential expansion forces. The performance
with petroleum jelly is again superior to the results
obtained with the aluminium/copper interface, which could
imply that the poor performance of the latter is associated
with mechanical effects (e.g. reduced friction) due to the
differential shear forces. It is important to note that, although
the aluminium to 60/40 brass contact, tested here, is superior
to aluminium/copper contact, the results were obtained with
good abrasion of the mating surfaces, and contact loads and
pressures consistent with what is now known to be desirable.
Several factors could contribute to the superior performance
of the aluminium/brass contact. Perhaps, most importantly,
the thermal expansion coefficient of the 60/40 brass is nearer
to aluminium (Al = 23 x 10"6/degC, 60/40 brass = 20 x 10~6/
degC, copper = 17 x 10 "*/degC) so that the differential
shearing force at the interface of the aluminium/brass contact
is only half that for the Al/Cu contact. Improved creep resistance
of the brass asperites my also be helpful and there is the
possibility that the changed nature of the oxide film, from
Cu2O for copper to ZnO for zinc rich brass, could render the
assembly less susceptible to oxidation attack following slight
interfacial movement. In contrast to the very satisfactory
results obtained with plain brass, tests using electrontinned
(l\ nm deposit) brass snowed high and variable initial contact
resistances similar to electrotin on copper. In consequence,
despite reduced thermal expansion forces, high temperatures
were produced and continuous degradation due to the soft
tin. ...
8 Conclusions
The 1350-grade aluminium/copper joint operating at
temperatures approaching 100°C is even more sensitive to surface
preparation and clamping force than the all-aluminium
joint. An important contribution to the enhanced degradation
of the bimetal interface is attributable to the shearing forces of
differential thermal expansion which are shown to depend on
area. Load-spreading plates are thus disadvantageous, but the
washers proposed for the aluminium-aluminium joint are
effective provided the overlap area is comparable to the washer
area. The improved performance of 6101 aluminium alloy in
contact with copper is ascribed to the superior mechanical properties
of the alloy asperities at temperature. Normal thickness
deposits of tin and silver on a copper substrate are helpful for
use in contact with 1350-aluminium.High initial resistances are
produced which increase further under current cycling. These
deposits are therefore to be avoided unless sufficient test data
is available on a particular arrangement to confirm suitability.
Instability of bimetal joints is not always accompanied by a
rise in the joint resistance measured off-load at ambient temperature.
This observation should be borne in mind when
routine checks are made of suspect joints. Nickel-plated
copper is more satisfactory and good stability was obtained
under load-current cycling. A brass insert with profiled surface
offers the possibility of avoiding the need for surface preparation
of 1350-grade aluminium, while at the same time
ensuring that this material can be successfully jointed to
copper and tinned-copper parts.
_______________________________________________
UNSUBSCRIBE: http://www.evdl.org/help/index.html#usub
http://lists.evdl.org/listinfo.cgi/ev-evdl.org
For EV drag racing discussion, please use NEDRA
(http://groups.yahoo.com/group/NEDRA)
