by Peter Yule
tated during snorting and when sailing on the surface, mixing the
oil with salt water, and it was almost impossible to avoid drawing
salt water into the engines.5
While the submarines’ designers concede that the fuel system
was complicated, they believe that the crews were poorly trained
in its use (training being an ASC responsibility as prime contrac-
tor). In their view the crews tried to operate the fuel system in
the same way as they had on the Oberons, rather than following
the procedures laid down for the new submarines. Olle Holmdahl
saw the main problem as being the crews’ practice of taking fuel
from any of the tanks rather than following the recommended
sequence. They were meant to keep the valves shut unless there
were exceptional circumstances, but Holmdahl believes they rou-
tinely opened them.6
Eoin Asker and others connected with the project office tend to
take a neutral view, conceding that the system was unnecessarily
complicated, but also seeing poor crew training as exacerbating
the problems. The problem of salt water entering the engines grad-
ually lessened as the crews became used to the fuel system, but it
was not until the system was supplemented with a navy-supplied
fuel coalescer in 1999 that the problem was overcome.
Crew members bridle at the claim that they were responsible
for the failures of the diesel engines. Peter Sinclair says that the
224
T H E C O L L I N S C L A S S S U B M A R I N E S T O R Y
crews were trained to follow procedures to the letter and ‘on the
odd occasion [they] made mistakes but the problems that continu-
ally occurred were due to design or manufacturing faults and any
engineer or manufacturer that blamed the crews invariably knew
he had problems’.
The problems with the fuel system were exacerbated in the mid-
1990s by persistent contamination of the fuel supplied to ASC by
the navy. Throughout 1995 and 1996 the project office regularly
reported that Collins’ trials were delayed by bacterial contam-
ination in the fuel. This caused many problems in the engines,
the most serious being damage to the fuel pumps. Swedish diesel
engine consultant Olle Person recalls that he was asked by Hans
Ohff to investigate the problems with the fuel pumps. After dis-
cussions with the sub-contractors he decided that it was not a
manufacturing fault but caused by a bacillus living in the fuel.
Some products of the bacilli were corrosive and this caused the
fuel pumps to stick. They had never had bacteria in diesel fuel
in Sweden so this problem was new to them. The problem was
eventually controlled by the addition of a biocide to the fuel.
Salt water and fuel contamination were the most serious prob-
lems with the diesel engines and lay behind many of the other
difficulties, which included cracked or broken parts and exces-
sive fuel consumption. However, there were other factors which
are widely regarded as contributing to the failures of the engines.
Quite early in the design phase, the decision was made to take the
500-kilogram flywheel off the engines to save weight. Experts are
divided over the effect this had, but many think that this made
the engines less reliable and, by changing the natural frequency of
the engine causing it to vibrate at its specified nominal operating
revolutions, led to problems of cracking and breakages.7
Greg Stuart, however, traces the excessive vibration to the fuel
problems. He notes that:
A number of studies were undertaken that showed the
removal of the flywheel had no real impact on vibration. The
proof of the problem was in fuel pump usage. The fuel
pumps and injectors were a standard Bosch design. The usage
and repair rate exhausted stocks of replacement pumps. The
reason for replacement was corrosion. The corrosion had two
sources; salt water and acid from bacteria. When fuel pumps
‘ T H E Y W E R E P R O B L E M S W E D I D N ’ T E X P E C T ’
225
and injectors corrode they can do two things, the amount of
fuel injected into the cylinder can vary and the point of
injection can vary. Both of these effects cause irregular firing
pressures which in turn causes destructive vibration. [This
was] destructive to the diesel and all engine driven
components including generator couplings and gear trains.8
Excessive fuel consumption first showed up when Collins had its
first full-power snort trial in May 1996, and meant that the sub-
marine could not carry enough fuel to meet its endurance specifi-
cation. Olle Person investigated this problem for Hedemora and
concluded that the main cause was a manufacturing problem with
the turbines leading to inefficiency in the operation of the turbo-
chargers.
The problems with the diesels led many to question the choice
of Hedemora engines. Hans Ohff thinks that the Hedemora sub-
marine engine was badly designed and suffered from numerous
manufacturing defects. He believes Hedemora should never have
been involved in the project because it was a small and declin-
ing company that lacked the resources to develop and support
submarine engines of the size required for Collins, or to remedy
any defects. Ohff believes that most of the problems with the sub-
marines were greatly exaggerated, but the diesels were – and still
are – a genuine weakness, and he was angry that ‘the navy and
DAO didn’t put their feet down and say “this is not the engine we
want”’.
On the other hand, Greg Stuart and others think that the
Hedemora engines got a bad name because ‘we were trying to
run them on salt water rather than diesel fuel’ and once the water
separation was modified they improved greatly. Stuart concedes
that the Hedemora ‘is a draught horse not a thoroughbred’, but
he and other defenders of the selection of Hedemora point to the
design advantage of being able to fit three abreast, the modular
construction, which allows for easier servicing, and the fact that
the engines ran with turbo-charged compressors driven directly
by the exhaust gases.
The diesel engines were a problem from the time Collins set off
down the Port River for the first time, but noise only became an
issue from the middle of 1996. Of all the problems the submarines
encountered, this was probably the most unexpected. Swedish
226
T H E C O L L I N S C L A S S S U B M A R I N E S T O R Y
submarines were extremely quiet and there was always the expec-
tation that the Australian submarines would have the same char-
acteristics. Further, the Swedish design had been assessed in 1987
by the Australian evaluation teams as meeting the noise require-
ments. In the Oberons the main noise concerns came from internal
machinery, and consequently the Swedish technique of isolating
all machinery from the pressure hull by mounting it on platforms
was attractive for its potential to reduce machinery noise
to an
absolute minimum.
During the Cold War the Swedish submarine force was
designed primarily to sit off the coast to attack a Soviet inva-
sion fleet. The submarines did not have to cover great distances
or run at high speed, so they were designed to be virtually silent
at low speeds, in the ‘quiet patrol state’. In contrast, Australian
submarines have long distances to travel to their operating areas
and they want to do that as quickly as is possible while remaining
undetected.
While the requirements for range and endurance were clearly
set out in the contract, the requirements for noise were less clear.
This is indicated by the fact that there were bitter arguments at
the time over what was actually required, and even today there is
nothing approaching agreement on what the noise requirements
really were.
There is, however, general agreement that the original noise
requirements and the way the requirements were expressed in the
contract lay at the heart of the disputes over noise. Andy Millar
suggests that the original specifications were vague because of a
lack of technical understanding of noise issues in the Australian
navy in the early 1980s. The Oberons were quiet, and when the
specifications for the new submarines were being prepared it was
decided to ask for them to be ‘twice as quiet’, even though it
was not known whether this was achievable. Millar suspects that
Kockums was uncertain whether the requirement was achievable
but assumed it would get close.9
However, even if the submarines had completely met the noise
specifications set down in the 1987 contract, this would no longer
satisfy the navy because expectations had grown. The contractual
noise requirements concentrated on noise levels at quiet patrol
state and when snorting, but were vague on noise levels at high
speed, yet by the mid-1990s the navy saw an increasing role for the
‘ T H E Y W E R E P R O B L E M S W E D I D N ’ T E X P E C T ’
227
submarines in directly supporting surface operations and wanted
them to be quieter when travelling fast.10 Eoin Asker, an expe-
rienced submariner who was involved with the project for many
years, says that:
The operators wanted to go faster and with less noise than
the contract specified. They wanted to go at eight knots with
a noise signature that was contracted for four knots but at
this speed they made more noise and it was this that gave the
perception that the submarines were noisy. The submarines
got very close to meeting the contract – what changed was
the operational requirement.
Greg Stuart emphasises that the noise requirements specified in
the contract were not changed, nor was the contractor asked
to provide more performance than specified – ‘the operators
may have wanted more but that did not flow to the contracted
requirement’.
The confusion about the contract requirements and the navy’s
expectations was exacerbated by the difficulties of actually mea-
suring the submarines’ noise. Ideally submarine noise is measured
by a noise range in a deep fjord where the background noise of
the sea is low, but Australia has few, if any, suitable sites, and
none near the trials area off Port Lincoln. However, under the
contract the navy was to provide a noise range and in late 1993
the DSTO began surveying Spencer Gulf to find a site for a ‘shal-
low underway radiated noise range’, eventually selecting an area
near Thistle Island.
The contractor for the noise range was a small Perth com-
pany called Nautronix, which had been founded in the early
1980s to develop marine signalling techniques based on research
into the acoustic signalling of dolphins and whales. The tradi-
tional noise range suppliers had no technology to measure sub-
marine noise accurately in shallow water with a high level of
background noise, but Nautronix worked out a way to adapt its
technology to do this – although it was found that it was impos-
sible to completely overcome the difficulties imposed by an inher-
ently unsuitable site.11
The first inkling that the new submarines might have noise
problems did not emerge until June 1996, when the project office
228
T H E C O L L I N S C L A S S S U B M A R I N E S T O R Y
expressed concerns with Collins’ noise levels. Six months later it
reported:
Investigation of problems relating to radiated noise are
continuing . . . The opportunity was taken to put Farncomb
over the [noise] range during December. Although defects
at the range prevented an accurate set of results, the data
collected suggested that Farncomb has a similar noise signature
to Collins, suggesting that noise problems are class related.
As this report indicates, the results of the noise range tests showed
that the submarines were noisier than expected but there were
arguments over whether the tests were accurate enough to estab-
lish whether the submarines reached the contractual requirements.
Hans Ohff agrees that ‘the submarines never met the hugely ambi-
tious specifications for noise’, but argues that ‘it was hard to prove
this because the background noise in the sea is greater than the
noise level specified for the submarines’.
While there is nothing approaching consensus on the noise
levels of the first two submarines during their trials in 1996 and
1997, the ‘median’ view is probably that of Peter Clarke, who
judged that:
The boats did not meet the noise specifications though they
were not as far away as the navy tried to make out. At slow
speed the boats exceeded the contract specifications for noise,
but above seven knots it was iffy and at high speeds it was . . .
over the contract.
Whatever the exact noise levels, there is no doubt that the sub-
marines were not as quiet as had been hoped and expected. What
were the sources of the unexpected noise?
Early in Collins’ trials there were some minor problems with
mechanical noise, notably from the weight compensation pump,
but these were quickly resolved. The main concerns were with
hydrodynamic noise made by the flow of water over the hull, and
noise and cavitation from the propeller. Critics like Mick Dunne
and Bill Owen argue that the design of the submarine is inherently
and irreparably noisy. While this view has been taken up by the
general public, it is not shared by either the submarines’ builders or
their operators. Peter Sinclair, as the first skipper of Collins, bore the brunt of the ‘first of class’ faults, but he is adamant that ‘she
‘ T H E Y W E R E P R O B L E M S W E D I D N ’ T E X P E C T ’
229
is super quiet, really, really super quiet at slow speed, quieter than
anything else in the world’. However, most agree with Sinclair’s
observations that the submarines’ flow noise increased greatly as
their speed increased
and share his view that the shape of the
casing was the main cause.
When the Type 471 was first designed the casing was smooth
and even, being virtually indistinguishable from the Swedish
V ästerg ötland class.12 A one-sixteenth scale model of this design
was extensively tank tested in 1986 during the project definition
study, showing hydrodynamic flows and noise levels nearly iden-
tical to those of the V ästerg ötland.13 However, after 1987 the
designers at Kockums were forced to make changes to this tested
design. The most important of these was the sonar dome in the
bow, which Kockums wanted to place low down, but the project
office insisted that it be high to minimise the ‘blind’ area behind
the submarine. The initial plans had a low bow but the changes
resulted in a large and bulbous bow that is generally believed to
be the major cause of turbulence and noise.14
The changes to the bow were made with what appears to have
been remarkably little consideration of the consequences for water
flow. Although the issue was raised by members of the navy team
in Sweden, they were told money was not available for a new
model or test program.15 Consequently, the revised design with
the larger bow was not tank tested. This reflects an early failure
of communication between Kockums, the project office and the
Australian navy. It was not until after tank testing and air flow
analysis was carried out in the late 1990s that some relatively
simple modifications were made to the casing that appear to have
reduced the flow noise.
The other noise-related concern was the cavitation from the
propeller. Peter Sinclair recalls that during Collins’ trials they gradually became aware that at some speeds there was no cavitation,
but at other speeds the water flow over the control surfaces onto
the propeller did cause cavitation. Although the propellers and
cavitation were hardly mentioned in the project office’s reports
of the early trials, by 1997 they were among the major concerns.
The chief of the navy, Don Chalmers, and American submarine
expert, Admiral Phil Davis, both saw excessive cavitation as mak-
ing the submarines unfit for combat. Don Chalmers recalls that
the contract did not specify cavitation levels but:
230
T H E C O L L I N S C L A S S S U B M A R I N E S T O R Y
What it said about cavitation was along the lines of ‘the boat
