A rail odyssey
64The journey of the train that collided with Mr. Eldridge's truck at Illabo commenced at Nyngan on 1st November 2006 at 1 p.m. The train was carrying a load of wheat to be shipped by sea from Inner Harbour, Port Kembla. This first stage of the journey was allocated train number 8952. The route to Inner Harbour was via Parkes, Stockinbingal and Goulburn. On the way it passed through a number of different train control regions, some more than once, commencing with Orange, Junee South, and finally Metro when it entered the Port Kembla region. Four different crews in succession manned the train. Messrs. Bell and Perry manned the locomotive from Stockinbingal to Goulburn on this inward leg of its journey.
65As I have said above, prior to the commencement of this journey, GrainCorp provided ARTC with its train consist. The relevant extract from the train consist for this inward leg may be found at page 69 of Ex. S3. Once again, as has been pointed out, the consist complies with GrainCorp's obligations under its access agreement with ARTC. I repeat, it provides in respect of each of the two locomotives three phone numbers, being the CountryNet number, the "on train telephone", and GrainCorp's 24-hour number. There are therefore four phone numbers in all. Moreover, the consist showed GL102 (OT60) as the leading locomotive. A footnote said:
This consist is correct after train reverses at Narromine.
At the point of origin GL101 (OT59) was the leading locomotive, but it was necessary for the order of locomotives to reverse at Narromine. This necessity is illustrated on page 77-8 of Ex. S3. Negotiation of the junction of the Nyngan-Parkes lines requires the locomotives to be uncoupled from the wagons at Narromine. The locomotives then continue past points that connect with a parallel line. The crew leave GL101 and enter GL102, which is then driven, in reverse of the previous order, through the points into the parallel line, past the stationary wagons, and through a second set of points back on to the main line. The locomotives are driven in further reverse and re-coupled with the wagons at what was formerly the back-end of the train. GL102 then departs as the leading locomotive.
66The train crew from Nyngan to Parkes were Mr Glenn Pollock and Mr Steve Kennedy, neither of whom were called by any party to give evidence. The call data collated by Mr Aitken in Ex. S6 demonstrates that at Nyngan the driver logged on to CountryNet, but completed registration by telephoning Orange train control via the CDMA phone in accordance with what seems to have been the preferred practice. It should be pointed out, as Ex. S7 makes clear, that the CountryNet logs produced to the Court by ARTC on 30th May 2012, during the trial, covered only 2nd November 2006, and not 1st November 2006. However it is common ground that the CountryNet equipment was "logged on" by the driver and registered by the controller at the commencement of the trip. I infer that when the reconfiguration of the train was completed at Narromine and the crew left GL101, neither of them turned off the CountryNet radio. This meant that this unit remained registered on the CountryNet train controller's workstation as the train moved through the various control regions. This was a mistake which probably had a "masking" effect, and to that extent contributed to the failure of Junee Control to warn 9351 of the danger at the Olympic Highway level crossing at Illabo, as I said at [36] above.
67Between Narromine and Stockinbingal where the plaintiffs joined the train, two more calls were made from the CDMA satellite phone: one at 7:14pm and the other at 7:21pm. Both were of short duration.
68As there is no polling data available for 1st November 2006, there is no independent confirmation that when GL102 (OT60) became the lead locomotive, the driver logged the CountryNet radio in that locomotive on. On the evidence I have, there is no record of a CDMA telephone call at the time the train left Narromine (at 4:20pm on 1st November 2006). There was a change of crew at Parkes. Mr James Martin and Mr Michael Tucker manned the train from there to Stockinbingal. Neither of them was called by any party to give evidence. They were the crew responsible for the two calls I have mentioned.
69The plaintiffs joined the train at Stockinbingal at 10:20pm on the same date. The evidence of each of them is that the amber light was emitting the characteristic double flash and a familiar "chirping" sound was heard during the handover from one control region to another. This occurs automatically: Ex. A1 [32], [66] - [75]; Ex. F1 [58], [65], [67] - [79].
70It is convenient to interrupt this narrative to deal now with the question of whether the CountryNet equipment in GL102 was on when the plaintiffs joined it at Stockinbingal; that is to say, whether after the reconfiguration that equipment had been logged on. There is no evidence that it was registered because no witness was called from either ARTC or GrainCorp relevant to whether GL102 became registered at Narromine, which was subject to train control out of Orange. However I am satisfied that the CountryNet radio in GL102 (OT60) was logged on when the plaintiffs joined the train at Stockinbingal, when Mr Bell rejoined it at Inner Harbour for the return trip, and when Mr Perry rejoined Mr Bell on the train at Goulbourn also on the return trip. In making this finding I record that I accept the evidence of each of Mr Perry and Mr Bell contained in the evidential statements to which I have referred, and their oral evidence under cross-examination in this regard which was unshaken.
71Mr Perry was cross-examined to suggest, essentially, that he may have been mistaken about the CountryNet radio being on. At 56.15T he said:
Can I say to you with them radios, when you changed territory, say we are going from Junee control area into Port Kembla control area, the radio makes a dominant beep like you know it has changed. It has gone from one to another. And when we come through Cullerin, which that normally happens between Yass and Goulburn, you can hear the radio change over. On the way home it done the same thing in the same area, so you know the radio is working. So I don't have any problem with the radio being working.
72Mr Perry denied ever experiencing any difficulty with the CountryNet radio working, except with voice echoing (66.15T). He also gave evidence that Mr Bell used the CountryNet phone to make an emergency call after the collision. The documentary evidence analysed by Mr Aitken in exhibits S6 and S7 confirms this call.
73Mr Perry confirmed that the logging on procedure was only required at the start of the trip "when you start up an engine or start the beginning of a journey. Not half way through it" (92.35T). He also said, although he was not on the train at Port Kembla on this trip, "they would not let you leave Port Kembla harbour until you were contacted on that network" (91.30T). In his experience, it was only necessary for a driver to log on, in this case, at Narromine, "because they've swapped the engines around". And leaving Port Kembla "because the train number changed" (92.40T). This evidence in substance is the same as the evidence of Mr. Baczynskyj I have referred to above. Mr Perry said that on the return neither he nor Mr Bell "touched" the CountryNet radio. By this I understood him to mean that neither of them did anything to change its status in any way. Mr Perry also explained, although in error, that because the NEC handset operated both the CountryNet radio and the WB radio, using the CDMA phone to contact train control when necessary left the WB line open to receive calls from signal boxes (104.25T). His error does not undermine this evidence of the preference for using the CDMA phone. Mr Aitken's logs confirm the employment of the practice on these trips.
74At 60.20T the following exchange occurs:
Q. What I want to suggest to you Mr Perry is that at least for a period of time before the accident occurred that the radio in the lead locomotive may not have been switched on?
A. I am telling you it was.
75In his evidence, Mr Bell said that when previously he had driven a train that was reconfigured at Narromine, as all trains heading for Parkes must be, he would turn off the CountryNet radio in what had become the trailing locomotive after the reconfiguration, by turning "the circuit breaker off in the cabin". He clarified that by this he meant turning the power off. He also said that it was part of the procedure to switch on the CountryNet radio when you entered "what was now going to be the lead locomotive" (177.30T). When this was done, that unit would be logged on by pressing the "P1 button" (178.5T). Mr Bell said at 178.20T that:
[P1] registers to control that I wish to speak to him and then he would come on and speak to me and then I can just talk to him.
He said he would not leave Narromine until he received contact from train control.
76Returning to the narrative, Mr Bell and Mr Perry left the train at Goulbourn, but because of an unexpected need to change crew Mr Bell was sent by road to Port Kembla to rejoin the train as the observer on its return trip.
77The wheat was unloaded at Port Kembla before Mr Bell rejoined the train, and the registration of 8952 was manually cancelled at 3:03pm. A call was made using the CDMA phone at the same time, I infer to the Metro train controller, who at the conclusion of the call or during it cancelled the registration by touching the relevant icons on his train controller workstation. This requires the controller to select the "CANCEL REG" button, and the relevant train button. But "the locomotive radio is still logged on and calls via the 'UNREG LOCO' button are available": Ex. 2D11 at p34.
78As I have already said, the unladen train was re-designated 9351 for the return trip for "stabling" at Junee. It did not depart Inner Harbour until 4:20pm. There is no record of a CDMA phone call to train control or any other contact with train control logging on and registering for the commencement of the new journey. However, although the phone call cancelling the registration earlier had been made from the CDMA mobile phone which was then situated in the leading locomotive GL102 (OT60), because of the confusion engendered by the continued operation of the CountryNet equipment in the trailing locomotive GL101 (OT59), it seems clear, and I infer, that what was cancelled, of course by computer, was the registration of the CountryNet equipment in the trailing locomotive. But it must follow that that equipment remained logged on because no one had turned off the power by operating the circuit breaker. And I would also infer that the CountryNet radio in the lead locomotive also remained on.
79Clearly the crew who had finished the trip at Inner Harbour believed that CountryNet was on because the driver took steps to cancel the registration. There is also a presumption of continuity in my mind in that I accept the evidence of the plaintiffs that it was showing all signs of being on, right up until they left the train at Goulbourn. I find it was on at that time. As I have said, Mr Perry gave evidence about the normal practice at Port Kembla. Mr Bell gave evidence to the same effect, although he could not recall observing the driver, a Mr Edwards, using the CountryNet radio to contact train control. And there is no log of that having occurred, the logs being available for 2nd November 2006.
80Mr Bell said it was necessary to make verbal contact with train control responsible for Inner Harbour (Metro train control) to obtain permission to leave. That is the normal practice and Mr Bell assumed that Mr Edwards had observed it (185.40T). Mr Bell's assumption is not evidence, but his evidence and that of Mr Perry about the usual practice is circumstantial evidence that Mr Edwards followed the usual procedure. I infer 9351 could not leave somewhere like Inner Harbour unless it had "permission". Presumably a working harbour is a large complex, leaving it may require favourable signals and the opening of points. The evidence discloses that neither Mr Bell nor Mr Perry were qualified to drive at Inner Harbour, although they could act as observers. This suggests to me that Mr Edwards had higher qualifications and experience. It seems very unlikely that someone with his expertise could, or would, ignore proper procedures with impunity. And there is no suggestion in the evidence that the departure from Inner Harbour was somehow irregular.
81Mr Bell said that the practice was to avoid using the satellite function on the CountryNet handset. He described this as a cost issue. He confirmed it was preferable to use the WB radio function. At 190.35T, he said that he would use the WB radio function wherever possible to make whatever communication he had to. He confirmed that the CDMA telephone was the preferred method of contacting train control (192.10T).
82That there is no CDMA phone record or CountryNet log, or poll, of a communication, it must be acknowledged, is some evidence that the normal practice may not have been followed. Mr Bell thought the relevant contact could not be made by WB radio because permission had to come from the yardmaster: 259.20-25T, which may have been obtained by the personal attendance of Mr Edwards.
83Mr Bell denied that he could be mistaken about whether the CountryNet radio was turned on as the train approached Illabo (196.5-40T). I accept his evidence. If it was on at that stage, it follows that it must have been on when the train left Inner Harbour, given what I have said already about the improbability of the train departing Inner Harbour, in the hands of a senior driver (in terms of his superior qualifications and experience) irregularly. To this might be added the additional consideration that it is highly improbable that all of eight qualified freight train drivers, each of whom would have taken turns at serving as driver and observer during their shift in the cab, must have completely disregarded proper operating and safe working procedure with regard to CountryNet. One's experience of life suggests perhaps that one or two of them may have in some respect, such as the crew at Narromine failing to turn off the CountryNet unit in GL101, but not all of them in all respects.
84Along the same lines, to my way of thinking, is the evidence that after the collision, which derailed the locomotives and some of the wagons and severely threw the crew around in the cab, Mr Bell and Mr Perry notwithstanding the trauma of the collision performed their duties in the most professional manner. Mr Perry spoke to train control on the CDMA mobile phone, and Mr Bell using the CountryNet radio, at the same time, as a satellite phone, attempted to call GrainCorp. Mr Perry walked back and beyond the crossing putting out detonators behind the train to warn approaching trains of the obstruction. Mr Bell did likewise ahead of the train. He put out detonators and a required warning light for trains approaching from the opposite direction (see GrainCorp statement, Annexure A to Ex. F1).
85The "GrainCorp statement" made by Mr Bell on 9th November 2006 contains the following narrative (Ex. F1, Annexure A, p13):
A normal journey was experienced from Goulbourn to Harden. At Cullerin I noted that the CountryNet radio sounded several pips as it normally does at this location which indicated to me that the radio was working. Throughout the journey I observed the double flashing lights of the CountryNet radio. At about 21.50 hours at Harden William Perry telephoned Junee Train Control on the CDMA telephone to find out what time we were expected to arrive at Junee so that he could make arrangements with Thomas Martin and James Martin to sign on duty and relieve us at Junee. He also advised Junee Train Control what working was required at Junee and then rang Junee South Box to advise that our train was going to be placed to Junee Railway Workshops.
That telephone call appears on the log prepared by Mr Aitken, being Ex. S6 (referred to further below). It was of 1 minute and 41 seconds duration, to Junee Control. I should point out that in the same statement Mr Bell says:
William Perry told me to get on the train working radio and call out 'emergency'.
I take this to be a reference to the WB radio. I will refer again to some of these details below.
86A difficulty which arises on the evidence is that Mr Foley, the train controller, gave evidence that he "[felt] there was only one [icon]" for 9351 on his screen when he commenced his shift. Leaving to one side the uncertainty in that response, it is argued if there was only one icon, and there was no dispute that the icon related to the unmanned GL101, then the manned GL102 (OT60) must never have been initiated (see Foley 403.35-404.20T). Moreover, the consideration that, as Ex. S6 demonstrates, GL101 continued to be polled from the time it left Inner Harbour (variously as 'UnReg' or OT59) until it was unilaterally registered without contact at 7:24pm, and GL102 (OT60) did not, is said to require the conclusion that GL102's CountryNet radio must have been off, and therefore not connected to the system. It is true enough that the experts could not explain what does appear to be an anomaly. And OT59 polling before train 9351 was registered, is taken as confirmation of the undisputed fact that after leaving Inner Harbour its radio was initialised and continued to be logged on after the previous train number registration had been cancelled. However that it was polling on the CountryNet system does not necessarily mean it was always displaying an icon on the appropriate train controller workstation prior to its registration as 9351. The log attached to Ex. S6 shows that its registration occurred at a changeover location and certainly at that point an icon in respect of it must have been displayed to enable registration by the train controller.
87At [54] above, I have set out the contents of the manual for answering an incoming call from an unregistered train. The icon represents the incoming call not the train per se. In Ex. S3, Mr Aitken said (at p52 [5.2.2]):
While the unregistered radio is accessible to the train controller, it is usually not shown on the train controller workstation. It will be shown if the locomotive has made a call to the train controller and the train controller has not registered the locomotive. It will also be shown when the train moves into a new train control area. (Emphasis added).
88This is confirmed, to my mind, by the contents of the manual at page 31 as follows:
Initiating call - Unregistered train
To initiate a call to an unregistered train:
- select UNREG LOCO button twice
A QWERTY keyboard is displayed.
- enter the locomotive identification number
- select OKAY to place the call.
The TCWS will attempt to call the unregistered train.
I find on the probabilities that an unregistered, but logged on or initiated, CountryNet radio will not constantly display as an icon on the train controller workstation. It will remain connected and can be found by the computer system to which the train controller workstation is attached (see Annexure A, Ex. S8), and it ought to continue to participate in the global positioning system. It is worth recording that there was no further changeover between Mr Foley's commencement on shift at 10:30pm and the collision at just before 11:25pm. Given that GL101 had been erroneously registered as the train by a previous train controller, that is the only icon that one would expect to be displayed continuously on the screen or to be displayed at all during the time the train was under Mr Foley's control, unless a call was initiated by the train by one of the crew depressing the "P1" button, which we know did not occur.
89This analysis does not deal with all anomalies. But they have in large measure, with a degree of irony, been generated by the late production of part of the polling data only. Bearing in mind the application of the civil onus I find it permissible for me in the circumstances to apply the Blatch v Archer maxim to my reasoning approach, leading me to make the finding expressed in [96].
90According to Ex. 2D11, the CountryNet users manual, there is a distinction between the "UnReg Loco" button (or icon) always present on the screen as part of the bar along its bottom, and an "UnReg train" button or icon which appears when a call is received from an unregistered train.
91The masking effect of the continued registration of the CountryNet unit in the unmanned GL101 (OT59) all the way to Inner Harbour is capable of accounting for train control not picking up the anomaly to that point; that together with the considerations, first that voice communication was rarely required after registration, and, secondly, the permissible practice of communication over the CDMA phone. But to my mind this masking effect does not displace what I would regard as the high probability of regularity in operation arising from the force of these other circumstances, which, in my judgment, is not displaced by the competing consideration of the lingering irregularity of the registration of the unmanned trailing locomotive. This consideration supports the finding that the CountryNet radio in GL 102 was logged on.
92There are other matters which support this conclusion. I have already indicated that Mr Aitken compiled a log of the communications from 9351 in Ex. S6, with additional analysis in Ex. S7, drawing upon the various contemporary documents recording the trail of communication emanating from the train. That objective documentation in significant ways supports the conclusion I have reached on the basis of the evidence I have discussed so far that the CountryNet radio in GL102 (OT60) was logged on at all material times including during the return trip. As the CountryNet system is largely automated through satellite and computer functions, not requiring much by way of human input, the communications between the train equipment and the train control equipment are recorded or "polled". Despite an early request for this information, it was not provided by ARTC until the third day of the trial, 30th May 2012. The explanation proffered by Mr Talintyre of counsel, which I accept, which is also supported by the affidavit of his instructing solicitor, Ms Jacqueline Remy, affirmed on 31st May 2012, is that the record was maintained not by ARTC but rather by RailCorp who are, obviously, not a party to the proceedings, and that despite that organisation's goodwill, ARTC were not able, by their own efforts, to have the documents produced as expeditiously as they would have wished. When they were produced, only records for 2nd November 2006 were available. The limited nature of the material hindered the ability of Mr Aitken to provide complete analysis (see Ex. S6 [1.1], and Ex. S7 [3.1]). In assessing these proofs, therefore, I have borne in mind the principle discussed in Blatch v Archer (1774) 1 Cowp 63; 98 ER 969 at 970 that:
It is certainly a maxim that all evidence is to be weighed according to the proof which it was in the power of one side to have produced, and in the power of the other to have contradicted.
93These additional matters are of significance. First, the material produced by Optus, the satellite operator, and ARTC, show that the equipment on GL102 was operating correctly at 5:37am on 2nd November 2006 (see Ex. S7 [5.2] - 5.3]). After OT59 had automatically changed over from Junee train control to South train control, OT60 is recorded as attempting to perform the same automatic function. Indeed its position was also polled at 5:36:40am. Initially the "changeover" appeared to be successful, but at 5:37:06am its registration is recorded as being "cancelled" at Junee. The evidence does not disclose why the registration was cancelled especially as it must have shown up on the train controller's workstation as an unregistered icon. (See the description above at [54], and [87]-[88]). For these pollings to have been recorded at all, the CountryNet equipment must have been on, and must have been logged on. The most likely explanation for the rejection of registration is human action at the train controller's workstation. Having said that I acknowledge there is force in Mr Talintyre's argument that given the competing demands of a train controller, it is unlikely that in the 35 seconds covered by the polling it would have been noticed by the train controller. But this is not impossible. Indeed, given that its companion locomotive was already registered as 8952 at that time, the cancellation or rejection of OT60's attempts seems a logical response even if it should not have occurred without making voice contact with the locomotive, or at least attempting to. But, I repeat, to my mind the real significance of this is that, however briefly, OT60 showed on the train controllers workstation. The fact that it did proves, to my satisfaction, it was turned on, and logged on. Thus this circumstance provides an important thread in the evidence of continuity.
94The second matter is the call Mr Bell made from the CountryNet unit on GL102 after the collision in an attempt to contact the GrainCorp 24-hour number. In the event the wrong number was dialled. But there is no suggestion in the polling data, as I understand it, that the CountryNet unit was, for the first time, logged on immediately before that call was made.
95Thirdly, there is the important consideration that the CountryNet equipment on GL102 was operating normally on 3rd November 2006: see table 4-3 page 11, Ex. S7.
96I find that the CountryNet equipment in GL102 was logged on, but not registered at all material times after the train left Narromine on 1st November 2006. I find that it remained logged on until the collision. The train driver can do no more than log on. Registration belongs to train control.