I started working at the MFB in December 1983 after leaving Sontron Instruments. The head office was (and still is) at 456 Albert St East Melbourne but the Radio and Electronics Services (RES) department where I worked was located in a cream brick factory/office at 613 Victoria St Abbotsford, right next to the old MFB Training College and just west of the famous Skipping Girl neon sign. RES had been created to provide support for the Brigade's IT systems and Abbotsford had been chosen simply because it had some free space. This building housed a facilities fitout and maintenance department that looked after the firestations and other MFB buildings. It had a full workshop that could fit out a fire station kitchen, build bathrooms and undertake general renovations. The Abbotsford building also housed a fire investigations department plus fire alarm manufacture and maintenance for alarm boards in the fire stations and break-glass alarms and sprinkler alarms used in premises. At that time, the MFB was sole supplier of alarm equipment. The mantra at that time was to do everything "in-house".
I was one of two engineers employed to work on the MFB's new Computer Aided Despatch (CAD) system, called Firecom. The other engineer starting with me was Lester McClure. When we arrived, RES already had electronics support technicians for Firecom. Lester and I were there to provide higher level engineering support.
The MFB had been the first emergency services organisation in Australia to embrace CAD technology when they announced the project in 1977, led by District Officer Alan Richards who, some 23 years later, would become Chief Fire Officer. The requirement specification for the Firecom project took a couple of years to be developed to a point where Philips Telecommunications in Clayton Victoria was awarded the main contract to design and supply the system, on condition that Philips subcontracted the equipment to be installed in the fire stations to a local company called EDL Australia, insisted upon by the then MFB Board President (who I was told had personal connections with the company).
The CAD consisted of a six seat call centre and a main computer room at Eastern Hill with corresponding equipment in each of the fire stations, unimaginatively called the FSE (fire station equipment). Emergency Triple Zero calls were taken by call takers in the call centre, with all call details entered into the CAD computers, including the caller's description and location of the incident. The CAD would verify the location to a map reference then apply a predetermined set of rules based on the call details to determine the type and number of the required appliances and their corresponding fire stations. The system could have easily gone on to despatch the firestations automatically but one of the key design principles was that the despatch would be presented to an operator for review and to potentially make changes based on other information not available to the CAD computers.
On despatch, details would be sent by the CAD to the FSE in the required fire stations via dedicated telephone lines and printed out automatically. In addition, an audio channel would be set up to the firestations to sound alert tones and allow the operator to announce the call details to the responding firefighters via the firestation public address (PA) system. The interior station lights would also be turned on. Later on, the FSE would also trigger station control functions to open the front doors, turn off power to the kitchen, turn on the engine bay exhaust fans and even change traffic lights in front of the fire station.
A further major component of Firecom was to modernise the alarm system. Under the then existing system, fire alarms in premises were connected via a telephone line to an electro-mechanical alarm panel in the watchroom at the closest fire station. When an alarm was received, that fire station turned out to the incident and the remaining watchroom firefighter would notify the communications centre who would then determine additional appliances and turn them out by telephone. This could lead to significant response delays. In the new system, alarms were transferred to the FSE which automatically passed them on to the Firecom computers as Eastern Hill. Firecom would then find the right resources and despatch them together, all within a few seconds of the alarm occurring. This capability was vital to meet new response standards for the MFB.
By 1983, Philips and EDL had implemented the project to a point where Firecom was ready to go online for call taking and despatch, though the alarm component was still not ready. The system went live in October 1983, but it was soon apparent that there were major reliability problems with the FSEs, which would go off-line almost as fast as MFB technicians could restart them. This was the impetus to employ Lester and I.
Lester was a few years older than me and was definitely my senior professionally. He had extensive experience in microprocessor hardware and software having come from working on traffic light controllers for Victoria's road authority. We quickly determined that the overall quality of the FSE hardware was not up to an acceptable standard. We also had no source code for the complex software and therefore no capability to fix or improve it. The FSE had no less than five microprocessors, two main redundant system processors (or SPs), a control processor (CP) to abitrate between them and two for the dual printers. Apart from the CP, everything was duplicated. The SPs were constantly checking operations and alarm monitoring with each other. For the time, it was quite advanced.
The FSE software was designed to detect errors, arbitrate between the SPs and identify and shut down the faulty SP, so it was important to understand how it worked. At that stage, the MFB did not have software development capability but fortunately, Lester had a well equipped workshop at home so we spent two weeks there disassembling the software, creating working source code and ensuring we could independently re-assemble the software to exactly match the code that had been supplied by Philips. We were then in a position to properly debug and address the problems we were experiencing with the software. Parallel to that, we embarked on rebuilding all the component boards in the FSE to bring them up to an acceptable quality standard.
So went the first few years at the MFB. Once we had got the FSE sorted, we started on the alarm system. Each FSE could monitor over 1000 alarms, requiring an equally unimaginatively named alarm interface unit or AIU at the premises to convert the alarm activation (essentially a switch open) to an electronic signal sent back to the FSE via the existing telephone line. Up to 8 AIUs could happily co-exist on the same line. Philips had supplied a few thousand AIUs that were well designed and built but did not have any magnetic shielding. Many AIUs were installed in places close to magnetic fields strong enough to affect their operation. After some fierce negotiation, Philips agreed that this was a "latent defect" and recalled all the AIUs to add the required shielding - a conductive paint sprayed on the inside of the plastic boxes.
An equally severe problem with the alarm system lay in the design of the alarm receiver in the FSE, called the alarm termination equipment or ATE. The FSE software was designed to scan up to 16 ATE cards installed in the FSE and like the main processors, these were duplicated too. Each card had 8 channels and each channel could have up to 8 AIUs connected, giving a potential total of 1024 AIUs. The entire scan was designed to be completed in just over 1 second, so that any delay in detecting an alarm would never be more than this. However, along with all the other things that the software was trying to do, we quickly discovered that the FSE processors simply didn't have enough time, especially if there were multiple alarm changes to process in the scan. This was a fundamental design flaw.
We made major changes to the software to properly scan the ATEs and over time, added a full suite of diagnostic tools so that the support technicians could fault find in the field. We also rebuilt the ATE hardware. The first live AIUs were installed in October 1984 and were progressively converted over the next year. There were around 7,000 AIUs connected across 47 Fire Stations by the time the last alarm was converted in October 1985.
The FSE printers used rolls of 60mm wide electrostatic coated paper using a 7 by 5 dot matrix to print capital letters, numbers and punctuation. The character was printed by applying a charge from the dot needle to the paper which darkened the paper from light to dark grey. This had the big advantage that no ink was required. All firefighters had to do was change the paper roll. It also had the unplanned advantage that at night, when you put the paper on a lightbox or a small torch, you could clearly read the printout. Firefighters are multi-talented and a number of firestations came up with their own lightboxes. Unfortunately, the printer didn't have any provision at all to contain the paper once it came out of the printer. This was fine for the despatch printer as the paper was always torn off to take for the turnout, but printout from the other printer needed to be collected as a long term record. Most firestations just put a small box on the FSE and manually rolled it up as needed. Others put a simple clamp on the roll and let it dangle over the edge of the FSE where it dropped under its own weight as printouts came through. Once it hit the ground, a firefighter would manually wind it up and refit the clip. A few stations came up with elaborate wind-up roll holders. The electrostatic paper was quite unique and became harder to get over time. The MFB stores bought up huge amounts of it when they could to ensure ongoing supply.
The only part of the original system that was never fully implemented was the front door "passerby" alarm, called the Front Door Fire Alarm or FDFA. This consisted of a break glass fire alarm at the front door for any member of the public "passing by" to signal an alarm if the firestation was unoccupied. Historically, the alarm went to another firestation which would hopefully be occupied and able to respond. With Firecom, this would now go automatically via the station's FSE to the Communications Centre. There was also now to be an ability to open up a voice communication directly from the passerby at the firestation to a calltaker at the Communications Centre. Again, the design was never fully tested and in practice, turned out to be flaky and very poor quality. Rather than try to make this work, we opted to use a standard AIU for the break glass alarm and replace the voice component with a standard telephone on a direct private line that rang at the Communication Centre when the passerby lifted the handset.
The Firecom computers and associated equipment at the Communications Centre were duplicated too. Philips had chosen the Control Data 1700 for the main computers and had engaged some extremely talented programmers from New Zealand company Progeny to create the main software. When the project was handed over, the MFB contracted one of these programmers, Dinh Ho Ta, to maintain and continue to develop the software. Dinh Ho made many improvements and updates, including upgrades of the main computers to Perkin-Elmer Interdata 7/32s, then Concurrent then Digital Vax. He stayed with the MFB over the entire life of the system and made an enormous contribution to the success of Firecom.
Apart from the CAD aspect, the Firecom project was also important in the history of the MFB because females were brought into the brigade for the first time as communication centre operators, pre-dating the first female firefighters by 5 years. The first intake comprised 9 females and 3 males. The operators worked in the communications centre at Eastern Hill and it became usual to see firefighters from the adjoining fire station visiting the centre to say hello. The six consoles in the centre were custom built by Philips. They were generally assigned as two calltakers and two radio operators, a supervisor and an overflow calltaker. Each console had two Intecolor colour terminals worth at the time around $20,000 each. One terminal was used by the operator to interact with the CAD, using two letter commands followed by the appropriate details. This proved to be very fast and efficient. The second terminal was used to display one of the available colour coded status screens - call display, appliance display and call map. The terminal supported 16 colours, 24 or 48 line text and character based block graphics. Due to their cost, only two other terminals were available, one for software development and one as a spare.
Over time, we upgraded equipment in the communications centre. The original analogue tape machines that recorded operator calls were upgraded to digital to allow much easier re-call and transcription for reviews and enquiries. The main computers were upgraded and various capabilities were improved or added. One such upgrade involved the alert tones. In the original design, these came on at a set volume level. Almost as soon as the system went live, we started to receive feedback from firefighters that they were very unhappy with the sudden loud tones coming on in the middle of the night. There were also complaints about the quality of the audio in the fire station for operator turn out announcements. I redesigned the audio system, upgraded the PA and speaker systems in the firestations and added ramping volume to the alert tones. We received very positive feedback from firefighters following these changes.
Another area of work was the operator terminals. As Firecom matured, firefighters started to see the value of having a CAD terminal at their firestations. However, the cost of these terminals was still prohibitive and would have required an outlay of around $1M for 50 fire stations. Personal Computers had started to appear in the MFB as early as 1985 and we had been at the forefront, using the PC for software development and documentation. By comparison, a PC at that time cost around $3000 which was a much better proposition. Terminal emulator software for the PC was available but the Intecolor terminal was relatively obscure and so complex that no emulator supported it. Always up to a challenge, I set to and developed a full featured emulator (mostly out of work hours) and also added extensive help screens and some further capabilities to integrate with the PC and simplify it's use for firefighters. This made it possible to provide PCs in the firestations connected to CAD. When the real terminals in the communications centre reached end of life, we could no longer obtain new terminals so we replaced them with PCs and I added some further capabilities to the emulator to satisfy some outstanding improvement requests from the operators.
The operator consoles were originally layed out as 4 positions with 2 positions behind, all facing in the same direction (towards Albert Street). This was decided based on an expected two calltakers and two radio operators in the front positions with a supervisor and a spare in the rear positions. It was initially expected that radio operators would despatch the firestations but it was soon discovered this was impractical due the time required to alert the station and announce the incident details. This left the radio channels unmonitored, so it was decided to have calltakers despatch the firestations instead.
As call volumes grew, it became more routine to have three or even four calltakers, meaning that the rear consoles were being used more regularly. As a result, a request came through from the Comm Centre Manager to rotate the back pair of consoles to improve the communication between operators. This required a great deal of planning because the two consoles were constructed as a single unit and could not be simply rotated individually. The complex array of cabling needed to be re-routed and re-established in the reverse configuration. I spent several weeks in planning and preparing for the move, constructing and live testing some necessary extension cables, and getting approvals from operations. I then spent most of a weekend with one of the FSE technicians rotating the console unit and testing while operations continued in our backup centre. The planning turned out to be good, there were no surprises and everything went smoothly. Feedback from operators after the move was very positive.
Around 1987, RES was expanded to become the Systems Engineering Group (SEG) and a service centre was established to take faults and service requests from operations for all systems, not only CAD faults but radio, facilities and alarms. To support the centre, a multi-user fault system was needed to record and update faults through to completion and manage the overall operation of the service centre. In the late 1980s, local area networks were pretty new. PCs predominantly used DOS with Windows barely released and neither supported a LAN. After a pretty extensive search, I came across a multi-user service system called Atlas that could run on PCs using the Pick operating system. At the time, this was pretty obscure and it took some convincing to get approval. Technically, Pick was an excellent foundation for a service system. It was inherently designed to support databases at the operating system level and could connect up to eight terminals via a standard multi serial port PC card. This made it possible to provide terminals in the service centre and manager offices within the same building. Atlas had great call managemnt and reporting features and we had access to the source code, written in Pick BASIC, so we could add our own reports and enhance the capabilities as needed.
The year 1987 was also the year that mobile phones were first introduced into Australia by Telecom. I remember going to the Melbourne launch, held in the playboy pad of man-about-town and former racing car driver Peter Jansen at the back of the Rialto Tower on Flinders Lane. The red brick and blue cobble stone building had once been the stables for the old Rialto Hotel. As part of the launch, Peter made a call to the Savoy Hotel in London using the shiny 1G handset that looked very much like a World War 2 walkie talkie. Amazingly, the call went through successfully, leaving the very surprised staffer at the Savoy nonplussed.
In the early 1990s, I had the opportunity to work on an interesting side project involving the Brigade's breathing apparatus or BA as it was called, working with the Senior BA Technician, Alan Smith. Alan had been charged with evaluating tender offerings for replacement BA and came to me about an idea he had to build an artificial lung for testing the sample BA sets. He had hooked up a "lung" cylinder to a "driver" cylinder driven by compressed air and controlled by valves on either side of the driver piston. The lung cylinder was hooked to a dummy head that the BA facemask could be strapped onto. Alan needed to computer control the driver cylinder so that various breathing rates and lung sizes could be simulated.
Borrowing from my earlier experiences at Sontron, I asked Alan to build a rod with holes drilled down its length attached to the driver cylinder. We fitted optocouplers at 90 degrees phase to measure the displacement, speed and direction of the cylinder with computer control of the driver valves. I wrote a control algorithm in 8085 assembler using a parallel port to communicate with the driver valves and read the opto couplers. It took some work to refine and tune the system but eventually we built basic breathing rhythms and a range of work cycles that proved instrumental in evaluating the sample BA sets. Alan used the system for many years to test operational BA sets before and after servicing and to evaluate subsequent tenders.
As the 1990s progressed, we establised internet based networks to all the firestations using microwave radio links. This allowed us to provide firefighter access to all the Brigade systems and the general internet. I was keen to expand the service system to allow firefighters to directly log service requests and check progress, so I modified ATLAS to support access via Netscape, the standard internet browser at the time. This involved installing an additional card to expose the serial ports as internet addresses. This could support many more users than the number of ports because access was only required when a query was made, rather than tying up the port for the duration of a login terminal session. I also enhanced the ATLAS software to service html requests from the browser and return html responses to the browser. These improvements allowed for new requests to be logged, list the the current active requests and view the current status of a given request.
One of the cultural aspects of working for the MFB related to the difference between uniformed and non-uniformed staff. The MFB was a paramiltary organisation and the heavily unionised firefighters were supported extremely well. It was routine for firefighters to easily move around the organisation and get support for interstate conferences or to obtain secondments to like organisations interstate or even overseas. On the other hand, career development for non-uniformed staff was virtually non-existent. You were there to support in whatever capacity you were originally engaged for and you couldn't really expect to progress. Professional development that involved anything other than time was hard to come by and after nearly 15 years in the Brigade, I had only been out of Melbourne in a professional capacity once. After applying for a promotion in SEG and being unsuccessful, I was unexpectedly offered an opportunity to attend a course at the Australian Institute of Police Management in Manly, NSW.
The AIPM course was designed for police but the principles are of course much broader and the Institute staff were always keen to enrol some non-police participants to broaden the range of thinking and ideas. So, on a Sunday afternoon in August 1997 and with a very large amount of trepidation, I arrived at the Manly campus ready to start the three week live in course. There turned out to be about 20 people on the course including police from all the state juristictions, the federal police and New Zealand. The only other non-police attendee was from NSW Fire Brigade. We spent the first week getting to know one another covering a broad range of management topics, interspersed with food breaks of dangerous proportions. As the studies progressed, we each started work on identifying and developing a major managerial initiative for our respective organisations. This was to become the major take home output from the course.
We had a number of high profile identities come to talk to us. I remember David Murray, then CEO of the Commonwealth Bank, talking about organisational transformation at one of our weekly guest dinners. I also remember our final graduation dinner with special guest, the Chief Commissioner of the Indonesian Police Force, head of an organisation with no less than 100,000 members. He spoke only through an interpreter and as part of the course, one of us was selected to be his host at the dinner. Thankfully, this wasn't me. Instead, a participant from the Australian Federal Police was selected, one Simon Overland. I must say that Simon had impressed me greatly during the course as someone with good principles and the ability to think clearly. I wasn't surprised when he was chosen and was impressed by his easy ability to engage with the Indonesian Chief through the interpreter. Simon was later appointed Chief Commissioner of Victoria Police and became embroiled in the Melbourne gangland wars and the subsequent informer scandal and Royal Commission. I'm still at a loss to understand how that all happened under his watch.
In 1995, the Metropolitan Ambulance Service (MAS) had put out a contract to build it's own CAD system. There were suspicions that the contract had been biased towards a particular vendor and when that vendor was awarded the contract, the suspicions turned to accusations and ultimately led to a Royal Commission into the tender process. Despite the turmoil, the government of the day decided that the system should be expanded to service not only MAS but all the major emergency services in Victoria - Police (VicPol), Ambulance (MAS), Fire (CFA and MFB) and the State Emergency Service (VICSES). In the process, this would render Firecom obsolete.
The contract was duly expanded and a timetable was struck to bring MAS on first, followed by the Fire Services in 1997 then VicPol and VICSES in 1999. I worked with the Operations led project team and others to migrate our Calltaking and Dispatch to the new Intergraph CAD and rework the FSEs and other turnout systems to integrate with the new CAD. This was a major change to the original intent of Firecom and meant that the MFB no longer managed its CAD system in house. However, the great advantage of the Intergraph CAD was that multiple emergency services could now be turned out simultaneously, based on the best response needed for the incident.
Like all organisations at the time, identification and remediation became the major buzzwords as the year 2000 approached. One of my collegues had been given the task of running the Year 2000 Project and as the scope and breadth of the required work was better defined, it soon became clear that more hands were needed. I was seconded onto the project in about March 1999 and we worked to identify and assess every technology system in the Brigade for potential Year 2000 vulnerabilities, then determine a remediation plan for each vulnerability. I remember spending the 2000 New Year's eve at the MFB's Emergency Operations Centre, ready to respond not only to issues that might arise within the MFB but also more general issues and possible emergencies in the broader community. As it turned out, nothing of any significance occurred, leaving to a belief amongst some parts of the general public that the Year 2000 threat had been either a hoax or greatly over-hyped. In reality, I believe it was actually due to the enormous effort by large numbers of people across multiple sectors to make sure nothing of any significance occurred. I still have an MFB badged clock that was given to each member of the MFB Year 2000 team to commemorate the event.
Towards the end of my time there, the MFB began work on building an independent dark optic-fibre network driven by it's emergency services status and enabling Commonwealth legislation. The plan was to commandeer a few fibres from any telecommunications company installing new fibre and thereby build the network independent from the telcos. The fibre entry and exit points from the host telco network would be brought to the closest MFB firestation or other MFB owned building and joined into MFB systems like Firecom and business/HR systems. Over time, the network expanded into regional areas to support other emergency services across the state.
My final big project would come from the earlier Integraph CAD project and ultimately lead to a decision to leave the MFB. The original Integraph contract was let for seven years and crucially, it called for Integraph to not only provide the CAD system but employ the staff and manage the calltaking and dispatch operations. They had never run this type of business before, yet despite many setbacks and ongoing criticism from the emergency services, by the end of the seven years they were providing a far better calltaking and dispatch service than those previously provided by the separate emergency services. However, this was not their core business and as provided for in the contract, Integraph notified the State Government that they were happy to continue to provide the CAD system but no longer run the Calltaking and Dispatch service. Some in the emergency services still like to think that the Government terminated this part of the contract but in reality, they never got over the fact that they were no longer able to run their individual calltaking and dispatch services using their own uniformed staff. It was better to believe that Integraph had failed.
A multi-agency team was established to deal with the significant change that the end of contract brought and the CAD Transition Project was established in late 2000. I joined that team in early 2001 to represent the technical interests of the MFB, along with a uniformed commander to look after the operational aspects. The project was divided into over a dozen major components. The overall task was to transition the entire calltaking and dispatch operation from Integraph to a new government entity. The government's earlier outsourcing model was turned backwards and a single state owned body, "Emergency Communications Victoria" was established to essentially buy Integraph out. The key success indicator was that there would be no loss of service at any time during the transition.
Due to the efforts of the team and many other people in the agencies, ECV successfully took over the operations on the day after the contract terminated on 07 Sep 2002 with no loss of service. All except two of the Integraph employed calltaking and dispatch frontline staff became ECV staff on that day. Over time, ECV became a statuatory authority, now called Emergency Services Telecommunications Authority (ESTA).
I left the MFB towards the end of 2003 to join the Bureau of Emergency Services Telecommunications (BEST) in the Victorian Department of Justice. It was a logical step for me, having been away from the day-to-day work in the MFB for well over a year. Some years later, the MFB completed the retirement of the FSEs after over 23 years of service. I was invited to Altona Fire Station 45 in February 2007 to take part in the official shutdown of the last operating FSE. For the record. the final FSE System Status Report, printed just before I shut down the last FSE is here.
Final note:
Missing on that last day was my old colleague, Lester Maclure. Lester had retired some years earlier and moved to northern Victoria. I lost contact with him but later heard that he had tragically committed suicide on his property. I don't know what happened or why he decided to end his life, but I did think of him when I shut down that last FSE.