Working years

see copyright notice. Page created 15-Aug-2020. Use the button groups above to navigate quickly around the site.

BBC Engineering (1976-1989)

I applied to the BBC during my final year at Cambridge, and following two interviews I was offered a Graduate Trainee post in their Engineering Designs Department. This would be based at Western House (later to become Wogan House, the home of Radio 2) near Broadcasting House in London. But first I had to attend a 5-week residential induction course at Wood Norton, the Corporation's training centre near Evesham, Worcestershire. There was a distinctly mischievous atmosphere among the students there, which nicely bridged the gap between college and the workplace, as we learned all about the BBC's way of doing things.

My apprenticeship started in Recording Section, where I was tasked with building and testing a timecode-driven "digital stopwatch" to replace the mechanical ones which were widely used in production galleries at TV Centre. As well as being a fairly meaty project technically, this provided an ideal excuse to explore TVC - and I soon discovered what a wonderful place it was for an evening out with friends. In those days access to staff (and friends) was quite relaxed, a far cry from the fortress-like atmosphere of later years. Only once did I ever feel a tad awkward passing through the Main Gate; on that occasion, in December 1979, I was on my way out - trying, instinctively, to conceal a reel of 2-inch videotape bearing some Christmas entertainment for my colleagues at Western House!

Returning to 1977: my next job, in RF Section, was to build an IF combiner and separator to feed R2 and R3 over a microwave link from Mendip to North Hessary Tor, in the West Country. I got an intriguing taste of working "on site" when I accompanied a couple of colleagues from TCPD and Transmitters to install and test the equipment. Back home, in Audio Section, I did some comparison tests on commercially available voltage-controlled amplifier modules, to recommend which were suitable for use in sound desks. Then, in March 1978, I went on a 5-week attachment to Equipment Department, the BBC's "factory" at Avenue House, Chiswick.

Next, I spent a further 18 months away from base, as part of a diverse team building a new CEEFAX origination system. My computer interest was rekindled by this project, based on a triple PDP-11/34 system, running RSX-11M and programmed in RTL/2 (a structured, real-time language) and MACRO-11 Assembly Language. The team, drawn from Research, Designs, Capital Projects and TV Operations, was led by two talented individuals from Logica, who taught me more than anyone else has, before or since, about software engineering. My work on the project involved synchronising the PDP-11 system clock to an MSF time-data receiver, and coding the display of current time on the top row of the CEEFAX pages. Amongst other things this entailed writing a device driver for a DR-11K parallel interface, which was a real baptism of fire in system programming. When it didn't work, the last thing I suspected was a hardware fault. I had to suspend the team's access to the system while I ran XDT (Executive Debugging Tool) and jabbed a screwdriver across the external data inputs of the DR-11K, trying to find out why my driver wasn't responding to the interrupts. Eventually I concluded that there weren't any. We replaced the DR-11K, and then of course it all worked fine! I look back on the CCS project as one of the most enjoyable of my career, but I think that was in large part because I was still a "trainee", with sufficient lack of responsibility to leave some room for fun...

In September 1978 I successfully applied for a Design Engineer post in Transmission Section (transmission, in the sense of getting sound and vision from the central premises out to the regional centres and thence to the actual radio and TV transmitters). Initially I did some odd jobs on video equalisers and CEEFAX inserters and regenerators, receiving on one memorable occasion a phone call from Head of Engineering Information Department, castigating me in no uncertain terms for increasing the number of data lines from 2 to 4 before we'd officially announced such a change. My defence that I'd only been obeying orders did little to quell his ire. We live and learn!

My other early achievements included designing a "Modular Audio Storage System" (MASS) using digital memory to replace the tape loops which were widely used for line identification (and station "jingles"). I also did some design work for a 68Mbit/s coder/decoder that did for PAL video roughly what NICAM did for audio, and designed a general-purpose Z80 microcomputer module (ADZE - we so loved our acronyms!), which was later incorporated into many other designs. These included my "Tiny Tim" speaking clock, used to make time-stamped cassette recordings during an investigation into digital link glitches, and a half-card computer running a variant of BBC BASIC ("ABBOT"), which provided a user-friendly platform for all sorts of monitoring and control applications.

But the bulk of my work in Transmission was on the evolving NICAM-3 digital sound distribution system (not to be confused with the later NICAM-728 standard for TV sound broadcasting). NICAM-3 uses 14 to 10-bit near-instantaneous companding to pack 6 audio channels into a 2048kbit/s bitstream. It was destined to replace the 13-channel PCM system, which had fed the UK radio network via analogue video circuits since the early 1970s. I designed much of the digital processing hardware and software in the first-generation NICAM-3 coders and decoders, which amounted to custom 16-bit computers built from small- and medium-scale TTL chips and ROM/RAM memory.

In January 1984 I was promoted to Senior Design Engineer, and embarked on the design of our first ever semi-custom chip, to carry out some of the more awkward computations of NICAM coding and decoding. I was instructed, from on high, to do this in collaboration with a major British research company. But, after signing off my design (which I'd had to do partly at the physical, silicon level as their CAD software had, er, some limitations) it turned out that there were insurmountable problems with their chip fabrication. Time being of the essence, I finally got permission to explore the commercial market-place; within a few weeks I and my colleague Duncan Nightingale (who did a brilliant job of porting the design to a completely different device architecture) received our first batch of working chips from LSI Logic Corporation. The LSI device worked as a co-processor, with two Z80 CPUs, to convert audio samples into NICAM data, or vice versa. The initial application was for stereo sound-in-syncs distribution of TV sound (licensed to a commercial manufacturer), but later my design was re-engineered, substituting a high-speed state machine for the Z80s, to provide a 6-channel package that superseded our first-generation equipment for radio sound distribution; it remains in use to this day.

In 1986 our department was renamed Design Group, part of a merged Design and Equipment Department (D & ED). The 8 specialist sections were reorganised into 4; I was placed in "Audio Section". Working with a team developing the Radiodata System (RDS) to label VHF radio programmes, I designed a data multiplexer and demultiplexer to enable a NICAM-3 signalling channel to carry RDS data, together with various transmitter control functions. I also wrote a large part of the software for the central RDS computer (a MIRA system with two micro-PDP-11 processors running RSX-11M, still our favoured solution for real-time processing owing to its closely-specified latency). My Schedule Editor, coded in PASCAL, enabled the required programme information to be entered through user-friendly menus and forms. The adjacent photo appeared in the staff newspaper, Ariel.

I developed much of the terminal-screen menu software under VAX/VMS, on a system which had grown from the single VAX-11/730 (compatible) that I'd first used to develop the NICAM gate array. Seeing another application for this software, I set about replacing our long-established, paper-based Management Information System with one in which staff entered their weekly work summaries online. With the enthusiastic support of my boss Richard Lawrence, I developed AMIS, as it became known, to the point where it replaced not only the staff input forms, but also the aggregated time accounting information derived from them, and was brought into use throughout the department.

In April 1988 we left our central London premises, and joined the "other half" of D & ED at Avenue House, Chiswick. There I took on the role of Computer Systems Manager, upgrading our system to a Local Area VAXcluster housed in a newly-built computer room (photo from BBC Engineering Progress and Achievements, 1988). I enjoyed this work, and became proficient at VMS system programming. But the daily commute between Upminster and Gunnersbury, plus the loss of West End shopping and recreational facilities, made my life outside the office rather miserable. I'd recently bought a house in Essex, and suspected that this downsizing and relocation of our patch within BBC Engineering wouldn't be the last (I was later proved right!). So, with a heavy heart, I resigned from the BBC, wrote the advertisement for (and helped to interview) my replacement, and embarked on a new career in academic computing - a bit closer to home.

During my time at the BBC, I co-wrote papers on NICAM-3 for the International Broadcasting Convention, on dual-channel Sound-in-Syncs for the Commonwealth Broadcasting Conference, and on RDS for IBC-88; I wrote a paper about the NICAM chip for the IERE, and wrote and presented papers on MASS for the IEE, and on NICAM for the Institute of Broadcast Sound. I also ran several half-day courses for sound studio managers at Broadcasting House London, explaining and demonstrating the principles of digital companding.

Goldmiths' and King's Colleges (1989-1998)

On a cold day in January 1989, I started work as Systems Programmer at Goldsmiths, University of London. On arrival at the New Cross campus I was given a warm welcome, and found myself among a bunch of people that I instantly knew would be fun to work with! The main central academic computer, an ageing VAX-11/750, had run without an experienced system manager for some time, and after spending several weeks assessing the problems I set about putting them right. The work included a wholesale clearance of disused accounts and disk files, a complete rewrite of the daily backup procedures, rationalisation of the user groups within college departments, and the construction of a secure, menu-driven environment for the operator (who was then a technician doing the job part-time).

Later in the year, we were offered a pair of VAX-11/785 processors from the London School of Economics. These promised a 5-fold increase in computing power, but required much more space in a purpose-built, air-conditioned computer room. For this, the College agreed to repurpose a ground floor area where there had previously been toilets - I remember suggesting, as a title for a newsletter article, "Goodbye Two Loos", which led to the new rooms being known, unofficially, as "the Lautrec Suite". I drew up plans for the equipment layout, printers, networking racks etc, supervised the technical side of the installation, and helped with the recruitment of an experienced, full-time operator. I moved all the user accounts and files to the new VAXcluster over a weekend; on the Monday morning some users were unaware that they were using a totally new system - but they did notice it was running a lot faster! The picture (from a faded Polaroid print) shows the new installation; the processors were named SILVER and BLACK, continuing a "smiths" scheme that started with the original GOLD.

Despite a long-standing reluctance to drift from technical work towards management, I found myself instigating Monday-morning progress meetings to improve communications among Computer Centre staff, and getting involved in discussions about the next major upgrade. Unfortunately for me, the DEC VAX systems in which I'd become an expert, and which had been the mainstay of academic computer centres for many years, were falling out of favour. In their place, servers running variants of the ancient UNIX operating system were being installed. I considered that to be a step backwards, but could see that the political drift in that direction was unstoppable. At the same time, I was starting to feel technically under-utilised.

Goldsmiths was part of an academic group known as the King's Cluster, which gave us access to a wealth of experience in running larger computer systems. Indeed, our own Director of IT, Brian Meek, was "on loan" from King's College - the adjacent photo was taken on his last day at Goldsmiths. We had meetings with fellow institutions from time to time, and I got to know my opposite number, Geoff, at King's. Over a pint or three, we established that we were each envious of the other's work environment, and so we persuaded our respective managements to sanction a job swap, initially for a trial period of 3 months starting June 1991. It went well, we extended the arrangement to the end of the year, and in October the exchange was made permanent. I resigned formally from Goldsmiths at the end of January 1992 - for the time being!

The Computing Centre at King's College London was about four times the size of that at Goldsmiths, and contained a long-established structure of specialist internal divisions. I was based at the Strand campus (with occasional trips to Kensington) as Senior Systems Analyst, sharing with a colleague, Andy, the system management of a cluster of VAX 8000-series machines, plus batch servers, printers, plotters, graphics workstations and hundreds of user terminals. Our Systems Division also included admin computing and networking specialists, and we worked alongside Operations, and User Services - or so the theory went. In practice I seemed to spend more of my time trying to improve communications and co-operation with these other groups than actually solving technical problems!

I spent 3 happy years at KCL; it was nice to be back among the pubs and restaurants of central London, and we had a great social club with the addded luxury of a snooker table! There was a choice of lunchtime eateries, including a good in-house canteen (and an even better one over the road at the London School of Economics). Covent Garden provided a handy place for a stroll after lunch. Workwise, I took on the role of Systems representative on the User Services Committee, and joined the editorial board for the Computing Centre's monthly newsletter. However, the writing was on the wall for big, central, power-hungry computers; most of our workload ended up being served by a table-top VAX 4000-100, and once again I found myself having to deal with an influx of UNIX boxes whose personalities displeased me. So, in January 1995, I did some calculations, and decided that I could afford to "retire" - or at least cease daily commuting in favour of part-time employment.

Initially, I found a niche back at Goldsmiths, where a system manager was needed for their VAX 4000-300 admin computer. From May 1995 until mid-2001 I fulfilled this role, working mainly from home with visits to the college about once a week. There was also a surviving VMS (Alpha) system in the Library, where I was occasionally able to help out.

Digital Terrestrial TV (1998-2001)

One evening in 1998 I received a phone call that was to change my life for the next year and a half. It was from an old Cambridge colleague who was then working, along with my ex-BBC boss, for Castle Transmission International - originally the BBC division that ran their transmitters, now a private company based in Warwick. They were in the mammoth process of building the Digital Terrestrial TV (DTT) network that was later to become Freeview. Was I free to do some contract work with them? You bet!

DTT was excitingly new to virtually everyone involved, and as dozens of new transmitters were commissioned, engineers were needed to test them, put right any problems that were found, and get signatures from the "customers" (at that time, BBC and ONdigital) accepting that the performance met the specification. The job required rapid familiarisation with new jargon and test equipment, together with the ability to work safely and reliably - and often alone - alongside the existing analogue TV transmitters. I guess that's why a handful of us veterans were coaxed out of retirement!

I met the team at Warwick, with a few anxieties: I had little practical experience of UHF power engineering, or of the PCs which had become an integral part of most items of test gear - and to cap it all I don't drive, so would often depend on public transport and Shanks's pony to get to and from remote hill-tops. No matter, I got the job and, after shadowing a two-man team at two major sites, became an active Stage One Tester. I was, effectively, on a "zero hours contract", which suited me well, as there were usually gaps of a week or two between the fortnights at work, when I could catch up with things back home. Getting to the sites turned out to be rather fun; I usually went as far as possible by train and bus, then hiked the last mile or three. Often some useful signposts were available, in the form of rooftop TV aerials! We were encouraged to stay in the cheapest possible hotels and guesthouses, but I have to admit that as time went on I managed to book into some rather nice country house hotels, on the pretext of having to be as close as possible to my workplace. The table summarises the sites I helped to test.

I could write a book about my adventures, but for now here are a few random notes. Saddleworth straddled Christmas and New Year, which meant several trips between home and Holmfirth - a lovely town enhanced by a thick covering of snow, and a warm welcome at the Old Bridge Hotel. On the ONdigital handover day, I made a ridiculously early start to travel by bus, via Huddersfield, to the nearest access point [as an aside, I found bus drivers in places like rural Yorkshire and Wales most obliging when it came to request stops in the middle of nowhere!]. I reached the transmitter building soon after 9am, and sought shelter behind same. Meanwhile Doug, from ONdigital, quietly drove up the track, saw no other cars, concluded I wasn't there, and also wandered round to the leeward side. We both nearly had heart attacks when we saw a shadowy figure looming out of the mist!

At several sites I stayed over a weekend, and enjoyed some fabulous country walks. Fenton comes to mind (Peak District), as does Carmel (South Wales). For the latter, I was based in Carmarthen, and setting out on a Saturday morning in glorious sunshine I passed several smiling folk who greeted me with a cheery "Bore!" ("Morning!"). After a while I made the mistake of replying in kind - whereupon my new friend launched into a torrent of Welsh! I smiled inanely, nodded, and moved swiftly on, resolving to have another go at learning the language. Also at Carmel, on a rare working Sunday afternoon, we needed to make a scheduled changeover of one of the analogue services to half-power working. This deceptively easy-sounding task involved shutting down two transmitters, then unbolting four large flanged connectors and rearranging the pipework between them into a different configuration. The Transmitter Manager couldn't do it single-handed during the 2-minute gap we had available between programmes, so he and I each took responsibility for half the task, working simultaneously. Waiting, wrench in hand, for the interval trailer to appear on the monitor screen was probably the most stressful moment of my working life! We finished just seconds before the start of the next programme (a rugby match) which duly went out, as scheduled, on reduced power. When I entered the hotel bar that evening, the first remark the locals made was "What you done to our rugby then?".

At Hemel Hempstead I stayed in a very nice hotel, from which the daily "commute" was a 3-mile walk along a canal towpath, then another mile to a farm - and back again in the evening. I did most of the work there single-handed, including chasing the farm dog out of the transmitter compound! Despite the moments of drama, these were enjoyable times - and lucrative too. I was paid by the hour, including travel time, and still working part-time for Goldsmiths, courtesy of a Nokia Communicator (an early type of smartphone). That meant I could earn two wages simultaneously, by doing Goldsmiths work during the long train journeys. Back in New Cross, disgruntled students began picketing some college buildings, including the computer suite, so that staff couldn't get in to do the daily disk backups. To help out, I copied as much data as possible onto the Library system (in a different building) - working through the night from a hotel room in Newcastle!

All good things come to an end, and although I was invited to do some further roaming contract work with Crown Castle, as it was now called, I decided to have a rest. Then, early in 2001, ONdigital recruited me as a consultant, to take over the development of a software system, VIA (Vehicle Input Application), for surveying network coverage. Written in Visual Basic, it had grown into something of a monster, which was in need of restructuring to make it maintainable and extendable. I worked on this for some months, occasionally attending meetings at their HQ in Battersea. During this time the company was re-branded as ITV Digital, but falling customer numbers forced them into making severe cutbacks, including to my contract. So I never got to complete the software, and as 2001 drew to a close I opted for a second attempt - this time successful - at retirement. ITV Digital subsequently collapsed, driven out of the subscription market by competition from satellite and cable services.

top ▲