Career History

You can read this on my LinkedIn profile too. This is a rough brain dump of my experience to date. Though it is in need of updating.

Vacation training, Summer 1990

Work started for me with a summer placement at GEC Electrical Projects in Rugby, UK. This was in the summer of 1989.
Great fun. I was working on an 8051 based board programmed in assembler which performed a bar code decoding function for the laser scanner for an automatic guided vehicle. These were big pieces of machinery, weighing a couple of tons each, and very capable of demolishing large parts of the lab – allegedly. The end result was a successful upgrade from an 8049 to an 8051 micro, with lots more memory and processing power.

Finished University, started at Philips Semiconductors : 1990

I graduated from Birmingham University with a BEng (Hons) 2.i degree in Electrical and Electronic Engineering.
I started at Philips Semiconductors in Southampton, UK in October 1990.

This was in the PCALS, Product Concept and Applications Laboratory.

We followed an IEE approved graduate training program for two years. I still have all the weekly records ready for donation to the Science Museum.

My first project was to fix the EASI development board. EASI was an FPGA prototype of a Dynamically Re-definable Character Set (DRCS) generator for Teletext (or Ceefax to give it the BBC trademark name). It turned out that the several 10s of Kbytes of un-commented Pascal driver code were not at fault. It was (as ever) the prototype socket for the FPGA on the ISA card. You just cannot get enough pressure onto those things to get all the pins well connected.

Next stop was a Teletext character editor. This was written in C (Borland C if I remember correctly) for MS-DOS on an 8086 Compaq ‘luggable’. This character set editor was revisited over the years. I added a postscript output feature, useful for pretty data sheets. Later on someone else (Winston Bond) made an X-Windows port. The first real use was for design of a Thai character set. Also spent many long evenings debugging various customer issues with our ICs.

Around 1992

After that I spent a while cooking up new ways of demodulating the data signal off the vertical blanking interval of the composite video signal. The existing DC loop based system was very very good, based on the old VIP solution designed by a colleague Paul Summers. My feeble attempts ar new architectures like “integrate and dump” weren’t as good. Though I did get rather good at using an old VMS VAX Lotus 123 spreadsheet to simulate it. I also used a Philips specific simulation tool, PSPICE to check some analog circuits I used to model the theory. Long and short of it was that the existing system was was doing the theoretical maximum possible without adding lots of extra expensive memory.

Meanwhile the US Line21 Closed Captioning system came into existence, so I did several projects to go with the LITOD SAA5252 IC development (single chip Line21 decoder). Such as modify a VG Electronics Calibrated Distortion Unit to generate the Line21 format VBI data. I also added a load of IIC controlled relays to automate testing. Worked with VG on this. Many thanks to Danny Payea for all your help ! Chebychev filters on the outputs for waveform shaping etc etc. Labview hookup. Programming an IBM compatible in C and C++ to automate collection of results via the parallel port. Also designed and wrote a Borland C application to give a “nice” Windows type GUI IIC controller to test the LITOD with. Used a tool called CSCAPE which was a nasty thing indeed. This little lot resulted in much faster and more repeatable characterisation of our data demosulators in the face of co-channel, noise and echo distortions.

I even knocked up some AHDL (Altera’s early VHDL) code that generated all the VBI formats from one ~10 MHz crystal and some SRAM.

My new found expertise at demodulating the VBI was put to work debugging design-ins of various teletext ICs. SAA5246 was the first, IVT1.0. We used to sell bucket loads of these into the various TV factories in South Wales, and also in Belgium and Germany.
The IVT replaced the analog demodulator with a digital DSP based one.

We literally sold hundreds of millions of these, and for a good price too, and the TVs with Teletext had a significant price premium too. Great times !

1993 and a quick tour of the Antipodies

My girlfriend and I then decided that cold, damp England wasn’t the place to be for a year. So we took off to Australia, New Zealand and a bit of Thailand and the Philippines. In Sydney I specified a video based vehicle detection system for AWA Traffic for about 2 months. How the other engineers laughed when I said I was just “popping to the shops” one lunchtime in North Strathfield, Sydney. Mid summer lunchtimes are about 40 degrees C out there. Took me the rest of the afternoon for my clothes to dry out when I got back. The rest of the time was spent traveling. Especially enjoyed spending November afternoons sat outside the Coogee Bay Hotel . Also the Sydney Flying Squadron in Kirribilli was another favourite ( known as The Skiff Club as they sail 18ft skiffs there, the F1 of dinghy sailing). My kind of sailing. I have an Ozzie residency visa, so if anyone needs to hire some crew, drop me an email !
But the Chicken Balti wasn’t as good in Sydney so we headed back. And upon our return, one of my colleagues, Noel Hurley had just left to join a little startup called ARM where he would be working for Robin Saxby. Needless to say, he has a lot bigger, nicer house and a faster car than me now ! (And is currently Director of something or other at XMOS )

1994: Back to mixed signal ICs

So I took Noel’s job (now in the Systems Laboratory, Southampton or SLS) and got on with a spot of market analysis whereupon I concluded that there was space in the market for a multi-standard VBI data demodulator for the PC. I did the feasibility report, got the OK from product development. Then did the specifying, designing, testing and application of a device, the SAA5284. Great days when you could still create a mixed signal full custom IC for less than a million dollars ! Albeit with only a few tens of thousands of gates. Lots of VHDL work, Cadence LeapFrog, scan chains, test specs, application circuits, data sheets (including those damn timing diagrams), application manuals, user manuals etc to create. And a project to manage.
Studiously ignored approaches by Noel to come and work at ARM. I like sailing, hence the South Coast is the place for me. Bit of a problem that one, as the vast majority of jobs in this industry seem to be located in the areas of Cambridge or Bristol. Apart from NDS, Tandberg, IBM Hursley and NXP (was Philips). Hmmm…

1996 ish – promoting SAA5284 around the world

At this stage, Philips Semiconductors Systems Laboratories were a very strong organisation with labs in Eindhoven, Hamburg, Southampton, Taipei and Sunnyvale. With over 750 engineers in total.

I lead a demo at COMDEX 1996. Plenty of interest in the little SAA5284 and attendant Windows teletext software we called PCText. Great show too. Just at the start of the PC entering the consumer market. Just before setting off, I had set up a website for my cousin’s business using my Demon Homepages. Spent most of the show testing them out on bits of kit around the stands. The AMT Marketing website has grown since then, and so has the SEO industry. We moved from HTML 3.0 to 3.1 (with frames), then added Javascript. Then I moved to PHP and MySQL. With many, many tweaks over the years. We are still up there in Google SERPS though.

Anyway, back to describing the day job. I supported SAA5284 design-ins, criss-crossing the US which was great fun. We sold more than enough SAA5284s to make a tidy profit. Then the design was subsumed into our desk top video products where it lives to this day.

In 1995 / 1997 we looked at carrying “The Internet” over the VBI. Using IP over VBI. This reached an RFC for NABTS (US Teletext) and a BOF in IETF for WST (World System Teletext). I defined and tested and implemented the FEC for this. There were other similar activities from Intel (InterCast), and of course PointCast. We defined an ETSI (TODO: dig out the number) spec for IP over the VBI. This was used in Teleweb, an HTML lite system for the TV/VBI market.

Meanwhile, a lot was happening in the Digital TV area. The ISO15322-5 (MHEG-5) spec had been completed, and plenty of my colleagues in Philips Research (Redhill) were getting excited by the new standard. After many long discussions in the DTG standards working groups, this emerged as the MHEG UK profile now used in the Freeview (was On Digital) service.

1997 MHEG-5 and Digital Terrestrial TV

So I lobbied inside Philips for a team to build our own new Teletext decoder to do MHEG-5. Big companies can be very political places. It took a lot of doing. And we got one of the first MHEG-5 decoders up and running on a Trimedia STB for Cable and Satellite 1999.

My team was the first in our part of Philips to license some software to a third party. Just getting the legal side sorted out for this took ages. It was a great achievement. Years later though, the powers of business models meant that the we sold this code to a third party who are now selling it as an ISV (independent software vendor). Meanwhile we got back to our core business. Making silicon. Let Cabot Software , OceanBlue Software and S&T and others play to their strengths and amortize that software cost across as many platforms as possible.

I started giving one of the lectures (on Receiver Software) at the University of Surrey Digital Terrestrial TV course .

2000 DVB-MHP and Java for TV

I gained Chartered Engineer CEng status in November 2000.

Around 2001-ish I started a project looking at partitioning personal Java execution between the MIPS and Trimedia of the new Nexperia platforms being designed in Philips Semiconductors. We built on a load of great research in Redhill. We invented a few patented techniques. Then started looking around for a DVB-MHP software stack. DVB-MHP being the standardisation of the APIs used by the Java VM to control and access the TV functionality. This built on prior work by DAVIC and HAVi. A good book about all this is written by one of my ex-colleagues, “Interactive TV Standards: A Guide to MHP, OCAP, and JavaTV” by Steven Morris. So I headed up a team porting JVMs and defining requirements for a DVB-MHP stack to go on our silicon. Part of the issue was scoping the requirements. We couldn’t finalise the APIs we were providing on our platform in time, and a lot of our ISVs couldn’t finalise their required APIs in time. We were (as usual) a bit too early. 8Mbytes RAM was too expensive and processor clock speeds were hovering around 80MHz (32 bit MIPS).

2001: Java hits mobile : J2ME CLDC and MIDP 1.0

As that roadmap emerged and was embarked upon, it was time for our innovation lab to move on to other Java areas. We worked on ahead of time compilers, just in time compilers, hardware bytecode translators (like ARM’s Jazelle), even pre-compiling straight out of the DSM-CC cache. See the recently granted “Method of compiling bytecode to native code” by Ian Willis who was in my team. Ian runs a very comprehensive site digvid.info .
(I have had about 6 patents granted over the years, despite many ideas, not all of them great !).

At one point I managed 18 engineers. My team were providing a large chunk of the Philips Semiconductors demos in various trade shows like Cabsat (now MediaCast). But with loads of help from Research in Redhill, guys like Jon Peising, Richard Houldsworth, Richard Miller-Smith (RMS), Paul Bristow (OK, not in Redhill, but affiliated), Stephen Tickel, Jon Foster, Stuart Butterfield, Simon Turner. Along with Steve Turner driving the sales effort. Sometime around now, Philips bought VLSI. That was interesting. I remember a long conversation with lots of “and then we screwed this up… oh so did we” points :)

2002: GPS, network resident applications, asset tracking

So I set a new direction for my team, and we started work on mobile Java. Namely J2ME CLDC and MIDP1.0. This coincided with development of our CARPET GPS device, SAF3100. Note the Nokia 6103i had just hit the shops as the first Java programmable phone.
We had stiff internal competition from a loosely affiliated company called SmartFusion (no link because they didn’t survive for long ). But they are still here in archive.org (I love that internet “afterlife”).

So we produced quite a few demonstrators combining GPRS modems (Wizmo, based on our GSM baseband), plus GPS for trackers. Then a back end server running on a Linux PC with a few Perl scripts and serial port to another Wizmo and a broadband connection and the kind people at Multimap. Then with a carefully crafted Midlet on a Nokia 6310i (first midp phone) we had a demo on a phone. The demo allowed the phone user to press a few buttons and get back a map pin-pointing the location of the thing being tracked (me going home on my bike one day, Robert Dobson traveling to a customer another time).

In my spare time, I used PHP, MySQL and a variant of amtmarketing.com to allow a colleague to submit a log of his progress on a Monaco to Glasgow cycle ride from his Nokia camera phone. I haven’t got the data for this live at present, but look in www.archive.org for www.touringwebcams.co.uk here and you will see a “Web 2.0″ application from about 2003.

2003: The Advanced Systems Laboratory

The dotcom collapse meant a re-organisation of the business. In fact things had become a little crazy. Our R&D spend had mushroomed, sales did during the boom, but then dropped back. So something had to give. We were re-organised into a smaller laboratory called the Advanced Systems Laboratory, spread between Southampton and Eindhoven. Many of us lost our teams. We were now consultants in the company. Twelve wise men to chart the course of the businesses. Evangelize, have ideas, influence and generally map out what lies just beyond the horizon for the product roadmaps. My area being interactivity and execution models. This was tough. I had to figure out how to add value to the company with minimal effort spend. Meanwhile I was offered a nice looking (but temporary) job in The University of Southampton eScience dept. In the end, swapping a permanent post in industry (still with final salary pension) for a lower paid 2 year contract wasn’t tempting enough. Had the mortgage been paid by then, who knows I would’ve taken it….

Most of the work in ASL was centred around tech demonstrators in the area of connected homes, and their interaction with mobile phones and cars.

Obviously I can’t be very specific on the work now, as it is still rather company confidential.

NXP Semiconductors

We are now NXP Semiconductors “founded by Philips”. Owned by amongst others KKR.
Today I am running several projects in diverse areas. One project is with smart RFID tags with bio-sensors and interaction with NFC enabled phones (like the Nokia 6131). This includes some early discussions with IBM on MQTT applications in this area (and connected homes), visits to events to develop the business into new directions, e.g. The forthcoming Toyfair. Another hot-topic is the support of Adobe Flash in multi-heterogenous processor architectures. The speed of the ascendancy of Flash based video services is incredible. The diversity of companies operating in the content provision area is amazing. TV meets Social Networking feels like another mini dot com boom. It’s a nice challenge to have.

This period in ASL has been productive, we opened doors into several new mega corporations. Our sales guys followed us in and made quite a few wins on the back of our technical business development work. The product architectures have all benefited from our work, being able to hold their own against the stiff competition. However, there hasn’t been another “mobile” or “PC” boom market that the silicon industry so needs to build margins back up again. In fact this lack of differentiation will push silicon into a commodity (already has done some say) business. Added value is now elsewhere in the value chain.

MSc in Renewable Energy Systems Technology at Loughborough University (started Sep ’08)

An ongoing activity. See the course program here. To date (June ’10) have passed Water1, Wind1, Biomass1, Sustainability and Integration modules. Water2, Biomass2, Solar1 and a project to go. Aim to complete by summer 2011.

Business Manager : Southampton Solent University. Feb ’09 to Jun ’10

I was Business Manager for the Technology Faculty at Southampton Solent University from Feb 2009 to June 2010. I divided my tenure into several phases.

1. Observe and learn university systems [2 months].

2. Skills, facilities and capability audit [3 months].

3. Report to management and seek mandate for change [2 months].

4. Report results of change management process and further actions [9 months].

The principal deliverables of the 16 months were:

  • Capability audit report for the faculty.
  • New finance reporting systems (crawl chart based).
  • Installation of new set of business resources. NDAs, contracts, quotation template, facilities hire agreements, business T&Cs. Developed in conjunction with Bond Pearce LLP.
  • Audit of the University IPR systems and implementation of new policies, procedures and guidelines, with a roadmap to full IPR exploitation systems.
  • Analysis and reporting of future shape of the enterprise activity to make optimal use of HEIF4 funding.

During the course of the 16 months, I learned a great deal about the differences between private and public sector operations. Also built a large network of contacts in the HEI sector in local and other UK universities.

Managing Director : Winchester Innovation Ltd. Sep 2010

Running my own business is something I have always wanted to do. So Nick Thorne Ltd is up and running with several projects in the pipeline. A mix of technology development, business and IP consulting and some in-house product development in the energy area.

For the latest, go to my  LinkedIn  profile.

theme by teslathemes