'Ferranti is to launch a microprocessor chip early next year. The F100L , as it is to be called, has now reached the final stage of development and is expected to go into production at the Electronics Division Gem Mill plant in the near future'.
So starts a story in the Oct 22nd 1975 edition of Electronics Weekly just four years after Intel's Ted Hoff made the first microprocessor, the 4004.
The story continues: 'The 16-bit processor will be fabricated using Ferranti's unique collector diffusion isolation technology and is capable of addressing up to 2,000 words of memory directly and 32,768 words indirectly.'
'The single chip device is to have a repertoire of 64 instructions, and the clock rate will be 10MHz giving instruction times of between 300ns and 2.1 microseconds. The microprocessor is particularly well suited to multiprocessor systems,' concludes the story.
I remember working on this processor as a Product Engineer in the mid-1980's, by this time the device could be clocked at 100MHz (just as the first 8MHz 8080/8086 were available). It could address over 1MB of RAM via a DMU support chip (640kB was the PC max at the time) and had a co-processor (F101L). But CDI was a bipolar process hence consumed almost as much as a modern day Pentium, and required the whole device set to be mounted on a cermaic carrier. However, the thing that prevented Ferranti taking over the PC world were their entrenchment in the military market. This chip-set had been designed as a missile guidance computer and when I asked about creating a PC platform for it was told in no uncertain details that this device was not for commercial release.
Martin, that is such an indicative story of how the UK mind-set was in those days. The manaqements seemed to think making things for the commercial market was somehow inferior to making things for the MoD or Post Office. How things have changed!
While systems engineering for BAC, I started to plan new missile guidance systems that would employ the F100L in the latter half of the 1970s. But got so excited by the prospects of building new consumer products using CDI that I rapidly headed north to join Ferranti. Of course we didn't realise the potential of that ASIC approach for the various reasons discussed in David's blog in recent weeks (excessive power consumption, insufficient consumption of capital etc etc)... But UK folks have built some pretty good things over the years. Should we make sure we don't miss the chance to build some good products, businesses and fortunes on top of capabilities like the XMOS technology?
Jon, you are so right. But how do you make sure people use it? I think ARM got used becasue Saxby's a great and persistent salesman and Nokia took a fancy to it. On the other hand the Transputer didn't get used enough. So how could we make sure XMOS's technology gets extensively used to build things?
David, Very interesting debate could start here given that XMOS chips can, as they seem, be easily used to build commercially viable products without NREs and without the need for a lot of extra engineering effort. That should make the job of selling the XMOS platform to product builders around the world less of an uphill struggle than ARM had (and thus less need for a full "Robin" of sales capability within XMOS - NB. Just inventing a unit of sales power there). However my interest is also in the UK product building startups or innovators who can get out there themselves and start building things using the XMOS chips...
Jon, Now you’re talking. I remember XMOS’ CEO James Foster saying at an IET Conference:
"What if you could start a semiconductor company with $100,000 again? XMOS is focussed on developing the fastest way to design an electronics product. Our chip enables a new business category - the fabless, chipless start-up. Our NREs are less than $100,000 and our prototype lead-time is 30 seconds.”
So, the return of the $100,000 semiconductor start-up is at hand. Anyone with an idea for a chip, who can programme in C, who can get their hands on an XMOS development kit and an XMOS SDS chip, can bring an IC to market.
Some of the government money going to support the tech sector could be channelled into XMOS apps start-ups. It would be a brilliant way to encourage start-ups and spread the gospel of the XMOS technology.