Former Ferranti Electric engineer [Martin Mallinson] recently posted a 1980s documentary on YouTube (see the video below the break). It shows in some detail the semiconductor plant at Gem Mill outside of Manchester UK, as seen through the eyes of the ghost of founder Dr. Sebastian Ferranti. This dramatic device seems a little silly at times, but the documentary still provides a very interesting look at the industry at the time.
The Gem Mill plant was one of the first semiconductor facilities, having begun operations in the 1950s by Ferranti. In 1959 they made the first European silicon diode, and went on to commercialize Uncommitted Logic Arrays (ULA) in the early 1980s. Most famously, Ferranti ULAs were used in many home computers of the day, such as the Sinclair ZX81 and ZX Spectrum, Acorn Electron, and the BBC Micro. Much of the factory tour in this documentary is depicting the ULA process, and they hint at an even more advanced technology being developed by the (unnamed) competition — an FPGA? CPLD?
In a series of events worthy of a mystery novel, Ferranti finally closed its doors in 1993 after acquiring a company that was involved with clandestine agencies and illegal arms sales (see Ferranti on Wikipedia). But through a series of acquisitions over the years, many of their products outlived the company and were available under the labels of future owners Plessey, Zetex, and finally Diodes, Inc. The Gem Mill facility was decommissioned in 2004 and in 2008 it was demolished and replaced by a housing estate.
Thanks to [Cogidubnus Rex] for bringing this video to our attention. A couple of other Ferranti documentaries of the same era are also included down below the break.
Collateral damage from three letter agencies meddling.
Some time ago I found two PDFs from Ferranti explaining the inners of the ULAs: the design of the cells, number of them for each device model… very interesting.
https://archive.org/details/FerrantiULA2000SeriesDatasheet/page/n5/mode/2up
https://archive.org/details/FerrantiQ.RefULA1984/page/n1/mode/1up
Another user found this one:
https://datasheet.datasheetarchive.com/originals/scans/Scans-002/Scans-0052254.pdf
I worked for Ferranti (Cheadle Heath) in the 80’s up to the point they went bust. Business lesson kids, don’t get in bed with shady international arm dealers.
Great technology company, terrible management, but that was British industry then.
Although we used some of their chips in our Sonar systems, I was not aware of the plant. (People of a certain generation may think that is weird, but before the internet and Google it was amazing how much you did not know going on around you)
Thanks for the nostalgia
You sound like the most likely person to know this. I have a Ferranti DIP28 ULA 1045E 8412 read write driver for a Tandon TM65 2L floppy drive from the 1980’s. Any idea of where to find the data sheets on such an IC?
The interconnection layer on the ULA will be a proprietary design, but it’ll be less than 100gates on such an old device. The 8412 is a date code i.e. the chip was manufactured in week 12 of 1984 and E simply designates a plastic package
Great info. to add from what I found from the quick reference guide on line. It was extremely generalized and mentions to reference the data sheet, which of course, has been lost to time it seems. https://archive.org/details/FerrantiQ.RefULA1984/page/n3/mode/2up?view=theater
£20 on eBay for a replacement – see https://www.ebay.co.uk/itm/363675339768?mkcid=16&mkevt=1&mkrid=711-127632-2357-0&ssspo=c5LmtnKlRay&sssrc=2349624&ssuid=i6vg_Ld8R3S&var=&widget_ver=artemis&media=COPY!
Yup, seen it! I just bought the FDD just future guessing it doesn’t work. The thing looks clean but the seller didn’t have a way of testing it. Thanks, just in case inquiry since I have another identical drive but with a different PCB. It has a faulty Cherry Semi 977509-001. Cherry Semi is another chip with a forgotten history nobody knows anything about. Thank you. I’m just a bit paranoid about Ferranti now.
The ULA1045 was, if I remember correctly, from an earlier family of devices than those listed in the link on your earlier reply. The datasheet would not have given any details on the logic or functionality of this particular device, but simply the topography of the underlying array of cells, environmental spec etc
You shouldn’t be paranoid about Ferranti – the engineering was very solid. Being a bipolar semiconductor process (CDI) the Ferranti ULAs were actually pretty rugged devices e.g. not as sensitive to static as CMOS. I hope the eBay replacement does the job, but we are looking at nearly 40 year-old electronics here!
My dad worked there too at that time and was devastated by the demise of the company having given his entire working life to them. He has passed away now, and I’d love to find out more about this time of his life. Do you know of any groups or forums online?
Thanks for this, I went for an meeting here in 1984 when the company I was working for was wanting some custom ULAs, I remember the factory tour looked like SciFi set at the time, its amazing how quickly things date.
I was an apprentice at Ferranti, 1980s. Spent a short time at Gem Mill. Sadly I remember little about it; Most of my time was spent in Moston and Cheadle Heath. If I’d known they were making ULAs for the very machines I had at home, I might have tried to grab a few!
A lot of UK plant still is run by Ferranti kit, several of the AGR power stations and apparently the McVities biscuit production line.
The Ethernet cards we used had the MAC address configured via 3 physical dials and it picks the last 32 bits to use as the IP address, interesting approach!
Wow, id like to see such an ethernet card
“and they hint at an even more advanced technology being developed by the (unnamed) competition — an FPGA? CPLD?”
anyone know who this company was or what they were designing?
” In 1959 they made the first European silicon diode, and went on to commercialize Uncommitted Logic Arrays (ULA) in the early 1980s. ”
Soon followed by the Recently-Divorced Logic Arrays.
Although they also made 16 bit microprocessors the F100L for example. I knew them for good performance cheap transistor, DACS, voltage references. I also had an early version Digital Panel meter IC ??450 or something. It used fewer external parts than another DPM IC because it used Delta-sigma A-D conversion. I suspect Crystal bought the IP for that?
A wonderful trip down ‘memory lane’ from my first job working in the ULA marketing department in Gem Mill. Happy days!
Hi Martin, I have a project to reverse engineer a Ferranti ULA (dont know which device yet) do you have any information that is relevent?
Hello David
The digital ULAs are essentially an array of NOR gates, so in principle given a good quality image of the single metallisation layer it would be a tedious but ‘doable’ task to deduce the logic configuration of the chip. However, from a commercial standpoint, customers expected to have a level of security akin to any custom chip, so you need to be careful that you’re not going to face any legal implications from breaching design copyright. In addition, some of the ULAs also featured analogue circuitry which might well be more problematic to reverse engineer, and even the digital ULAs offered ‘peripheral cell’ configurations to facilitate different IO requirements
In summary, tread carefully!
Martin
I worked for Ferrari designing ULA’s in the 1980’s. Still have one of the design manuals.
I too managed to sneak out a design manual and some other bits and pieces. A reminder of the amazing time that we had with Ferranti back then having a world lead with the ULA
Small addition the ULA was a particularly influential development but the part was organally designed by Signetics a Philips company in USA.
My understanding is that the Ferranti ULA concept was ‘invented’ by the late Technical Director David Grundy while he was watching an Oldham Athletic soccer match. He always finished the story with the throwaway line that, “It was a very boring match”!
The ULA may not have been unique as a ‘sea of gates’ customisable chip, but Ferranti’s CDI (Collector Diffusion Isolation) process was well-suited to the implementation in being able to provide Vcc and Ground contacts in every cell across the chip surface
Commercially the Ferranti ULA enabled (and protected) many ground-breaking products of the time e.g. Acorn BBC and Sinclair home computers. It was an amazing time!
I was Brought into Ferranti’s by Clive Briggs via the student apprentice scheme based in Hollinwood in 1978. I eventually ended up at Gem Mill and after the required 6 weeks in each and every department worked in I.C. Test for quite a few
years under Peter Barnes (Legend!) What I learned during my time at Gem Mill was amazing, both from a technical perspective and also from a (internal) political perspective. Everyone amongst my peers were hot-shots and wanted to prove themselves, none of us were willing to admit that we didn’t know it all, however we did all learn from many of the longer in the tooth Engineers and boy, did we learn,.
I met so many knowledgeable people and learned so much during that time that I only wish that I could pay it forward as much as they did with me. Those were the real Engineers. They were a company with a great system for innovation and creating a new generation of skilled Engineers and workers. In my opinion Derek Alun Jones should have been stripped of his knighthood, Alan Shepard was an arrogant idiot who could tell people what to do, but did not stand up for what was right. That is where Ferranti failed both the employees and the shareholders.
Do you remember my dad, Brian Woodhead?
Looks like most of the Sinclair ULAs have already been reverse-engineered – see https://www.vretrodesign.com/collections/products
Isn’t technology a wonderful thing!
Don’t know whether this will work, but the attached links are to a scan of the Ferranti ULA brochure of the late 1970s and an FT article of 1981. Of those pictured in the latter I understand that David Grundy died quite young many years ago, Bryan Down in 2020 and if Alan Shepherd is still alive he’ll be well into his 90s!
https://www.dropbox.com/s/yb8tezo3t43dlvt/Ferranti%20ULA.pdf?dl=0
https://www.dropbox.com/s/9ud26vizaf6ar1w/FT%2011JUN81.pdf?dl=0
I worked in Gem Mill (for Thorn-Brimar, though, not Ferranti) for a short period in 80/81. We moved out to Greenside Way shortly after I joined. I wasn’t at the Mill for very long (<6months) but I have happy memories of the place, all the same. Sad to see it being demolished. I used to love wandering around the Mill looking into the glass-sided rooms, watching the ladies doing what they did.
As an electronics hobbyist, the stores window was a great boon 😀. And it was staffed by a very attractive young lady by the name of Elaine.
I spent rather too much time chatting to Elaine than I should have, really 🤣, but i did come away with a few chips that I’d never have been to afford otherwise, notably ZN1040 counter ICs, the DVM chip, the part # of which I can’t remember, and ZNA134 TV sync pulse generator ICs. I also had innumerable E-Line transistors from the ZTX range (slightly out of spec, and bagged in 1000s, IIRC). Happy memories
I too used to enjoy wandering around Gem Mill, though I don’t remember meeting Elaine! You reminded me of how worn the steps were on the lower floors (there was only a single ‘industrial’ lift), presumably eroded by the clogs of earlier cotton workers rather than itinerant electronics engineers
I came here to find out when Gem died. This has been a fascinating diversion. My dad worked for Gem in the early 60s around the time I was born, and then was let go during the first great scandal when Ferranti had its fingers heavily on the scales of cost+ for the MOD