So, who wants to see a proper computer ?

[Bassoonbloke]

Hi All,

Yes I have an interest in old computers as well and I thought someone might be interested in my old beastie.
Its a 1970's GEC 4070 Minicomputer (just the processor box and Memory storage box). I have built power supplies for this and have had it running basic internal diagnostics.
I think it's a great bit of British engineering, but now totally obsolete (even the maker). It is totally TTL logic based and consumes a large amount of power from SM power supplies.
It has a bit of a story attached to the thing which i'm happy to pass on if anyone is interested (as to why I have it and where it came from).
Hope you like it.
Comments please....

Alan.

[pamphonica]

In 1978-80 I was in BT and we were developing the UK's first radiopaging network using one of these beasts. My job was to develop test equipment to simulate the interfaces between this beast and the telephone network line interface, while each was being developed, and in subsequent manufacturing.

My clear memory was of a GEC engineer turning up at Burne House exchange building (at the junction of Edgeware Road and Marylebone Road) each month to do "routine maintenance" on the 4070.
He would power it down, remove and re-seat each card, and end up doing a zylophone rattle across the backs of all the cards with the handle of his screwdriver.
After each of these visits the minicomputer would become highly un-reliable and keep crashing for a few days. This seemed to be due to dust being shaken into the edge connectors.
We ended up agreeing with GEC that we would continue to pay for maintenance on the strict understanding that their field engineer only visited in cases of actual failure!
Peace descended once again...

Incidentally, that National Radiopaging Network ended up as the largest in the world at that time, capable of serving a million pagers.

Jeremy

[Nick]

Why not see if the National Museum of Computing or the Science Museum are interested?

[Bassoonbloke]

Hi Nick.

I have been in touch with a lot of people connected with this computer including Michael Melliar-Smith who was one of the primary designers of the GEC Systems. Also with people who were using and programming them during the 70's and onwards.
It has been a very interesting and enjoyable process collating all the information that I have now accrued.
I certainly have no plans to get rid of the system as I think it's rather special and I really enjoy putting it all together once In a while to mess around with it.
It was originally running 24/7 for approx. 30 years and was replaced by a couple of 486 pc's.
Not bad I guess.

Alan.

[peter scott]

Hi All,

It has a bit of a story attached to the thing which i'm happy to pass on if anyone is interested (as to why I have it and where it came from).
Hope you like it.
Comments please....

Alan.

Yes, please.

Peter

[Bassoonbloke]

OK, so Peter say’s ‘Yes’, so here we go……

On a Planet far far away – Oh wait I can’t use that here.
 
I have been interested in old computers for quite some time and I spotted someone on e-bay selling a ferrite core memory board, it looked interesting and I was intrigued enough to want to know how this technology worked, so I bid and won it.
When I tried to find more info on the board I contacted the seller who said he had more other  boards for the computer which turned out to be a GEC 4070 mini.
Long story made short I bought all the rest of the boards from the seller, so I had a complete set of computer boards. The system that the boards came from was part of the chain that ran Prestel – excellent.
Next thing was to try to put together a Bus board to connect everything together. Now by chance one of the people I contacted still had a 4070 and was kind enough to photographed his system’s bus board (both sides and in detail) and also told me which boards needed to go where which I was very grateful for. It then became obvious that this was not a straight through bus board.
It was after all this and many e-mails I was contacted by Bob Mannix at RAL in Oxfordshire who basically said that their last working GEC 4070 had been scrapped and was going to be skipped as junk, so would I like it – Hell YES !!
So I arranged transport to Oxfordshire (I don’t drive) and picked up the Processor and memory box, which turned out to be a Silicon Ram version (about £12000 at the time !).
All of the documentation for the computers had already been skipped and all the cables connecting anything up had been cut at the box entry, so that wasn’t helpful, but what the hell.
My idea was to get the basic system going and then swap out the Silicon for Ferrite memory and see if the original data was still on the boards – hopefully an OS and some of the programs could be saved and used later on.
So my current computer was running the monitoring system at RAL’s Accelerator plant (along with a few others) and had been running for approx. 30 years, 24/7 and had finalyl been replaced by a couple of 486 desktops. All of my spare boards and Ferrite memory boards came from a Prestel system GEC and hopefully at some point in time run an OS again.
For those of you who don’t know Ferrite memory data is not lost on power down and can be retained for many, many years fully intact and on power on will run as though never turned off. There are stories of GEC’s being re-started and users losing connection for a short while and then continuing work as though nothing had ever happened as everything came back on exactly as it was left.
So there you have it.
Those of you who have fallen asleep at the back – fair enough, I can understand that.
Hope that is of some interest.
Alan.

[Nick]

Great story - thanks for that.

Many years ago, in the early/mid 1980s, I worked on Prestel and even wrote a Prestel emulator for xterminals and ANSI VDUs (DEC VTxxx and others).

My company provided tour operating systems to major travel companies, e.g. Rank Group, Thompson, Butlins etc. and we had terminals in every travel agent in the UK...

[pwdrive]

I remember Prestel, used to service Philips TV's that had that system fitted, G11 chassis as I remember.

Lawrence.

[Terry]

Ah! What's in a name?

When we were developing our own system at the Stock Exchange, we had to use the rather clumsy term 'wired teletext' (to differentiate it from 'broadcast teletext') because BT used the term Viewdata for their own service.

However, as development proceeded, BT decided that they ought to register Viewdata as a trade mark - but were turned down!

The reason, apparently, is that because viewdata says exactly what it does, it must be generic, and you can't register a generic term as a trade mark! Thus Prestel was born and we could call our TOPIC (Teletext Output Price Information Computer) system a viewdata system! (The TOPIC name stemmed from an existing system to which it was essentially a bolt-on extra and that was known as EPIC - Exchange Price Information Computer)

Our system differed from Prestel in a number of ways. Firstly it primarily operated over private circuits (with dial-up as a back up) and was also much faster in the City area because we used 9600bps baseband modems - 8 times faster than the 1200bps dial-up speed - but these required DC circuits to operate and circuits which extended beyond BT's City boundary were AC coupled. However, the ability to refresh a page in a second or less, dependent on the amount of data, helped it compete alongside the 22 channel TV system which had been operating for 10 years and which had 'instant' channel change via a push button selector.

We also added a few extra wrinkles to the standard specification by using the CAN code, for exampler - CAN means delete to the end of the row, carriage return & line feed, all with the one character!

This was particularly useful on the price pages because each line consisted of Stock Name, previous day's closing price and then changes occurring during the day. Many stocks trade at quite low levels, so those rows had a lot of blank space on the right.

It also completely confused any decoder following the Viewdata Specification, in addition to a different password structure, this making it difficult to attempt to high-jack the service for free!

[peter scott]

Hi Alan,

Thanks for relating the story. It's really nice to hear of such a system nearly up and running again. I worked in Hewlett-Packard's R&D lab in South Queensferry for most of my working life and when I first started in the early 1970s the single lab computer was a rather similar machine, an HP 2116. We also had an HP 9100A programmable scientific calculator on a trolley shared by about 70 engineers. It had a flag pole mounted on the trolley so you could see where it was in the open plan office. It too had a core memory holding up to 50 register operations. I subsequently acquired an HP9100A, currently gathering dust in my attick and no doubt still remembering what I last programmed it to do despite having not been played with for a decade or so.

Peter