Happy Birthday LEO

[RichardS]

LEO, probably the World's first business computer first ran on 17th November 1951.

It was (of course) British.

The Lyons Electronic Office (LEO) was originally built to help schedule work and deliveries for the (once) highly successful Lyons Tea Shops.

http://www.telegraph.co.uk/technolog...puter-age.html

But somehow, Britain's once world-beating computer industry became stale and crumbled.

[jacksjacobs]

Originally Posted by RichardS LEO, probably the World's first business computer first ran on 17th November 1951. It was (of course) British. The Lyons Electronic Office (LEO) was originally built to help schedule work and deliveries for the (once) highly successful Lyons Tea Shops. http://www.telegraph.co.uk/technolog...puter-age.html But somehow, Britain's once world-beating computer industry became stale and crumbled.

LEO used ultrasonic mercury delay lines as a data storage medium.
ATC Secondary Radar certainly did, storing a single scan for comparison with the next scan to cancel images the same in both - hence leaving only the moving targets visible.
I contracted a very heavy mercury count once, when working on them.
Delay lines were contained in heavy steel tubes. I found the big delay lines heavy to carry one at a time, but there was a young woman, in my job before me, who would pick up two at a time. Wow.

I seem to remember in Asimov's Caves of Steel the protagonist, Elijah Bailey (?) passes through the room where there were tanks of mercury used as data stores for the computers used in the city.

The technology of those days used such varied devices as CRTs, ultrasonic delay lines, and magnetic cores to store data - not counting magnetic film drums and discs storing maybe a kilobyte of data. When these devices became obsolete, replaced by solid state electronic devices, manufacture of these became the province of American, and Far East companies. We still made mainframes for a while - but who could compare with the USA for investment in development.

[RichardS]

Originally Posted by jacksjacobs LEO used ultrasonic mercury delay lines as a data storage medium. ATC Secondary Radar certainly did, storing a single scan for comparison with the next scan to cancel images the same in both - hence leaving only the moving targets visible. I contracted a very heavy mercury count once, when working on them. Delay lines were contained in heavy steel tubes. I found the big delay lines heavy to carry one at a time, but there was a young woman, in my job before me, who would pick up two at a time. Wow. I seem to remember in Asimov's Caves of Steel the protagonist, Elijah Bailey (?) passes through the room where there were tanks of mercury used as data stores for the computers used in the city. The technology of those days used such varied devices as CRTs, ultrasonic delay lines, and magnetic cores to store data - not counting magnetic film drums and discs storing maybe a kilobyte of data. When these devices became obsolete, replaced by solid state electronic devices, manufacture of these became the province of American, and Far East companies. We still made mainframes for a while - but who could compare with the USA for investment in development.

Mercury delay lines apparently provided a handy option: A separate transducer could be fitted some distance before the end, to give prior information about the data. I was involved in a proposed scheme to recreate Pilot ACE which also used these mercury delay lines.

Later, the knowledge was useful while creating a special machine which used electronic shift registers for a similar purpose: Long before microprocessors became common, RAM was far too expensive.

Somehow, the British designed ARM based family of CPUs now outsells even Intel CPUs.

There are also some great British software companies.

But, we once had a greater lead.

[jacksjacobs]

Originally Posted by RichardS ...Mercury delay lines apparently provided a handy option: A separate transducer could be fitted some distance before the end, to give prior information about the data. I was involved in a proposed scheme to recreate Pilot ACE which also used these mercury delay lines. ...

I spent a couple of years in the late 1950s developing nickel wire type ultrasonic delay lines. These were used in IFF applications, for ATC.

Deploying several receiver transducers along a line meant that a corresponding pulse train could be decoded using a multiple input and gate to sense the simultaneous reception of the pulses at all transducers. The first pulse of the train would reach the last receiver just as the last pulse reached the first receiver. Only if the (antenna) received pulse train matched all the line receiver positions would you get an output from the line circuit.
A coded pulse train could be similarly generated, with spaced transducers each feeding an or gate to the output. As the transmitted pulse reached each receiver, a pulse would be fed to the output.

Reflections from the ends of the wires were the bugbear of the device, and assembly of the line and its end damping "pads" was an art. Not an easy technology.
The wires, actually we used fine tubes, had to be heat treated as the ultrasonic temperature dependant characteristics were modified in drawing the tube. We used a long quartz tube furnace with a hydrogen reducing atmosphere to anneal the tubes, and adjusting the annealing temperature profile was the big problem to be solved.

We measured the delay time temperature coefficient in a little laboratory oven. It didn't have refrigeration, so we put a tray of dry ice in the top, and did the measurements as it warmed up, turning the heater on as we went through room temperature.

It was quite a dull task. It was part of the "apprenticeship" to the profession, and I believe it was a necessary part of my engineering education. Helping to design spacecraft is at the other end of the profession, but I am grateful for the experience leading to that elevated activity.

[Meg]

There's probably some previously unwritten history being recorded here!

[RichardS]

Originally Posted by jacksjacobs I spent a couple of years in the late 1950s developing nickel wire type ultrasonic delay lines. These were used in IFF applications, for ATC.

Thanks. I'd not come across that type of nickel wire delay line. Some later versions are shown here and on the links... together with "retro" fashions.

http://en.wikipedia.org/wiki/Delay_l...ve_delay_lines

[jacksjacobs]

Originally Posted by RichardS Thanks. I'd not come across that type of nickel wire delay line. Some later versions are shown here and on the links... together with "retro" fashions. http://en.wikipedia.org/wiki/Delay_l...ve_delay_lines

For long delays, a coiled wire in a torsion mode is probably more appropriate.

However, for the short delays relevant to IFF pulse coded words, both receiving and transmitting, a nickel alloy tube about 6 inches long, with a longitudinal compression wave is appropriate.

Transducers are about an inch or so apart. The pulse width is adjusted to obtain a receive signal in a bipolar triplet form, say, a low amplitude -ve going pulse preceding and following the main pulse, +ve going, combining the two +ve going pulses as the compression pulse passes through the transducer front and back end.

The torsion mode wire long line can be easily damped by foam pads, and I would think is an easier device to make and calibrate.

For our short tube lines, we had to fit a tapered soft concentric slug. Pushing the taper onto the tube, taper first was an art. If you greased it, it slipped on, but there wasn't then enough friction to damp the waves reflected from the hard-mounted tube ends.

Much of the device was stainless steel. involving many 8 BA and some 10 BA threads. Workers were reduced to tears looking upon a pile of broken 10 BA taps, our design was probably faulty in having such small threads applied to long holes.

We used to magnetise tiny ferrite doughnut cups using a wire through the central hole. A huge Ignition valve was used to fire the magnetising pulse, and I could tell if it was in use well down the factory by hearing the mains wires "clicking" in their trunking.

Having described so much. I should state the work was done at Cossor Radar, Harlow Essex. Before it became Raytheon.

[RichardS]

Originally Posted by jacksjacobs Having described so much. I should state the work was done at Cossor Radar, Harlow Essex. Before it became Raytheon.

Thanks. Cossor, another once great British company.

[jacksjacobs]

Originally Posted by RichardS Thanks. Cossor, another once great British company.

If you watched Dr No you might have recognised two rows of Cossor 1035 scopes. All with Lissajou squiggles, I think.

Apart from using them as the standard at the Northern Polytechnic - round the corner from AC Cossor, in Highbury - I had one at home. Time base wasn't linear - never could correct it. They used a gas filled valve in the time base - ancient technology even then!

Those funny things spinning at the airports - they were Cossor radars, at the time.

[RichardS]

Originally Posted by jacksjacobs If you watched Dr No you might have recognised two rows of Cossor 1035 scopes. All with Lissajou squiggles, I think. Apart from using them as the standard at the Northern Polytechnic - round the corner from AC Cossor, in Highbury - I had one at home. Time base wasn't linear - never could correct it. They used a gas filled valve in the time base - ancient technology even then!

I think it was a 1035 'scope I considered buying when a certain Portsmouth Airport based company was disposing of them. But they were rather too battered.

So, I have only my simpler Cossor 339a 'scope.

[leo]

Gee, thanks

[RichardS]

Originally Posted by leo Gee, thanks

Glad you enjoyed it.

BTW. Why do those signs on the M4 motorway for "Reading Services" make me think of friendly libraries?

[jacksjacobs]

Originally Posted by RichardS I think it was a 1035 'scope I considered buying when a certain Portsmouth Airport based company was disposing of them. But they were rather too battered. So, I have only my simpler Cossor 339a 'scope.

Have you looked at http://www.thevalvepage.com/testeq/scope.htm ?

I am reminded of many CROs that I have used and owned.
The scopes in Dr No may have been 1035s, but the one I owned, from what I can see on that site must have been the 1049. The 1035 used miniature valves and had a Puckle time base. The scope I had used old fashioned big bottles. I think it did have the big panel on the side.

I once owned a miniature Cossor 1039M scope, and even a home made version I bought from a Polish engineer back in 1957 - I still have parts of that.

I had a Hartley 13A, I bought from STL in Harlow. That had a lovely fine trace, and a time calibration that put a beautiful damped pulse on the screen.

I also had at one time a Solatron CD1014 dual trace scope, and I still have a coffin-like long Tektronix dual trace portable scope.

While in memory mode, I remember converting a Loran unit to a scope. That had a 5 inch CRT - that was something! Playing with a Loran unit in a laboratory while doing vacation work is what set me onto a predilection for digital electronics.

[RichardS]

Originally Posted by jacksjacobs I am reminded of many CROs that I have used and owned.

Thanks for that link. My Cossor 339a has its slot-in graticule, and its manual.

Looking at those photos, the company's redundant 'scopes might have been Cossor 1049s rather than 1035s... or might even have been a mixture. But they were in poor condition. This was probably in the late 1970s or very early 1980s.

In the labs, we usually used those large Tektronix units, possibly the 545s, with whatever plugins and probes we could scrounge.

During one particularly hot spell - while the highly paid astronauts in Skylab - were whinging about the heat - I remember being bathed in the heat from the fans of three of those hot beasts.
-----

Returning gently towards delay lines:

The shift registers storing data in that special equipment needed a precise number of different length clock pulses in order to maintain integrity and synch. It was vital that there were no extra spikes etc. It took me quite some time to understand the problem, then to calculate exactly how many pulses were necessary and in what sequences. Eventually, I came up with a theory, built the electronic circuits and needed to test my theory... using only the meagre test equipment available.

I ended up having to measure & count the precisely 4332 pulses (or whatever the number was) by peering at the waveform on the 5 inch display on one of those Tektronix 'scopes.

Not something I'd recommend if you're to remain sane!

[jacksjacobs]

Originally Posted by RichardS .... Returning gently towards delay lines: ...I ended up having to measure & count the precisely 4332 pulses (or whatever the number was) by peering at the waveform on the 5 inch display on one of those Tektronix 'scopes. Not something I'd recommend if you're to remain sane!

I was working on thin film stores for some time at STL Harlow in about 1964, and to check whether a store location would hold its data while being nudged by contrary pulses in adjacent locations, I had to look for a 2 nano second pulse every several seconds. This with a state of the art HP scope, but with bright sunlight in the lab. I suffered at the time with hay fever, and my eyes really suffered.

Could it have been a sampling scope for that task - I can't remember. STL was well funded. One occasion, I was told to immediately spend some thousands of pounds (equivalent at least) as an end of budget year spend, so we looked through the HP catalogue - remember them? I ordered a beautiful nano second pulse generator and a sampling scope. That equipment still has me drooling at the thought of it. Old hat now!

I think I have mentioned this before. As a young engineer, I used to read the HP advertisements, seeing their address at Page Mill Road, Palo Alto, in the SF Bay area, wishing I could be there in the fabulous USA. Many years later, on a secondment to a firm in the locality for a few months, I used to regularly go along Page Mill Road, pondering that you never can tell what will happen in life.

One sometimes envies, or used to, life in the USA - it was less austere than here in the UK for so many years after war. But I now understand that it would take the holiday of a lifetime for the average US citizen to come to Europe, with the shorter annual leave allowance common in the US, and the cost. That's why we meet mainly older Americans when we visit the places of interest in France to Americans - especially the Normandy beaches.

Having been in the space business, I have visited many European countries, and several outside, and it has made for a most interesting life. I could never expected have that as a child. The continent was cut off by an iron wall when I was young, and interestingly, it was only when we visited Calais that we realised how visible the "White Cliffs of Dover" are from there.

[RichardS]

Originally Posted by jacksjacobs I was working on thin film stores for some time at STL Harlow in about 1964...

There used to be contests between designers, to slip personal signatures past the managers and inspectors... rather like Slartibartfast and his signature fjords on planets.

Some of these later figured in court cases about pirated designs, helping to prove which company actually owned the design.

Some designers created individual registration targets for aligning the different layers of the photographic processing masks... like the registration targets in PagePlus.

The result could only be seen using a high-powered microscope, and only before the integrated-circuit or electronic module was packaged. But the designer knew that his signature was there.

Anyway...

I saw one design for an electronic delay line which when viewed through a microscope looked like a toy railway track snaking around the chip.

So, the designer had managed to print a wild-west loco puffing around this track... and his design still worked.