This article recreated from way back machine, its no longer available at the museum of Scotland

During the 1970s, John Ames helped to upgrade the telephone exchange system in rural Grampian communities. Thirty years later, he visited the National Museum of Scotland and was surprised to see a very familiar piece of technology – the Glenkindie telephone exchange! Here he tells us more about working on the project. 

I had never imagined that I would one day become a museum-piece, but on a trip to the National Museum of Scotland one day a few years ago I came very close!  I was wandering around the technology area of the museum and was attracted, because of my background, to the telecoms exhibit. As I rounded the corner of a display case I was suddenly transported back to my time over 30 years ago when I worked in a laboratory at BT’s Research and Development centre at Martlesham Heath in Suffolk. There, prominently displayed in a large cabinet, was a piece of equipment that I had helped to develop.

I started work at BT Labs after graduating in 1972 and joined a team doing research into what was then the new technology of digital telephony. After a few years we moved more towards the development of demonstrators and prototypes and here was one right in front of my face: a small digital exchange that we had developed for use in rural communities.

The UXD5 (Unit Exchange, Digital, No 5) at Glenkindie in Aberdeenshire was the first digital local telephone exchange brought into public service in the UK. It was intended to replace ageing small electromechanical exchanges, serving between 10 and 150 customers often situated in remote communities, especially in Scotland.

A completely new line interface had to be designed to feed the current needed to work the subscriber’s telephone and to survive a lightning strike, this in the very early days of the evolution of the Subscriber Line Interface Circuit (SLIC). In effect, BT Labs had to develop a SLIC for the most difficult lines before the technology had been proven on some of the shorter and better-protected lines.

Whilst Monarch was very reliable, it did not have the resilience needed for public service, where the telephone can literally be a lifeline in an emergency. The exchange is operated by large software programmes running in the exchange processor: BT Labs duplicated the hardware and software and developed the means to detect if a processor was misbehaving and to hand control over to its partner whilst also raising an alarm in the remote management centre.

The UXD5 also needed to provide information to bill subscribers and this is where I first came into the picture. I developed software to run in the exchange processors to monitor the destination and duration of every call. It would also send signals to operate mechanical meters, one per subscriber, that tick round at a rate depending on the type of call. This was an old way of charging but the first UXD5s had to be backwards-compatible with the old exchanges where the meters were photographed every quarter to give the information to work out the bill. I also developed the hardware and software that interpreted the signals and worked the meters.

Later I went on to manage a group of people who were developing the SLIC circuits; this led me to visit the Glenkindie site several times with members of the team to make electrical measurements of the variety of subscriber’s lines to inform the development process. 

Manufacturing of the UXD5 by BT suppliers was organised by another part of the company and the UXD5 went on to be a huge success, serving smaller communities throughout the UK. It has proved remarkably flexible and a small team at BT Labs continued to add features and services to the exchanges as technology and subscribers’ needs evolved.

Two units from the Glenkindie telephone exchange will be on display in our new Communicate gallery, which you’ll be able to visit from Summer 2016.

- Posted 9th December 2015

• Peter 19th September 2016
  Any names for the other 4 in the picture?Reply

  • PETER SMITH 28th August 2017
    Names in Picture
    Left: Elsden
    3rd from left: Pete Clark
    Reply

• NIgel Dyer 22nd December 2018
  I was fascinated to stumble across this. I also worked on the UXD5 with John Ames. I particularly remember travelling up to work on the UXD5 in Kinlochard where there was a problem which we believed to be caused by earth currents from a local hydroelectric power station interfering with the meters on the exchange.
  Reply

• Tim Regester 11th May 2019
  I spent 3 months working in the UXD5 team at High Holborn in 1985. I recall the mod to the circuits where tippex was applied to all the VRAM chips so they wouldn’t get wiped by the flash from Photographers when the Press came in for publicity.
  Reply

• Steve Reynolds 10th August 2021
  I have worked at Glenkindie many times. I work on AXE10, System X and UXD5B. The UXD5B exchanges are the most reliable of the three and very easy to work on.
  Reply

• Keith Wilson 5th April 2022
  As an exchange Construction T.O, part of one group based in Aberdeen, I recovered the UXD5A cabinets at Glenkindie and replaced it with the (current ?) UXD5B.
  With one of the UXD5 support group, Doug Savage.
  A ‘baby’ UXD5B supplied by Doug Savage (I remember a ‘half height cabinet’ self contained unit with Control shelves & one 96 port shelf (and a 32 port for Junction cards)) was used as an intermediary to which all the ‘Subs’ transferred, allowing the 5A cabinets to be recovered.
  The UXD5B cabinets were then installed, cabled up, data loaded, and the Subs transferred from the ‘Baby UXD5B’ on to the (then) new UXD5B exchange unit.
  I spotted the UXD5A at the museum in March 2020 just prior to the first Covid ‘Lockdown’
  In the photograph, 2nd left (brown suit) is John Pike (Planning ?), and far right, in front of gate is Graham Gerrard (maintenance T.O).
  Reply

Document supplied by Museum - The UDX5 Glenkindie Exchange John Ames 27.10.15

September 2025

"I have also attached a Word document with some further information that is not in the blog. Unfortunately, I cannot find the images mentioned in the document that are not in the blog above."

The UDX5 Digital Telephone Exchange

In the mid-to-late sixties interest had grown in the possibility of switching telephone calls carried by the newly developed pulse code modulation (PCM) technique directly in the digital domain without decoding them back into their original analogue form.

At that time a PCM system could carry 24 calls over a copper pair originally designed to carry one call in its analogue form.  PCM transmission systems were being introduced to carry calls between telephone exchanges as this made more efficient use of the copper cables connecting those exchanges.

Originally, analogue calls would be mechanically switched at the originating exchange to cables heading for the next exchange on their journey, each call occupying one pair of wires in the cable.  With the introduction of PCM, 24 calls could be converted to digital form and carried over a single pair.  On arrival at the destination exchange, the calls were converted back into analogue form to be forwarded through more electro-mechanical switches to the next step in its journey, possibly via another PCM link.

Research focused initially on the development of a digital ‘tandem’ exchange to eliminate the need to convert calls back to analogue form for switching.  In a tandem exchange calls arrive in PCM form and are routed to the next exchange on their journey also via a PCM system.  Conversion back into analogue form for electro-mechanical switching in the tandem added cost and also degraded the quality of the speech slightly.  The research culminated in BT’s development of an experimental digital tandem exchange that was installed in the ‘Empress’ tandem switching centre in London, opening for service on 11 September 1968.

Conversion from analogue to digital form was initially very expensive which is why early research focused on switching inter-exchange traffic where a single link carried many calls.  The next stage was to convert calls into digital form on a per-subscriber basis as they arrive at the edge of their local exchange so that they could be switched digitally for the whole of their journey.

BT research focused on cost reduction of the coding and decoding (‘codec’) function and worked closely with its suppliers (then GEC, Plessey and STC) to develop a digital local public exchange switching system called System X for national roll-out.

Around the same time BT also worked on development of a small digital private automatic branch exchange (PABX), named internally as CDSS1 (Customer Digital Switching System No 1).  A PABX is similar in function to a public exchange but is technically much simpler.  The CDSS1 was launched in 1980 as the ‘Monarch’ exchange, a small PABX serving up to a few hundred lines that was very popular with small companies.

Early in the development of System X BT realised that it would not, at that stage, provide a cost effective solution for very small exchanges such as are found in the Scottish highlands.  These exchanges can serve as little as a few tens of subscribers to a maximum of around 100 and were known as Unit Automatic Exchanges (UAX).  Several variants of UAX existed with different maximum capacities.  This development focused on the small UAX5 that was commonly found in small highland communities: UAX5 was a Strowger mechanical exchange with a minimum economic capacity of 10 lines and maximum of around 100.

BT decided to design in-house a digital replacement for the UAX5 that became designated the UXD5 (Unit Exchange Digital No 5), based closely on the CDSS1 design.  Whilst the UXD5 was based on the CDSS1 considerable development was needed to make it suitable for public exchange use:

• The one-per-line subscriber’s line interface circuits that feed current to the line and convert the analogue signal to digital had to be modified to handle longer lines and to resist lightning damage.

• The central computer that operated the exchange (the exchange processor) had to be duplicated to enhance reliability.

• Extra programme storage memory had to be added to accommodate more complex programmes.

• Calls had to be timed and charged for.

• Electronic circuits to connect the exchange to the analogue lines (‘junctions’) running to other exchanges had to be developed.

• Many other smaller adaptations.

My Role in UXD5 Development

I was initially member of the small team that developed much of the electronic hardware; another team was responsible for the software.

Firstly I was assigned to develop the charging capability.  The UAX5 had one mechanical meter per customer: as a call progressed the digits on the meter were incremented at a rate determined by the appropriate charge for the call.  The meters were arranged in a closely packed array: once a calendar quarter someone visited the exchange and photographed the meter array; back at the headquarters the previous quarter’s reading was subtracted from the latest one and a bill generated based on the number of ‘units’ used.

The UXD5 had to be completely compatible with the UAX5 so I developed software that ran in the main processor to monitor calls and to generate messages to increment the meters; I also developed an electronic card that took those messages and converted them into 50V electrical impulses to operate the appropriate meters.

I also developed the exchange status system: this controlled a set of lights and switches set into a panel that informed technical personnel about the status of the exchange and also allowed them to conduct simple tests.

Later I moved to run a specialist group concerned with research and development of subscriber line interface circuits: this group was, amongst other things, charged with the responsibility of developing the modifications to the CDSS1 line circuits and also developing the interface electronics to communicate with other exchanges.  This work needed an understanding of the complex electrical characteristics of the long subscribers’ lines found in Scotland and I made several visits with members of my team to the Kildrummy and Glenkindie UAX5 exchanges to make electrical measurements.

John Ames

25 October 2015

Photographs

The photographs show the interior and exterior of Glenkindie and Kildrummy UAX5 exchanges prior to installation of the UXD equipment.  We made several visits, some of them with local managers, to these exchanges to make practical arrangements as well as to make electrical measurements of the subscribers’ lines and the junction circuits.

Photographs at Glenkindie were taken on 28 October 1978, those from Kildrummy are from around the same period but are undated.

Photographs not specifically mentioned below are self-explanatory.

Glenkindie 2: John Hovell, an expert in subscriber line circuits.

Glenkindie 3: A team in London was responsible for arranging production of the UXD5 by BT’s suppliers; the figure on the right is Mike Elsden, head of that team, others are local managers from Scotland.

Glenkindie 4: Mike Elsden with members of his team and local managers; I am 4th from right.

Glenkindie 5: Interior of Glenkindie – I can be seen in the background and behind me is Dr Peter Deighton, the overall head of the development team.

Glenkindie 6: Showing the mixture of old (‘pre-2000’ in grey boxes) and recent (‘2000-type’) electromechanical equipment deployed in a typical UAX.

Glenkindie 7: The most remote subscriber from the Glenkindie exchange.

Glenkindie 8: On the way to the most remote subscriber.

Glenkindie 9: The local exchange technician carrying out a manoeuvre that would today not be permitted by the health-and-safety rules.

Kildrummy 2: Dr Peter Deighton and Mike Hill, a colleague on the development team.

Kildrummy 3: Dr Peter Deighton

Kildrummy 4: Mike Hill, John Hovell and Dr Peter Deighton with visiting local managers.

Kildrummy 5: A mixture of pre-2000 and 2000 type equipment at Kildrummy.

Kildrummy 6: The original subscriber meters at Kildrummy.

Kildrummy 7: The main distribution frame at Kildrummy where all the customers’ lines are terminated.

Kildrummy 8 and 9: The grounds of Kildrummy Castle Hotel where the team often stayed.  We were not allowed to claim for any alcohol on our hotel bills so we made an arrangement with the manager that the total alcohol bill would be spread amongst all the rooms and the final individual bills would simply show ‘accommodation’ and a total.