- In 1964 there were still companies who would employ Technical
Students, ie allow an 18year old to join the company on a salary, and
then allow then to do a University degree!. The degree course I went on
was a thick sandwich course, 4 years long, 6 months at University and 6
months back at the Steel works. The subject being Electrical
Engineering, the business requirements being the possible conversion of
the Steel making process from oil based furnaces to electric arc. My
final year subjects were Digital Computing and Control Theory. The
first being about the electronics, and electronic binary gates that are
the heart of the von Newman computer architecture, the second being
mathematical representations of real world control, 3 term controllers
using analogue computers being some of it..
- After graduation, back at the Steel works, the investment in
Electric Arc Furnaces never happened. However management opportunities
did. With my final role being the maintenance manger for the Hot Steel
Rolling Mills. This required supervision of 7 x24 hour service and
maintenance crews, as well as many large equipment upgrades. To
re-shape a 7 ton hot ingot to a 3” dia bar took two 4,500HP
electric motors. Over the 10 year period several automation projects
were instigated, including a computer based measuring system, based on
a Digital PDP 8.
- With the ‘retraining’ packages available to
redundant Steel workers I was able to complete an MSc in Power
Electronics. Computer architecture had not changed. With the advent of
the ‘microprocessor’ they had just got smaller. One
technology I was behind on was the electronic speed control of electric
motor. High powered semiconductors were now making much
‘better’ variable speed controllers.
- One of the companies on the ‘milk round’ visits
was British Leyland. They had just invested in a highly automated
manufacturing facility at their Longbridge site. I joined the new
Control Engineering department, which was responsible for providing
second line maintenance support to the site electrical maintenance
service. I was now among several hundred robots and electronic
controllers, PLCs (Programmable Electronic Controllers), which had
replaced the electro mechanical control systems previously used. Some
of the more difficult challenges were in some micro computer controlled
facilities. An automated Panel Store (for the steel panels that the
robots weld together to make the car body). A painted Body Store, which
could re=order the sequence that the Painted, but not yet assembled,
vehicle body could be introduced to the assembly line. Both these
stores were controlling High Bay Stacker Cranes which received
positional instructions from a 16bit computer. Each store had a Digital
PDP 11/34 computer holding the store map and required sequence of
output.
- There was the first vehicle electrical test system, based on a
Ferranti computer. These were able to take many analogue signals and
compare the readings while the operator switched the electrical systems
on an off in a set programmed sequence.
The Control Engineering department then merged with the site
Electronics Department. They had several ‘home built’
control systems around the factory. They also provided second line
support to the site Electronic technicians. Electronics, from valves to
transistors and then microprocessors, supported many measuring
functioning the Engine and Gearbox factories, as well as proving the
support for the Private Voice Telephone exchanges (PABX). These private
exchange were all electro mechanical , strowger, technology, which
required very regular maintenance.
- With the enormous investment in the new metro vehicle (some
£250M at 1980 prices) the IT department had introduced some
‘manufacturing systems’ to control vehicle assembly
sequence (orders to the painted Body Store) and on site ‘line
broadcasting’. Line broadcasting used PDP 11s to send real time
print outs to all sequence areas in the factory. This meant engines,
seats, wheels and tyres etc, were placed on conveyors , in the right
sequence to match up with the body, and order, the were required for.
This facility used a CATV ‘Broadband’ cable distribution
system for its data communications. In 1980 computers were all RS232
not Ethernet connected. Over the years the Cable Broadband expanded to
run a sort of Ethernet service as well as Ascii terminal devices.
- In 1980 all the design engineering had been done on drawing
boards. In 1982 the first ‘CAD’ stations were introduced.
Our department became involved in the installation for those on
Longbridge site. These were sill standalone 16 bit computers with
arithmetic processors and graphic (mono) screens. Each had a many
copper cables running from the computer to the terminals. By 1985 our
first Ethernet connection was installed to allow two physical
locations, some 400mtrs apart, to pass files between them. The first,
of several hundred, fibre optic cables was installed.
- Other computer systems introduced included engine test beds, to
set the now obligatory electronic control and test the engine
‘worked’ before it as put into the vehicle.
- As time went by, I moved from being a technician, to being a
manager of technicians. By 1986 the department was 25 strong, with 6
technicians and 3 control engineers on 24 hour shift patterns.
- During the 1980s industry was very focused on ‘lean
manufacturing’. The Japanese were becoming much more efficient
than their ‘Western’ counterparts. One of the themes was
trying to achieve a batch size of one. This meant a factory could
produce any number of variants in any order and be as efficient as line
production. Engine parts were machined on large sequential process
transfer lines. Tool changes were lengthy, and resetting the line for
slight variations in the product was also a lengthy process. The first
‘FMS’ (Flexible manufacturing System) was installed for
prototype and pre production engine parts. Link to the 3D CADD tools,
it was possible to engineer a component, create the NC tool
instructions, download then to the CNC machine, and create your part.
In practice this was rarely done. The CNC operators were more adept at
programming manually, or editing the CADD tape, than relying on
complete automation. However it did produce a very good example of
using technology for a valuable business benefit. Prototype and
preproduction parts were always very expensive, and more importantly,
long in time. Some 12 weeks wait for a component was not unusual. The
same parts could be made in days.
- The second and third facilities were on a much grander scale.
Aimed at real production components. Cylinder heads for a <1000 a
week engine, and Brake Discs for several 100 a week. The one system
aimed at machine rectangles on 3 axis CNC machines, the other aimed at
round objects on CNC lathes.
Each facility had fully automatic material handling (Auto Guided
Vehicles, AGVs) t least one shift of ‘lights out’ working.
The complexity is all these facilities was the management and control
software. A computer system that gave real time instructions to the
machines and material handling. Schedule tool changes, waste clearance,
and, the complicated bit, a production schedule to maximise the
throughput.
Commissioning complicated software while production engineers were
trying to commission the machines provided most of the challenges. In
the end the systems did ‘work’ but not with the complete
automation and integrations that was envisaged. People were still
needed!
- While under the ownership of the Government of the day,
investment decisions would be made in the cabinet. One of these was the
development of whole new range of Aluminium Engines. The
‘Government’ wanted the company to buy its engines from
Honda, who were becoming a close partner in vehicle design. However
these would cost a great deal more than producing in house. Equally the
new engine design would be very cost effective in its manufacture.
Sense prevailed and some £200m was ‘allowed’ for a
complete new engine and associated gearbox facilities.
More details can be found at:-
http://www.austin-rover.co.uk/enginekseriesf.htmBoth
facilities were to be strongly influenced by the pervious FMS
facilities. Initial thoughts were to just scale up the designs used for
the first two production facilities. This however proved some very
costly numbers. For the engines cylinder head and block facilities a
more traditional approach was taken. The ‘transfer line’
approach was still the most cost effective. However CNC machines were
introduced at strategic locations in the line. This enabled changes to
‘bolt-ns’ and other more variable features to be easily
made. The gearbox facility did however use many CNC machines, even
though they were laid out between more conventional material handling.
The other major change was no central control system. Scheduling would
be done at a ‘simpler’ level.
- The input form the Manufacturing Systems Department was to
provide a central monitoring system for faults and production figures.
The final assembly lines were also provided with a ‘batch’
type schedule control system. This could be loaded up with the next few
hours production, and indicate o the various machines and operators
which variant they were to make.
The other big input was into the size and working of the material
handling systems. With very different tool change times and set up
times to swap derivatives, ensuring all machines could work to their
peak efficiency off line simulation tool ( Witness http://www.showflow.com/default2.htm is a similar modern product) was extensively used.
- As with all businesses re-organisations were a continuing theme.
Throughout my working life there has been a major redundancy programme
running at least once every 3 years. Also the IT function within Rover
was evolving. In 1980 the then ‘central IT’ function was
formed into a separate holding company. This was finally ‘sold
off’ to an in-house management buyout. This group of IT
specialists provided most of the programmers and architects for the
business applications. Within ‘Rover’ there also grew
‘in house’ IT function, mostly business analysts, but also
technical specialist.
IT strategy departments were created, the Engineering, Manufacturing,
Personnel, Finance, Sales, Marketing , Service, Warranty all had
separate IT functions at one time or another.
As IT became more and more central to many Business Functions,
reorganisations, mergers and acquisitions within IT were inevitable.
Another major feature of ‘Rover’ IT was the outsource model
it used / acquired. The ‘sold off’ function continued to
flourish and eventually was purchased by AT&T. Then some of it
moved to the BMW software house Softlab, and then some of them moved to
CSC. However, more or less, the same people who new each other over the
years carried on developing and supporting the systems regardless of
the company providing their salary.
- From and IT Infrastructure view there evolved 3 different
technologies. Engineering based on SUN and Solaris, Manufacturing based
on Digital OpenVMS, and Sales and Marketing moving towards an HP UX
environment, but still having a broad collection of other systems ,
STRATUS, HP 3000s, Mainframe, Lotus Notes on Windows NT. The Mainframe
did provide many services to other functions, it was
‘atypical’ of mainframe services, reasonably reliable, but
very expensive, and, very inflexible. Several images existed, but it
was only on Sunday afternoons that changes could be implemented, and
only a few on each occasion. It could take 12 weeks to have some
changes implemented. Hence, over a period, this became the main
architecture to be ‘removed’.
Office services faired a little better. IT strategy departments had
funded a Digital All-in-One email system, and voice and data networks
had a single management structure for all intersite connectivity.
PC networking had the benefit of the IT Strategy Group who facilitated
the ‘selection’ of the PC networking OS for all the
company. So Netware 3.1 was introduced. Along with the site to site
network and the affordability of the IBM PC, this service exploded.
Concepts like ‘locked down desktops, user identification and
viruses were all ‘learnt’ on the way. With Microsoft
joining the party with NT, there was the never ending discussion over
how much longer NetWare would last. And, as to be expected NT4
eventually replaced it.
Wide Area Networking started with a few the centralising of the
mainframe service to Redditch, and hence the need for many 3270 type
connections. These were initially provided by individual BT lines.
However a more modern Multiplex solution was introduced to solve the
management headache.
The ‘real’ WAN infrastructure was installed for
Engineering. With the introduction of Computer Aided Design, the
software evolved to mostly run of UNIX platforms. Hence Engineers would
have a SUN workstation of their desk long before PC’s appeared.
The challenge for IT was the management of these CADD files. Each
should have a centrally issued part number, each should be associated
with a component design within a project to be associated with a
vehicle programme. Engineering Data Managers or Parts Data Managers,
EDM and PDM, were introduced to provide some control over the engineer.
A central repository for all engineers, rather than replications on
each site, was chosen as the preferred solution. The UNIX workstations
existed on several sites around the Midlands. Longbridge and Solihull
in Birmingham, Canley in Coventry, Gaydon in Warwickshire, Swindon and
Cowley in Oxford. An office on the Warwick University campus also
‘appeared’. The first solution consisted of 2Mb/s circuits,
3 per site, with at least one of the circuits being provided by BT or
Mercury. BT via the ground and Mercury via Microwave links. These were
terminated on Infotron LanSpan Ethernet bridges. Routers could have
been used, but these bridges were far less expensive, and would load
share down the 3 links, achieving all that a router could do.
This WAN service quickly took on all the other departments WAN
connections. All ASCII Terminal connections (usually via terminal
servers with a LAN connection) and the embryo PC NetWare service. After
an incident on Longbridge site where the manufacturing of vehicles was
disrupted due to a faulty terminal server (not used by manufacturing),
an new design was introduced. To keep the manufacturing service
‘away’ from the ever more complicated business world, a
separate WAN was introduced to connect the manufacturing sites. This,
now, could use the much more affordable Router technology. Over this
time Data Centre consolidation had moved to two main computer rooms, on
on Longbridge site the other at Solihull. The then Birmingham Cable
company (now Telewest) provided ‘virtually’ dark fibre
services between the two site. This allowed 2 FDDI rings to be
installed. The WAN design then switched to Router technology, but all
circuits coming from either Solihull and Longbridge, with at least one
circuit coming form each site.
- From 1990 – 1995 the Structured wiring explosion occurred.
From an initial 300 ordered in 1990 some 8,000 existed 5 years later.
Not all UTP outlets were in service, thank goodness, but it indicated
the swift move from text based IT systems, to windows based, and, more
importantly, personal productivity tools, the PC, Word Processing,
Spreadsheets and Email.
More things to say
Data Centers why were they were they were, CDC etc.
Influence of Honda
Influence of BMW
Quality process, Belbin and Baldridge. Rover Tommorow, IT Operations
and outsourcing
Back to Top