This text is reproduced solely for the limited academic use of students in MBA 665.
Numbers
in brackets indicate the start of a page in the original text.
November 16, 1994
This case was prepared as the basis for class discussion
rather than to illustrate either effective or ineffective handling of an
administrative situation. “Pilkington Float Glass — 1955” uses materials from “Pilkington
Float Glass (A)’ (672-069), “Note on the Flat Glass Industry -1955” (672-070),
and “Pilkington Brothers P.L.C.” (see references herein).
In April 1955, the Board of Pilkington Brothers faced a
critical decision with respect to its float glass project. The three small
scale pilot plants already constructed had cost $1.48 million;[1] while
the results were encouraging, many difficult problems remained unsolved. The
proposal before the board called for modification of a redundant glass furnace
at a cost of $1.96 million. The new line was to be a full scale production line
capable of producing a 2.5-meter wide glass ribbon.
The Board’s decision was critical to the fate of the
project. As inventor of the process, leader of the development team, and a
company director, Alastair Pilkington was in an
unusual position. Commenting on decision-making in the project, he noted the
attitude within the float glass team, and his role:
On the team’s attitude: We try to assess day-by-day
whether or not the whole project should be continued. We try not to feel either
discouraged because we are tired or have met a new fault, nor to feel that we
have just got to keep going because we have already spent so much effort. This
does not mean that each day we record that we have decided to continue, but it
is an attitude of mind which is important to maintain in spite of the sound of
crashing cullet in our ears day and night.[2]
On his role: As a scientist I have to be an optimist, but
as a director I have to make the sober financial judgments. A lot of people in
the company are highly skeptical.
In 1826, William Pilkington and two well-known glass
makers founded the St. Helens Crown Glass Company in Lancashire, England.[3]
St. Helens (11 miles east of Liverpool) was a logical site for glass making:
its rich coal deposits were a ready source of fuel, there were plentiful reserves
of sand, and there was limestone and dolomite nearby. In 1849, the company came
under the sole control of the Pilkington family. At that time William was
joined by his brother Richard, and the company was renamed Pilkington Brothers.
Although it began as one of many flat glass producers in Britain, it was almost
alone in surviving the industry shake-outs that followed – first the repeal of
import duties in the 1850s, and later the mechanization of the manufacturing
process, both of which inundated Britain with cheap continental glass imports.
Flat glass was a generic term covering sheet, plate and rolled plate glass. By
1903, Pilkington was Britain’s sole plate glass manufacturer; after World War I
it also emerged as the dominant sheet glass producer. The company was managed
as a private firm under the direct personal control of the family, with family
members occupying all top management positions and controlling all of the
shares. A strong tradition of secrecy governed the company’s affairs. The top
officers performed middle management duties rather than allow important
financial information to become common knowledge.
By 1955, Pilkington Brothers was one of the largest
private companies in the UK, employing over 23,500 people. It was the world’s
fourth largest flat glass producer, manufacturing window glass, plate and
rolled glass, glass tubing, TV tubes, optical glass, and fiber glass. In
addition to its domestic monopoly, Pilkington had a virtual monopoly position
in British Commonwealth markets (with factories in Canada, South Africa, New
Zealand, and India) and substantial interests in South America (with factories
in Brazil and Argentina). Pilkington also operated seven finishing plants for
plate glass in overseas locations, and maintained warehouses and sales agents
in most countries of the world. In 1955, about 81 billion each of sheet and
plate were produced worldwide. Projected growth in the flat glass market was
estimated to be 8% for the next several years, with most of the increase coming
from the post W.W.II reconstruction efforts in Europe and the growing demands
of the automobile industry.
While it demonstrated little initiative where sheet glass
was concerned, Pilkington devoted considerable attention to the development of
continuous process, mass production technology in the manufacture of plate
glass. (Prior to the 1920s, plate was manufactured by pouring pots of molten
glass onto a large iron table and rolling the glass into shape. After cooling,
the glass was cut and polished.) It was Ford that first demonstrated that plate
glass could be rolled continuously, but it was only through cooperation with
Pilkington that continuous plate became a reality in the early 1920s.
Pilkington contributed continuous grinding (one side at a time) and applied its
substantial knowledge of furnace technology to develop the tank furnace from
which molten glass could flow continuously to the rollers. In 1923, Pilkington
installed the world’s first continuous plate manufacturing process at St.
Helens.[4]
Recognizing the importance of research to its future
prosperity, Pilkington first set up a technical committee in 1933 to focus on
cost-saving innovations. In 1936 a Technical Department was formed under the
committee and in 1938 a central research lab was completed. Even after the
technical committee was established each individual Pilkington plant retained
considerable responsibility for process development since it was impractical to
recreate in the laboratory the conditions inside a 420 square meter glass tank,
heated to 1480-1560° C.
Simultaneous with the new organizational emphasis on
research came the second major plate technology innovation in as many decades,
the “twin grinder,” which made the plate process fully continuous. The machine
gave Pilkington world technological leadership in the manufacture of quality
flat glass.[5] The twin
grinder, and somewhat later the twin polisher, made it possible for the first
time to grind the rough ribbon of glass produced by continuous rolling on both
sides at once.
By improving precision and parallelism, it decreased
grinding waste and made possible thinner plate. It also helped to increase
economical plant size. Pilkington’s first twin grinder was set up in 1935.
Pilkington licensed its twin grinder to most of the world’s plate manufacturers
over the next 15 years.
In addition to developing its own proprietary processes,
Pilkington also purchased technology from abroad — specifically, from Owens
Corning, Corning Glass, and St. Gobain. The company
customarily improved and adapted the techniques it licensed from others. In
Pilkington’s view, the major inventions for which it has been responsible are
no more important than regular attention given to improving productivity in the
melting, manufacturing, warehousing, and handling of glass.”
Given its growing history of innovation, in the late 1940s
Pilkington was thought by industry experts to be the most likely of all major
plate manufacturers to bring out further major innovations. Pilkington was the
leader in plate technology, had large potential markets, and kept tight secrecy
around its operations. However, there were indications that competitors were
seeking to combine the optical qualities of plate with the fire polished surface
and economy of sheet. Pittsburgh Plate and Corning Glass in the US as well as
St. Gobain in France were all rumored to be working
on new flat glass processes by the late 1940s.
From a chemical perspective, glass was a giant matrix of
silica atoms in which were locked atoms of sodium, calcium, magnesium,
aluminum, and oxygen. The raw materials that provided these elements in the
manufacture of glass were primarily sand, soda, and limestone or chalk, which
were mixed with broken waste glass (“cullet”) in a melting furnace or “tank.”
Because the materials used in the tank’s construction were expensive and
because 14-21 days were needed to bring it up to full temperature, the glass
tank was run every day of the year for economy (see Exhibit 1 for process flow
diagrams, Appendix A for more process detail).
Sheet glass was made by drawing the viscous molten glass
vertically from the surface of the tank’s working end, and cooling the ribbon
to the point where its surface was hard before it came in contact with the
drawing rollers. The ribbon was cooled slowly (“annealed”) to relieve any
excessive stresses by passing the ribbon through an annealing oven, or “lehr,” which was essentially a large steel box comprised of
controlled temperature zones. The three sheet glass processes in use were Fourcault, Colburn, and Pittsburgh, which were essentially
variations on the same underlying concept (see Exhibit 2 ).[6]
This production method yielded good quality, cheap glass
suitable for domestic windows. Since the process relied purely on temperature
control and speed of draw to control the thickness of the sheet (the rollers
only provided traction), it generally produced a sheet that had some optical
distortion because of localized variations in thickness. However, because the
sheet was not touched by the rollers before it was hardened, its surface had a
glossy fire-finished appearance — a much sought after feature. Some of the
better quality sheet glass was used as side windows for automobiles. Pilkington
was one of 50-60 sheet glass producers worldwide. In 1955, the British company
manufactured about 27 million square meters of sheet glass at three locations:
St. Helens, Queensborough Kent, and Pontypool Wales, selling the product for an average price
of $.97/square meter.[7]
The worldwide market for sheet glass in the 1950s was approximately 1 billion
square meters.
Plate glass offered substantial optical improvements over
sheet. Plate was made by flowing molten glass horizontally from the furnace
between two rollers which caused surface markings on both sides of the ribbon
(see Exhibit 3). These markings were removed by grinding away 15-20% of the
annealed ribbon from both surfaces (see Exhibit 4). In this manner an optically
perfect ribbon was obtained, but at such a cost that the end consumer had to
pay up to eight times sheet glass prices. Irrespective of the rolling,
grinding, or polishing method used, the basic melting furnaces were the same as
those used in sheet glass manufacture. In addition to being more costly, the
plate process required additional equipment and space. Most plate glass lines
were about 760 meters long. Plate glass was used in automobile windshields,
mirrors, and commercial and prestige domestic windows. In 1955, Pilkington had
two plate lines in operation: CH3 at Cowley Hill and D5 at Doncaster.
While plate glass accounted for only 4.6 million square meters of sales, it
represented more than 50% of Pilkington’s profits, with prices averaging around
$6.35/square meter. Approximately half of the $1 billion in plate glass sold in
the 1950s was ¼ inch (6.35 mm) plate.
Rolled plate was made by a process similar to plate glass
except one of the rollers (usually the bottom) through which the glass was
rolled had an embossed pattern that was transferred to the soft glass. No
polishing operations were used after annealing. This product was used mainly
for decorative screens and domestic windows.
Although Pilkington’s development of a twin grinding
process significantly improved the efficiency of the plate process, it still
had substantial drawbacks. Equipment investments of up to $30-$40 million were
needed for a single glass furnace and its associated plate line, and as many as
800 people were needed to keep a line operating continuously. As much as 15-20%
of the glass ribbon was ground away in the finishing processes. A normal plant
could produce as much as 500 tons per day of grinding debris, such as a cullet
and water suspending finely ground sand, plaster, glass, iron, felt, and rouge.
The noise level of grinders, transfer machinery, and crashing cullet was
formidable. Also, repairs often necessitated shut downs or hazardous work in
the grinding pits underneath the glass ribbon.[8]
Many in the industry wished to combine the continuous
flow, fire polish, and low cost of sheet with the distortion-free quality of
polished plate. One such dreamer was Lionel Alexander Bethune (Alastair) Pilkington, who had joined the company in 1947
with a degree in mechanical engineering and service in the Royal Artillery
during World War II. After working in both the sheet and plate works’ technical
development groups, in 1949 he was named production manager at the Doncaster works, a plate glass plant.[9]
The three year assignment at Doncaster gave Alastair experience with molten metals, since molten tin
was used as a frictionless surface in the bottom of a small experimental
furnace. Using tin in place of the usual bricks increased the amount of time a
furnace could operate without rebuilding.
Following his Doncaster
assignment, Alastair moved to the Head Office to work
as an assistant to the Production Director, James Meikle.
It was during this assignment that the basic idea for the float process took
shape. Sir Alastair later described how he arrived at
the basic idea for float glass:
One quickly became aware that grinding and polishing
was an extremely cumbersome way of making glass free from distortion. [You
could see] that the window glass process produced a beautiful surface, which
glass naturally has because it is a liquid. What you wanted to do was preserve
the natural brilliance of molten glass and form it into a ribbon which was free
from distortion.... A large part of innovation is, in fact, becoming aware of
what is really desirable. [Then you] are ready in your mind to germinate the
seed of a new idea.”[10]
In June 1952, while washing the family dishes, Alastair got a more concrete idea of how this could be done
in a “sort of ‘bang.’” Alastair’s new idea was for a
continuous ribbon of glass to move out of the melting furnace and float along
on the surface of a molten metal at a strictly controlled temperature. The
ribbon would be held at a high enough temperature for a long enough time to
melt out all irregularities. Because the surface of the liquid would be dead
flat, the glass would be dead flat too. Natural forces of weight and surface
tension would bring the glass to an absolutely uniform thickness. In order to
gather support behind his idea, Alastair quickly drew
up some sketches and got the head of Pilkington’s Engineering Development
Group, Barradell Smith, to convert them into working
drawings. The Pilkington Board decided to give the project the highest priority
so that either success or failure would be determined as early as possible, and
within three months a $70,000 pilot plant was built and made operational. By
October 1952, tests were being conducted.[11]
Following the emergence of an idea, three milestones
typically defined the progression of a development project at Pilkington: 1) an
investigation to see whether it was worthwhile at all; 2) a decision to make it
a project and carry out pilot plant work; and 3) a decision to install
full-scale production facilities. Alastair, with a
team of several engineers, a supervisor, and workers sworn to secrecy,
performed the preliminary investigations. They tested the concept in practice
after a short period of laboratory experimentation.
One of the first decisions that needed to be made was
which metal to use as the support medium. The metal had to be liquid over the
range 590-1050° C (the range of temperatures involved in the manufacture of the
glass ribbon), dense enough to support glass (which has a density of 2.5
grams/cubic centimeter), and with a high boiling point (a low vapor pressure at
the 1050° C end of the range) in order to avoid excess vaporization and
contamination of the glass or the process. Finally, the medium could not
chemically combine with the glass during processing and had to be available at
a reasonable price.[12]
Tin was chosen from several alternatives, including gallium and indium, since
it showed the least interaction with glass and was the most abundant. Also, Alastair had learned from his experience at Doncaster that clean molten tin would not stick to glass.
Decisions had to be made as well about the composition of
the glass ribbon itself, because the composition would affect the processing
parameters. Since the 1920s it was well known that the rate of production
depended on the composition of the glass; glass formulation had a direct affect
on the softening and setting points of the glass ribbon and the speed of
transition through the spectrum of physical states. As a result, the float
glass team iteratively fine-tuned the composition of the glass and accordingly
adjusted processing parameters. Float opened new chemical challenges to
Pilkington’s glass technologists, testing and expanding the knowledge they had
accumulated in material science.
The original float tests were conducted in a pilot plant
built from makeshift equipment. The molten glass was provided by a rolled glass
furnace with a hole knocked in its side. According to Alastair:
We got the cheapest flow of glass we could find in
the company. At the earliest possible moment we made a box for the molten tin.
The first one leaked like a sieve because we heated the tin by immersed tubes.
We had to make gland joints at the end, and I can tell you molten tin goes
through any gland joint. It just poured all over the ground. But it showed you
could take a ribbon of glass, pass it over tin at a relatively high
temperature, and produce bright parallel surfaces.[13]
In addition to the leaks, the team witnessed problems with
tin oxidation inside the tin bath. The oxide upset the glass/tin interface so
badly that the glass surface became severely damaged, a result that could have
been found in the laboratory. Although the process was far from producing
commercial quality glass, there appeared to be no insurmountable barriers to
achieving salable results. One member noted, “The only answer we wanted at this
point was: Did the process look promising? Or would we crash up against some
basic chemical or physical laws which would prevent the process from operating?”[14]
Following James Meikle’s
retirement in June 1953, Alastair Pilkington took
over his duties in plate production and was appointed a subdirector.
One of his principal responsibilities was to provide information and advice for
flat glass manufacturers who had licensed Pilkington’s twin grinding process.
He was forced to allocate his time between development and production, spending
about 20% of the day at the former.
In early 1954, a new pilot plant was built to produce 30.5
cm ribbon in longer runs than had been possible in the first pilot site. The
longer runs would make analysis and debugging easier by generating more samples
for study. Work continued to be performed under conditions of maximum security.
Those working on the project were sworn to secrecy, and many other Pilkington
employees were not even told of the existence of the float project. Should word
of the new process leak out prematurely, Pilkington was fearful of undermining
its major licensed technology, the twin grinder.
The technological problems encountered in this stage, many
in areas where completely new ground had been broken, were formidable. As one
engineer noted, “In attempting to control the possible chemical interaction
between the glass, the molten metal, and the atmosphere in the float bath, we
had no guides, no precedents.”[15]
For example, even though the team never questioned that tin was the best float
medium, it did not prove trouble-free. Because of irregularities in the pool’s
temperature, as the tin cooled, its oxygen content reached its saturation point
in some of the coolest areas. This caused stannic oxide to exit the solution
and damage the glass ribbon’s lower surface. The early oxidation problem was
solved in the pilot plant by flooding the bath with an inert atmosphere of
nitrogen and hydrogen, but much more would have to be learned about the
chemical characteristics of the interaction before the surface of the glass
would be of commercial quality.
A frustrating aspect of the experimentation was that
solutions to one problem seemed to surface other problems. For example, once
the stannic oxide issue was treated, problems with sulfur in the tin appeared.
Although pure tin was chosen initially as the medium because of its low vapor
pressure, sulfur from the glass contaminated the tin and caused stannous
sulfide to vaporize into the atmosphere of the bath. Later condensation of
these vapors formed microscopic specks of tin (about 120 microns in diameter)
on the surface of the glass that could potentially cause visual distortion.
Because the sulfur originated in the glass, the source could not be removed.
Instead, the team found a way to control the flow and condensation of the
stannous sulfur once it vaporized into the bath’s atmosphere.
Fortunately, not all surprises encountered proved to be negative.
The team made an important discovery: when the glass was held for one minute at
the 1040° C temperature needed to eliminate its surface irregularities, a
combination of surface tension and gravity effects caused it to form at an
equilibrium thickness of 7 mm (0.275 inches). By applying a tractive
force as the ribbon emerged from the annealing kiln (lehr),
the glass could be thinned to 6.5 mm and sold as nominal ¼ inch glass. As Sir Alastair later said, "This was a fantastic stroke of
luck” — about 60% of Pilkington’s plate sales at that time were of the ¼ inch
thickness.[16] Subsequent
analysis using mathematical models showed that this phenomenon was the result
of a balance achieved naturally between the force of gravity and surface
tension effects in the bath.
Although the output of the 30.5 cm pilot line was not yet
commercial quality, the Board continued to have confidence in the project.
Senior managers at Pilkington believed that judgment had to play a particularly
important part in projects with high risk. With open and honest lines of
communication between the Board and the team, the Board knew enough to make
such appraisals. In autumn 1954, the Board approved a new pilot plant capable
of producing a 76 cm ribbon of glass. This experimental new line was designed
and built in 3 months at a cost of $140,000.[17]
The 76 cm plant taught the float team about the
difficulties of moving to a larger facility. Levels of uniformity that had been
achieved on the 30.5 cm line could not be achieved with the same ease on the
newer, larger course. In order for the 76 cm process to work, the furnace tank
had to be larger to keep the glass hot enough for a long enough time to degas, debubble, and homogenize. As seen earlier with the oxygen
saturation problem, homogeneity was crucial to the process. A member noted that
“Every time you made a move, you needed to optimize the plant for that
particular thickness, width, or speed. It was a long, long learning process.”
Despite the setbacks and frequent problems, team members’
dedication remained high. Alastair observed, “Everyone
who worked on float was highly motivated — it was almost a crusade. Chaps were
literally taken off on stretchers from heat exhaustion, yet came back for more.”[18]
In addition to his responsibilities as head of plate production, Alastair continued to head the float glass experimental
team. He discussed progress three times daily with the development team and met
each morning with the chief project engineer, Barradell
Smith, to lay out strategy for the day: “I took [Smith] away from the noise of
crashing cullet so he could have a chance to think. A large pilot plant running
24 hours a day creates great stress and urgency.... We discussed results, what
was needed ahead, and how the morale of the people was holding up.”[19]
Many of the problems at this stage were thought to be a
product of the rolled plate tank that supplied the molten glass rather than of
the process itself. Because bubbles in rolled plate glass could be masked by
the embossed patterns, higher bubble counts were acceptable in rolled plate
manufacture than was the case for plate glass. The team members felt hopeful
that the high bubble counts would be minimized in the more suitable melting
conditions of a plate glass tank.
During the operation of the 76 cm pilot line in early 1955, two important events occurred at the Board level. First, on March 1, 1955, at the age of 35, Alastair Pilkington was made a full director of the company after having served as a subdirector since 1953. Second, the Board decided that float glass would only be launched on the world if it could replace plate glass.[20] If float merely provided an improved sheet glass, it would occupy a peculiar position between two glasses with well established positions. According to Sir Alastair, a participant in the decision making process,
The forum for the decision was the
Executive Committee of the Board. There were no detailed calculations of such
things as the ultimate capital implications of the process or its effects on
our overall capital structure. Nevertheless, over a period of time a consensus
crystallized. This evolved from a series of formal and informal discussions
among the members of the Executive Committee and the Board.
Once arrived at, I don't think
anyone had any doubt this was the right decision. On the other hand, as
technical director I was very disturbed to be expected to make such a
tremendous jump forward in one enormous leap. It would have been easier for the
technical group to learn about the process while making a better quality of
sheet, and then launch ourselves up the ladder from sheet to plate.[21]
By April 1955 the three pilot facilities had cost the
company about $1.5 million.[22] Neither
the 30.5 cm nor the 76 cm line had produced commercial quality plate. While the
team had solved a large number of difficult problems, issues of glass flow,
ribbon formation, and oxygen and sulfur contamination persisted. These and high
bubble counts from the rolled plate source kept the glass from approaching
commercial quality. Yet the team remained optimistic. At this time Alastair Pilkington presented a requisition to the Board
for $1.96 million to modify a redundant plate glass furnace and go to a full
scale production line capable of producing a 2.5 meter ribbon. On it he hoped
to achieve float glass of commercial quality. The cost of operating this full
scale line would be $280,000 per month.[23]
Everyone involved in the float project recognized the
difficulty of achieving commercial scale and quality. Simply scaling up the
process would be difficult enough, but there remained significant uncertainty
over the basic chemistry in the process, as well as issues of mechanical
design. The team had yet to resolve, for example, the design of the system to feed
glass onto the molten tin. The primary approach — used in the pilot plants —
was the "roller method," which formed a ribbon between water-cooled
steel rollers and fed it onto the bath of molten tin (see Exhibit 5).
Pilkington's years of use of rollers in its other flat glass facilities made
this a logical first choice, but in its float application, the method created
quality problems that had yet to be overcome. Most notably, tin oxide condensed
on the water-cooled surfaces of the rollers and became imprinted on the glass.
In addition, the rollers took a substantial amount of heat away from the glass,
creating the need for additional heating later in the process.[24]
The other ribbon forming process under development was the
"direct pour" method (see Exhibit 6). In this approach, the molten
glass flowed directly from the tank, through a refractory spout, onto the bed
of molten tin. The obvious attraction of such a process was its simplicity,
lack of additional mechanical parts, and relatively "gentle" handling
of the glass ribbon. Although it eliminated the problems with the roller
method, tests of direct pouring showed that it created contamination on the
bottom surface of the ribbon (because of erosion at the interface between the
glass, the spout, and the tin bath), which produced an optical distortion
called "music lines." Based on experience in the pilot plant, the
team believed that if it could be controlled, direct pouring would be a
"very attractive solution to the problem."[25]
If the new line were to be approved, Pilkington’s
investment risk on the promising but so far unproven process might go much
higher than the amounts committed so far. Yet even conservative estimates of
the possible savings involved should the project succeed were impressive. A float
line was much smaller and less complex than an existing plate facility. Capital
costs for a float line were projected at two-thirds the cost of a comparable
plate line, while operating costs were expected to be much closer to the cost
of sheet than the cost of plate. It was not unreasonable to expect that a
successful float process would displace plate production over a ten to fifteen
year period, and could become the method of choice for all flat glass
production.
The opportunity inherent in a successful float process was
thus high motivation for the team and the Board.[26]
But the sheer difficulty of the problems they confronted was a sobering
reality. As the Board contemplated the decision to build a full scale
commercial float line, Sir Alastair put the challenge
they faced succinctly: “My own greatest fear is that float will drag on for
years being a near success, interesting enough to justify further work, but
never quite achieving satisfactory results.”[27]
The basic problem to be overcome in vertical drawing is
the tendency for the ribbon of glass to “waist” or lose over-all width. This
problem can be demonstrated readily by immersing a rod in a dish of syrup and
then withdrawing it slowly upwards, while keeping the rod level. The Fourcault process overcomes this difficulty by drawing the
glass through a slot in a fire clay shape (called a debiteuse
block) which floats on the surface of the glass in the drawing chamber. The debiteuse is forced down until the level of the slot is
below that of the glass in the furnace; hydrostatic pressure is then sufficient
to cause the molten glass to well up into the slot. The drawing operation is
started by means of a metal frame (“bait”) which is lowered vertically into the
slot. Sufficient glass becomes adhered to the bait so that when it is withdrawn
upwards a steady ribbon of glass follows. Adjustable water coolers are then
installed on either side of the ribbon, resting on the top of the debiteuse block, which also help to maintain ribbon width.
Further assistance is provided by a pair of knurled rolls which grip the sheet
edges a short distance above the debiteuse. Driven
asbestos rollers convey the ribbon vertically through the lehr
or tower to the cutoff floor some thirty feet above glass level, where it is
cut into sheets of a predetermined length. The average drawing speed for 3.2 mm
glass ranges from 100 to 150 centimeters/minute depending on the process
modifications employed.
Like all forming processes the product is only as good as
the die; in the case of Fourcault the glass surface
is susceptible to damage from hundreds of tiny crystals which grow along the
glass/fire clay block interface. After seven or eight days the draw must be
stopped and the crystals melted away before the draw can be restarted. The
thickness of the sheet depends on temperature control and the speed of draw;
product thicknesses range from 0.8 to 12.7 mm while machine widths are normally
2-2.3 meters.
The two other sheet processes, Colburn and Pittsburgh, are
similar to the Fourcault in theory of operation. The
Colburn differs in that it does not use a debiteuse,
and it has the ribbon bent horizontally over a bending roller shortly after
leaving the bath. The Pittsburgh Process was developed about 10 years after the
Colburn and Fourcault, and incorporates minor
variations that remove many of the inherent defects of the other sheet forming
processes.
The “modern era” of plate glass technology began in the
1920s when Pilkington Brothers introduced the first continuous grinding and
polishing machine. Instead of being mounted on circular rotating tables, the
blanks of rough glass could be laid on rectangular tables coupled together as a
continuous train, passing under the disk grinders and then under the polishers.
The next logical step was to develop a machine that could
grind the ribbon on both sides simultaneously as it came out of the annealing lehr before it was cut into plates. Pilkington developed
such a machine during the early 1930s and put it into service at their Doncaster works plant in 1933. This machine, called the
twin grinder, was a masterpiece of large scale precision engineering. A
continuous ribbon of glass about 270 meters long was ground on both surfaces at
the same time, using enormous grinding disks fed by sand which was made
progressively finer as the glass proceeded down the line. The machine was
driven by 1.5 million watts of power. A remarkable feat was that the bottom
grinding wheels were kept perfectly flat and level while they were actually
wearing away. The standard width of plate glass made by Pilkington Brothers was
2.5 meters, and the average speed of producing 6.0 mm glass was 4
meters/minute.
*From Note on the Flat Glass Industry - 1955.
[1] Conversion rates for all cases taken to be 1 Pound Sterling = 2.80 in U.S. dollars.
[2] Sir Alastair Pilkington, "Float: An Application of Science, Analysis, and Judgment, Turner Memorial Lecture." Glass Technology, August 1971.
[3] See "Pilkington Brothers P.L.C." by James Brian Quinn in Henry Mintzberg and James Brian Quinn, The Strategy Process: Concepts, Contexts, Cases. Englewood Cliffs, NJ: Prentice Hall, 1991, pp. 826-844. (Hereafter referred to as Quinn, "Pilkington Brothers P.L.C.")
[4] Quinn, "Pilkington Brothers P.L.C.," p. 827.
[5] Quinn, "Pilkington Brothers P.L.C.," p. 827.
[6] In the mid 1950s, the Fourcault process was used in 72% of worldwide production, the Colburn in 20%, and the Pittsburgh in the remaining 8%.
[7] Average gross margin for sheet glass was about 10%. Plate glass had a gross margin of about 27%.
[8] Quinn, "Pilkington Brothers P.L.C.," p. 827.
[9] Quinn, "Pilkington Brothers P.L.C.," pp. 827-828.
[10] Quinn, "Pilkington Brothers P.L.C.," p. 828
[11] Quinn, "Pilkington Brothers P.L.C.," p. 828.
[12] Quinn, "Pilkington Brothers P.L.C.," p. 829.
[13] Quinn, "Pilkington Brothers P.L.C.," p. 828.
[14] Quinn, "Pilkington Brothers P.L.C.," p. 828.
[15] Sir Alastair Pilkington, "Float: An Application of Science, Analysis, and Judgment, Turner Memorial Lecture." Glass Technology, August 1971.
[16] Quinn, "Pilkington Brothers P.L.C.," p. 831.
[17] Quinn, "Pilkington Brothers P.L.C.," p. 831.
[18] Quinn, "Pilkington Brothers P.L.C.," p. 831.
[19] Quinn, "Pilkington Brothers P.L.C.," p. 833.
[20] Quinn, "Pilkington Brothers P.L.C.," p.
831.
[21] Quinn, "Pilkington Brothers P.L.C.," p.
831.
[22] Quinn, "Pilkington Brothers
P.L.C.," p. 832.
[23] Quinn, "Pilkington Brothers P.L.C.," p.
832.
[24] Quinn, "Pilkington Brothers P.L.C.," p. 832.
[25] Quinn, "Pilkington Brothers P.L.C.," p.
832.
[26] Licensing agreements in the past typically included about $1.2 million as an information transfer fee for each manufacturer, $120,000 per glass line, and 3% of sales. If a successful float process were to follow Pilkington's practice of licensing its technology to competitors, the arrangements would involve a one-time up-front payment on the order of $1-1.5 million, a fee per line installed (generally 10% of the up-front fee), and a royalty on sales of 2-4%.
[27] Quinn, "Pilkington Brothers P.L.C.," p. 828