THE 1953 GEORGE BRAY MEMORIAL LECTURE

THE MANUFACTURE OF

PLATE GLASSby SIR HARRY PILKINGTON

Presented to a General Meeting of the Institution of Production Engineers at the University of Leeds,

9th November, 1953.

Sir Harry Pilkington entered the well-known family glass manufacturing firmof Pilkington Brothers in 1927. He became a director in 1934 and is nowChairman of the Company and a director of its subsidiaries at home andoverseas.

Sir Harry is responsible for the commercial policy of his firm ratherthan the technical aspects and he has travelled widely and made frequentoverseas visits in connection with Pilkington Brothers3 activities in othercountries. From 1944 to 1952, Sir Harry was Chairman of the ExecutiveCommittee of the National Council of Building Material Producers.

He was appointed President of the Federation of British Industries in 1953.Sir Harry Pilkington

W HEN Sir Cecil Weir did me the honour ofapproaching me a few months ago to ask me to

deliver this Lecture, I made it clear that I was not anengineer in any shape or form and that I could notgive a highly technical Lecture. In my own firm Ihave always been connected mainly with the sales sideof the business and most of the terms in which youtalk are far above my head. I thought, therefore, thatit might be of the greatest interest to you if, insteadof a detailed engineering talk centred on a particularsubject such as I believe you have had in the past,you had one that was a great deal more general inscope, dealing with one part of the British glassindustry that has had an interesting evolution duringpast generations, and that I should ask one of mycolleagues to provide me with the technical facts thatare material to the subject.

Development of the IndustryFirst of all, however, I think that you ought to

know something of the position of the plate glassindustry (since plate glass is the subject of this talk)in this country and the world. In the 1870's when myfirm went into plate glass manufacture, half a centuryafter it had started sheet glass manufacture, it wasquite a venture as there were already well establishedplate glass manufacturers in this country, and manycontinental competitors. In a series of technical

revolutions, however, manufacture has changed fromcompletely discontinuous to completely continuousprocesses, and this has meant that the size of unit foreconomic manufacture, and its output, has grownvery much more rapidly than the demand for theproduct, with the result that now the whole require-ments of the British Empire for ordinary plate glass,in the thicknesses most usually purchased, can easilybe met from two manufacturing units only—one atDoncaster and one at St. Helens. There are nowfar fewer manufacturers overseas than there were 50years ago and for some 35 years there has been onlyone manufacturer of plate glass in Great Britain orindeed in the British Empire, and who, during thattime, has so much improved productivity that plateglass for ordinary purposes is now manufactured fromonly two units; all other manufacturers in this countryhave long ago disappeared through not having keptpace with the growth of technical development,particularly of engineering. This has lessons anddangers for the future, and inconveniences for thepresent on which I will enlarge later.

Because it can only be undertaken economically ona very big scale, plate glass manufacture for com-petitive supply to the world markets is virtuallyconfined now to the U.S.A.; to this country; and toBelgium, France and Western Germany; only in theU.S.A. are there more than two manufacturers and

even there, probably 95% of that sold in competitionis made by two firms. In a few countries smalleruneconomic factories operate behind high tariffs fortheir own market, and one day we will probably findJapan and Russia in the world markets too.

Flat glass, whether plate or sheet, is used mainly inthe building, furniture and motor industries; on thewhole people do not buy glass for its own sake, andthe level of demand can only be slightly influencedby anything the producers can do over a short period;The fact that the main use of plate glass is in themotor car industry, one that all over the worldis particularly subject to fluctuations, brings usproblems that are made extra difficult by our largescale inflexibility, made still more inflexible by thefact that ours is essentially a 168-hour a week process;so much for the general economic background.

The Basic ProcessesGlassmaking and, in particular, the making of flat

glass in its several varieties, is an activity about whichrelatively little is known to those who are notpersonally engaged in the industry. It may, therefore,be of interest to describe the basic processes involvedbefore dealing with the historical development, in atechnical sense, of methods employed in making plateglass and their effect on efficiency and productivity.

With few exceptions which, for the purpose of thisLecture, are unimportant, flat glass in its severalvarieties is known, colloquially, as a " soda-lime"glass. This means that in addition to the silica whichis by far the largest constituent, soda and lime comenext in order of importance. There may be, andusually are, smaller quantities of other constituentssuch as magnesia and alumina, but for our purpose itis sufficient to say that flat glass is made of (1) silica,(2) soda, and (3) lime.

The raw materials corresponding to these con-stituents are, respectively, (1) sand, (2) soda ash andsaltcake, and (3) limestone and dolomite. Thesematerials, finely divided as found in nature like thesand, or mechanically ground as the limestone mustbe, are assembled in their correct proportions andintimately mixed by special machinery to form thebatch—or, as it is known to glassmakers, " frit".

This frit goes into a furnace where it is subjectedto a progressively rising temperature reaching amaximum of 1500° Centigrade or thereabouts, and inthe process a succession of chemical reactionsbetween the constituents results in the formation ofmolten glass. Continued exposure to high temperaturehas the effect of removing imperfections, such asbubbles, after which the molten glass is cooled downto a temperature appropriate to the particular processwhich, in converting the glass from a plastic to a solidstate, will also deliver it in the required form.

There are three main forms of flat glass. Theseare—sheet glass in universal use for the windows ofsmaller houses; rolled glass in its many patterns foruse where translucency is needed or is sufficient; andplate glass for shop windows, mirrors, motor cars,larger domestic glazing, and innumerable otherpurposes. In choosing plate glass as my subject, I

have been influenced by the fact that, because of thegreater diversity of the operations which lead to thefinal product, it provides a broader field of interestthan either of the other two. It will, however, helpyou to form in your minds a more comprehensivepicture of the technical aspect of the matter if, byway of an introduction to plate glass in greaterdetail, I give you a brief description of the methodsemployed in making sheet and rolled glasses.

Sheet and Rolled GlassesIn the modern method of making sheet glass, frit is

fed into one end of a long continuous furnace whereit is converted to molten glass. This, in passing alongthe furnace, experiences a cycle of temperaturevariation by means of which it is refined and thencooled. At the exit end of the furnace the glass, stillmolten, and at the correct temperature for thepurpose, forms a bath from which a ribbon of thedesired width is drawn vertically upwards. For thesame thickness the rate of production is dependent onthe viscosity and therefore the temperature of theglass in this bath. The thickness itself is dependenton the speed of draw; the higher the speed thethinner the glass is drawn. As soon as the ribbonis formed, rapid but carefully controlled cooling isbrought about and when the ribbon has becomesufficiently solidified to ensure that its surfaces willnot be easily injured, it is gripped and propelledupward by rotating rollers through a vertical towerin which it is annealed and then cooled to a tempera-ture at which it can be handled. It is then in its finalform; at the top of the tower, sheets are cut from themoving ribbon and in due course find their way to thewarehouse, where there may or may not be furthercutting before despatch to the customers.

As a finished product, sheet glass differs from plateglass which, as will be explained later, is polishedmechanically, by the fact that its polished surfacehas a natural " fire finish ", as it is termed, and it isfor this reason that the ribbon after it is formed mustnot come into physical contact with a roller or othersolid body, until sufficiently solidified to be incapableof taking an impression. The process is, however, avery sensitive one and in spite of temperature controlto the highest possible standards of precision, it ischaracteristic of even the best sheet glass that thereare slight departures from flatness and uniformity ofthickness which manifest themselves in optical distor-tion. While, therefore, it serves its purposes well andcheaply, there is this characteristic of distortion whichdistinguishes sheet glass as a commercial product fromthe much more costly mechanically polished plateglass.

In the case of rolled glass, the molten glass havingbeen made in a continuous furnace in very much thesame way as that described for sheet glass, arrives atthe delivery end at a suitable temperature and fromthis stage the process becomes quite different. Theribbon, instead of being drawn from a bath, is formedby passing the molten glass between two horizontalwater cooled metal rollers, whose distance apartdetermines the thickness. If either roller has a surface

•4.

8

pattern, this is imprinted in obverse on that surface ofthe ribbon which has been in contact with it and sowe have in great variety the figured rolled glasses.The ribbon of glass is then annealed and cooled in itspassage down a long horizontal lehr.

If the rollers are plain instead of patterned, theywill produce glass with plain surfaces but they will notbe polished surfaces. Having been formed by physicalcontact while still soft, they lack what has alreadybeen referred to as the natural fire finish of sheetglass and, if they are to be converted to polishedsurfaces, they must be ground first and then polishedby mechanical means and this, in principle, is howplate glass is made. The process of rolled glassformation also differs from sheet in that the rate ofoutput is much less dependent on the viscosity of theglass and much more on the speed of rotation of theforming rollers. Much higher speeds can therefore beachieved. Quality is less important and the processless sensitive.

Changes in Manufacturing MethodsThis brings us to a more detailed consideration of

the technical aspects of plate glass manufacture and Ishall begin by tracing the history of its recent develop-ment, so that, in due course, it can be shown howtechnical progress has contributed to higherefficiencies and productivity.

It is unnecessary to go back more than about 50years, when methods which now seem primitive hadfor generations served the needs of glass-makingcountries whose export markets were more easily wonand preserved than they are to-day. During the lasthalf-century or so, there have been revolutionarychanges in methods and I shall try to describe to youthe course of technical progress in this country andthen relate it to results in terms of economic progress.

Earlier in this Lecture, I told you how few manu-facturers of plate glass there are in the world and thatthere is only one surviving manufacturer in theEmpire. This might seem to imply that plate glass isin a very comfortable monopolistic position and thatthere is no need to concentrate on further improve-ments in order to preserve or increase the demand.Such an impression would be completely false andunjustified. During this century the abandonment ofthe old method of manufacturing sheet glass (underwhich it was made as a cylinder and then laterflattened), has given place to a method where sheetglass is drawn continuously from the tank. Theresult is a very great increase in flatness, together witha cheapening of the product, by the completeomission of one of the major processes of manu-facture; but the greatest advantage that plate glasshad over sheet glass lay in its flatness; the more thatsheet glass improves, the less does the advantage ofplate glass become. Moreover, under the newermethods of manufacture sheet glass can be made inthose same thicknesses that used to be regarded as thepreserve of plate glass.

Apart, therefore, from competition with othermanufacturers in the export markets of the world—and since we export, directly or indirectly, half of

our total production, foreign competition in theworld markets is a material factor—the need to fightfor the survival of plate glass against the encroach-ment of sheet has been a constant driving force tocompel simultaneously a further improvement inquality and a vast reduction in cost; a reduction incost usually achieved by greater output from thesame plant and therefore only materialising fully ifdemand also increases.

Plate glass will never intrinsically be as cheap assheet glass. Sheet glass inherently is made in oneprocess only; plate glass inherently is cast rough andmust have the surface subsequently ground andpolished. This is more expensive and must producea finer quality to justify its price.

Moreover there are some sizes which it wouldbe quite uneconomical to manufacture in sheet glass.The motor trade, for instance, in particular, is verysensitive to price and plate glass cannot be manu-factured economically in large sizes alone. We musttherefore continue to fight to reduce costs incompetition with sheet glass, whether manufacturedby ourselves or by anybody else, if we are to keepplate glass in commercial and economic productionon its present vast scale. I do not think this has beenthe only driving force, or even the principal one,behind our inventions, but I do say that if the plateglass industry had not had the benefit of three majortechnical revolutions during the past thirty or fortyyears, it would survive now only on a small scale,at very high prices, for shop windows, high qualitymirrors, thick portholes, and other special uses.

Fifty years ago, plate glass was made, not incontinuous furnaces in which frit went in at one endand molten glass came out at the other, but inrefractory clay pots (Fig. 1). These pots occupiedthe floor of a furnace for a period of about twenty-four hours, during which they were filled with frit,the frit was turned into molten glass and afterhaving been refined and cooled, each pot, containingabout a ton of molten glass, was withdrawn from thefurnace, emptied of its contents and returned to the

Fig. 1A clay pot in the furnace preparatory to filling with the

raw materials

Fig. 2

Pouring and spreading the molten metal prior to rolling

furnace to undergo the next cycle of operations. Thenumber of pots in a furnace was anything up totwenty and if a pot survived twenty cycles, or inother words, lasted twenty days, it had earned afavourable obituary.

The emptying, or casting, process consisted oftipping the pot in such a way that its contents, nowfairly stiff in consistency, formed a bolster across oneend of a large iron table (Fig. 2). The bolster, whilestill plastic, was rolled out to form a plate by a heavyiron roller which, near both ends, was supported bysteel strips equal to the desired thickness anddepending upon this thickness the plate might beanything up to 150" wide x 250" long. In shape itwas roughly a rectangle with parallel sides butirregular ends. Thermal distortion of both table androller was unavoidable and with an average thicknessof 10 or 12 millimetres there might be a variationover the area of the plate of 2 or 3 millimetres. Theplate, substantially solidified, was then transferred to aseries of annealing chambers constituting what isknown to glassmakers as a " lehr " from which, afterit had been annealed and slowly cooled over a periodof some hours, it was withdrawn in readiness forgrinding. This operation was carried out in amachine in which a flat-topped circular iron table,22 ft. in diameter (Fig. 3), carrying a patch-work ofrough-cast glass plates laid in Plaster of Paris, revolvedhorizontally. Supported by the top surface of theglass were heavy iron runners, having their circularunder-faces intersected by grooves through whichwater-borne sand could flow to distribute itself overthe area of the table. The runners, free to revolve,did so in virtue of their frictional contact with therotating glass-topped table and so the glass wasground until all traces of its original rough castsurface had disappeared. Continued grinding withsuccessively finer grades of sand until the glass hadacquired a frosted appearance of very fine texturecompleted the grinding operation.

The table was taken from the grinding machineand transferred to the polisher where runners

operating on the same principle as the grinder runners,but consisting of a multiplicity of felt-covered discsfed with water-borne rouge, converted the finelyground surface to the highly polished one which weassociate with plate glass. The table wasthen withdrawn from the polishing machine, theglass plates lifted and relaid with the other side up,and the processes repeated.

For the next fifteen years or so the methods whichI have described remained substantially unchangedin principle and progress consisted in the use ofbigger pots and furnaces, the substitution of a castingprocess in which, by the use of two revolving rollerswhich received the molten glass from a shapedcontainer, plates of better shape and greater areawere produced, and a change-over from a grindingtable of 22 ft. diameter to one of 36 ft., whichbrought about some improvement in efficiency andproductivity.

The Economic AspectAt this stage I suggest that we take ourselves back

thirty-five years and, without attempting to assesstheir significance in numerical terms, consider someof the more obvious shortcomings, from an economicpoint of view, of the methods which have beendescribed.

With negligible exceptions, plate glass is sold inrectangles and a circular grinding table is not of theshape which one would choose to accommodaterectangles to the best advantage. To complete thecircle, and its whole area must be filled, a substantialproportion of the glass must be in small andirregularly shaped pieces which, when they reach thewarehouse, have little or no usefulness. The irregularends of the rough cast sheet also involve wastage.Variation in the thickness of the rough cast platemeans that not only do the thicker portions consumeglass which, had there been less variation, would haveincreased the useful area of the plate, but theyconsume time, wages and power during their removalby the grinding process, and which all on one tablemust be ground till the thinnest sections have been

Fig. 3Disc Table in position beneath the grinding heads

10

00 CVKVLQQCVlPOT FURNACE

CASTINC

GRINDING

ANNEMING

JL JL

POLISHING

Fig. 4

fully smoothed. The wide temperature variationnecessary for the glassmaking cycle must be followedby the massive furnace structure over every twenty-four hour cycle and this is reflected in a high fuelconsumption. In terms of electrical power the samething applies to the grinding cycle for which theconsumption curve is not very far from a trianglewith, as a consequence, low average operatingefficiencies for the electrical machines. The pictureis therefore one of a sequence of processes which couldhardly be further removed from what is regardedto-day as a continuous process and which with all itssources of loss and wastefulness of manpower, fueland glass, was an ever-present challenge to plate glassmakers to look for progress along radically new lines.

The chart shown in Fig. 4 gives only a veryelementary indication of sequence but shows fivedistinct processes.

By the end of the First World War a good deal ofthought had been given to the possibilities andfollowing the results of experimental work whichseemed to justify adventuring further, trials on a pilotplant scale led to commercial production in a smallway of a continuous rough cast ribbon and also of afinished plate glass, ground and polished by methodsmuch more nearly continuous than the old circulardisc process.

The rough cast ribbon process has already beendescribed in outline in connection with the makingof rolled glass and, indeed, the first trials on a

Fig. 6Grinding and polishing machine

commercial scale were devoted to this product. Butrolled glass, with its obscured and patterned surfaces,does not call for the standards of glass qualitydemanded by polished plate glass, and it was a year ortwo before this was achieved by suitable design andoperation of a manufacturing unit consisting ofcontinuous furnace, rolling apparatus and annealinglehr delivering a continuous ribbon of rough castplate glass which, after grinding and polishing, provedto be of excellent quality and uniform thickness.This was revolution No. 1. (Fig. 5.)

Simultaneously with the development of continuouscasting, work was proceeding on a grinding andpolishing machine with many of the features of acontinuous process. This machine in its latest form(Fig. 6) consists of a travelling table, 650 feet long,made up of a succession of cast iron slabs constructed

CONTINUOUS FURNACE ROLLING ANNEALINGFig. 5

11

CONTINUOUS FURNACE

*> >

1mil linnni

with great accuracy and coupled together to form alevel platform of the right width to take glass platescut from the rough cast ribbon, so that, in effect,there is a continuous glass surface. This passes undera succession of revolving iron runners fed, as in thecase of the circular grinding machine, withprogressively finer grades of sand followed by asimilar arrangement of felt polishing runners fed withrouge. At the end of the machine, the glass plates,polished on one side, are lifted and transported to thebeginning again to be relaid, and their other surfacesground and polished. This was revolution No. 2.(Fig. 7.)

The " Twin " ProcessRevolution No. 3 was to follow about fifteen years

later. In the interval, much thought and experi-mental work had been devoted to the possibility ofgrinding simultaneously both surfaces of the travellingribbon, while still in the form of a continuous ribbon,and in 1937 my Company brought into successfulproduction the first full-size commercial unit con-sisting of continuous furnace, rolling machinery, lehrand grinding machine. By this date, therefore,manufacture of plate glass by what, in the fullestsense of the term, could be called a continuous processup to the end of the grinding stage, had beenachieved; in addition the combination nowoperates successfully on an output which, in con-sequence of relatively minor adjustments, is at least

POUSHINCJ

Fig 7

three times that for which it was originally designed.Due to this and to the expansion of manufacturingcapacity, all ordinary plate glass produced in thiscountry, as distinct from the considerable amountof special types or thicknesses, is now made by thistwin process (Figs. 8 and 9).

The " Twin " even now requires a plant nearly1,300 feet in length, and as a straight engineeringproblem you will understand that it is not easy tokeep a continuous ribbon of glass, a hundred incheswide but only a quarter of an inch thick, subject togreat pressure and many stresses, so straight anduniform that it does not wander a quarter of an inchin the course of nearly a quarter of a mile.

It was originally designed to do the polishing onthe same ribbon and in the same way, and for atime this was done : the wartime increase in output,and therefore in speed, from the tank and throughthe lehr was, however, such that more grinding headsbecame necessary and in the space available this wasdone at the expense of the polisher and the glass isnow polished subsequently on a separate high-speedpolisher, one side at a time. This, on our scale ofoperation, is not, as might appear, a step backwards.There are three main reasons why this is so—first,the increase in capital cost of making the building andequipment for a ribbon, say 2,000 feet long, so that onour estimates the saving in cost might be more thaneaten up by capital charges; second, the greater likeli-hood of breakage at the polishing stage and the

12

consequential damage caused by more scratches toglass following the breakage; and thirdly, that thepossibility of taking advantage of sudden technicalprogress in any one sector of this continuous line maybe limited if others cannot match it, and being at theend of the line the polishing is the most likely tolimit advance in any of the other sectors.

Although, therefore, if we were starting a newplate glass works completely afresh in a flat openspace we would think again very carefully beforeomitting the Twin Polisher, and although we wouldcertainly leave room for its later adoption, at presenton existing plants its advantages are outweighed by itsdisadvantages, including overall cost.

The production line therefore is now as shown inFig. 10.

To complete the manufacturing picture, largerectangular plates, after polishing, pass through awashing machine and go into the warehouse whereeach is placed vertically in a frame which can beraised or lowered at will. The glass, illuminated byspecial lighting, is then examined for manufacturingfaults which are marked on the surface by theexaminer. He is followed by one known as the" marker" who, with the faults before him and acomprehensive knowledge of the sizes called for bythe orders, decides how the plate can be cut withleast loss of glass and indicates his decision in writingon the plate. This is now transferred to a horizontaltable where it is divided up by the cutter inaccordance with the marker's instructions, after whichthe cut plates are finally checked for size and qualitybefore going for packing. -

I hope that this account has enabled you to form,in outline, a picture, not only of how our plate glassis made, but of how the methods in use to-day havebeen developed over the last few decades. To you,as Production Engineers familiar with modern trendsthroughout industry generally, the economic signifi-cance of each step forward will be apparent and itcan be shown that the advantages gained have beenvery real.

Let me give you a few examples. In the first, themodern trend of making molten glass and casting acontinuous ribbon is compared with the old potprocess producing individual plates. The twovariables of greatest significance are output per man-hour and per lb. of fuel consumed and in both cases

Fig. 8Twin grinding and polishing process

they are related to the area of glass delivered ready forgrinding and polishing to finish at the same thickness.

Square feetMan-hour

Square feetFuel consumed

Pot Process Ribbon Process

1 11.2

1 8.0

The second example compares the three grindingprocesses, namely :1. The 36 feet diameter circular table machine2. The continuous table machine3. The latest machine in which both surfaces of acontinuous ribbon are ground simultaneously and inthis example the significant factors are man-hoursand consumption of electricity.

No. 1 ' No. 2 No. 3S'quare feetMan-hour 1 1.8 10.8

Square feetElectricity 1 2.1 4.0

The third example is on somewhat different lines.In connection with licensing negotiations, it wasnecessary to make a detailed analysis of the processcosts to show how the latest manufacturing unit

Model of " TwinFig. 9

process of plate glass manufacture

13

CONTINUOUS FURNACE ANNEAUNQ EXAMININq

Fig. 10

consisting of continuous furnace, rolling machine,lehr and Twin Grinder (as the machine which grindsboth surfaces of the ribbon simultaneously is called)compared with the immediately preceding combinationconsisting of a similar continuous furnace, rollingmachine and lehr but with the corresponding grindingcapacity in the form of continuous table machines.This analysis showed that, by comparison with itsimmediate predecessor, the more modern combinationreduced the works cost by 21.4%.

Incidentally, this process, developed in England,has been licensed in U.S.A. where the first plant hasjust started up and in a number of countries inEurope where it is in commercial operation in France,Belgium, Germany and Italy.

The fourth and last example is based upon acomparison between a complete manufacturing unitof the most modern type and its equivalent capacityin the form of the processes in use just before theFirst World War. In seeking figures so far back ithas been necessary to fill in some gaps from laterinformation on the operation of the older methods,but the comparisons may be regarded as fair and, ifanything, conservative.

YearSquare feetMan-hours

Square feetFuel Consumption

Square feetElectricity Consumption

1913-1914

1

1

1

1953

7.6

8.3

4.5

Effect on the CustomerFinally, as the eventual criterion I think you would

be interested to know what all this has meant to thecustomer, bearing in mind all the time that in 1913(which we have taken as the starting point) plateglass was by no means a new industry but one thathad been established here and on the Continent fora very long time. Here is the price, taking 1913 as100. of a normal shop window about 10 feet x 7 feet

in19131920193019381953

.100:240145148208

I will not elaborate on the comparison between thisindex and a similar index for the cost of coal or

wages, or transport, for the general picture conveyedby such figures is well enough known and will nodoubt approximately be present in your minds. Butit does show that the benefit of the inventions hasbeen passed on to the customer and that our rewardhas been the consequent and continuing increase inthe use of glass. Another aspect of this, that toProduction Engineers I am sure is obvious but nonethe less impurtant, is that every labour saving deviceor improvement, every fundamental development,and every reduction in cost, has reduced the numberof man-hours required to produce a foot of glass, butit has before very long increased and not reduced thetotal volume of employment in the industry. Theindustry has had its ups and downs in employmentfor other reasons—usually economic factors such asthe 1929 slump brought these about, but on balancemuch more glass has been produced from about thesame labour force, rather than the same amount ofglass from a smaller labour force.

And now, if we take ourselves back again to thirty-five years ago and re-examine the list of short-comings, we shall find that they have all disappeared.With this state of affairs you may think that oneshould be well satisfied, but I hope that it won't bethirty-five years before you hear in retrospect of thelist of shortcomings as it stands to-day. Let me tellyou of one now.

Due to the combined effects of high temperatureand chemical attack, the interior areas of a continuousglass-making furnace wear away and at intervals oftwo years or so, it becomes necessary to put thefurnace out of operation for repairs. The stoppageof glass-making, including the rolling and grinding ofthe ribbon, lasts from six weeks to three months andas it affects a very large proportion of the totalmanufacturing capacity, the interruption may beregarded as a major disturbance of production.

An obvious necessity accompanying the creation ofthe large integrated units has therefore been theprovision of a large stocking capacity to act as aflywheel, absorbing glass slowly while the unit is inproduction and releasing it quickly during thestoppages. This is, of course, an undesirablenecessity, as not only does it entail additional costsbut money is tied up in stocks. Skilled glass-makershave to be found alternative and usually less profit-able employment for the duration of the stoppage,and it is not too easy to plan maintenance workthroughout the factory in such a way that a staff oftradesmen sufficiently large to deal expeditiously with

14

very big repairs at infrequent intervals can be keptfully and usefully employed at other times.

Every technical advance brings its own problemsand this is admittedly a conspicuous example of thosethat have accompanied the solution of the largeintegrated glass-making unit. By comparison, it is aminor problem which may or may not be solved byhaving two furnaces to each unit and moving one intoposition for production while the other is moved outfor repairs. As a modern continuous plate glassfurnace with its contents of molten glass weighsabout 7,100 tons, it should provide an interestingexercise for the engineers.

Importance of QualityI would like now to refer in some detail to an

aspect of plate glass manufacture which is of greatimportance in its effect on productivity. It concernsglass quality and, as a continuing perplexity, it owesits existence to the fact that, stated simply, it is noteasy to make good plate glass.

Glassmaking, especially the making and refining ofmolten glass is not, or at any rate is not yet, an exactscience. The sequence of thermal, chemical andphysical events inside a continuous furnace is of greatcomplexity and is incompletely understood. Therefractory materials of which the furnace is madehave been much improved over the last thirty years,but they are still capable of contaminating the moltenglass. Minute quantities of some insoluble mineralsin a wagon of sand can survive the processes withinthe furnace and appear as faults in the glass, andsince glass is transparent, blemishes in the body ofthe metal, that have no other significance than thatthey are visible to the eye of a keen inspector, are notpermissible.

In the case of grinding and polishing, the basicprinciples are more fully understood and surfaceblemishes can be more easily traced to an identifiablecause but faults of sufficient importance, whether inor on the glass, entail cutting, and cutting imposed bythe presence of faults invariably entails wastage ofglass.

That the percentage of glass wasted must be verysensitive to the prevalence of defects is almost self-evident. Even if the quality of the large platesentering the warehouse were perfect, some wastagewould be inevitable, because the order sizes, in alltheir variety, would not be a precise match for theglass sizes but, clearly, the larger the glass sizes thelower the wastage. If, on the contrary, the glass sizesbecome substantially smaller, and it needs only onefault in the middle of a plate to halve its size, thegreater the wastage, and it is not difficult to see whywastage, expressed as a percentage, can rise verysteeply with quite a small increase in the averagenumber of faults per plate.

When it reaches the warehouse, the finished glasshas, of course, borne nearly the whole of the pro-cessing costs and a 25% wastage means that foursquare feet have been made to sell three. It might beargued that this overstates the case because all thewastage goes back into the furnace but, regarded asa raw material, broken glass, or " cullet" as it is

called, is no more valuable than frit, the cost of whichis very small by comparison with that of the finishedproduct.

Wastage at all stages of manufacture due to faultyglass is, of course, nothing new, but it assumes a muchgreater importance when it is associated with theoutput of one of a few large manufacturing unitsinstead of one of a much greater number of smalland unintegrated plants. On the other hand, thereare, from the standpoints of technical control, com-pensating advantages in the modern continuousprocesses in which, after establishing suitable condi-tions, the aim is to keep these constant in contrast tothe old processes where the objective, less easy toachieve, was to control the varying conditionsthroughout each of a succession of cycles.

Moreover, efforts to achieve better glass for moreof the time go on unceasingly. Since the First WorldWar, technical development has been accompanied bya great increase in the number of scientific andtechnical men engaged in the industry. Researchwork is providing a better understanding of basicprinciples and much that was once obscure is nowexplained. Improved instrumentation has not onlymade more extensive technical control of the processespossible but it has enabled it to be more precise andmore informative. To take just one example, theapplication of polarised light to the moving roughcast ribbon reveals certain irregularities which, with-out it, would be quite invisible until the glass hadbeen polished. And as a last reference to thecampaign for quality, every plate of glass, with itsdefects, if any, entering the warehouse can be tracedback to the conditions which prevailed at the time atwhich it passed through any one of the processes inthe course of its career.

Keeping the LeadIt must also be plain that successful maintenance

of a highly competitive productive industry in thiscountry depends on being technically in the lead; thatthat lead can be achieved by invention, by imag-inative thinking, by team work, as has just beendescribed, by acquiring and improving processesinvented by others, by co-operating with others—whether competitors, manufacturers or customers innew methods of manufacture or distribution—andeven by acquiring a process and by agreeing tocontinue to acquire future developments from thesame source. All these methods we have tried in onekind of glass or another. What is certain is thatinventive genius is not the monopoly of any onenation or firm and that it is essential to preserve anopen mind and to use all these methods to keep in thelead.

In plate glass, generally speaking, we have beenthe inventors and in due course we have made ourinventions and knowledge available to others. Insheet glass it so happens that we have more oftenacquired fundamental inventions under licence fromothers and have improved upon and developed them:and similarly in many uses of glass, of secondaryprocesses such as the toughening process used forturning plate glass into safety glass.

15

I do not think, moreover, that there is any oneanswer as to which of these methods is the moreeconomical. I do not think that in licensing someof our more important processes we have ever re-covered fully the cost of development. On the otherhand, by developing them we gain a lead and astart on others for the next advance, and we gainadditional knowledge that is of great use in theactual operation, and over a wider field. Moreover,firms that do not themselves invent or develop, arelikely to find it difficult to obtain other people'sinventions indefinitely on reasonable terms.

I think there is one matter that is important inour industry, and in everybody else's, that is wortha reference. That is the very long period betweenthe conception of an invention and the putting ofit into commercial use. In the case of the TwinGrinder which I have described, I suppose com-mercial full scale production first came at least fifteenyears after the main outlines were fairly clear in theminds of the inventors. This is partly due to theextreme difficulty in our industry of operating anykind of a production machine on a pilot plant or onan intermittent scale, and this means that unless thereare subsequent patents or a great deal of specialisedknowledge, the original invention has a very shortperiod of protected full scale commercial operation.However, the fact remains that in the plate glassindustry we have, in the case of the Twin, led theworld by many years and that even now (someseventeen years after it has been in commercial oper-ation with us as a main process and some thirtyyears after it was first conceived) we are licensingit to the American manufacturers, and of course ournew licencees will be getting a much better plantthan we have since it will be built in the light oftoday's knowledge, whereas we have had to incor-porate improvements as we went along.

Future DevelopmentThat leads me to say a few words about future

lines of development. The ideal time to produce anew invention is as soon as you have exhausted thepossibility of licensing the last one and when youhave sufficiently written down the tremendous capitalinvolved in the invention to be superseded. It iscertainly not economical if inventions supersedeinventions so rapidly that there has to be a constantcapital loss to be provided for, or if the consumerhas to bear an extremely heavy obsolescence chargedue to the very short life, but competition does notalways let inventive genius conceive and bring forthat the psychological moments.

For some time now the number of fundamentalrevolutionary developments has been small, but atleast equal contributions to productivity have comefrom the improved operation of existing processes.The Twin has, in fact, operated within the last twoor three years at over three times the speed forwhich it was originally designed, and I think thatfor some short time further our competitive positionmay well be kept by continual further refinementsand improvements on present processes; refinementsthat mean constant attention to detail, not only to

those parts of our cost that form part of the mainmanufacturing processes, but to those parts, such asthe handling of raw materials, or to the better useof fuel, that come before manufacture, and certainlyto those parts such as warehousing and distributingtechniques that come after. For the main manufac-ture, since we are one of the few industries in whichthe plant already works 168 hours a week and 52weeks a year, we can only get more from thefurnaces by increasing speeds; by increasing theamount of glass that we can make within a givenperiod; by reducing losses; consuming less fuel, lesspower, and only to a very small degree, less rawmaterials. I am, however, certain that well withinthe next ten years we must and will see clearly theshape of further fundamental advances. The prob-lems to be solved are known to us. They are beingtackled and by others as well as by us. I hope thatwe shall again get there first.

Use of Work Study MethodsPerhaps this is a point at which I might say a few

words about productivity as between us and theAmericans. We have not sent a Productivity Teamto the United States and we and our Americancompetitors do not know all each other's secrets,but I am sure that despite the fact that we haveconstantly led in invention in plate glass, there aresome matters in which they lead in cheapness andin satisfying the customer with his requirements, andthere are some matters in which they have a definiteadvantage, and paradoxically enough, one of themost important is that they have the benefits ofgreater flexibility through possessing smaller units inrelation to the size of their market. Their market,with the enormous American motor trade, is so largethat they could not possibly supply it from the out-put of only two tanks as is the case with us. Thisgives them a much greater opportunity for massproduction for individual trades, for example, forthe motor trade. They can therefore turn the wholeof one furnace on to manufacturing thin plate glassfor laminating for motor windscreens; they can makewidths suitable for that trade and no other; theycan indulge in automatic cutting of the glass forlarge orders of one size, to a degree that would bequite impossible with our more varied market.

A very great deal of our labour is used in thewarehouse, in examining, cutting and packing; asfar as we can judge from such employment figuresas we see in the United States, the labour force usedin plate glass manufacture over there per square footof output is less than two thirds of our own; withall the valid reasons why this should be so, I amstill sure that we are not fully streamlined, or pro-ductive, yet. Of course, we have for years tackledthis question with the aid of time study and workstudy methods, and we believe we have got very farin their sensible application; there is further to go,certainly; but in twenty years it has been system-atically applied throughout the works—includingthe control of maintenance—and costs have comedown steadily. We know we have certain naturaldisadvantages and certain other natural advantages;

16

but we know too that only by getting our produc-tivity up yet further can our plate glass hold its ownagainst competition from low price sheet glass, lowlabour costs on the Continent, and from Americanmass production. This cannot be achieved hap-hazardly : it requires leadership and team work. Itis because, throughout this century and before, wehave been blessed with men of imagination, visionand drive, real leaders, and with loyalty and teamwork throughout the factories, that our manufac-turing people have been able to give people like mesuch an easy task in selling plate glass in the marketsof the world.

From what has emerged in the course of thisLecture it will be concluded, rightly, that in the searchby the glass makers of the world for new and moreefficient methods, the plate glass industry in thiscountry has not been at the tail of the procession.In so far as it has led the way, its success has beenattributable to the imaginative thinking, perseveranceand courage of an ever-growing number of peopleat all levels working together as a team. In generalterms, there are now as many hundreds of scientificand technical people as there were of individualsfifty years ago and while one hesitates to associatewith imaginative thinkers what is called, and some-times miscalled, " organisation", there has to besome sort of orderly framework within which eachmember of the team will be happy, not only to giveof his best but to give it to his fellow workers.

That then describes the manufacture of plateglass and attempts to show that there are plenty of

problems for us to look forward to in the future aswell, and that we have the will to tackle them, andthe men and the experience to give us a chance ofsuccess. The eventual criterion of our progress iswhether our competitive position in the world hasimproved, and whether or not we are able to offerbetter value for money now than we were beforethese revolutionary changes took place. As to theformer, in the 1920's imports of plate glass into GreatBritain were twice our exports; now our exports aremore than three times as great in volume as then,imports only a third of that level, and exports havebeen more than double imports every year since theWar, and this year will probably be four to five times-that level.

Value for money is compounded of price, qualityand service. I have shown you that the result of allthese developments has been a great reduction inprice in comparison with wages and most materials.Quality, too, has improved, so that with an enormousgrowth in the proportion of high quality glassdemanded and a great reduction in the market forthe low quality by-product, we have still been ableto raise the general standard. Good service hasbecome immeasurably more difficult to give, thanksto these inflexible mass production methods; but ithas been maintained, at the cost of carrying muchheavier stocks. In total, unquestionably this progresshas resulted in the consumer getting much bettervalue for money than ever before, and this tendencywe confidently expect will continue in the future.

REPORT AND DISCUSSION

AT a General Meeting of the Institution of Produc-tion Engineers, held at the University of Leeds

on Monday, 9th November, 1953, the first GeorgeBray Memorial Lecture, entitled " the Manufactureof Plate Glass", was presented by Sir HarryPilkington, Chairman and Managing Director ofPilkington Brothers, Limited, and President of theFederation of British Industries. The President of theInstitution, Mr. Walter Puckey, occupied the Chair.

Among the principal guests were Mr. Clifford Brayand Mr. George Bray, in memory of whose father,Colonel George Bray, the Lecture has been estab-lished.

Before the Meeting, the President gave a sherryparty at the University Staff House, where membersof the Yorkshire, Halifax and Sheffield Section Com-mittees, together with leading personalities in theglass industry, were able to meet Sir Harry Pilkington.

The CHAIRMAN, in opening the proceedings, saidhe felt greatly honoured in being allowed to take theChair. His temerity in coming to Yorkshire to do sowas due solely to the fact that this Lecture must beregarded at a national one, and the platform as anational platform. The Institution were very grateful

to the authorities at Leeds University for providingfacilities for the Lecture to be given in such asplendid room, and to the Yorkshire Section for theirinvaluable help in organising the Meeting. It wasonly right that Leeds should be the venue for whatwas hoped would be the first of a series of lectures tothe memory of a very distinguished man.

A Distinguished YorkshiremanIn 1951 the Council of the Institution was asked

by General Appleyard, one of his own predecessors inthe office of President—and he was happy to seeGeneral Appleyard that evening—to accept a sugges-tion put forward by Mr. Nurrish, who was at thattime President of the Yorkshire Section. He wasdelighted that Mr. Nurrish also was present andwould take the opportunity to say how great a debtwas owed to him for his work as President of theSection. The suggestion put forward by Mr. Nurrishat that time was that a fund should be set up tocommemorate a distinguished Yorkshireman, ColonelGeorge Bray, who had himself done a great deal forthe Institution as well as for the Yorkshire Section of.which he was at one time the President.

17