5x THE DESIGN OF VALVES AND FITTINGS By G. H. PEARSON AMIMech.E, AMID F, A.1MarE, LONDON SIR ISAAC PITMAN & SONS, LTD.Messrs. Cockburns, Ltd., Cardonald, Glasgow. Messrs. Crosby Valve and Engineering Co, Ltd., Wembley. Messrs. Dewrance & Co. Ltd., London, The Edward Valve Inc., East Chicago, U.S.A. Messrs, Hopkinsons Ltd., Huddersfield, oer Messrs, Richard Klinger & Co, Led., Sidcup. Messrs. Newman, Hender & Co, Ltd., Woodehester. 1 Messrs. R. L. Ross & Co. Lid., Stockport. Messrs. Ronald Trist & Co, Lid., Slough. 2. ‘The National Boiler & General Insurance Co, Ltd., Manchester and to the Editors of The Mechanical Word & Engineering Record “he Stam Engineer, The Indutrial Heating Engineer, and Compresia Air © Hyieautc Enginering for kind permission to make extract From the Author's past contsibutions to these Journals ‘Acknowledgment is also due to my wife for her eneouragemen throughout the compilation and for her invaluable assistance ii checking the Mi ‘Also, that this work has even been conapleted is largely due t the enthusiasm and energy of my Secretary, Margaret Whitson whose desi to see the work in print was ever a great incentive fustained effort. Her asistance in typing what must have been # very dificult MS. is beyond assessment. also take this opportunity of recording my grateful thanks Jean A. Bird, BSc. (Eng.), for her neat execution of some of the ‘graphs, to Mr. P. Field Foster, BSc. (London), M.Se. (Wales) AMLUMedh.E., Wh.Ex,, for kind pension to include the result of some of his researches on the behaviour of cylinders of nom circolar section under internal pressure, to Mr. Ee M. Bywel Curator of the Railway Museum, York for his git of photography ‘llustrating Hackworth's fist safety valves, to Dr. Ing. Guide Ucelli, President of the Museo Nazionale Della Scienza e Della Tecnica, Milan and Managing Director of S.AZ. Costruzion Meccaniche Riva, Milan, for his gift of photographs illustrating the stop cock which he salvaged from Emperor Caligua’s fis galley, to the Secretary of the British Standards Tnsttution for hi kind permission to draw freely upon those British Standard Speci fications having a bearing upon the work; last but not least to those many friends and acquaintances who, by their sincere encourage: ment, have been @ constant inspiration during the process 0 compilation. GHP. ‘Newport, Mon. ebay, 195¢ 5 7 CONTENTS Preface Isrronuarory . - CooKs Ground-plug coc, ssbetospacked cocla;, plug design; com ‘tions goversing fdighines; Tubdeated‘vaes and) cocks; fleewe packed code Senzw-pows Stop VALvEs—Szarmsos aND Booies Definitions globe ad junction valves feed check valves; variations ‘eteaucdon: famed and rated seasos foes ang facet determination focal al free to coe valve eilecuvely; curves for rapid de Termination of width of sea coasting faces globe valve bodies Scxew-pows Sto VaLves—Vatve-acTuaTina SCREWS. Single» and ruliple-threaded serewe lation between load and ‘Sorts ficiency of squaredizeaded screws; eficleney curves for Squaretareaded screws cooficiet of felon’ reyrsoefciency EiSguarestrenced sews, seme and vee treadey elBelency of ‘ee threads; fiency curve for sharp-angle ends Screw-pown Stor vauvss—Bourep Covers Conditions governing efztvenest of jint; eandvich joint with Selig pacing apd with (pitas Sroed oop ae al, illic pas rough bota flanger; spgote! Jens wh grow AES cRigtel joint nay bol taterasy pe of los tle nos afcoves fangs mma operating mecbanion attachment SFSplndle to valve members one and fwowpice spindle cgure Secber inpeseds length of thrend. engagement? lubeleation; ‘ieeuucatte hendwneels crouead; andheel;stufing boxe, ‘lands snd packings; land bols or suds Paracter Stipe Vagves . Gonilions govering Auidightnes;” deternation, of width of cating sonlcts ype ef seatngy.dtsign of ds; determination ‘Stn oece wees cloeures cei of actuating spindle, proper- lS efopindle sections; spade collar dimensions; tarquerequres to overcome, ‘cillar freiony advantages of employing bypases; GEegia of spindle ead cngagentar; play stepmember: ‘Sothead; sling bow and gana Sarery avo Reuter Vauves : Porpoe of efey and elif valves: basic eypesy evolution of the Eid cher, matings chaterpronenes and Ss prevention: ‘Shnltions alecuag”fecwue leadings sean contac presure: 5 & 8 135yh of senting fies; Board of Trade Rules; British Standard Roles. "Spring safety valve ordinary highlit and fl safety valves; Miebeit featuees; minimum aggregate sey” accumu on retention of eating sings; ea of docharge over, wing deatances; spindle design; spring designs properins oF waste ‘tram pasage; compresion sews, Ramborom valves Pop" safety Yale; easiag’ gear Lever salty valves mechanis Tot Speraton; proportions of checeweights) arachients of cheer ‘Weights to lever spindle and pin Sealy den of levers age ‘uti righ steam and low-water sey valves fimctions snes ‘toperations determination of centeal excweght; eterainstion Dofeheceyeight; determination of feat lance weighty deta: ation of maximum Mowing off preware fora given Boat system: albmetal floats and caleaatone fnting thet, din of con traction, ‘Beading salty waves: open dcharge pes seatnge and thee, method of retention? stem desig deternination et Fequted deadweight; contollad ducharge 8, Now-rerurs VaLvEs . - Functions types. seaings and clacls; arca of opening; ante Snzaning dvi, ean of dy coer aes Spiel Satis ‘offi pave cover design iting seangements| 9. Repucie Vatves anp Surpies Vauves Function everand.weigh type; diaphragm and spring controlled ‘yper design; determination ef value that cissesions to, Boner Warer Leven ALas . 0. Desisbilty of maigtazing constant wate level; single beam alarm; desription cfactions ype offoas; mothanict ef tperations Timing prestres postioning of foes dud beam alti; vataton in depth of immehon; machancr of Gual betta alarny whisde ey exer alas 11, Warer-tpve, Gavors,- Presume Gavors ano ‘Watertevel gauges: fonctions, cleaning plugrp. operating Hanes; Aisining arrangement; protectors gla ube Gametey Automate ‘sing evi column’ mounting; altemative types of slums Sudlary wp valves or cel salty plugs rex Ewuges Peers |Troges ad ceserias Bourdon ype presure gauge siphons 12, Hypravue Stop Vauves . 0. * : Condition affecting closure ghtaea; mechanics of prestietealings iechanic of the actuating ser tongue reqised ts oyetcome sae fee; tctional restance of "U lethery cover ange dein ‘mpariton of BSW. and BSF. studs o belay deternaton of hat dices 13. Wepor-care Stuice Vatves . Principe ofthe wedge; fares acting oa gate and determination of ‘Ela lve to lo vats tongusregured to cee canary piel disign examples spindle deals; wedge msts wedge gate} dine sf hesoy saing box and glands ‘gest opeaon ©” 2 25 ay} 15, Vauve Cinesrs : 14. Equuserio FLoar Vatves : 334 feos a eh Hedy Hasina eee eel eee . 396 Tinpertance of employing clinical configurations; thiciness of he of ean cece Porm yori er ‘eideations Tor stel casing, eh tings, sod see orpngtt Hike Be ental lec ec pens i acy an Boiler Formulas wioughtstee and wrought ton seampipes} cat el arnt pipe and bends; castivon asaight pope tod bends 16, Gearixc FoR MANUALLy-ormrareD VALVES 38 ‘Limitations of dzet operation and conditions of tooth engagement [Engpur geass analy of frees imposed; worked example, Double ‘edueton gearing: combined spurcond:werm ruedon gearing; design of worm fevtion gvarng Appendix . 397 Appendisa Se gt. Indes + 493PLATES (Betiven pp. 138 and 139) - Barty Roman Srop Cock IL, Tororey Hackwortn’s Spric-Loapep Sarery VALVES IIL, ‘Tesr-uc ror Invesnioanine Frow Turmunence my 4 Grose Vauwe IY. Tesr-aio vox Isvesnieanine Inpucep Sransses wv 4 BOWLER Teotarna Stor Vauvs INSETS 617. Cocxouns Srazascive Panatit Supe Vauve. 132 rot, Truss “Ho” Avan 275 480 nia. Ware Gaver of Mopsrw DesioxCHAPTER 1 I i INTRODUCTORY ‘Tus origin of valves is lost in antiquity. Any device which enables an aperture to be opened or closed for the purpose of controlling fluid flow may be deemed to be a valve, however crude may be the manner in which this is effected. When man contrived to - fashion mechanical devices to meet his everyday needs and to administer to his comforts he must have been contronted with the problem of devising valves in some ferm or another. Eariy workers in metals foand that a blast of air was essential to the melting process; thes the bellows was born. The fundamental idea of the Lellozis was conceived by different peoples, despite their separation by dhe oceans anid other impenetrable natural barriers. ‘Thomas Ewbank in his Hydraulic and other Machines for Raising Water, publisined in 1847, writing on the subject of bellows states « “But by far the most important improvement on the primitive ‘yellows or bag was the admission of aie by a separate opening, a conirivance that led to the invention of the valve, one of the most essential elements of liydraulic, as well as pneumatic, machinery... . Now a forcing pump is merely a syringe ‘with an additional orifice fer the liquid’s discharge and having oth ite receiving and discharging orifices covered by valves or clacks. Ctesibius,* therefore, did not invent the piston and cylinder, nor was he the first to discover the application of these to force water, for they were in previous use for that purpose. Was he the inventor of velves? No, for they were ‘used before he lived, and appear always to have been an essential part of those instruments.. They were employed in . clepsydra (water clocks) and were most likely used in the hydraulic organ of Archimedes... .” ‘The earliest valves probably comprised a leather flap arranged to cover and uncover an aperture alternately, particularly in sal Treasury published by Longman, Brown, Green, and Longman:, London, in 1839, gives the following relerence SCiatbian of Alexandeia, aa eminent Tathemasican. He ie said to have fgveaied the pump: and he tonstructed a clepaydea, or water-cock, in which the moving pawer was felling column of water. A work otis entitled Gandeioy 4 7 the Art of diciding and measureig Bodies, is to be found in the Vatican Libeary. Fre fourishet in the tOsta Olympmad, or about 120.0.” 1 Naunders Bio- tae wastes UF VALVES AND FITTINGS pumping or bellows mechanisms, where regular opeaing and closing was an. essential requiremen: (nen-rerurn valves). It is probable that the cock came very much later in man's development, say when the necessity for controlling liquids became urgent, and the earlier models were of wood, giving ‘way to metals as progress ‘was made in the craft of metal working. According to Horus Apollo (or Horapollo)," an Egyptian of the fourth century A.v., the priests of that time were wont to give the forms of a lion to “the mouths and stops” (cocks) of conseczated fountains because the inundation of the Nile occured when the sun wes in Leo. The early workers in metal, castor wrought, were prone to ornament their handiwork, various devices being commonly fashioned to represent human or animal form. Thus the cock gata its name from the male fowl of tha: name since it was customary to fashion the termination of the plug into the head and comb of this Dird, the water issuing feom its beak. Bronze or brass cocks were common in Ancient Rome where the more wealthy were accustomed to provide themselves with baths and fountains, and some device for controlling the flow of water was essential. From Vetruvius we leara that every main pipe, pasting through the strect had a large cock by which water was “let in oF excluded” and these cocks were manipulated by meaas of an iron Key much in the same way as they are to-day. Several speciniens of ancient cocks are still in existence amongst which raay be cited the one discovered in the rains of a temple built by Tiberius at Capri, and preserved in the muscum a: Naples. This finz example was found attached to a reservoir but the method of attaching it (o the connecting pipes is open to speculation. ‘The upper termination of the plug is interesting and might be copied with advantage in piesent-day examples; it consists of a zenewable member dove- failed into the top of the barrel and arranged to zeceive a tommy- barf. Most cocks fail by shearing of the squared termination of the plug intended to receive the operating key but the Romans antici= pated this weakness by making the termination renewable. Plate I and Fig. 1.1 depict a very fine example af an early Roman bronze cock of unusual interest. This cock was taken out of * Meandn?'s Biographical Teeasay, published by Longman, Brown, Greco, ad Longmans, London, in 1651, give ie following retcrnce ““Horplle ot Horus Apolo, a gcemnoatian of Panoplus ic Egypt in the fourth centaty wkotaaght hse ih Alcsandeas and nore is Comsetvenies Phere fre extant under his mame two boola’ concerning ‘the heroslypbics_of the Egyptians, pgnied in Creek by Aldana tgoge bus thee athens cgseeuomble The plug termination is similar to that shown in Fig. 1.1. pub PAYEE at it | | tv tg mrronverony | slo: Emperor Caligula’ first galley, salvaged from the bottom of Lake Nemi in July, 1929. Caligula was born in 12 A.b. and was Enaperor Esvom 37'to 41 A.D. Despite this particular cock’s immersion for close on 2,090 years the surfaces of the body and the plug were excellently preserved, being ground ag smooth as glass, and the ‘opinion has been expressed that the finish could scarcely be excelled to-day. With the advent of the steam age little or no thought appears to have. been lavished~—or apparently considered necessary—on the k | Fie. ta carller designs of dtcam stop valves, if the scant illustrations of early Firtings are any criterion, and leakage at the seatings and elsewhere must have been common. Tt must be remembered that steam presiuées were low in the earlier boilers, about 10 Ib/in? in the arly part ofthe last century. It|was not until about 1850 that as a result of long awaited improvements in materials and processes-and the nevessity for a wider appreciation of the underlying principles of design became Sppireat—any really serious attempts were made who may then modify his original plans se aa t0 provide the ideal body’ contour. “The value of such preliminary investigations prior to embarking upoa large-scele rmatufacture wil be at once apparent, iS mpee aioe Plate IV illustrates an Edward 10-in. goo Ib/in 1,000°F pattern globe type non-return valve (employed as ¢ boiler isolating valve) being subjected to a pressure test for the purpose of carrying out stress analysis. The equipment used for stress analysis is the Baldwin type SR-q clectric strain gauge which is capable cf measuring strains of one millionth of an inch. ‘This analysis is made subsequent to a preliminary overall valve body stress analysis making use of the Wstrenscoat” brittle lacquer method. I the background will noticed che Edward air-fiow pressure-drop testing equipment for measuring the pressure differential across the valve which wll accurately zecord oressure changes as minute as one hundredth of ‘a pound per square inch, "The operator depicted on the extreme left of the photograph is6 THE DESION OF VALVES AND SITTINGS a ego a rn sepa Ge, ter ee Fee a a eee Se en ne ee ale CHAPTER 2 (| j COCKS By fir the simplest device ever conceived for the manual control of ‘luid flow is the. common stop cock. ‘To whom eredit is due for its invention is lost in antiquity and its origin must be ascribed to those communities alive to the necessity of conveying their drinking ‘water in pipes as distinct from the alternative of repeated perambu- lations with pitchers or similer vessels with which to replenish their supplies, a task which would appear to-have been the sole prerogative of the womenfolk of those times, judging by the evidence exhibited in early Egyptian artistry, whether in admiration of their skill or poise it is difficult to decide. It is equally difficult to decide whether it was man or woman (probably the letter) who prompted the idea of conveying water in pipes or conduits, as distinct from the alternative irksome and precarious mode of conveyance referred to, or whether it was the Sutcome of much heated argument between the two. At all events something apparently had to be done in the matter by one party or the other, or both, and the fact that a certain Sarah Dixon, on the untimely death of her husband, John Dixon, undertook the sole ‘management of his business in Leeds devoted to the exploitation of his patent granted in 1822 for “Certain Improvements on Cocks,” is scarcely admissible as evidence of woman's interest in labour- saving devices having only manifested itself within comparatively recent times. ‘| . Tt is somewhat of a paradox that the natural inventiveness in ‘man| which is usually directed towards reducing manual exertion should often result in providing him with additional problems of an entirely different nature. The introduction of pipes demanded gome form of controlling device, to wit,.a valve or cock. ‘The common e4ck must have appeared to its inventor to be imbued with essentially simaple characteristics (as indeed it is) yet the, problem of iaintaining fluid, tightness in service must have provided’ him with something more to think about. Perhaps he did not worry unduly about a leak or two and there were no tell- tale *neters to bother about. || Tie earliest ‘were of wooden’ Construction, xamples of which are in use at the present time, notably in the beer, wine and spirit trade since they are not contamination-prone to any great 7 iextent and, being cheap, can be discarded in favour of mew ones when they are unfit fo: further use. ‘The simplest cock consists essentially of a body and a plug, the latter conical in form and arranged to rotate in a complementary Puve. PLUG IN OPEN POSITION Bre. 2.1. Groupee Gpcx conical hole formed in| the body. | The plug is provided with a transverse port or opening which, by rotation of the plug; may be brought into line with similar ports in the body, permitting flow. By rotating the plug through 90°, flow is arrested, since|the ports : \ ) i | : Two ways ope BhE MACE tS oy : Two wave open She WF LO ES ss Fro, 22 } ALL WAYS CLOSED i} . Ro. 2g ALL Us cLoseo‘will no longer register. This witl be apparent trom Fig. 2.1 depicting cock in its simplest form. ‘There are many variants of this one basic idea; for example, multiway flow may be obtained by a.ranging tne ports in T fashion as depicted in Fig. 2.2, this type being gencrally termed a three-way Fro. 24 cock. A similar variant is the three-way cock depicted in Fig. 2.3 wherein alternative port positions are clearly indicated. in this design use is made of L ports situated in the horizontal plane, whilst 2 change of flow from the horizontal to the vertical plane (or vice versa) may be provided Ly inverted L ports, as in Fig. 2.4 ‘Thus the cock is a most obliging device, being so readily adaptable to provide for multiway flow in a manner which cannot be equalled for simplicity by any other member of the valve family. h cocks a Ground-plug Cocks From an apptication point of view, cocks are basically of two kinds (apart from the numerous alternative port arrangements); ial OX’ Fic. 2.5, Aspesrotacken Cock they ate cither of the'gromd:plug (metal-to-metal) pattern, or they are asbestos packed, Ground-plug cocks are thove in which the plug is iu intimate metallic: contact with the body at every; point on its ing furlace (Fig, b.1) whilst asbestos-packed cocks|zre those in which the plug is a1 ‘to bear only on;the surface of the asbestos : Hl hagare “nme ayy grooves formed in the cavity of the body (Figs. 2.5 and 2.6). The latter type is intended primarily for steam service whilst the simpler ground-plug type is more usually empleyed. for the less arduous duties of air or liquid contrel, although tars and ‘viscous substances—even-cerhent slurry—are often controlled by cocks of the ground-plug patt.:n. Se Se! IGM SLOG Bic, 2.6. SucmoN op Asmesroeracs> ‘Cock stiowane Pace Gnooves Notwithstanding the -wide diversity of patteras employed, the ideal included angle of taper of the plug for all types of cocks is 10°, ‘This has been found to give the most satisfactory compromise since too slow a taper will promote seizing of the plug and the body, rendering the cock difficult to operate, whiist a quick taper will offer no resistance to the lifting effect of the line pressure acting on the unbalance area of the plug. Since it is somewhat difficult to match a given taper, the ideal combination is that in which the hody cavity and the plug are both ‘turned in the same lathe and special lathes are available designed solely for this particular line of manufacture. ‘The plugs of ground-plug cocks may be prevented from lifting under the influence of the line pressure, or from other causes, in various ways; they may be restrained by means of a holding-down plate abutting against a shoulde: formed on the plug, as shown in Fig. 2.7 (a), by means of a serewed ferrule or gland, also abutting aginst a packing ring ititerposed between the ferrule and a shoulder formed on the plug as in Fig. 2.7 (5), or by means of a nut and washer arrangement at the base of the plug, as in Fig. 2.7 (c). The | | \ | latier i a neater arrangement end one zauch fevoured, although in ‘Bi taample ths bearing surfaces between the washer and the body o © © Fee. 2.7 rust be ground:-in to ensure a seal being made. The washer plate torte atranged to rotate fm union wth the plug and ti is THasted by arranging a square hole in the wasber plate to engage Stquated shoulder on theplug Alternatively, the hole and shoulder. sine DESION OF VALVES AND FITTINGS’ may be ogee shaped (Fig. 2.7(d)). Were this proviso ignored, and a circular hole employed, repected operation of the cock would, tend to slacken or tighten the base nut. The latter, for preference, should be secured against relative rotation by a split pin or other approved fastening. Asbestos-packed Cocks We have seen that 2 disadvantage of the plug cock is that the plug is apt to become jammed, generally as a result of infrequent ‘operation, and this constitutes something of a danger since itis not an uncommon occurrence for the shank of the plug to shear asa result of excessive force being applied in a vain effort to rotate the plug. Glanse 5 of BS. 759: 1950 (Valves, Gauges and Similar Fittings for Land Boiler Installations) states: ‘The bodies of cocks 1 in. bore and ‘over shall be packed with asbestos or other hect-ressting racking and shal! have provision, other than the gland, for securing the plug. ‘The clacise refers specifically, of course, to cocks destined for steam service. It is customary (and should be obligatory, if adherence to the fore- going clause is rigorously observed) for the bodies of cocks to: be asbestos-packed, that is to ray, they should include an arrangement. of grooves within the body as Gepicted in Fig. £.6 (and also in Fig. 2.5) each groove encircling the ports and stemmed with shredded asbestos packing. In cocks destined for steam service indurated packing should be employed, this being a type of packing which is self-vuleanizing under the influence of the heat of the steam, although it is customary to bake the cocks thus stemmed piior to their leaving the maker’s works. In a correctly-packed cock the plug should actually be supported on the packing itself and not be in metallic contact (although nearly so) with the body. Tt wilt be appreciated that this makes fot easier operation. In fact, the asbestos-packed cock represents the most practical interpretation of the basic principle of the cock. Reverting to Fig. 2.5 i¢ wil! be appreciated that dus to the conical form of the plug there will be a tendency for this to lift under the influence of the line pressure acting on the difference in exposed areas. This tendency will require to ke countered in some way and is best accomplished by a holding-down plate A (arranged to abut against a shoulder conveniently formed gn the plug) and two holding-down bolts C. ‘TLe holding-down plate in the design shown serves the additional purpose of a locking-guard for preventing withdrawal of the operating key D excepting when the cock is in the closed position. ‘This featare is in compliance with Clause 35 of BS. 759: 1950 which states: . . . Any hey or similar device yor ny BB Bee t COCKS | 4 5 hog the ck hal uh at cate rene les ‘galoe or cock is fully elosed. i a tie cover ‘dae IE eorves only to compress the groove-and-header | acting Pan to seal the body, no reliance being placed on This eee Folding down the plug B. Te will be noted that this design | itm for holding (ow enqad gland G, fendered exental by the i Buea Panacine Sue Beowore VALE 4 , sting ingtes of voviion of a small communicating port AY permitting ingress of provision of op aide ofthe plug and so wholly of partially counters Tees the prewure upthrure on the plug. ‘The plug may be balancies fihaut disturbing the groove packing (8 coma fear). Not cers resent examples Ereound practice and in the sic coc call for every refinrentincuded in sles the design portrayed, however, is representative se Pie atetent compleaon i er of sound pace Se eons sgn i sometimes favour 2 en cal pee he co ora al 2 Boies blew. down och Qavrowed ip nme eile where Seite although. 1 Megson operated valve ofthe parallel I ee purl de Bowaheaoe (g 2)6 THB DESIGN OF VALVES AND FITTINGS ‘Ac all events the cock described complies in every respect with | Clause 35 of BS. 759: 1050. ‘The practice of making cocks pressure tight by groove packing is attributed to Mess. Dewrance & Co. Ltd, who claim to have introduced this method in 1875, A furthe= improvement was the introduction of their indwated asbestos packing which is resistant {o the absorption of water. This packing is selehardening and vulcanizes under working conditions if the steam pressure exceeds 50 Ibjin.? Consequently, it should not be used for the stuffing boxes of valves or cosks. For the purpose of identification it is beoloured red. Another form of indurated packing, coloured blue, is intended for the stuffing boxes of valves and cocks and does not vulcanize, Ordinary white asbestos fibre should nct be used for packing cocks as it has no pronounced resilience and calls for too frequent renewal ‘Of all the devices ever conceived for the control of auids the cock is the one more open to design by ge than by any other means There is so lite which lends itself to analytical investigation with any hope of reasonable reward; the body is at best an arrangement of varying configurations, is subjected to combined stresses resulting from the application of pressure and from the wedging effect of the plug so that experience and intuitive con‘ec-ure must olay their part in the design of such devices. Als, the plug cannot be designed in the strict sense,of the word since it is a complicated affair with varying diameters and rapid changes of section. The presence of the ports cutting through the zones of greatest stress distribution complicates the issue and so it is as well to forsake mathematics in favour of judgment backed by experience and with one eye, as it ‘were, focused on the machine shop with en occasional glance dinected towards the faundey. Plug Design ‘One of the essential requirements in cesign is that of providing fat least the same amount of area through the ports as is presented bby the connecting pipes, or bores, of the cock, athough many examples fall short of the mark in this respect. Where the design. ‘of a complete range of cocks is in contemplation it is important to adhere to a strictly proportional relationship among -he three basic {simensions, namely, bore of cock and width and length 2 pots in Tehe plug. In what follows the width and length dimensions are taken co-planar with the horizontal and vertical centre lines respectively of the cock body; not that this makes much difference but it is as well to define clearly the manner in which these18 ‘cue pEsrax dimensions are acen, Tn Fig. 2.9 the width w, and the length A | fare clearly and unmistakably delineated. (For the sake of clarity Sharp-edged ports are shown and the angle of taper exaggerated) Tt wil be noted that all the inclined surfaces, e. ouside and inside sunfaces of the plug ard the edges of the port, converge upon the common apex 0, It i particularly desrable that the port | eelges should converge upen ‘0 in this manner and for restons ‘which will be best understood from Fig. 2.9 studied in conjunction | with Fig. 2.20 in which latter illusestion the port edges are | Aeligerately arranged parallel tothe vertical axis, a most undesirable | arrangement, as willbe apgarent late. Tt wil be seen in Fig. 2.9| that selected points X and ? shown inthe front elevation ie on the | straight line X0 and remain on the same line XO projected in| plan view. Consequently, the two lines XO and 70 in plan view both subtend an angle @ 20 the centre line C From this we may in‘er that on any transverse plane (such as / 2c’, DD’; etc) containing the port, the ratio of length of contact! {othe complee—or uninterrupted—perimeter of the plug. is| constant on any plane taken normal to the vertical centre line.| Expressed mathernaticaliy, if d, denotes the diameter of the plug inva plane containing X (se Fig. 2.9), dy the diameter of the plug| 4 plane containing 7, p, the arcuate length of contact of one side of the plug in a plane containing and p, the corresponding jength in a plane containing 7, then fy Pe I, ms, ~ nd, where is any diameter and p, the corresponding arcuate length ‘This leads us at once to adopt the stratagem of determining the requisite port width, amount of cover, ete, on the central plane CC’ in the Knowledge that a similar relationship wil obtain for any} other plane parallel to CC’ above or below this convenient datuan In short, that lationship which applies to the central plane wil apply <0 any otter parallel plane. provided all inclined edged converge to meetin the common apex O. ‘In Fig. 2-10, wherein the port edges in a lengthwise direction ar arranged paralel to the verdcal axis ofthe plug, it wil be seen thal ihe foregoing relationship does not apply. Instead, the arbitrarily selected points X" and "(corresponding to points X and T in the preceding example) do not le nthe same straight line converging| Gn O' ‘plan view); corsequently, the angle of contact will be varying one throughout the length of the plug and the ratio arcuate length to uninterrupted perimeter of the plug will suffer «Bq 2a Similar variation. The impact of this on design shouldbe at once apparent Tt curtah the amount ofcver so exental for ensuring fuédtightnes and particularly atthe point of minimum diametes the bate of the plugand this desirable feature is more pronounced in the smaller sae, Having thus oudined the geometical procedure to be adopted in the design of the plug we tay proceed to the next step, that of suiking a satsactoy relationship Bemeen the bore with, ed Tength of the port For al practical purposes the area of each port ray be taken athe product of ay ad hin Figs 25, Tn Pgs 0 this product wil accurately detemine the ares. Due allowance should be made forthe lot of area occasioned by the necesty of raising the comers of the ports ftom considerations of mauling anu asa precaution against the development of cracks, ‘The mean width of te port taken on the central plane OCT and denoted by'ty in Fig. 2.9 and 2.10 should be made equal tone half the bore ofthe cock, an eaty rule to remember. ‘Then equating the area ofthe mean rectangle (asiming sharp corners) tothe ares of the bore ” a w,h = + where d= the bore diameter whence - Bg aa But since Eq. 2.2 may be written Ey. 23% ‘The length derived from Eq, 2. will require to be increased slightly according to the comer radius adopted. This radius need @ not exceed Be 6 Conditions Governing Fluid-tightness Fluid-tightness between the plug and the body contacting surfaces is an essential requirement but whether this is achieved by accident or design it is difficult to say. ‘The peculiar adaptation of the wedge principle so admirably exhibited ia the cock is in every way conducive to the attainment of fluidtightness. The sl taper sun)enabling relatively high cont y high contact presure 10 be exerted om the sirtace ofthe plog fr only 2 modeate applicator af eat Paes the high degre of nish ofthe contacting prtces volte te Son afforded by the working su fact osha eee, be) all operate in aver of uantning Soe Ro. aur Wat ac th anion to be leven ne eet wee Suse fa Seg 9 renin ped tacting surfces (achieved by grinding and lapping) snd (0) tie rnb sanigrars 2? S00Ns 2 ei) onde circumferentially when the plug is in the closed position. This will ian ie MEMES Mena on. TA value of ¢ = ~# but this may be exceeded with advantage. Lubricated Valves and Cocks ‘The frictional resistance of ordinary ground-plug cocks and their proneness to binding fast after periods of inoperation have often Caused them to be shunned in favour of some other fitting equally unsuited to the task in hand and many-and varied have been the ideas brought forward in an endeavour to obviate these disadvan- tages., Eliminated to some extent by groove packing in the case of the qabestos-packed ¢ock, the liability of attack by corrosive media renders such a cock stiil unsuitable for many applications, particu- larly in the chemical industry. | Since any macl.ine works all the betier for a modicum of Iubrication it was natural to suppose that if the plug and body surfaces of the ordinary cock could be lubri- ‘cated @ much better performance might be expected. Credit for being the first to put this idea to the test goes #0 a Swedish engineer, S.J. Nordstrom, who was engaged in the design, construction and operation of cyanide plants in Mexico, and who was obtaining indifferent performance from the many gate valves ar plug cozks handling cyanide solution and slurries, His experiments soon ‘proved that by interposing a film of lubricant under pressure between the machined surfaces ofa plug cock, considerable resistance to the ingress of che working fluid was set up and friction and corrosion were considerably reduced. 1 will suffice to describe two such types of lubricated plug cocks manufactured in this contr)’ and representing the practical embodiment of Nordstrom’s original idea, together with certain additional refinements dictated by subsequent experience with this class of fitting. : ‘These are (i) the Audco Valve manufactured by Messrs. Audley Engineering Co. Ltd., Newport, Shropshire, ard (ii) the Newman- Milliken Valve manufactured by Messrs. Newman, Hender & Co. Ltds, Woddchester, Glos. (The term value is preferred by both these manufacturers, although the fitting is essentially a cock, Hereafter both typer will be referred to as calves in deference to the designation preferred by both these firms.) Fig. 2.12 depicts in exploded view a standard Audco valve suitabie for the moze common applications of fluid control. Its ction will be apparent from the illustration. The standard type is constructed on orthodox lines, the plug operating stem and tapered portion beiig arranged san integral unit, a gland and its packing bearing on the plug to retain this in its ordained position. With appropriate modification this particular design is available for ail pressures np to goo Ib/in. but for higher pressures (from 600 to 5,000 Ibjin.t) the “Hyprescal” valve is the one advocated.This is illustrated in Fig. 2.13, the exploded view again being employed for the better understanding of its constructional features. In the “Hypreseal” lubricated plug valve the plug is assembled through the base of the valve-and positively seated on a ball thrust, and threaded pressure screw, the latier passing through the steel cover plate. This arrangement is slightly resilient but the paris are unaffected by changes in pressure and temperature and se render Ber rerin aire unlikely the necessity of subsequesit adjustment. ‘The orthodox type of spindle and packing arc eliminated; instead, the detached operating stem is screwed into the ody, the plug being rotated through a floating equalizer rivfg. In this way the plug is immune, from disturbance arising out of externally applied forces, particularly those applied in a lateral direction as are occasioned by the thrust, of the operating media. : In the Newman-Milliken design depicted in Fig. 2.14 the plug is cylindrical, as distinct from the conical plug featured in the two previous examples, Furthermore, no gland or packing is incor porated since i. is claimed that these refinements are unnecessary. r account of the shouldered-head construction of the plug which i | i ice cocks 23 Intricated in addition to being a ground fit in the body cavity against which it bears. ‘Some users favour the tapered plug design of the Audco valve and some the parallel plug of th Newman-Milliken valve but in both designs the znode of lubrication is substantially the same; an C—O a soreiacee wares Ee ee SSeUSsa RS. Fro. 29, insoluble lubricant in stick form is forced through the various ducts, ‘communicating with the rubbing surfaces through the agency of a Jjacking-screw, the simple ball check valve shown preventing the lubricant from Eeing forced back by the pressure of the working fluid, Te will be appreciated that the grade of lubricant is dependent upon the nature of the working fiuid itself, numerous grades being available to suit a wide range of appiications.a . cocks 25 Sleeve-packed Cocks Before dismissing the subject of cocks there is one particular design which merits special mention, namely the Klinger sleeve- packed cock (se Fig. 2.15). It has aircedy been pointed out that cocks employing a, conical plug depend ‘or their fluid tightness on the intitaacy of fit hetween the contacting surfaces of the body and the plug., Tt has long since been recognized that if the plug could bbe made cylindrical whilst yet providing fluid tightness between the plug and the body the stiffness more usually associated with the sconical type of cock might be eliminated. This has been achieved ——wepicent ey yuaricany——__ Becerracue |_—eate cneck sear open HonizonTaL, | SS norrom cover: | SS prckime sueeve Fro. 25 in.the Klinger sieeve-packed cock, the principle employed being that of introducing a renewable sleeve of compressed asbestos between the phig and the body. ‘The sleeve is inserted as a single unit (see Fig. 2.16). . ‘The cock plug is cylindrical and ground to a fine degree of surface finish, whilst the renewable packing sleeve is made from resilient Kiingerit compressed asbestos built up in layers as shown. On one side of the sleeve is formed a ridge which engages with a corres- ponding groove in the body 20 as to ensure tha’ the ports are always | inline. Thissleeveis compressed endwise by the action of the screwed 4 cage and has the effect of forcing the asbestos into close contact t with thé body cavity and with the plug, thus ensuring pressure- ‘Lightness. Steel eyelets recessed in the sleeve prevent the asbestos from ‘being squeezed into the ports during the tightening-up process and at the same time prevent the line ffuid—which may have a high velocity—from scoring the asbestos packing. These feyelets are conical in section so that there is a tendeney during i tightening-up to squeeze the asbestos more tightly around the plag Fro. a4. dar, Newean-Mitiniex Lusnicares Puvo Vauve“v Hi DESIUN UF VALVES AND FITTINGS : and to push the eyelet towards the cock body, thus avoiding any scoring of the plug when this is rotated. Klinger cocks employing this principle are also. available in two-way, three-way, four-way, and angle patterns. Klinger sleeve-packed cocks have been successfully employed on saturated steam at 2,000 Ibjin, super= heated steam at 800 Ibjin.* and gpo°F and én benzine at 3,000 Ibjin.*, Calculations on the torque required to operate a cock of ordinary construction are best left unattempted and would probably lead to disappointment and conflicting results since there are so many, indeterm.nate factors. Fo: example, the frictional resistance to rotation of the plug is dependent upon the extent ‘to which the ° hholding-down bolts have been tightened, the nature of the metals in contact and the lubrication (if any) afforded by the working fluid, the frictional resistance. of the gland, where fitted, and so on. In_a device of this nature experience and intuition must be the, ~ guiding factors. To what extent these expedicnts will serve! the designer will be gathered from the chapters which follow in which ~_ ‘more elaborate forms of closure devices are deseribed and which ‘call for a greater display of mathematics than the example chosen as a fitting intre-luction to a work of this nature. CHAPTER - SGREW-DOWN STOP VALVES—SEATINGS AND BODIES Next to the common stop-cock described in Chepter a, in order of simplicity for the mansal control of fluid flow, ranks the screw down stop valve. Under this heading may also be included all thoce valves, large or small, variously rele-red to ia makers! catalogues bal Fro, $2 Grove Stor Varwe as globe stop valves, junction stop valees, controllable feed check valves, ete. Tn fact, any valve operated through the agency of serew-and-nut mecharism for actuating a closure snember may be classed as a _ssfevdovn stop valve [This ype i generally encountered ether asa gob alee (Fig. 9.1) | for insertion in a streight pipeline or at a jun tom valve (Fig. 5.2), 8 | pattein suitable for mounting on the standpipe ofa boiler to serve a28 THE DESIGN OF VALVES AND FITTINGS 1 the main stop valve, or it may be conveniently inserted at a point in the pipeline where it ir required to make ¢ right-angled change in direction, , a ‘Both Fig. 31 and Fig, 3'2 depict valves of the outside sorew pattern, that is to say, the screwed portion of the actuating spindle is outside Fa. Gia) Tyres, Sormw-nows Axor Stor Vatve— * ‘Ousinw: Scxew PATTER the pressure space of the valve. This is a most commendable feature, particularly in the caze of steam valves. But for cheapness the inside serew pattern is often employed, a typical example of the design of junction valve showa in Fig. 3.3. SCREW-DOWN STOP VALVES—SEATINGS AND BODIES . 29 Globe valves of the inside screw pattern are also commonly employed but are not to be recommended for steam service, Seatings and Bodies ‘The first iter to claim attention in the design of ary valve is the : seating; in fact it is the most important item of any valve, and, as Fro. 913: Tevteat Semewoowne Avot Stor Vatve— ‘hae Sontw Parse design proceeds, it becomes more and more apparent that the dimensions of many other components are linked in some way with the various diameters of the seating. A leaky seating means a leaky valve, and that much hackneyed slogan beloved of valve manu- |, -factisers proclaiming that the heart of the valve is ite seating + embodies for once every element of truth. It is important, ‘therefore, that caref 1 thought be paid to the K ‘design of the seatings and to the materials employed. In this latter [7 2 connexion the designer is free to choose from a wide variety of ”\ * :materials ranging from fibre composition (for the valve member)SCREW-DOWN STOP VALVES—SEATINGS AND BODIES 31 30 rue Drsion oF VALVES AND FiTTiNGs a3 to high grade nickel-chrome steel or cupro-nickel, depending tipon “the application. | Sedtings include not only the component that normally goes by \ ‘this mame but also it companion member che valve clack, per or i ‘isc as it is variously described; in short, the tw6 components’. | ‘which by mutual contact seal the valve against leakage may both | termed seatngs. yet ‘The point at which pressure-tightness of a seating is ultimately : setured is on a relatively narrow annulus usally described as the | eating face, and this should be ground and lapped to a mirror-like | ‘nish. The contact face of the valve member may be arranyed °° | ~ flat, mitred or radinsed and diagrammatic examples of these threc |. ‘he esanesovn inion pores spect | | axis of the valve (as in Fig. 3.7 (e)) co one presenting an obtuse ~ angle (as in Fig. 37 (9), the latter closely approaching the flat-faced type of seating. To close a screw-down valve effectively against che line pressure ‘wo separate and distinct resictances have to be overcome in addition fo that of the frictional resistance of the spindle in the stuffing box Gvaich will he disregarded for the inoment). ‘These are—(j) the Stlamping force at the contacting faces of the seatings for achieving © fluid tightness (chis fect is aot generally appreciated but it should ‘3 be obvious that mere contact at the surfaces is insufficient and experimental evidence points to a unit clamping pressure of at Gaus He : WY i i Fie. gf Fo 33 © ®) Scnew-pown Stor Vatve SeaTiNos - Fro, 36. Vatve Searmics +. Since all three types enjoy urifversal application it would appear © that there is little to choose between these three differing: designs. Each of them mv; be relied upon to give fluid’ tightness under 1 ' + favourable conditions if ground-in properly but it's interesting to 7 aonsider each one in turn with a view to ascertaining, if possible, what outstanding virtues may be-aseribed. to any particular type 7 Raterting to the seating arrangement sho fi such as would unhesitatingly commend fs adoption. | Sihimoe extent and tie wil eceason some small amause oaiding Consider frst the valve member with radiused face'as commonly between one face and the others consequent, frictional resistance ‘employed on deadweight safety valves and the smaller sizes of stop, . 3 i + + least one and a half to two times the line pressure being required lay the case of steam valves); and (ii) the force exerted on the valve 2 disc by the fuid pressure, ic. the product of the fluid pressure and the effective area of the seating. Referting to the seating arrangement showr: in Fig. 3.7 (a), the ‘ alles ines of stop: of che two materials in contact will require to be overcome and, as Nalves.. This gnay take the form of the tr aly hemispherical fac feictign tends to oppose motion, the resultant force R, (see the force shown in Fig. $.6 (2) or, with a much larger radius, as in Fig. 3.6 (6). Siar ie Ba ee) al ee lead Ban arc ot the ‘The radiused type of seating has at least two inherent disadvantages; «" gram in Fig $17 (e) will be deflected from the normal, shown it is more difficult to machine and lap than a mitred, or fiat one, dotted, by an amount determined by the angle of friction ¢ shown. and it is more wasteful of metal and consequently more expensive. = (Note that tang = 1 where is the coefficient of friction for the ‘The forces which are brought into play at the contacting faces of _ the seatings of a screw-down stop valve in order to ensure fuid ~~: © tightness will gow be considered. : Tn Fig. S740), 6), (0, (2, and (¢) are shown fve designs of seat } contacting edges varying from one presenting an acute,angle to the . materials in contact.) ‘The three forces brought into play are the axiai force P, required 4 © \t6 overcome all frictional resistances and maintain fuid tightness, H,, the horizontal component and K, the resultant of these two | forces,Fie. 37. (a = 30% SCREW-DOWN STOP VALVES—SEATINGS AND BODIES 93 5 Fig. $94 (6). (, (, and (© are similar to Fig. y-7 (0) with the exception that these depict seatings having increasing edge angles | eae ete, until in Fig 97 (0 we have an incereating case whesein the edge angle «, is equal to go” —g. In this instance the resultant o Fig. 9:7 70°. (== 90" (2) Compute Digran for he Five Varios of ‘force R, will be coincidental with and equal to the axial force P, producing this resultant, or Pua Ry sin (a + 8) = Risin (90° —$ + 4) R, sin go” R‘Thus for all edge angles in excess of 90° —¢ the horizontal com- ponent H vanishes and there is no tendency to expand the seating in the process of tightening down the valve. mnsideration of these five successive vases in conjunction with rig Uy) (on which all five force diagrams are combined) is mart revealing. ‘The value of R, the resultant minimum contacting force required for the maintenance of fluid tightness, has been purposely kept constant in all five force diagrams for the sake of comparison and it will be observed that the axial force P required to produce this force R diminishes with a diminishing angle a. This would appear to favour the adoption of a steep (small) edge angle but the apparent advantages arc offset by a progrestive increase in“ dg — \the value of H, the force acting at right angles to the line of action OP, and this is the force tending to burst the seating. ‘This bursting tendency is augmented by the fluid pressure acting, in the same direction and, although the combined effect of these tivo bursting forces is not Htely to be appreciable, and is very small in comparison with the major force required to close the valve against all resistances, it may have the effect of bursting the seating if ‘this is inadequately proportioned or if no restraint is placed on the free deflexion of the seating or, additionally, if the seating is made of brittle material. |. With flat-faced seatings the axially applied force required to | maintain fluid tightness is unaccompanied by any bursting force ‘other than that produced by the fiuid pressure itself which is, | pethaps, one sound argument in favour of the adoption of iat \faces, now gaining in popularity. The bursting force may be f | i \ SGREW-DOWN STOP VALVES—SEATINGS AND BODIES 35 | disregarded for all practical purposes in most cases unless the | seating is of exceptionally: thin section. At all times lheral pro | portions should be adopted for seatings so that they may be better able to retain their wue ciscularity and flatness under varying coniitions of temperature, presre, and the unequal body strains Lincurred. “The following empirical rules may be used in. propoctioning restrained seatings, i.e. seatings which are not fee to expand vol DIAMETER OF SEATING df Re. 39 . Jaterally, and these are ‘pased on values which have proved satis- factory in practice if lengthy trouble-free service is any criterion. ‘These rules are embodied in Fig. 3.8 and the accompanying graph (ee Fig. 5.9) enables a rapid selection of a suitable value for t,, the thickness of the wall of the seating, to ke made. This dimension is che one on which most of the others are based. {The width of the contact face w is, or should be, a variable | factor dependent upon the intensity of the line pressure, the navure * of the working fluid and the permissible unit pressure om the materials jn contact. In the lattee connexion the values for u, the maximum saie unit surface stress, given in Table 3.1 may be adopted. ‘The maximum force that could be transmitted to the contacting surfaces of a screw-down stop valve seating, by simple operation of36 ‘THE DESIGN OF VALVES AND FITTINGS | the handwheel and screw mechanism, would be that obtained when | the valve was operated rnder conditions of zero line pressure. This | will be obvious when it is considered that in a correctly designed valve the mechanical advantage of the handwheel and screw mechanism is sufficient to permit of the valve being closed against the pressure ‘Taste $3. Maxoane Rewosace User Sowace Srmess rox Vane Searines . ‘Max. Brine | Max. Unit Surface Seaing Material | Hardon Si Authority ‘poo ke Gain | a Sie Sposion ‘s Gunmen t 5 | Sees | | a | NST land aw): | 2B Risk ee aa a Basket: | for which it is designed, with an allowance for the clamping force at the seatings and for other resistances. Consequently, if the major resistance, that of shutting off against the line pressure is removed, and the valve eperated, the whole of the thrust of the valve spindle may be transmitted to the contacting faces of the seatings which must be of such proportions as to take this thrust safely without suffering any damage. : ‘This leads us at once to correlate the various factors previously outlined and to determine w, the width of contacting surface of the flat-faced seating, as follows. *" Referring to Fig. -10, let F, = total axial force to close the valve effectively axial force ‘o resist pressure load . effective area x fluid pressure pat, ‘the pressure being taken to act upor the whole of the area bounded by the outer edge of the upraised seating edge * "Monel" (Regd, Trade Mark) is a propriciory alioy containing roxy ‘wwo-thieds nickel and one-thid eopper- ee cae of scree dowe op valves nth tide sere the aa face pent ote ‘onbalzneed area ofthe spindle were it pases vhrowgh the st Sand the ine peesure, but this foree wil be supplanted by the substantially greater force F a: the instant of final closure. rae 2 OF TECHNOLOG: SonEW-DoWN gror VALVES—szar e Anne ATSNE! Do ALY! fF ee RE SEAS P = axial clamping force at the seating area of seating edge twice the Tine presstre (for steam)” 2ndgep approximately fiuid pressure hb | 45 safe unit working strew, or clamping pressure on seating | Then, total asia force | \ Fyn Fg+P ~ Fatp + andyup j 4 bs ] = op (2 + 20) tee Bq. 37 > load p area tag Fee cman of etn of he appropri value of wth ares gon in Fgh 5 (O00) oy nad aa ee Plotted Te is resonimended that these cores be canioned Prcrence tthe mula derived in Ber go he a el Nevis to incu the valus of Gs and hee crs smaller sizes of valves s0 as to be more in line with practical require. ments,werbice mrt mt tain‘ett Sau oF sca Fig. gat © weet ‘rena sauttn oF saris he Fig. g.rt. (@) “44 THE DESIGN OF VALVES AND FITTINGS The dotted portions of the curves fall through points determined by Eq. 3.2, whereas the fall curves embodying the above recom- mendations should be adopted. The disparity between the theor- etical and empirical values is no\so pronounced as might rst appear since the ordinates in every case are logacithmic: Bxaurie ‘What would be the total axial forre required co effect positive closure of a 6.in. screw-down stop valve having fat-faced phosphor hronze seating: against a steady steam pressure of 100 Ibjin.®? Determine first of all he permissible width of the coutacting face of the eeating assuming , = 63 in. Neglect gland friction. Souvnion From Table g.1, safe unit pressure wm 4,000 Ibjin.* whence * be We 100 x 6:25 3,000 ~ o0s6. ‘ ‘This may be amended to o-0g@ in. curve given in Fig. 3.11 (@) and secolaice vith the appropriate P= andsep 2X gr1416 x 6:25 x o-0L8 x rod = 228 lb (approximately) Also Feta jae = 07854 x 6-25 x 100 = 2.075 Ib (approximately) and Fea FatP = g.g0g Ih fapproximnctely) Note that P= 262100 = 6-9 per cent of the total force required to 3.3 close the valve, yet tis isa factor overlooked by the majority of designers ‘who invariably base their calculations solely on the force required to resist the steam load. SOREW-DOWN STOP VALVES—SEATINGS AND. BODIES 45 By way of comparison take the case of a mitre-seated valve. ‘The axial clamping force P will be determined from the relation P=Rsn(e + $) where a = edge angie of seating 4 = friction angle of the materials in contact {ote ‘hat tan p= p= the coefficient of fction) : = resultant force on setings Now” | R= 2ndgop very approximately serene Pm awihup sin (2+ 8) and i= que as previouly determined, ‘hes total axial force required to cove the valve sere by Re +P PRP + aedey sin (a+ $) = nip (+ ze sinte + 8) and load per unit atex of seating edge ‘eo ‘ (d_ adap (E+ aw sin (2 + 9) creel ma Od nen bd os wget PDT Ss) Bde _— wae ww enlgegy 2 Bg To make the comparison complete take the previous numerical example ‘of the Gin. stop valve but substituting a mitre-faced phosphor-bronze Seating fsa 43" edge ane, tnd aking the cocfcient of ction of phosphor-bronze on phosphor-bronze as 0°24. + Since p= tan gong then ¢ (from teigonometrical tables) = 13° go" 2 ae Ga 100 x 695 qfg,000 =a X too * sin’ (45° 0” F137 307}6 Note that this value is somewhat lee than that previously determined in she case of the fac faced seating, the aight wedging action induced by the hive having the eficct of aightly reducing the axial effort required to make a sal Contiing— Py (ue Fig. 8:7 (8) ‘adap sn (25 + 8) ax r1gi6 x 625 X o-05g2 X 100 x sin (45" 0" = 107 1 (approximately) 13° 30) and Raddy = 075! (approximately) as belore Now Fe= Fit Py 075 + 187 = 262 Ib (approximately) 187 x 100 ‘3202 [ote that in this case Ps = = 5:73 per cent of the total force equized to close the valve. ‘aking the cae of the undesirably acute seating depicted in Fig $7 (0) by way of a somewhat odious comparison, the ‘eam Salers a obtained ia sal manner to tev obtained Dy te foegoingexdolations— 70545 in. 471b 075 Ib (as before) 222 Ib is suied effectively tf P, being 4:57 per cent of the total axial force required effect ‘close the valve against the line pressure. Similar values for the ca Shown in Fig. 47 (@)s wherein the edge angle a, = go” — , are follows — o-058 in. , ing tough 6 pr cet ofthe totaal re egies notte afl wang. aes ee ee cat oft sping invests on wie SOREW-DOWN STOP VALVES—SEATINGS AND BODIES 47 differing edge angles, some feasible and others ludicrous, ia order to embrace all extremes, three inferences may he drawn—- @) The width of the seating face w varies but slightly for the whole range of edge angles between the two extremes of g0° (flat face) and 30°, so that one might reasonably base the width of contacting face on that computed in the case of fla-faced seatings. (Gi) The ratio of the clamping force P to the total axial force F, required effectively to close the valve diminishes with a diminishing edge angle « but this apparent advantage is offiet by the corres: ponding increase in the horizontal component, which is most ‘undesirable, ii) For all edge angles between go”—¢ and go”, (flat face) the clamping force is constant and the horizontal component H All things considered, flat-faced seatings are to be perferred, particularly for high-pressure steam service, ot for any other service ‘where variations in temperature are expected. Under such con- Baa from which n, may be more readily calculated than from Zq. Eq. 410 and 4.11 are expresvions for the eficiency of screws with single square threads Int they are also applicable muliple- Waurvethreaded screws, provided. (ead) Ww subwttuted for f, (itch). Tm certain cases the thread fe not cut to the fill depth SRown in Figs, 2 and. g.2 mainly in order to preserve the core Strength of the sctew. “This doesnot affect the foregeing expressions Thee iselways taken to be the mem diameter ofthe thcad. Vales of m, have boon plotted in. Fige sce corresponding to various lead (or pitch) angles of threads and various values of the Static coeficien of frieton for the materials employed. Tatermediats Walues may be interpolated, “Table 41 gives details of the static coeficiont of friction between materiak icely to be employed for valve spines and snating pars, | force $42 + w75mpsd, + Fy may be sandwiched between the .- SCREW-DGWN STOP VALVES—BOLTED COVERS 73 Py hap + 1d) +P, Camping force to be exerted ky sash bolt 5 Fa ae 2 + Ba. bet’ 2 xb L, 2 dd +74) Fn de + Ted) + “The width of the spigot s may be taken equal to 24 + oxtin. 047 6 © Burne '* Determine the maxicnuim total tensile force in the cover belts and pillars ‘of a 7-in. serew-down stop Valve, working pressure 250 b-in. from the following data: Distance be'ween bolt {or ouuide diameter of gasket) 1a} in,, diameter of spigot for locating cover gb im. axial fnce to effectively. lose falve 12,000 Ib. Pillars to be of the serewed-in type. Number of bolts ro, (This anexample of Care r and Eq. 5.1 should therefore be ured). [vgs (228)* ~o.8s(9H)* + 05 x 128 x of] + Adopting jin. diameter BS.W. bolts, cors area 0-422 in#, the stress 4319 oqae ‘Js in accordance with the values given in the graph (Fig. 5.2). Now consider the same valve but with the pillars projecting through both cover flanges (Case 2). ‘Total clamping force to be exerted by each bolt and pillar =Fa impoted = #22 11,680 Ib/in., or approximately 53 ton/in.! This value = (ogsb! ots +o =e 35D" 25d" + o5Dd) = SEZ gy at 05 x (OD! + 05 = gaze < 198 x ot] This isthe fore» contributed by each pillar and bolt in order to maintain floid-tightness of the joint and 19 resist the force imposed by the pressure of the working duid acting upon the plate.Sine te plan poet rough he cove ange ming ml ny wie ah Soe OS oe A TOES Kite Spel se prose ot Clatag "ave Auta States Essen . ont oc in each ptr = P+ 22 =s7704 8 770 Ib (os agninat 6,000 had they been simply sree. “Adopting ‘thin. BSW. the core aves & oGgpin® and the sires ane . 227° — sgocotbfne approximately, a somewhat Digh value for mild fied! fom which chete member are generally made, And indicating that the selected value of ry in. BSW, was on the low sides and deen we BSW would nave boca war Sltee Ady seein Le SW. Wwould have been chosen pravice with yer ehessicier Bing more in keoping with ihe ses of te a ok see aE diameter the age fs anal spose imceichescat ete ered ‘would have been pportenaecl Sige Tetmight be arged that te cover bali could be mad somewhat ares or thle number Inereased os toreiee the aa of echo aac Blea roments section wil revent tie teprecacs ha oe ae Bist be rely sessed by the vous egy the hes be Bl Where the pillars are simply screwed into the top cover plate conditions are somewhat different. ° ® Hee ond npr exc pia il be ade ail od which wwe have sc it Gyloolhs‘ASopigg the same working eeu’ abo, namely.14,000 Ib/in.*, we have id * 6.000 too Core area of pillar 0-428 in# corresponding to j in. B.S.W. (appro) Now consider Case (cover plate with netal-ovmetl spigot jit) Tie can dlmete ofthe sget tay be alent be ok sds width ins other data as ia he to previous Can placed ns “Total clamping free pot bel cap + pet +E (oe = Ries nay +B oata: 5) 14 % 250 4% 10 8801 ‘ ‘This result is close to that obtained in the frst example taken so we may conclude that {-in. B.S.W. bolts would be suitable. (erst $7 x x 13) | bottom portion. This, of course, is impracticable in the case of the “one-piece” spindle, from considerations of assembly. Léngth of Spindle Taread Engagement ‘The length of thread engagement of the spindle within the crosshead is not always determined from considerations of the ~ shearing forces imposed so much as upon the area presented te SCREW-DOWN STOP VALVES—ROLTED COVERS — 79 #2 TSist wear. If insaficient bearing area is provided, feilure may ‘ecur due to flaking or pitting of the surfaces in contact. Accord- © ingly, the intensity of pressure mist be kept well below the figure at which pitting will occur. . ‘Librication of the thread surfaces will appreciably increase their "wear resistance but where such prevision is made itis often neglected Gy those whose duty it is to maintain the valve in good condition. I U | Fre. 5.8/0 Lewil generally be found that where conditions of wear resistance zee fully met the resulting Jength of thread engagement will be Sufficient to provide ale resistance to dhe shearing force imposed, Although this elation should not be deemed to apply in every case, Experience point to adequate shear resistance. being obtained hore te length of thread engagement is not les than the outside iometer ofthe sre Rerring #0 Fig. 5.8 Projected area of bearing surface of oe complete thread | ani, : t | ghtn ‘ana i hoe Nubteatperigentee bE fl |. Therefore ‘Total projected area of bearing surfaces in length [, a, he Sty = Fah and ‘Maximum allowable thrust = = dy fle where f, = this must equal axial thrust F, previously determined, ‘Then Eq. 5-7 ‘The value of f, varies according to the. characteristics of the materials employed and to the amount of lubrication, if any, provided. With the usual combination of gunmetal spindle and steel crosshead, values for f, of 1,000-20°15500 Wbjin.® are recom- ‘mended. Where means of ubricaton i provided thesg values may Spindle End Attachmént to Handwheel ‘The attachment of the spindle to the handwheel may be accomp- lished in a variety of ways. At one time it was common practice to provide a square termination to the spindle arranged to engage with a correspondingly squared cored hole in the hahdwhee! (sce Fig. 5.9) and this expedient is still favoured by manufacturers in the case of very small valves where space limitations preclude the adoption of anything more pretentious. ‘The squared attachment, whilst effective, is uneconomical froma ‘a machining standpoint. It does not lend itself to mass-production ‘methods since the hole cored in the wheel boss is usually left in the as cast” condition and the spindle terznination is generally filed ‘up by hand to suit, thus negativing any attempts at interchange- ability. Additionally, from considerations of moulding, the square hole must be given slight taper, again adding to the difficulties of machining and, being invariably chill-east, is difficult to file om broach. The size of squared termination will need to be determined. ‘The polar modulus of a shaft of square section is given by 0-208! where ¢ is the length of side of each squars, whence the torsional, maximum allowable intensity of bearing pressure, and“ SGREW-DOWN STOP VALVES—BOLTED Covers 81 incrient, or torque Q= o-208st/, where f, is the maximum permiss- ible saear stress, Then (Note that, is equally a maximum at the centre of each side but is zero at each corner and, of course, at the geometrical centre.) Eq.5.8 Fro, 5.9: Sgvanianno Seieoue ano Haxownnes [A mote saisfgtory attoghment is provided by the cylindrical termination fited with sunk key, either sctandlar or square in section on if prefered, with tae Woodruff types this method lends Ia to mass production and intrchangeabiiy, the wheel boss Tring cast soli, dried and lepwayed fo suit the spindle teria: Shon as shown in Fg. gto in which a Woodraff key and s plain sun key are alternatively portrayed. Since the presence of & eyecat cannot be aocuraely acounted for in any mathematical teatmnent on. the strength of a shat mrbjected fo torsion, resort must be made fo the nslts of dinect Gsperimert to establish te weakening eect induced by the presence82 THE DESIGN OF VALVES AND FITTINGS ofa keyseat, (Note. A Keyseat isthe slot formed in a shaft to accom- modate a key; a kiyweay the slot formed in a wheel boss for the saime purpore.) Such experiments reveal that where a key is made to the usual proportions, ie. width of key equals one-quarter the diameter of the shaft, and depth of keyseat equals one-half the width of the key (one-eighth the diameter of the sha) the resulting shaft is weakened by about 19 per cent. Another enipirical rule in rere Fro. 5.0. Vatve Srixpur Tzmavanons very close agreement with this experimental evidence is that the strength of a shaf having a sunken keyseat is equal to that of a plain shaft whose diameter is less than that of the actual shaft by one-half the depth of the keyseat, Expressed algebraically or 2 Ba.5.9 SCREW-DOWN STOP VALVES—BOLTED covers 83, equivalent diameter of plain shaft diameter of shaft’ having a keyseat of standard “the poiar modulus of a shaft of circular section is “4? where 4, = diameter of shaft, whence the torsional moment, or torque , is given by Q + 4s Eq 5.10 or, making allowance for the, presence of a key of standard proportions 3/5°09Q 4 1066 5/5282 | i ‘Where a Woodruff key is fitted, the value of d, derived from Eq. st should be inctased somerrhat fo compensate for the tolatond weakening eft oceasoned by the deeper keseat 2s Eqgetr bbe assigned the following values é . , Ibjin2 f Brass 3750 ‘ Gunmetd 5.600 ' Nickelebroine (stainless) Steel, | + 12,000, Phosphior-bronze, cast : 4,200 (average) i Phosphor-bronze, rolled 6,500 } Steel: 5 Mild, o-t per cent carbon. 8,500 Medium, 0°3 to 0°5 per cent carbon ©" (annested) : ‘9,800 (hard) 0:5 to 0-7 per cent carbon. 9250 ; Han k 5 ‘The permissible shear stress f, in the material of the spindle may rr ° 4a se VESIGN OF VALVES AND FITTINGS The foregoing formulae should be used with discretion; for. example, for the sake of avoiding unnecessary machining, and ‘perhaps for the sake of aprearance, it may be desiratle to adopt a diameter in excess of that calculated, A minizaum width of shoulder should be left at the junction of the termination proper with that- of the adjacent portion of the spindle. This will be quite clear from Fig. 5.10. Crosshead Ie will be appreciated that this member is subjected to the full axial force of the spindle in the process of closing the valve and by the pressure acting upon the valve member when the valve is closed. Being virtually a beam loaded more or less at the centre of the span. (actually the loading is not concentrated at the geometrical centre ‘but on an annulus) itis desirable to keep the span as short as possible and so restrict the size of the crosshead to the absolute minimus consistent with the necessity of providing the reauisite strength. In this connexion those designs wherein the pillers are screwed into the cover plate permit of the smallest possible span being arranged, apart from all other considerations previously mentioned. ‘Whether or not the supports (the pillars, in our case) mey be deemed to be fixed (encastré) or simply supported (cf. the theo-y of ‘beams) is open to speculation. They are really neither one thing nor the other but in this particular insiance, and in the absence of any concrete evidence to the contrary, we may regard the: conditions of support as being equivalent to that of a Leam simply supported, Which assumption errs on the right side. i Since the actuating spindle is positively attached to the crosshead, by virtue of being in threaded engagement, itis reasonably accurate to assume that the crosshead is loaded at the centse, Applying the usual relationship for a beam 20 loaded KR ' where M = the bending moment f= the stress at the remote surfaces (fe= tensile stress and | fa compressive stress) Z= the section modulus at the point where the bending: moment is at maximum, i. + Thien BE Mag at the centre of the span SCREW-DOWN STOP VALVES—BOLTED covers 85 ‘here , = the maximum axial force transmitted by the spindle jand = the apan (or pillar centres). FL i Transposing GF “The mid-section containing the threaded hole engaging with the ‘actuating spindle is our chief concern since this is situated where she maximum bending onset, and consequent manimam sre, secure The vanovere ssction on ths plane wil be ae denoted ih Fig: gr andthe sco ml for Such seoon may be taken = 2 5 Oda) where dy — the mean diamecer of the spindle threads, EL 2» ‘Then BOs whence Eq 5-12 Note that /: should not be less than J, the minimum length of “thread engagement derived frem Eq. 5-7. In calculations such as {theie, whece there are a number of arbitrary variables, some amount of adjustment may have to be made before « satisfactory combina- tion of dimensions is ultimately determined, ‘Thus iff, the depth of_ 86 THE DESIGN OF VALVES AND FITTINGS the crosshead, works out lass than the depth derived froni considera. tions of effective length of thread engagement, it will be necessary to adopt the higher value and calculace the ‘width 5 from the following transpositior of the preceding treatment / a eo) Since crosshead are usually carbon steel forgings or staxapings, or cut from mild steel bar stock, f, the maximum permissible tense Fro, 5.02 Stress may be taken equal to 20,000 Ib/in# for steel forgings and 10,000 Ib/in.? fer mild steel. If sand-cas: gunmetal crosskeads are employed (as is sometimes the case in smail salves) f, may be taken as 6,ooodbjim? Gos. Fy ‘On no account should cast iron be used. ‘The usual combinaticn of steel crosshead and ‘gunmetal spindle is quite satisfactory where the length of thread engagement is of the right order, but where a one-piece steel spindle is employed the better alternative is to fit the crosthead with a non-ferrous bush preferably of hard-wearing gunmetal, an exaraple of which s shown in Fig. 5.12. ‘The crosshead is also subject to shearing forces in'the vicinity of the pillars but unless the crosshead is thinned down excessively at Lo. isorew-pown stop vanvzs—sorrup covers | 87 "2p: these points the shear stress imposed is amply catered for by the ‘amount of material required ip provide “‘cover” for the pillar-end ovate, Where any doubt exists tne strength at these points should be investigated. Handwheel ‘From the wide variety of handwheele encountered in valve practice a diversity of opinion would appear to exist regarding the thanner in which these should be proportioned so as to effect the Best compromise between the two essentials of utility and simplicity of manufacture. “" Many text books on engineering design generally devote a section to the proportioning of such common details as pulley wheels, gear wheels, ele., but are, for the most part, silent on the subject of the commen handwheel, che authors, no doubt, believing these to be incignificant items of construction best left to the whim of the designer. ‘As an example of the imagination, or lack of it, exhibited in the design of some handwheels may be cited the case where these are so liberally proportioned as to prevent anyone except a:giant from affording: a grip of the rim. One such handwheel discovered on a . vaive taken out of the Severn Tunnel had a cim 2 in. thick, yet Fe should not be necessary to draw attention to the fact that whilst ‘2 Froportionate increase in certain dimensions of the handwheel may be expected:to accompany an increase in the diameter of the valve, the hand of the operator suffers no corresponding change in > dimensions. ‘The tim of the handwheel should be comfortable to the grip; fa fact, it should present an inviting appearance and be carefully exainined before assembly with a view to removing any excrescences likely to camage the hand. a the design of a bandwhcel no reliance can be placed on the method prevalent in some circles to assign to it a diameter that tones with soine Key dimension of the valve to which it is adjunct, The correct diameter is that which will provide sufficient leverage to enable the valve to be operated against the intended pressure swithout undue manual effor'. Moseover, auy valve, whatever its purpose, should be fitted with a handwheel of such a diameter as will render it capable of being operated by one man unaided at any time, irrespective of the pressure conditions prevailing, zince occasions may arise where a speedy shut-down is imperative in the interests of safety. The. spokes of the handwheel are each subjected, to bending3CREW-DCWN STOP VALVES—BOLTED covars 89 » forces; the effect of which is more pronounced at the point where they join the eentcel bos, but if these are to be proportioned in } accordance vith the bending stresses imposed, the resulting hand- Wheel would be inordinately weck from a casting point of view | And its consequent frailty woutd not fit it for any subsequent harsh treatment, zither in transit or in eperation. "Accordingly, such details are proportioned more from considera tions of moulding and treatment meted ont to it in service than from purely theoretical considerations. ‘Sica valve handwheels ue usually of tae form shown in Fig. 5.33, employing straight radial spokes or ams of quasi-cllipical section. «| No explanation is offered for this particular preference for straight spokes, but it seems to be established custorn to fit steam valves with Fandwheeis of this pattera, the.cu.ved spoke variety, by the same tokee, being ostensibly reserved for valves of the wedge-gate and “ther types of low-pressure hydraulic valves. ‘The Gata tabulate’ in Fig. 5-13 will provide the necessary © additional information when once the required outside diameter jas been determined, although the dimensions given may be modified according to individual preference. The sim of the hand- Wheel depicted is somewhat unique ia that both the inside and dutside surfaces are notched in a manner calculated to afford the fnaximum Fandgrip, @ featare worthy of adoption, especially in © Ghose applications whese an oily atmosphere prevails (6g. ships’ ‘engine rooms). ‘For smallebore high-quality steam valves, bakelite or similar composition handwheels are oft favoured, being cheap and some- what more prepossessing thaa cast iron -nore commonly employed, {An example of such a handwheel is given in Fig. 5.14 together wwith dimcasions of the mors popular sizes. "A govin. diameser handwheel should be regarded as the largest that can be satisfactorily maospulated by one man unaided. Hand- wheels of larger diameter are not conducive to the best effort being Sbeained by she operator who would be compelled to adopt an awkward stance with arms outstretched in an attempt to provide the masimum torque effect ‘Having thus described the basic principles underlying the design, of screw-down valves, itis fiting to conclude wish an example of outstanding merit, namely the “Tried” junction stop valve manu factured by Hopkinsons Ltd., Huddersfield (Fig. 5.15). This feveals 9 number of features not included in the usual assortment "of serew-doyn valves encouncered. ‘The angle, or junction, valve has been chosen by way ofillustration Rim | No. With | Sp nt By a He an z t 1 i i # 1b Dimensions (in) Spake Bow ab a a ab aE a 2 a 3 at at at 4 a a a a Ao oF 3 st 3 4 a ide | Bow Dia, Diameter ‘Minor | Major io. 53. Vatwn Hastatents Cast Tnon90 THE DESIGN OF VALVES AND FITTINGS since this is the one moxt favoured as a boiler stop valve, particularly g fon boilers of the Lancashire, Cornish, Economic, and “shell” EN sme amet types generally. ‘The seme essential features, however, are retained in the globe pattern, the only difference between the two being ia the bodies, the glebe pattern often being favoured as an engine ~ stop valve. oe | | a i — SS ‘The valve member is of simple design, being devoid of wings or aa | other guides likely to cause obstruction to free flow. ‘The seating : a faces are flat and taus permit of some slight amount of relauive a Wa VA lateral displacement uncer varying temperature conditions, whilst : | the spindle is permitted some amount of axial displacement under the influence of high temperature, thus relieving this member of Fro. 5.15 Horkassor’s “Traap” Jonertox Stor Vawvi much of the stress to which it would otherwise be subjected. lepend y ray of Ma. hin, Li Hurl)o sae URsiUN OF VALVES AND FITTINGS Valves of 4-in. bore and over are fitted with a rotate spindle, as illustrated, the: purpose being provide means of rotating the valve member on its seating independently of rotation of the actuating screw. ‘This enables the position of the valvemember relative to its seating to be varied at will in the final closure position in addition to affording means of wiping the faces clear of any foreign matter. The conditions of the seating faces is thus main- tained over a longer period than would be the case if the valve member always came to rest in the same position cach time the valve was operated. “The valve member and its seating are of Platnam, a nickel alloy of outstanding durability under the severest conditions of pressure and temperature. ‘The locking device featured in the example illustrated is fitted as an extra when specified, and a closure indicator may be likewise included. Soot-blower and bypass connexions may be sim‘larly arranged. ‘The valve is designed for pressures up to 2 maximum of 450 tbjin.t steam and 600 Ibjin.® water (or air) and for temperatures up to 800°F. “Stuffing Boxes, Glands and Packings The depth of the stuffing box cavity is dependent upon the number of turns of packing required to give effective sealing, and this in turn depends upon the pressure and temperature conditions. ‘The best form of soft packing to employ for steam. service is good” quality square-section graphited asbestos, preferably of a type ‘ reinforced by anti-friction wire. Arcinforcement of lowmelting-poirt would be no good, of course, at high temperatures. Cheap packings are worse than useless and night easily condema what is otherwise a well-designed valve. On superheated steam a packing containing rubber or tallow is similarly worthless; in fact it should be shunned for all steam duties irrespective of temperature. Many new valves require to be repacked after only'a few hours of service due to the inferior nature of the packing fitted and this is a distinct reflection on the maker’s integrity - ‘Table 5.2 gives recommendations to guide selection of the grade best suited for given conditions. : ‘Assuming the’ unly block packing will be employed, the depth of the stuffing box may be given in terms of the number of turns of packing required for the pressure conditions obtaining. Allowance should be made for accommodating a few tims of soft header material above and below the segmental packing. < ‘tana $i) Renan Ea oF Grae Phos? Standard |_ Dimensions _ Description ‘Applicaton | proofed asbestos cloth | in. sauase ri a | Ee - peng” | ition | gor cal ia eee sag See ato the oth and Ee = |e mea Die seomsane, | Sie se | incomes |i “or de of anti-friction | “ing b 10, Pacing" | Re covered wit | Tai 20 | Bas sgrusscan |atege ng | SOE foe oe | tome Someee [ses che | Seer et ae foe Teas” REE ge Packing, | ‘Sheen’ asbestos) Proofed asbestos cloth | 2. square rise = st Prato dae Bebe aed FEB. | packing Bee Saat | Ten cle | ee \ packivg F . Serene eam |Saamantect | ASSES tp, f| mgm ‘: tematic feria Pad ee Peoples ao cton| Tatra ; |S eS cm | i 9 Fro tation supplied by Bells Asbetos & Engineering Co., Led. Fig, 416 embodies the esental requirements ofa valve shfing bor Sl gland whist the graph (ee Fig. 6.17) simplifies the tas PPhiteeluing Dy, the cutsde diameter ofthe stung box corres ornding soa known value o she dstneter ofthe spindle (ov tank, Poth case may ba) and the width of packing we. these values are PRY GoM cedatal practise rather than upon any mathematial premive. ° : uve &, represented by an unbroken line, gives the outside aismetcr of the rating box D, i termes of. Curves 2 and 9 have ean added to increae the usefulness of the diagram siuce by the iiiiulon of curve g representing the diameter of the spindle d; Hedy smcane of determining the wideh of the packing annulus i rotided, Semply by ting a vertcel reading between curves and’ and halving the gure thin obtained, 4 Obviously we