SESSION III

REPORT AND DISCUSSION

Chairman : J. B. TURNER, M.I.Prod.E.

The Chairman : I am as keen as you all areto hear Mr. Puckle, of E.M.I. Ltd., expound themysteries of numerical control. To most of us, itis something about which we have read in thetechnical press or have seen demonstrated at theMachine Tool Exhibition, but of which we have littlepractical experience. The claims made for the timesaved and the accuracy achieved by this type ofcontrol are such that we must follow it up. Mr.Puckle is well known in the electronics world and is adistinguished member of the Institution of ElectricalEngineers. He is the right man to put the subjectacross to us.

(Mr. Puckle then presented his Paper, whichappears on pages 169- 186.)

Mr. H. A. Chambers, M.I.Prod.E. (GeneralManager, Rockwell Machine Tool Co. Ltd) : It isa pleasure to open the discussion on such a vital andinteresting Paper. We all use machine tools invarious forms, and this is research of a very highorder. The progress is astounding. The time savingswhich Mr. Puckle has given in his Paper are reallyworthwhile. They should encourage anyone toinvestigate these methods. I hope his Company willreceive the just reward of being encouraged in theuse of these machines.

We have not heard about forgings in thisConference. I feel that if the method could be usedto produce, cheaply and quickly, forging dies and ifthere could be greater use of heavy hydraulic pressesto enable us to produce accurate forgings to befinished by this method, it would be another verygreat step forward.

In regard to accuracy, I am wondering whetherfrom stage to stage in the use of this equipmentprogressive synchronisation takes place, so thatthere is no accumulation of error at theend of the run of the cutter. Perhaps Mr. Pucklewill deal with that later.

I am grateful to Mr. Puckle for going into detailabout numerical and digital systems. I am sure alot of people, like myself, would like to know moreabout these two systems and, indeed, other systems.I understand that the M.I.T. system has a greaternumber of channels than we have produced in thiscountry, even going to the point that they get audioinstructions for the operator, which appears to meto mean a great saving in cutter life in that you can

avoid damage due to the cutter being blunt or risingsteeply.

Mr. Puckle : I thank Mr. Chambers for hisremarks.

With regard to dies, some work is being considered,and I will ask Mr. Booth to deal with that part ofthe question.

With regard to cumulative errors, that is not likelyto happen except, possibly, as a result of wear of thecutter, because the analogue system is based onmeasurement of position. As far as wear of thecutter is concerned, cutter regrinding or replacementperiods can be coded into the tape.

Mr. R. H. Booth, B.Sc. (Technical Manager,Industrial Applications Division, E.M.I. EngineeringDevelopment Ltd.) : I do not think there is anyreason why dies for forging or any other purposeshould not be made by numerical control. We havetrfed to start with simple things like templates,intending to continue with more and more compli-cated cutting operations. Dies present no essentialproblems, but as you increase the number of controldimensions and get on to more complicated dies, theproblems of programming and specifying what youwant, in terms of numbers which can be put on a tape,increase rapidly. A lot of work is required before onecan take a drawing of a complicated shape for a dieand run straight into the tape with it. It is onething to make a die of a simple shape, but it isanother thing to make one for a complicated shape.It is easy in the case of a turbine blade, where thereare mathematical expressions for the curves. Machinesare being conceived all over the world with three,four, five, six and seven dimensions, but I think it willbe a little while before more than three dimensionsare used accurately, because of the difficulty of pro-gramming.

With regard to synchronising, if Mr. Chambersmeans to ask whether the progress of the work canbe synchronised with some change such as the pro-gressive wear of a cutter, the answer would be that inmany cases cutter wear is not a major factor, butwhere it is. the control cabinet can be adjusted forthe effect of the diameter of the cutter and that cando much to help. I am not clear about the extent towhich cutter wear of dimensional significance cantake place without first affecting surface finish. Ifeel that often surface finish deteriorates as adimensional change appears. In the future, one

187

should look forward to automatic sensing of the cutterdiameter. At the moment it is a matter of auto-matically checking the cutter diameter, and there isno means of automatically sensing the effectivecutting diameter of a cutter. Also, a cutter has notalways an obvious diameter. There is always somedeflection of it. Its effective diameter depends, onwhat it is cutting and upon the feed rate.

Mr. R. S. Brown, M.I.Prod.E. (Director, BristolAircraft Ltd.) : I read the Paper with great interest.In the aircraft industry we have been sold the ideafor some considerable time that the numerically con-trolled machine tool will be of great help to us inour model shops and other places. However, Mr.Puckle mentioned mostly machine tools of Americanmake. I wonder whether the people actually bringingout this control have sold it to our own machine toolindustry, because we cannot get the dollars to buyAmerican machine tools and at the present momentthere seems to be no progress in this country inmanufacturing machine tools which will take thiscontrol.

Mr. Puckle : We have made arrangements withthe Cincinnati Milling Machine Company. Theyhave a branch in this country which makes some ofthe smaller tools and, therefore, presumably one doesnot have to spend dollars to get them. The fact thatwe have an agreement with the Cincinnati MillingMachine Company does, not prevent us from pro-viding information in respect of, or fitting the controlto, machines made by a British manufacturer.

Mr. Booth : There is nothing in our arrangementswith any company which prevents our working withany machine tool company outside the United States.

Mr. Brown : I asked whether you in your spherehad made approaches to the British machine toolindustry to see whether they could bring out machinetools which would fit your control.

Mr. Booth : One answer to that is that in thePaper there are photographs of three machine toolsfitted with E.M.I, control, and all those machine tools,were manufactured in this country, one by an entirelyBritish Company, Research Engineers Ltd., one bythe Cincinnati Milling Machine Company inBirmingham, and one by Wadkin Ltd., anotherentirely British Company.

Mr. W. S. Hollis, B.Sc.(Eng.), A.F.R.Ae.S.,M.I.Prod.E. (Assistant Director, Aircraft Production,Ministry of Supply) : Colonel Clark, of the MachineTool Trades Association, has seen Mr. Goodinge andarrangements are being made whereby any require-ments in respect of machine tools and their develop-ment can be specified by the aircraft industry at thatlevel to the Machine Tool Trades Association, andthey will then consider the best means of approachingthe problem. If the Production Development Branchcan lend any assistance to either organisation, weshall be pleased to do so-.

Mr. Booth : The National Research DevelopmentAssociation is sponsoring work on machine tools inthe Manchester Institute of Technology.

Mr. R. P. Gardner (Consultant ElectricalEngineer, Asquith Machine Tool Corporation Ltd.) :I can assure the questioner that E.M.I, are workingwith the machine tool trade. The M.T.T.A. sub-committee has issued a report on electronically con-trolled machine tools covering the whole field, andmy Company is making for an aircraft company, askin miller which will be tape controlled, though itis not Mr. Puckle's tape. Therefore, something isbeing done.

Mr. T. A. Wake (Machine Tool ControlDevelopment, Vickers Group Research Establishment,Weybridge) : Mr. Brown is not quite up-to-date withregard to British-made milling machines. We haveone manufactured in this country and propose to fitnumerical control to it.

Turning to the Paper, I would congratulate Mr.Puckle upon producing the first comprehensive Paperon machine tools, certainly in this country andprobably anywhere else. There will shortly be aneven more comprehensive one, but this is the first andMr. Puckle is to be congratulated upon it. PreviousPapers on the subject have rather concentrated ondescriptions of particular systems.

Being in intimate contact with aircraft machiningprocesses, I would not agree with the general descrip-tion given of the way machining is done in theaircraft industry at the moment. The operator doesnot turn a handle, do a bit of thinking and then turnanother handle. The bulk of machining is done bycopy milling from templates. But as yet the speedand accuracy of these machines are superior to anythat I have seen electronically controlled.

The aircraft production engineer has a very acutemachining problem. He may have to remove up to90% of the material from his billet. Therefore, heis acutely interested in pounds per h.p. per minute.To achieve this, he is prepared to sacrifice accuracyto some extent, but he is not prepared to sacrificesurface finish, for reasons given by Major Teed inhis Paper. So the critical specification for him is notthe accuracy in so many thousandths of an inch, butthe accuracy of so many thousandths per unit ofcutter movement, and that is a different thing.

Perhaps I could have some comments on thefollowing. By manually operating a copying machine,we can get up to a speed of 15 ft. per minute, cutterfeed rate. I do not think this would be possible withan electronically controlled machine tool, because ofthe facility which the operator has for anticipatingthe movement of his cutter. Therefore, it may be aneconomic proposition to use one electronically con-trolled machine tool to produce the templates andto use the templates on existing machines, auto-matically or manually controlled, which would becheaper to make and no doubt faster in operation.

You mentioned that it was a pity that there wasnot in operation a machine which translates a picture

188

of a drawing into a machined workpiece. In myCompany there is such a machine in operation.

You also mentioned cutter compensation, which Ibelieve is a feature of your machine. It is alsopossible to do it on a digital machine, and we hopeshortly to prove that point.

There is a technical detail on which I should likeMr. Booth's comments.

In Mr. Puckle's Paper (page 8 of the preprint)there is quoted an informational tolerance of0.001", a cutter feed rate of 50" per minute,and a cutter radius of £". That would implya machine working at about 1" per second.To control this to 0.001", you would have to havea motor which is capable of responding inperiods of the order of a millisecond. I do notknow of any such motor, and should be interested toknow what you have done in this direction.

I thank Mr. Puckle for his Paper. It has given usa comprehensive review of the present state ofmachine tool development.

Mr. Puckle : I am, of course, fully aware of thefact that the repetition work in the aircraft industryis done by copy milling, but the Paper does,unfortunately, suggest that manual operation withoutmasters or templates is used for this purpose. Thiswill be altered in the final form of the Paper, andI am grateful to Mr. Waite for pointing out the lackof clarity.

I do not fully understand the question about infor-mational tolerances and time constants. If oneignores machine tool errors, the tool will follow acorrect straight line path though there will be a timelag, however small. Furthermore, there is no reasonwhy one dimension should not be highly accurate ifthis dimension is constant, even though a very highfeed rate is used in another dimension.

Mr. Booth : You have referred to a speed of15 ft. per minute. The Giddings and Lewis Americanmachine is operated at 100" per minute, and theCincinnati skin millers are expected to operate at upto 120" per minute under control.

As to compensation, I imagine that is a questionof the size of the computer that you feel you canafford to have on the spot. A large computer couldbe used immediately adjacent to the machine tool andcould carry out cutter radius compensation, but it islikely not to be economical. Most digitally controlledmachine tools have used a large computer, and it hasbeen felt necessary to move it to a central place andto share it with other machine tools.

The point about copying drawings is rather aquestion of what you mean by " drawings ". Therecertainly are machines which can accurately reproducedrawings. There is a Swiss machine which copiesdrawings. But that is not using the sort of drawingwhich we have in mind in this case. What Mr.Puckle means is that there is no means of directlytranslating the information, in the form in which it isfirst conceived, into our machine. There always hasto be some human processing to the entry to thenumerical machine, and that is programming in its

widest form. One of the advances that we must makeis to reduce the amount of programming well belowits present figure. I do not think it is any answer tothe problem to produce an extremely accurate copyof the drawing, as the Swiss machine does, because ofthe labour involved in producing the accuratedrawing.

Mr. Waite : In the case of the machine which Ihave mentioned, it is not necessary to have a speciallydesigned computer. For a simple design, it is possibleto use any suitable computer to produce the informa-tion for you. If the organisation with which you areassociated has one, it would not be necessary to havea specially designed computer.

Mr. Booth : There must be a certain amount ofspecial design associated with it. I do not know of anyexisting computer which could deal direct with anyexisting transducer.

Mr. Puckle : My words were " could be arrangedto do this work ". My answer is that you might haveto add a small amount to it to do it.

With regard to the business of copying drawings,the expensive and accurate work of making thespecial drawing is a form of programming for themachine.

Mr. R. E. Mills {Specialist Designer (Mechanical),A. V. Roe & Co. Ltd.) : It appears from thediscussion that even at this late stage there is stillsome doubt as to the value of machine tools con-trolled in this way.

I wish to put a question which I have asked onseveral occasions. I think there is only one case towhich this method of machining can usefully be put,and I should like to have Mr. Puckle's opinion on it.Whatever the shape, be it in two or three dimensions.I should imagine that the cutter can cut at optimumspeed all the time. This appears to me to be a mostrefined method of machining where machiningstresses are — I will not say " cut out" — keptconstant throughout the whole of the manufactureof the component.

I should like to hear Mr. Puckle's views on whetherthis machining can be considered as I have suggested,which, in spite of all the other doubtful points, isworth something — bearing in mind Major Teed'sPaper.

Mr. Puckle : It is not possible to keep the speedof cutting constant. When one gets to a straight part,one can in general run more rapidly than one canround a curve. One then has to catch up the velocityat the end of a curve where it joins a straight, sincesudden changes in velocity should obviously beavoided. That is to say, you may have to alter thevelocity of feed as you approach a bend, or come froma bend on to a straight.

Mr. Mills : You would do that in normalcircumstances. In fact, you might actually stop.

189

Mr. Puckle : Yes. We have developed somethingcalled a parabolic start and parabolic stop.

Mr. Booth : If with a simple form of E.M.I,control you assume that the time taken to go fromA to B is the same as that from B to C, but put Bother than in the middle, the speed can be varied

during the interval. For instance, if the distancesAB, BC are in the ratio of 3 : 1, you can make themachine move under a constant deceleration fromA to C and stop at G. This is a useful feature whenapproaching a corner. There can be any desiredratio of initial to final velocity by suitable choice ofthe position of the point B. In the case of theskin-miller discussed in the Paper, there is a pro-grammed choice of feed rates.

Mr. J. O. Mayer (Aircraft ProductionDevelopment, Ministry of Supply) : A contributorhas referred to the possible economies to be achievedby equipping several machines, or one, with tapecontrol. I thought it was generally intended thatone would have a master tape-controlled machine,which could cut master forms fairly accurately,although perhaps slowly, and copying machines whichcould run at the desired higher speeds.

With digital computers, it is possible to translatepolar information to cartesian and vice versa. If wehave an analogue-equipped machine tool, are we tiedto one form of drawing office information, or can weconvert from one to the other and, if so. how ?

Mr. Booth : I should have thought it was largelya question of the way in which the machine wasconstructed. If it is a machine with more than theessential free axis — in other words, you can rotateas well as go acrosswise — then it will naturally bepossible to put information for any of the axes thatyou have.

If, however, the information is presented in a formdifferent from the way in which the machine is con-structed, for instance, polar instead of cartesian, thena conversion may take place, either by conventionaltechniques of manual computing, or by the use of adigital computer. Generally speaking, the use of thiscomputer is independent of the type of control usedon the machine tool, and can as easily precede ananalogue control system as a digital one. It wouldbe possible to design either type of control systemto accept alternative types of input directly, but tothe best of my knowledge this has not so far beendone. The nearest approach is perhaps certain

American systems, where a general purpose computeris an invariable accompaniment to the system, andwhere this conversion can easily be done if the com-puter is large enough.

A much more important question than whetherinformation is presented in polar of cartesian form isthe way in which the shapes are specified, whetheryou must specify blend points or merely that youwant a | " radius to blend between two straight lines,which latter seems to me desirable to save pro-gramming time.

I accept what has been said about extremelycomplex cases in which, although you may make athree-dimensional model, under numerical control,you still have a very considerable waste of time inthe setting up of the various templates. I have seensome incredibly complex Cincinnati machines inwhich there are 100 templates to fix on the machine,and the time taken to put them on is much greaterthan the operating time, and I do not think that isan economic process. In more reasonable cases —which means the majority — it is right to producetemplates first, and so it may perhaps be in the caseof turbine blades.

Mr. S. W. Potter, A.M.I.Mech.E., A.M.I.Prod.E.(Development Engineer /Works, Royal AircraftEstablishment) : I was surprised to hear the lecturerplace the possibility of adherence to more kinematicprinciples of machine tool design last on his list. Ishould have thought that this was a fundamentalfirst requirement to obtain the maximum accuracyfrom any known reproduction system.

Along with some aircraft firms, we have theproblem of producing highly accurate aircraft modelsfor aerodynamic testing, and we hope eventually toproduce a machine tool which is capable of fullyprogrammed universal movement in three lineardimensions and possibly three angular dimensions,but this will take some time to develop, possibly fouror five years. We know there will be a gain inproducing a machine of that type. We are out forfundamental accuracy, and the systems of controlknown at the moment fall short of the ideal aroundwhich to develop a machine tool of that complexity.However,, as an interim measure, we are takingadvantage of an English machine tool which iscapable of positioning, using punched card control, tothis degree of accuracy. I say " positioning" asdistinct from " continuously moving" because thereaction of the machine tool during cutting is funda-mentally different.

We are producing models by machining a series offacets tangential to the required profile, for whichthe machine itself is set to this extreme accuracy. Toaid this machining of models, which includes millingthe facets on an existing jig boring machine, we shallhave a similar machine converted to programmedpositioning on three axes whereas the present typecontrols only two axes.

Reference was made by the lecturer to an inclinablefixture on a skin milling machine which was pro-gramme controlled. I should like to know what sortof accuracy of angular setting can be obtained on this

190

fixture. The degree of accuracy is interesting becausewe have a similar problem in the precise angularpresentation of workpiece to cutter, and so far wehave considered a linear scan well displaced from thecentre, rather than adhering to circular scale setting,as being essential to obtain the accuracy required.

Mr. Puckle : I am in agreement with Mr. Potterregarding the use of kinematic principles in thedesign of machine tools, and he is probably correctin suggesting that I have not sufficiently emphasisedits importance.

Mr. Booth : I agree with what you say inregard to the application of kinematic principles tomachine tool design. I have no idea of the accuracyof the fixture movement on the skin miller. Wemerely control, not design, the tool.

Mr. G. H. Taylor (New Development and ProcessEngineer (Aircraft), English Electric Co. Ltd.,Preston) : I agree that we have to walk before wecan run in these matters, and feel there is an urgentneed for such control, particularly in the manu-facture of wind tunnel models.

In the majority of cases, however, the draughtsmandraws the basic shape and leaves the rest to themould loft, from which models and templates aremade, but without dimensions. Dimensions arerequired for programming.

Would it be possible to produce tapes from themovements carried out by the hydraulic machine,cutting out idle time and saving on computations andprogramming, and so get us where we want to goa bit faster ?

Mr. Booth : It would be possible — in fact, ithas been done by several organisations — to recordthe movements of a machine either under copyingor manual control. It has been done by GeneralElectric and others, and also by Hewittic in France.But: T should have thought that the essential computa-tion had been done in the original model.

Mr. Taylor : What about starting off from themould loft ? We should want ordinates.

Mr. Booth : Given a series of ordinates, that isbasically what is needed for programming.

Mr. Taylor : But it would mean working out averv much greater range of them, and there would bemillions of calculations.

Mr. Booth : Fundamentally, that additionalprocess is one of interpolation which it should be thefob of the control system to do. Any control system,if given a long straight line to do, automaticallyintroduces a long line of points all the way along,there being vast numbers of intermediate points. Thesame process is undertaken round curves. As longas enough points are produced to specify what youwant, it is the job of the control system to do the

additional work, and it should not involve an expertin the drawing office.

Mr. Hollis : The success of the introduction ofelectronic equipment will depend largely on thepresentation made generally. This equipment iscostly — generally £15,000 or thereabouts — andwhile you can lose that sum of money in equipmentcosting, say, £100,000, it will be difficult to educatepeople who normally use plant valued between£2,000 and £10,000 to believe that they can affordan additional sum of £15,000, unless you can showthat they will get some considerable productiveadvance resulting from the use of the equipment, orthat they will get from the machine tool greaterutility and versatility, in that it will do jobs that itpreviously could not do.

We have hammered this matter out with a numberof the electronics firms, and we think we have gotsomewhere towards the answer with regard to specificapplications. We feel that electronics will be extremelyuseful in a case calling for a high degree of repetition.Yet we hear from electronics firms that they havethemselves introduced numerical control, simplybecause of the great number of different problemswith which they have to contend and that theversatility of the equipment justifies its cost. Iwonder what views Mr. Puckle or Mr. Booth haveabout this.

Mr. Puckle : I think these machines have onevery great advantage in respect of tooling. In relationto turning out small quantities for making suchthings as templates and jigs, they have the advantagethat they permit you to make rapid alterations. Youcan make another cam extremely quickly. One camshown in the Paper took 45 minutes to make, com-pared with something like three weeks. Thatrepresents a value which you just cannot afford tobe without. It is a case not of whether you canafford it, but of whether you can afford not tohave it.

When it comes to large quantity production, Ithink it depends on the type of industry you are in.If you are in the aircraft industry, particularly thecivil aircraft industry, I think that what you call" large quantities " are probably within the capabilityof the machine and the consistency obtained inrepetition is of the greatest value. If, however, youwant to turn out quantities of the 100,000 order.I would not suggest that you should try to do it witha numerically controlled machine. I suggest that youmight make a model of it from your drawing with anumerically controlled machine, and then put themodel on to a copying machine to make it inquantity.

Another great advantage lies in its use in the modelshop attached to a Research or Development Depart-ment. It becomes easy to obtain modified parts muchmore rapidly when a numerically controlled machinetool is available, and so the likelihood of " bugs "appearing in the finished product is generallyreduced. This should appeal greatly to the produc-tion engineer, because the value of the time and

191

trouble saved can be much greater than the costof adding numerical control. Moreover, the timesaved in research and development means that pro-duction can start earlier.

Mr. L. J. Bolton (Production Superintendent(Hydraulics), Sperry Gyroscope Company Limited) :In regard to cutter compensation, it has been saidthat one of the difficulties is that, due to the swarfand the coolant, it is difficult to take a measurementof any cutter wear. Surely, however, it should bepossible to measure the deviations of the form youhave just cut from the desired line, by either opticalor mechanical means, and feed that deviation backinto the system and so make an allowance andcorrection for it. Has that been done ?

Mr. Booth : Yes, in principle. One factor whicharises is that some of the errors of the work are notnecessarily repetitive. However, so far as the cutteritself is concerned, perhaps there it could be done.The worthwhileness of it depends on whether theerrors that you want to correct are due to permanenteffects such as wear of the cutter, or to factors suchas strain in the machine, which will be different on asecond cut, because the cutter load is much less.

Mr. L. G. Burnard (Chief Development Engineer,Vickers-Armstrongs (Aircraft) Ltd.) : I should like torevert to an earlier question about the use of drawingsand the possibility of a machine which would be ableto read a drawing.

Would it not be possible, with the development ofthis technique in the future, to dispense entirely withdrawings ? If a part such as a pipe or a wingcontour can be described mathematically ornumerically, it would be possible for the draughtsmanor someone at the draughtsman's desk to punch acard containing this numerical information and,possibly, information about the number of parts tobe made, and pass it to the machine shop, so thatwe might get a part or parts from the shop withoutthe intermediate planning, ratefixing, work office,production control, and so on. There is a possibilityof this happening in the future. It may be that weshall then see our production control work done bya computer of some sort. There are millions of bitsof information which could go into the memory orstore of a standard computer so that, with certainparameters, it could give you immediately the infor-mation you want about the state of production inthe shop.

It is said in the Paper that one of the advantagesof the analogue system is that if, for any reason,the machine is stopped, it will return to the positionwhen next switched on, the voltage analogue havingbeen determined. Surely that can also be done todaywith the digital system ? New binary coded scalesare available which have unique positions for everyincrement that you wish to put along the machinetool slide. Thus, if there is a unique position, the

machine upon being switched on, will return tothat position. Is that true ?

Mr. Booth : Yes, that is already dealt with.I think the statement in the Paper is true. The digitalsystems in this country and America have been incre-mental systems, but in principle it is certainly true,and 1 know of a rotary form of what you describe.

I would suggest that very often a qualitativedrawing will be wanted, because most people havepictorial minds long before, and to a much greaterextent than, they have numerical minds. I am surethat qualitative drawings will continue to benecessary.

Mr. Burnard : I meant that the drawing couldbe done in the shops because there was a numericallycontrolled inspectorate.

Mr. Booth : There is still the pictorial case tobe considered.

Mr. Puckle : Mr. Burnard is quite right in sayingthat a computer can be used for Shop PlanningControl purposes, but that lies outside the scope ofmy Paper.

The Chairman : It is a pity that we do not havemore time for discussion. It has been most stimulating,We now have plenty to think about while ourcompanies save up enough money to purchasenumerically controlled machine tools.

I should like to return to a point in Mr. Puckle'sPaper. He mentioned that an aluminium part wasprepared in half a day and machined in 14 minutes,but it took 12 hours to inspect. I suggest that anumerically controlled checking apparatus be devisedas quickly as possible, to obviate having to have ahorde of inspectors checking the things turned outwith the machine.

The advantage of the equipment must be apparent,particularly in the accurate production of compli-cated machine parts in expensive materials whichwe are now having to face in the new projects — thevery high tensile steels, titanium and similar materials.

I thank Mr. Puckle on behalf of the SouthamptonSection of the Institution of Production Engineersfor his excellent Paper.

The vote of thanks was carried by acclamation.

Mr. Puckle : The Chairman has remarked on analuminium part made in 14 minutes which took 12hours to inspect. I would draw his attention to thefact that this statement appears in my extract fromthe Paper read by Mr. Wood in Cincinnati. We arein agreement that an electronic inspecting devicewould have great value for many purposes. On theother hand, perhaps I might say that once the firstpart has been made and checked, there should notbe any need to check future ones made from the sametape!

192