During a long motor-less period – I have not touched a steering wheel for some 15 months – I have been thinking a little about the sort of car that I should like to own when the present nonsense is over. I find that my ideas correspond roughly with those expressed in the late Ft.-Lt. Scafe’s letter published in Motor Sport in June, 1942, in which he indicates that the solution is along the lines of the Type 328 B.M.W., but that I have gone into rather more detail than he has. The fundamental requirements on which the details of my design are based are as follows:–
(a) Comfortable accommodation for two people, with luggage space and emergency accommodation for a third person.
(b) Performance to give 75-80 m.p.h. cruising speed, and 100 m.p.h. maximum with a high rate of acceleration, to be provided at a minimum running cost.
(c) It is also assumed that a person buying the type of car envisaged would not want his motoring confined to smooth roads, and that he would probably enter occasionally for trials of the not too fearsome variety, and would, therefore, expect good ground clearance and hill-climbing capacity, as well as a fairly sturdy motor-car that would not break up too quickly.
These requirements can be met in a car of quite moderate dimensions – say 7 ft. 10 in. wheelbase and 4 ft. track – housing some 1 1/2-litres of engine in a rigid chassis with soft independent front suspension, and either independent rear suspension or a De Dion type axle, the whole being covered over by a streamlined body of the enveloping type (open, of course). This body, quite apart from the gain in performance resulting from its aerodynamic shape, offers the great advantage that the inevitable third passenger who always turns up when the car is already loaded with its normal complement of two, is comfortably accommodated behind the windscreen on the bench-type front seat, and does not have to be stowed away by himself in a draughty seat at the back. The merits of front-engined front-drives and rear-engined rear-drives have been considered, but the disadvantages resulting from these layouts more than outweigh the advantages when applied to a moderate-sized sports car, so that the normal layout is retained, except that the gearbox is in unit with the final drive rather than with the engine, for reasons which become apparent later.
Now to consider the design, component by component, in more detail: Firstly, the frame, which has to provide a rigid foundation on which the whole structure of the car is built, and which at the same time must only occupy space which is not otherwise required. For any closed car and, for that matter, for an open racing car, the correct form of frame construction is surely either the unified body-chassis construction as exemplified by the front-drive Citroen or the Vauxhall Ten, to name two of many examples, or, for cars not built in large quantities, the basically similar built-up tubular structure used on the original Burney Streamline some 12 years ago, on the Aston Martin “Atom,” and, since I have included racing cars, Robert Waddy’s very brilliant design, “Fuzzi.” On an open car, however, unless one expects the crew to climb over the side (which is rather a lot to ask of a fair passenger when dressed up for an evening festivity – more than I intend to ask, any-way), this type of construction is rather out of the question, due to the very small depth available where the door is placed. (Yes, I can remember the early Lancia “Lambda” which was open and which embodied combined body-chassis construction, but the “Lambda” was not a very low-built car, neither were the doors very deep, which makes it a rather different proposition.)
One is, then, left with the alternatives of a frame of more or less conventional layout, with either tubular or box-section side members, or the backbone-type of frame. Brief consideration will show that the backbone frame has many merits. First, there is considerable saving in weight and, secondly, it only occupies space in the passenger compartment that is already occupied by the propeller shaft tunnel, so that the seats and the floor can be placed as low as other considerations will allow. I lay great emphasis upon the lowness, as not only is it a means of reducing air resistance to a minimum, but also the lower the centre of gravity can be arranged, the softer can the suspension be without the car behaving like an American saloon when cornering, braking and accelerating – which is a very important consideration. Actually, the height of the seat cushion is governed by visibility over the front wings, but it is still considered necessary to have a flat floor as low as possible, as a driving position in which the feet are no lower than the seat cushion is not comfortable on a long run, and the nearer that one can approach to a kitchen-chair sitting position, the more comfortable does driving become and the more control one has over the motor-car. The desirable end of making the floor boards the lowest part of the car is achieved in the Morgan 4/4 without departing far from the conventional type of frame, but if the front seat is to be wide enough for three, this can only be done at the expense of an undesirably sharp and pronounced kink in the plan shape of each side member where it has to sweep in behind the seat to a width that can be accommodated between the rear wheels.
Of various examples of backbone frame construction, the Tatra arrangement, in which the crankcase and gearbox unit and the final drive unit form part of the frame and are connected by the tubular steel backbone to which both are bolted, offers rather serious objections as regards accessibility and makes impossible the resilient mounting of the engine or gearbox; but the R-type M.G. Midget frame, which overcame these objections and at the same time was very rigid and weighed only 57 lb., if I remember rightly (though I may be thinking of the famous 57 varieties), surely indicates the lines along which to work.
Briefly, then, the arrangement visualised is something very much like the R-type M.G., shaped in plan view like a tuning fork, but with a smaller fork at the rear end in which the gearbox is mounted, the whole affair being of welded rectangular-box-section construction.
Incidentally, I rather hope that since this is to be an open car the noise and vibration from a rigidly-mounted gearbox can be overlooked, and that the gearbox casing can be bolted into the rear end of the frame, where it will serve very nicely to brace together the two arms of the rear fork.
At the front I am afraid that modern practice will insist on flexible mounting of the engine, and so a large box-section cross member, either in line with, or just in front of, the front wheels, is needed. This cross member may be connected to the side members by a multiplicity of small bolts all drilled, and split-pinned or wired, so that the engine can be comparatively easily removed by sliding it forward when the cross member has been taken out. Another cross member will be required right at the front of the frame, which should end in such a way that a bumper can easily be fitted. (This, I freely admit, sounds rank heresy, and I would hate my own car weighed down front and rear with such a fitment, but in countries where it is normal for a car to be used as a shunting locomotive when parking, one’s wings definitely need protection. Fortunately there are other bumpers than the chromium snowploughs fitted to the latest American cars, and it should be possible to build a narrow bumper with a coloured rubber surface on to the front of a streamlined body with much less offence to the eye than mounting an ordinary one on to, say, a vintage Bentley.)
The detailed design of the frame is very closely tied up with that of the suspension, so that must now be considered. Of the possible arrangements at the front end, I rather suspect, perhaps unjustly, that the Lancia vertical guide system does not allow very large wheel movements and so rather restricts the softness of the springing. This difficulty definitely arises with the trailing link arrangement unless the links are made very long, in which case difficulty is found in making them sufficiently stiff in a lateral direction. In any case, both of these systems necessitate cross members running almost the width of the track and carrying the weight of the car, whereas the twin-wishbone system fits very nicely indeed on to a frame of the backbone type, is very rigid in a lateral direction, and is well able to resist brake torque and, if used with either a transverse leaf spring or torsion bars connected to the wishbones, does not need any cross member other than that bracing the two side members of the frame together. I am inclined to favour the torsion bar suspension rather than the transverse leaf spring, mainly on the grounds of weight saving, but partly because it requires less maintenance. To reduce unsprung weight to a minimum, the stub axles should be ball-jointed direct to the suspension links, as in the case of the Citroen front-drive or the Porsche suspension of the Auto-Union and the E.R.A. From the point of view of safety, it would appear desirable to connect the upper link to the torsion bar, but this is not essential if the torsion bar cannot be accommodated in this position on the frame. A point that I do consider rather important is that the shock absorbers, of the piston hydraulic type, should be built on to the hinge of the same link as the torsion bar, as I rather dislike shock absorbers connected up by rubber links, and also I think that all the control of the wheel motion should be vested in one link only. As an alternative, I am partial to the direct-acting, hydraulic shock absorber, but this needs fairly substantial struts and stays to provide a rigid mounting for the fixed end, which should be as far out from the centre line as possible.
As regards the rear suspension, I am endeavouring to keep an open mind between the respective merits of the De Dion axle and various systems of independent rear suspension. Much has been written and spoken regarding the respective merits of these since the two German G.P. cars switched over from swing-axle suspension to the De Dion type axle, and I am inclined to think that all types of independent rear suspension are being condemned because the swing-axle type has been tried and found wanting for very high speed work. There has been no evidence that the rear suspension of the G.P. Alfa-Romeo or the R-type M.G. Midget were other than satisfactory, and there is positive evidence of good results obtained on the Alta, since this feature was continued in the latest car completed at Tolworth during the first few weeks of the war. The same state of affairs holds with regard to training cars, though there is an undoubted decreased resistance to roll with independent rear suspension of types other than the swing-axle, as compared with axled suspension. (Incidentally, it has always been one of life’s more curious conundrums that where many makers found it necessary to increase the stiffness of their suspension against roll by adding torsion-bar stabilisers, Lancia, having adopted in the “Aprilia” an all-independent suspension, which is supposed to have a lower than normal resistance to roll, reduces this further by arranging that only half the rear suspension resists roll, and that the pivoted leaf spring only acts, presumably, when both wheels drop into a gully.)
Until definite evidence is produced that it is unsatisfactory, I propose to plump for a wishbone suspension system at the rear, very much the same as at the front end, with the same torsion bars used as the springing media. In case difficulty is experienced in accommodating a long torsion bar, either as regards the space that it occupies or if it means introducing a mounting stress on the chassis at an inconvenient place, it must be remembered that Vauxhall’s have shown that it can easily be doubled back on itself by enclosing the torsion bar in a torsion tube.
If the wishbone rear suspension really has unknown vices, then the De Dion axle must be used in its place. The use of long radius arms usually employed to locate this type of axle is ruled out by the low seating position, but as the reduction of unsprung weight to a maximum dictates that the brakes be carried on the frame-mounted final drive unit there is no brake torque to be resisted by the axle, which can, therefore, be controlled and located by shackling its ends to wishbones in the same place as the lower ones of the usual twin-wishbone suspension. The axle must be located laterally either by the Mercédès system of two triangular members, ball-jointed together at their apices and hinged, respectively, to transverse axes on the frame and on the axle, or preferably, to my mind, by Watt’s straight-line motion consisting of two transverse radius rods connected at their outer ends to the axle and at their inner ends to a short link ball-jointed at its centre to the frame.
The steering is closely connected with the type of front suspension. The B.M.W. and other continental cars show us that with a properly laid-out steering geometry, a simple rack and pinion system can be employed with perfectly satisfactory results and, this being so, there seems no excuse for adding weight and expense (and a la of ball-joints) in adopting one of the modern proprietary steering boxes, excellent though they are.
An integral part of the suspension is the tyres, and here I suggest that right from the start the car be equipped with different tyre sections front and rear, instead of leaving it to the enthusiastic owner to find that the front and rear wheels have rather different roles to perform, and that these are better performed by a smaller section at the front than at the rear. In the interests of lowness and of saving of unsprung weight, wheel diameters should be as small as possible; 6.00″ x 15″ rear and 5.00″ x 17″ front are suggested. These sizes have the same nominal diameter, so that with one spare the same size as the front, a journey interrupted by a puncture at the rear can be completed without the car being lopsided or the differential having to work overtime. Our pre-war tyres were very reliable and punctures were quite rare, but whether the same will apply to post-war tyres remains to be seen. Incidentally, I am not sure whether 6.00″ x 15″ tyres are made (5.00″ x 15″ and 5.50 x 15″ are used on the “1,100” and “1,500” Fiats), but if Mr. Dunlop will not oblige, then the rim sizes suggested above will have to be increased by one inch in diameter.
With such small rim diameters as 15″ and 17″, the Rudge type of wire wheel is heavier than a cast light-alloy or pressed steel wheel, so that, if the enthusiastic owner can possibly forgo the pleasures of rapid wheel changing, a 4-studded, bolt-on wheel is preferable. (The Austin Seven wheel construction saves a lot of time in wheel changing and gives no trouble if the nuts are tightened intelligently from the beginning of the car’s life.) If, however, a centre lock system must be used, then the wheel can be located on conical faces on four or six false wheel studs and can be held up against them by a large eared nut in the centre with a simple locking device of some kind.
The braking problem is solved by the use of one of our excellent proprietary systems, between which there is little to choose, but whatever system is used, hydraulic actuation is indicated. The disc type brake is now an accomplished fact and is giving excellent service, and is therefore worthy of consideration, particularly at the rear where the wishbone suspension imposes a certain amount of restriction on brake diameter, if the brake is to be dismantled without having to take down the suspension as well.
Before describing the transmission in detail, a few words are necessary in explanation. The need for a low centre of gravity and a bench-type front seat indicate that the propeller shaft must pass through the frame and not along the top of it as in the R-type M.G. Midget, and that the frame must be as low as ground clearance (say 7″) will permit. This in turn indicates that something must be done to bring the propeller shaft down to the low level required. This end is reached by using a two-piece shaft, with a centre universal and a steady bearing somewhere in the region of the front edge of the seat, and by changing the gearbox design so that (in effect) the primary shaft is thrown away and the propeller shaft drives the front end of the layshaft, the bevel pinion, of course, being on the rear end of the mainshaft.
This, of course, makes the provision of a direct drive in the normal sense impossible (though it is actually the same as the Frazer-Nash, which was claimed to have a direct drive on all “gears”), but I cannot see that this is a serious disadvantage, as what one loses in efficiency in top compared with the direct drive, one gains on all other gears by taking the drive through only one pair of gears instead of through two pairs, while with well cut helical gears (or straight gears for that matter) the noise factor is negligible. Also, one can quote at least two high quality cars which have indirect top gears, namely, 4 1/2-litre Bentley and 20-h.p. Daimler, and many not such good quality American cars fitted with overdrives. Also, to obtain effortless cruising and at the same time good acceleration at high speeds, I consider that two close ratios are necessary, the highest in effect an overdrive, too high to give maximum road speed, and the lower one such that the car’s maximum speed is attained at, or past, the engine’s peak speed, and as these two gears will be used as alternative top gears, there is little justification for discriminating and making one direct and one indirect.
To provide a bottom gear low enough for easy starting on a very steep grade will necessitate three further gears if wide gaps in the ratios are to be avoided, so that the gearbox will be called upon to provide a total of five ratios, in addition to a reverse.
And now to consider the transmission in more detail. Few, I think, will find fault with the selection of a Borg and Beck single-plate clutch mounted in the flywheel, especially since this is rendered relatively accessible by the banishment of the gearbox to the back end of the frame.
The propeller shaft can be relatively slender and light, as it now only has engine torque to transmit and, being in two short lengths, does not have to have a large diameter to prevent whipping, although it must be remembered that its maximum speed is limited, not by the maximum speed of the car, but to that speed at which the engine blows up. As it is desirable that the centre main bearing of the shaft be as low as possible, the front and centre joints will run at a certain amount of angularity and should, therefore, be of Hardy Spicer needle-roller type, but the angularity of the rear joint can be laid out to be nil, so that a Layrub rubber-bushed joint can be used here.
While on the subject of propeller shafts, the two universally-jointed half shafts can be mentioned. With the wishbone suspension, the wheel remains approximately parallel to the frame, so that Hardy Spicer shafts with the wide angularity-type, needle-roller joints, as used in the R-type M.G., can be utilised, as they can be laid out so that the variations in angular velocity from the two joints cancel each other out; but With the De Dion axle it may be preferable to use constant-velocity joints.
The general layout of the gearbox has already been dealt with and it only remains to detail this a little further. The chief difficulty with the type of gearbox suggested is that of providing an adequately low bottom gear without having to add a compound gear train. However, a 13 1/3 to 1 bottom gear, which should be quite adequate with the good power/weight ratio and the comparatively small wheels (27″ nominal diameter), can be obtained with 2 2/3 to 1 gearbox reduction and 5 to 1 crown wheel and pinion, and these ratios should be attainable without a crown wheel so large as to lack rigidity and thus be noisy, and without increasing the centre distances of the two gearbox shafts to an unwieldy figure. With an engine of 90-mm. stroke, developing some 70 b.h.p. at 5,500 r.p.m. and running safely to 6,000 r.p.m., I suggest that the remaining ratios be 8.8, 6.1, 4.5, and 3.6 to 1, giving maxima of 36, 54, 78, 106, and a very theoretical 132 m.p.h. at 6,000 r.p.m. (3,550 ft./min. piston speed) and 93 and 74 m.p.h. on the two highest gears at a piston speed of 2,500 ft./min. (4,220 r.p.m.). With these ratios, of course, both fifth and fourth are, in reality, overdrive ratios.
As regards the actual construction of the box, this should be an aluminium or electron casting, divided into four compartments by transverse positions, all of which carry bearings for the two shafts, the rear compartment containing the crown wheel and pinion and differential, and the other three the first and reverse gears’ second and third gears and fourth and fifth gears, so that each gear is adjacent to a bearing and noise will be reduced to a minimum. All gears should be engaged by dog clutches without synchromesh, to reduce to a minimum the motion required on the selector mechanism, except possibly for the bottom and reverse gears, where by sliding it along the splined shaft the same large wheel may be used for both gears. The two highest ratios should have double helical teeth for quietness and to avoid end thrust, but I think that well-cut straight gears would not prove unpleasantly noisy for the remaining two gears, though these should still run in constant mesh and be engaged by dog clutches. To ensure that all gears rotate as slowly as possible and thus reduce churning losses, the fixed gear of each pair should be smaller, so that the sliding dog will be on the driving shaft in case of the top two gears, and on the driven shaft for the others, which indicates that the selector mechanism should be along one side of the box.
As regards the gear-change arrangements, I am afraid that the consensus of opinion will be that I am becoming influenced by living in a land of American motors, as with a seat designed to take three people the only place that I can see for the gear lever is on the steering column. I do specify, however, that the mechanism must be very carefully laid out with a view to rigidity and correct feel, and, granted this, I am convinced that a satisfactory arrangement can be made. Many modern American cars definitely lack rigidity in their gear-change mechanism, so that it is doubtful if the gears could be engaged without synchromesh, but it must be remembered that they follow on in the tradition that used to provide a yard-and-a-half of chromium plated 1/4″ wire with which to effect changes of ratio!
It also becomes necessary to utilise a dashboard hand-brake and, here again, I think that the pistol-grip arrangement, as it already exists, could, with sufficient care in manufacture and in selection of the right leverages, be made to suit the rather fastidious taste of the sporting motorist.
And so at last to the power unit, which, frankly, has not been dealt with before because my ideas with regard to it are rather more tentative. To give an easily attainable maximum of 100 m.p.h. together with brilliant acceleration will require some 70 b.h.p., a very awkward figure, since to produce this will require a rather exceptional unsupercharged 1 1/2 litre, judging by pre-war standards, while 2 litres, I feel, is getting a little large for a high-speed engine with only 4 cylinders. My solution, then, is the unblown 1 1/2, with the alternative of the same motor bored a few millimetres larger, to a capacity of 1 3/4 litres, for the customer who is not fussy about being a few c.c. over the international capacity limit.
As a first step towards attaining the reliable 70 b.h.p. from 1 1/2 litres, the stroke can be reduced by breaking away from the almost standard dimensions of 69 x 100 mm. and adopting instead 72.5 x 90 mm., which gives 1,486 c.c. and will cost another 25s. per year (13.03 h.p. R.A.C. rating). (The 1 3/4-litre version, incidentally, has 78 mm bore, giving 1,720 c.c. and 15.00 h.p. R.A.C. rating.)
If we then reckon on the 1 1/2 litre peaking at 5,500 r.p.m., it has to develop some 47.1 b.h.p. per litre and to have a b.m.e.p. of 110 lb./in., which is a high figure, but not one which is inconsistent with reliability, given an individual type of engine with good porting and cylinder head layout, a high compression ratio (say 9 to 1) and a rigid bottom half, plus, possibly, a slight gain from the higher octane fuels which post-war conditions may give us.
To provide the efficient porting and cylinder head an inclined valve layout with hemispherical combustion spaces is indicated, which, in turn, means that a certain amount of complication must be introduced into the valve gear. However, while one may deduce from the successful operation of E.R.A. valves by means of a push rod and a bell crank rocker at up to 9,000 r.p.m., and of the inlet valves of the Type 328 B.M.W. at up to two-thirds of this speed through two pushrods and two bell crank rockers each, that valve mechanisms no longer cause designers the severe headaches that they did at one time, it is considered that the valve mechanism should be as simple as possible in the interests of long life. This end may be conveniently attained by a single camshaft, chain driven, mounted in the cylinder head just offset from the centre line so as to clear the sparking plugs, operating the valves through rockers mounted on two rocker shafts, one on each side of the head midway between the camshaft and the end of the valve stem. With a chain drive, the overhead camshaft offers no more complication than one in the crankcase, while it offers the advantages of a narrow crankcase (which fits neatly into the backbone frame) and a symmetrical cylinder block with its freedom from distortion.
The whole cylinder head will be covered over by one rather extensive head cover, the sparking plugs either being protected within a tube extending from the head through the cover, or dealt with in the Lancia “Aprilia,” in which they are contained within a sort of permanent “box-spanner” affair.
I must confess to being rather conservative as regards engine design, and for the sake of rigidity prefer to use a solid block of cast iron as a basis on which to build the engine. The Type 46 Bugatti construction, however, in which the main bearings are cast as part of the cylinder block, and the light alloy crankcase merely forms a box which encloses the lower half of the engine, indicates a method of combining rigidity of the integral block and crankcase, with the weight saving of the separate cast-iron block-light alloy crankcase construction. Since the crankcase is completely unstressed apart from the weight and torque reactions of the engine, one can even go so far as to use electron, which is lighter, but even less rigid, than aluminium.
Unlike Bugatti, however, the cylinder head will be detachable, as this component can then be cast in a light alloy, with advantages from weight saving and thermodynamic points of view, though great care will be needed to ensure that its “buttoning down” is very thorough, to avoid gasket troubles.
Three main bearings will be sufficient for the crankshaft, which will gain in stiffness due to the short stroke and also due to the reduction which can be made in the length of bearings by using high duty bearing metals, though at the expense of having to harden the shaft to prevent a high rate of wear.
I must admit a liking for the Aston-Martin camshaft drive, using a pair of two-to-one helical gears to drive the lower chain sprocket from the crankshaft, so avoiding a high linear chain speed. The intermediate spindle can then be used to drive the water pump, the dynamo can be directly driven from the front of the crankshaft, and by taking another leaf out of Bugatti’s book and driving the Scintilla Vertex magneto directly from the rear or the front of the camshaft, all of the auxiliaries can be disposed of except the oil pump, which can be mounted under the front main bearing cap and can be driven by a spur gear meshing with the gear already provided on the crankshaft to drive the timing chain shaft.
Dry sump lubrication has the advantage of reducing the engine height as well as of keeping the engine-oil cool during prolonged running at high speeds, while if undershielding is used, the amount of air passing through the bonnet can be reduced to a minimum. The oil tank can be accommodated between the radiator and the radiator grille, or else alongside the crankcase, in which case an oil radiator will have to be provided. I should very much like to see the induction system arranged as on the Type 328 B.M.W., in which the mixture drops straight from the downdraught carburetter into the cylinder head by means of inlet ports which lead into the top face of the inclined valve head instead of in at the side as is usual. This is, however, impracticable as it means raising the bonnet level to accommodate the carburetters and two horizontal or semi-downdraught S.U.s, mounted on a manifold on the side of the head, will have to be used instead.
As already indicated, the cooling water is circulated by pump so that directed flow cooling is available for the exhaust valve seats, and also because the radiator header tank will be rather low for efficient thermocycle cooling. A fan I very firmly bar, as I am convinced that most fans use up far more b.h.p. than is generally appreciated and that one is quite unnecessary if the radiator is properly designed. In order to accommodate the forward projecting dynamo and possibly the magneto as well, the radiator can be divided into two blocks, with a vertical gap some six inches wide between them.
The details of the body I leave to a better brain than mine, except for specifying a construction of aluminium panelling on a welded tubular-steel or light-alloy framework and for suggesting that the Bugatti that graces the cover of the March, 1943, Motor Sport illustrates the lines along which to work.
Having thus laid out the design for the car, it remains to consider with what effect it can be operated. 70 b.h.p. should suffice to accelerate its 15-16 cwt. to a speed of a mile per minute in not much over 12 seconds, and comparison with the 42-b.h.p. 90-m.p.h. Fiat saloon indicates a maximum of 106 m.p.h. if one assumes (not unreasonably, I think) that the better streamline shape of the aerodynamic saloon is compensated by the lower frontal area of the 2/3-seater. Its high top gear ratio, streamline shape, and soft suspension should make possible cruising at 80 m.p.h. with about 27 m.p.g., even on roads that are not ultra smooth, without fatigue to either itself or to its driver.
There, very briefly, are my ideas, and now all that remains is for my fellow enthusiasts who avidly devour Motor Sport each month to point out to me just why it cannot be done.
[Capt. Moon deserves the very greatest credit for setting down on paper the outline of a design which is so excellent as to require no excuse on the grounds that its originator is very busy indeed in the Middle East, and is away from his collection of data, although such is the case. Readers should certainly consider the points he makes and the layout he suggests with more than usual care. Suggestions for other “ideal” high-performance cars that would enliven the prospect of peace are invited. – Ed.]