A DESIGN FOR A 11/2-LITRE FORMULA CAR

Capt. John Moon presents

A DESIGN FOR A 1-LITRE FORMULA CAR

In September, 1943, Capt. Moon described his ideal sports car, which aroused widespread interest. Now he outlines a most intriguing design for a

I Formula G.P. car—it seems likely that if there is any more International Formula racing the IHitre limit will be used, and certainly 11-litre racing may be expected in this country, so John Moon's design, intended to beat anything known before the war, has a topical appeal, and the details of his proposed straight-eight, four-wheeldrive car will repay careful study.—Ed.

HAVING no motoring to do and not even any motoring literature to read—a few precious copies of MOTOR SPORT that have travelled round with me for the last two years have had to be weeded out in an endeavour to reduce my kit to the weight laid down-I have been passing a few enforced idle hours in hospital in considering what sort of racing car we may see when new cars are produced after the war, and it now occurs to me that these ideas may possibly interest other readers of MOTOR SPORT, and may invoke criticism that will make interesting reading.

I have confined my notes to a proposed 1 Flitre car, as voiturette racing was progressing from strength to strength before the war and there was every indication that the next Grand Prix Formula would take the form of a 1 flitre capacity limit. There is considerable reason to believe that things will continue in the same direction when racing recommences after the war. Incidentally, if all the 1939 and 1940 voiturettes have avoided the blitz (which I fear they will not have done) there is some very fine racing in prospect as soon as activities recommence, with the new E.R.A., the little Mercedes-Benz, and the 1940 editions of the Alfette and the Maserati, all of about the same performance and so far largely unmatched.

Of course, as John Bolster told the London graduates of the I.A.E., the racing car of the future may be driven by an internal combustion or a closedcircuit mercury vapour turbine with independent electric or hydraulic transmission to each road wheel, and with a system of regenerative braking, but as I do not understand these things I have left them out of my deliberations. In any case, however, I feel that these things will not be with us just yet.

And now to the racing car of 194? Engine.—As the power unit is the heart and soul around which the rest of the car

is built, it is logical to deal with it first of all.

Present-day 1 i-litre engines are generally reckoned to develop about 250 h.p. in long-distance racing tune, and they will probably be developed in a year or so, so that an output of 200 h.p. per litre, already exceeded by the E.R.A. for short events, can be sustained throughout a race of 200-300 miles' duration. This being so, it appears that their successor will have to produce a sustained output of about 350 h.p., or some 233 h.p. per litre. This sounds a pretty terrific output, but Laurence Pomeroy has shown us, in an article some two years ago, that an output of 450 h.p., or 300 h.p. per litre could be attained from a 12-cylinder 111litre engine without exceeding boost, mean effective pressures or piston speeds that are already used in present-day high-output engines. Fortunately, for an output of 350 h.p., it is possible to avoid the complications of 12 cylinders, by using eight cylinders with heads and *porting based on the layout of M. Lory's 1926 11-litre Delage, with a modified bore/stroke ratio.

The Delage, which still remains one of the most efficient engines produced, has a bore and stroke of 55.8 mm. by 76 mm., and as modified by Ramponi for Richard Seaman in 1936, produced 185 b.h.p. at 8,000 r.p.m. with a boost of only 7.5 lb./sq. in., which is equivalent to an unblown brake mean effective pressure of 134 lb./sq. in. If this engine is redesigned with a bore and stroke of 60.5 mm. by 65 mm. (which gives 1,495c.c. swept volume) the cylinder heads, ports and valves can be enlarged in proportion to the increase in the cylinder bore (particularly as the sparking plug and the valve seat widths do not need to increase in size). This will mean that the gas speed through the inlet port will bear the same relationship to the piston speed as previously, and if the same piston speed is used, then the gas speed, which has an important bearing on the efficiency of the top half, will also be the same. The piston speed of the Delage is 4,000 ft. per min., and with the reduced stroke of 65 mm., the corresponding crankshaft speed comes up to 9,350 r.p.m. With the increased engine speed and with the slightly larger combustion chamber, which may not be quite so favourable in shape, due to the reduced stroke/bore ratio of 1.08, it is reasonable to suppose that the equivalent unblown b.m.e.p. will fall off somewhat from the Delage's excellent figure of 134 lb./sq. in. to somewhere about 120 lb./sq. in.

To produce 350 h.p. from 1,495 c.c. at 9,350 r.p.m. requires that the brake mean effective pressure be 326 lb./sq. in., and to produce this the induction pressure must be boosted over atmospheric in the same proportion as this figure is in excess of the equivalent unblown b.m.e.p. of 120 lb./sq. in., which indicates that an absolute induction pressure of 40 lb./sq. in. or a supercharge of 25 lb./sq. in. is required. Summing up, then, our 60.5 mm. by 65 mm. 8-cylinder engine boosted at

25 lb./sq. in. develops 350 h.p. at 9,350 r.p.m., at a piston speed of 4,000 ft. per min. Horse-power per sq. in. of piston area is 9.8, a high figure, but lower than that of many racing engines operating satisfactorily. The horse-power per sq. in. of inlet valve area is again high at 32, but without figures to support me, I think that this is lower than the 1938 2-litre E.R.A., an eminently satisfactory performer. The maximum engine speed will probably be about 10,500 r.p.m., corresponding to a piston speed of 4,490 ft. per min. General Construction.—The keynote of the whole engine layout must be the necessity for obtaining as rigid a structure as possible in the interest of bearing life and in keeping friction to a minimum by reducing distortion. There is little justification for using a V-form of construction unless the number of cylinders makes it necessary, as lubricating oil is notoriously difficult to retain in these engines at high speeds—witness the troubles that afflicted Mercedes-Benz throughout 1938 with their Formula cars. The horizontallyopposed engine is a good proposition from the point of view of rigidity, but is rather difficult to install in a racing car, and may be no better than the V-type for retaining its oil. Thus one is left with the welltried straight-eight construction, which has been used, I think, for more racing engines than any other type. Because it happens to suit the proposed chassis layout rather than because of any anticipated difficulties with the crankshaft design, I propose to copy the arrangement of the 1940 V16 Alfa-Romeo :3-litre, designed by H. R. Ricardo, in which the drive is taken to the clutch by a downward extension of the central timing gears and a shaft passing underneath the crankcase. In my case, however, the drive is taken out sideways to a gearbox mounted alongside the crankcase. Incidentally, provided that a form of coupling is used that definitely prevents the transmission of torsional vibration between the two halves of the crankshaft, this arrangement, which is virtually the mounting of two engines end to end, offers immense possibilities in the multiplication of numbers of cylinders. For instance, a 24-cylinder 1 f-litre, consisting of a double V12, is presumably a practical proposition, and with its 02-c.c. cylinders, 800 h.p. per

litre should be easily surpassed. On the other hand, it would be rather a handful to dismantle !

Cylinder Blocks and Ileads.—The primary consideration in the layout of the cylinder block and heads is the arrangement of the porting. Efficient porting layout seems to me to necessitate the head being integral with the cylinder block. Study of a cross-sectional drawing of the Delage engine makes it quite clear that a detachable head could not be incorporated in the design without spoiling in a large measure the beautiful sweep of the ports. At the same time, use of a fixed head eliminates the possibilities of gasket troubles and does away with inconvenient stud bosses, and so on, that may cause distortion. In view of the high power output the head-cum-block must be made up of high-conductivity material—cast iron served admirably for the Ii-litre Delage, but in spite. of the excellent water spacing, it would crack, I am sure, if subject to the heat flow resulting from a 25 lb./sq. in. boost. As the cylinder blocks represent quite a mass of metal, light alloy castings are indicated, and these satisfy the previous condition also, as well as being quite easy to cast. It does, however, raise difficulties with regard to cylinder bores, to which there are three alternative solutions.

Firstly, dry ferrous liners can be pressed in from the bottom and locked into place. Secondly, wet liners spigotting into the cylinder head and sealed at the top by a Wills pressure ring (an annular ring filled with inert gas), and with a flange at the bottom, possibly separate from the liner, bolting to the underface of the water jacket, can be used. Finally, there is the Cross linerless cylinder in which the piston runs on special iron rings which bottom in their grooves without allowing the piston skirt to touch the cylinder bore. Of these alternatives, I am inclined to favour the second, as the first involves an extra joint face in the heat path, and the third is untried on engines of ultra high output. A further disadvantage of light alloy is that this material is much less rigid than east iron. This can be mitigated by extending a method of construction used by Bugatti, combining the cylinder block-cum-head with the crankcase as far down as the crankshaft, or even lower. That this renders it impossible to grind-in a valve without removing crank

shaft and pistons is, of course, not unexpected in a design by M. le Patron, nor need it be considered a great disadvantage in a racing engine that is probably rebuilt between every three or four events.

Reverting to the general construction, two cylinder block-cum-crankcase assemblies would be used, with a timing case made in two halves, each of which would be bolted to its respective cylinder block and the two halves then bolted and spigotted together when the timing gear train had been assembled. The two crankshaft halves, already connected together with the driving gear, would then be fitted up into the main bearings. The bottom will be closed by a sump which can be a ribbed electron casting, either in one piece or in two halves bolted together. The final power take-off gear which would be driven through an intermediate idler gear from the crank gear, will be carried in its bearings in a separate casting bolted to the side of the timing case. If a further tie between the two cylinder blocks is required at the top of the engine, single piece valve chests and covers can be employed.

Crankshaft, Connecting Rods and Bearings.—In order that the largest possible proportion of the power developed in the cylinder head reaches the clutch, roller bearings are used for both main and bigend bearings, since the increase in mechanical efficiency resulting from their use compared with plain bearings is apparently quite appreciable, particularly at high rotational speeds. In view of the short stroke which helps towards a rigid crank, three main bearings for each will be sufficient, making a total of six. As there is no external thrust at all on the crankshaft, lips on both sides of one main bearing will be quite sufficient to locate the crank endways. By taking the power from the centre of the crank, slight reduction in the size of the journals can be made, slightly reducing bearing friction.

As a certain amount of wangling is already necessary in assembling the engine, I think that pity had better be taken on the mechanic to the extent of specifying split big-ends and mains, with removable caps, rather than a built-up shaft with solid connecting rods, which would be a real jigsaw to assemble into the cylinder blocks.

One difficulty arising with roller bigends is supplying oil for piston-cooling. With the high power output it is desirable to oil-cool the underpart of the piston head, and this is most satisfactorily achieved by a jet in the top of the connecting rod to which oil is fed from the crankshaft up through a rifle drilling in the rod. Obviously, though, oil in sufficient quantity cannot be fed radially through a roller-bearing big-end. The only solution appears to be to widen the big-end until there is room for two sets of narrow rollers side by side, in cages with very little radial clearance, and to pump oil through the space between the two races. The problem of delivering oil into a crank carried on roller main bearings can be solved by feeding it in through pressure bushes at the two ends. As the induction arrangements make it desirable to treat the engine as two 4-cylinder units, the 4-4 crankshaft

arrangement has to be employed. The resultant unbalanced secondary couple is of no importance in a racing engine. I believe that all Alfa-Romeo racing eights have used this construction, and so does the Isotta Fraschini tourer.

Valves and Valve Gear.—In writing this, so far I have assumed that the normal form of poppet-valve gear, as we know it, will be employed. It may well be, however, that by the time this car is laid down, a satisfactory rotary valve system will have been evolved, or that the Aspin rotary head arrangement will be in common use on racing ears. The latter, at any rate, now appears to have reached the stage at which a car incorporating it could give most unblown cars a run for their money.

However, in the absence of a more satisfactory alternative, there is little doubt that poppet valve gear can quite satisfactorily cope with the conditions envisaged. Valves with • sodium-filled stems will obviously be used, and it may be they will be large enough (33.5 mm. diameter inlet and 31.5 mm. diameter exhaust) to enable the heads to be hollow as well. A further dodge towards easy elimination of heat from the valves is to follow the modern German aero engine practice of allowing the coolant direct access to the outside of the valve guide.

Valves will, of course, be operated directly, by the usual twin overhead camshafts. A possibility that occurs to me is that the thick stem of the modern sodium-lilled valve may have sufficient bearing area in the valve guide to take the side thrust if the cam operates directly on the hardened end of the valve. This would mean that the usual sliding tappet or swinging finger could be eliminated, so reducing reciprocating weight and cutting down the spring pressure required. A further artifice for reducing reciprocating weight is the use of hairpin valve springs so commonly found on high performance motor-cycle engines, and for which there is plenty of room on a twin-Cam engine. With these springs the only reciprocating weight is one-third of the top .arms of the hairpin, whereas with the normal coil spring one-third of the total weight is reciprocating. Due to this, surge is practically eliminated with hairpin springs. Induction System.—We have already determined the supercharge pressure required as 25 lb./sq. in., or possibly a little more to allow for contingencies. This pressure is rather more than can be effectively generated by a single-stage Roots blower, which becomes inefficient at over about 18 lb./sq. in. This leaves us with the alternatives of a single-stage vane-type compressor, which can easily handle pressures far greater than that required, or else a two-stage Roots blower. Just at present there appears to be a strong tendency towards the use of twostage Roots blowers yather than highpressure vane-type compressors. Quite why this should be is not readily apparent. and it would appear that the vane compressor is the more efficient alternative. It is possible that the two-stage Roots blowers give a better performance low down in the speed range (this is a characteristic of the single-stage Roots blown B.R.A.) and, of course, it is possible to adopt a measure of, intercooling at the intermediate pressure of the two-stage arrangement. On the other hand, it may be that efficient vane-type compressors are only manufactured in this country. Perhaps one of our more technical readers can give us a few words on the subject or, maybe, someone would care to raise the question at the next Motor Racing Brains Trust ? [The Zoller originated in Germany and the Cozette in France, of course.—En.]

Anyway, until I am given a good reason for adopting the alternative, I am suggesting the use of vane-type, one per block to keep them within manageable dimensions, mounted rather high up on the near side and delivering directly into the four adjacent ports, with the usual blow-off valves in the manifolds. As vane-type compressors have objections to ultra high speeds, they will be driven at 60-70 per cent. crankshaft speed, either from the timing gears or by chain from the front end of the power-take-off shaft, which will be directly underneath the blowers and will itself run at less than crankshaft speed (although I am not very keen on chain drives in high-speed engines).

The S.U. constant-vacuum carburetter has given excellent service on nearly all high-output engines produced in this country—so much so that a blown engine without an S.U. seems almost wrong— and there is every reason for specifying these instruments. One per blower would be used, of the type most convenient to fit in the available space.

Cooling.—For the cooling system, the use of glycol, or some other fluid having a higher, boiling point don water, in a pressure-tight system, will permit the engine to run at a rather more efficient temperature than with water, and also permits of a reduction in radiator dimensions and so cuts down head resistance. Directed cooling from internal ducts on to the plugs and exhaust valve seats will be employed, water pumps being mounted on the camshaft ends or on the outer ends of the blowers. The radiator would be situated low down in front of the front suspension assembly, maybe with a separate header tank. Ignition.—There is little to improve upon one or two magnetos of the Scintilla "Vertex "type, driven from the camshaft ends. When he designed the Delage, Lory had no alternative to the 18-mm. plug, which he fitted deeply masked. Nowadays 14-mm. and 12-mm. plugs are

usual practice, and there will be room in the head for these to be fitted without masking.

Lubrication.—An engine fitted throughout with anti-friction bearings is comparatively easily satisfied with quite a simple lubrication system. In this case the principal requirement is an adequate oil radiator to dissipate the heat collected from the piston crowns. A dry-sump system is desirable, I think. The pressure and scavenge pumps, of gear-type, can be located under the two centre main bearings, and their common shaft driven by yet another gear from the main driving gear on the crankshaft. A very high oil pressure is not required.

Clutch.—The conventional dry-plate clutch is perfectly satisfactory and there is no reason for departing from the usual design except that, with the power takeoff alongside the crankcase, diameter is somewhat restricted and the multi-plate variety may prove more compact.

Four-Wheel Drive.—As the transmission layout and, indeed, the chassis layout in general is largely influenced by the adoption of four-wheel drive, it is relevant to insert here a few words of explanation of the whys and wherefores of this system.

It is pretty clear that the next big advance in roadholding must come from the reduction of unsprung weight by the removal of the braking mechanism to the sprung portion of the chassis and connecting the wheels thereto by universallyjointed half-shafts. Having thus bridged the most difficult gap, so to speak, there seems very little reason for not going a little further and connecting all four wheels to the engine in view of the important advantages to be gained thereby. What are these advantages ? They can be demonstrated mathematically, but only approximately, owing to the impossibility of allowing for time lost in gear changes, for the fact that gear ratios are only correct at one speed, and for air resistance. However, taking the ideal case and assuming a 350-h.p. car weighing 1,800 lb., with a weight distribution of 40 per cent. front, 60 per cent. rear, then with rear-wheel drive the back wheels can be spun up to 122 m.p.h., and not until it reaches that speed can the full power of the engine be used for acceleration. With four-wheel drive, the wheels can be spun up to 73 m.p.h., and above

that speed all available engine power can be used. These two speeds are not affected by air resistance. Another way of explaining the advantages accruing from four-wheel drive is as follows : In the case of the rear-driven car mentioned above, acceleration up to 122 m.p.h. is limited by wheel adhesion to a rate of about 60 per cent. of g (the acceleration due to gravity), or some 13 m.p.h. per second. The four-wheel-driven

• car's acceleration is also limited by wheel adhesion up to 73 m.p.h., but, owing to all wheels being used for driving, the maximum acceleration will be about equal to g, or 22 m.p.h. per second. Between 73 m.p.h. and 122 m.p.h. the acceleration will be limited by the engine power and will therefore fall off. At 122 m.p.h. it will be equal to that of the rear-driven car, and thereafter acceleration of both will fall away equally. Thus the four-wheeldriven car has an advantage up to 122 m.p.h., the 'order of the advantage being such that about two seconds can be gained in attaining 70 m.p.h. from rest. These figures, although approximate, do give an indication of the advantages accruing from the use of four-wheel drive, particularly on a circuit with many slow corners connected by fast straights. What about the history of four-wheel drive for racing ? The first four-wheeldriven racing car was a Belgian Spyker, constructed in the Edwardian era. I have read its description in the Motor Car Journal of th0 date, but do not recall

anything of its subsequent history. It was evidently not such as to cause a clamour amongst rival manufacturers anxious to copy its design and reap the reward of its advantages.

Next on the scene, in 1932, I think, appears the four-wheel-drive Bugatti, built in an endeavour to obtain full benefit from the urge of the supercharged 4.9-litre engine. In this it was not over successful—after one or two appearances on the Continent it was brought over to Shelsley Walsh, where it crashed in the hands of Jean Bugatti. As far as I know, it was then carted back to Molsheim and probably decently interred—anyway, I do not think that it has been heard of since. In passing, it is of interest ias being at least one example of a Bugatti with independent suspension, the two front stub axles being supported by two transverse leaf springs.

From about the same time right up to the beginning of the war, four-wheel-drive cars were being made in America by Miller, who already had considerable experience with front drive, and in view of their continued manufacture they were presumably at least moderately successful, though, like most American cars, they are rather of the track-racing variety. However, full details regarding them would be of great interest. Possibly the most promising four-wheeldrive car to date made its appearance in, I think, 1935. R. Waddy's " Fuzzi," which embodied two 500-c.c. J.A.P. engines,

each driving one axle through motor-cycle transmissions. Like most amateur-built "specials," its rather short career was beset by mechanical troubles, but in one or two ascents of Shelsley Walsh it did prove that its constructor had the right idea.

Since then, of course, John Cobb's Railton, the present holder of the Land Speed Record, has been built, embodying the four-wheel-drive arrangement, but that is rather outside the scope of these notes. It has been suggested that a four-wheeldrive car will require some sort of torqueapportioning device so that the torque applied to the front wheels can be suitably reduced when conditions render it desirable—whether this device is to be automatic or controllable by the driver I am not sure. " Fuzzi " did, in fact, embody such a device in an arrangement whereby the proportionate opening and closing of the two throttles could be varied, and according to Waddy it was very useful. However, I still remain unconvinced that such a device is essential. In parenthesis, I am quite prepared to admit that this view may be influenced because I cannot see how this torque apportioner can be worked except hydraulically or electrically, and I am not too keen on having the efforts of my car's horses transmitted either by a stream of oil or of electrons I think that no racing driver would demand a little lever in the cockpit to switch off the braking effort on the front

wheels when he wants to do a spot of steering—on the Multi-Union he is lucky enough to get such a lever, but it is a refinement and is not used for the above purpose, anyway. The driver is given a fixed proportion of braking torque, front and rear, and if it does not please him, then he has to "lump " it, at any rate, until the next race. In the same way, I think that he should be given a fixed proportion for the driving torque, and if it does not suit him, then he must make the best of a bad job. My view is that satisfactory results with a four-wheeldrive car will come from the driver adapting his technique to the new transmission—it may even take a driver who has never driven anything else to show us the final polished result, just as it took Rosemeyer to show us how the unorthodox central-engined Auto-Union could perform. I must say one word in sympathy for the poor driver. Progress has taken away the "side brake" with which he was wont to control the "dreaded sideslip," and now it is proposed to deny him the aid of the engine in achieving the same end. . . .

Reverting to my previous remarks, I can see no reason why the torque proportion between the front and rear should be variable, but it is quite clear that the torque should not be equally divided between the two axles, partly because the rear axle will be more heavily laden and partly because of the rearward transfer of weight which occurs on acceleration. The proportion, I suggest, should be somewhere about 40 per cent. to the front axle and 60 per cent. to the rear. Fortunately, this fixed proportioning presents no difficulties, as it can be done by quite a simple adaptation of the common differential which is described later.

General Chassis Layout.—It is now possible to describe, very roughly, the general layout of the transmission and chassis. The engine is mounted in the normal position, offset slightly right, with the drive coming out on the lefthand side from the centre of the crankcase. The drive passes through the clutch to the upper shaft of the all-indirect-type gearbox mounted between the rear half of the crankcase and the left-hand frame member, with the upper driving shaft on or just above crankshaft level. Universally-jointed propeller shafts take the drive, front and rear, from the ends of the lower driven gearbox shaft to the framemounted final drive units, which are also offset left. On the right of these and connected to the differential cage are the brake units, so that braking torque is equally divided by the differential. Universally-jointed half shafts connect the differential pinion shafts to the road wheels. The driver sits behind the engine, offset right, and with the rear propellershaft on his left. It will be noted that I have fallen into the same trap as Lory with the 1925 Delage, in that I have placed the driver's feet in rather close proximity to the rear exhaust stub, but as in my case the feet can be dropped down to the level of the undertray, I am hoping to get away with it, without having to turn the cylinder block round. The alternative is to rearrange the components along the same lines as on the Auto-Union, with the engine behind the driver, but still retaining the gearbox Moon bases his calculations as to engine efficiency largely on the results the late R. J. B. Seaman obtained from the famous modified 1926 if-litre straight-eight Delage

alongside the engine and the transmission along the left-hand side.

Gearbox.—This has already been described as being of the all-indirect variety mounted alongside the engine. It consists merely of two parallel shafts, one driving and one driven, on which rotate on needle roller bearings five pairs of meshing gears, one for each forward ratio, and one pair of gears connected by an intermediate gear to provide reverse. Each gear train is then connected to the upper and lower shafts by dogs which slide on splines on the shafts to provide the different gear ratios. Each train of gears is so mounted as to be alongside bearings in either the end wall of the box or one of the two internal transverse partitions. By allowing both gears of each train to run free when not transmitting the drive, instead of only one, churning losses in the gearbox oil are reduced. Each train of gears as it is selected by the selector mechanism is lubricated by an oil jet delivered from a small gear-type pump on the rear end of the driving shaft. The driven shaft is hollow to permit the front axle driving shaft to pass from the centre differential, which is behind the gearbox, to the universal joint of the front propeller-shaft just in front of the gearbox.

Centre Differential.—This device is required to compensate for the extra distance covered by the front wheels when rounding a bend and also to act as a torque apportioner. This latter end is achieved by making the driven gears of differing diameters-in proportion to the amount of torque that they are required to transmit, instead of being equal in diameter, as in normal construction. With a bevel differential, this involves machining the differential cage so that the planet pinion pins are not at right angles to the main axis, and also involves machining bevel gears whose axes are not at right angles. This is not a complicated process, but is probably not possible with the gear-cutting machines normally available in an automobile factory. On the other hand, if a spurgear differential is used, all axis pins are parallel to the main axis and all gears have straightforward spur teeth, so that this is a much simpler manufacturing proposition. The differential is mounted at the rear end of the gearbox and is enclosed in an extension of the gearbox casing. The cage is driven from the hollow driven shaft on the back end of which it is mounted, and as has already been ex

plained, the drive passes from the forward driven gear through the hollow driven shaft of the gearbox to a universal joint in front of the gearbox casing, and the drive is taken from the rear driven pinion to the rear propeller-shaft universal joint just behind the differential casing.

If considered desirable, the centre differential can be fitted with the selflocking device described in connection with the axle differentials.

Final .Qrives.—The final drives are of normal straight bevel-and-pinion variety and call for no particular comment.

Obviously, some form of self-locking differential must figure in the axle assembly. The type so far almost exclusively used in racing cars is the German Z.F. mechanism, which operates on a cam principle. It is clear, though, from a recent series of 'articles by Raymond Mays, that this device is not entirely free from snags, particularly a tendency to lock itself rather suddenly when wheelspin does develop.

It appears that rather better results may be obtained by using the form of self-locking differential patented by L. M Ballarny. This device consists of a normal differential of spur or bevel type, with the addition of a multi-lobed earn mounted on either the inner end of a half shaft or an extension of the differential driven gear. This cam operates springloaded pistons in radial closed cylinders in a casing attached to the side of the different ial cage. The space in the ends of the cylinders is filled with oil admitted through a hi-directional valve arranged so that the oil enters easily, but is only released slowly. The result is that slow relative movement between one half shaft and the differential cage, such as occurs during normal 'cornering, is unchecked, but faster relative movement (i.e., wheelspin.) meets with rapidly increasing resistance.

The universally-jointed half shafts, at any rate at the front end, will need to embody some form of constant-velocity joint, of which there are now several types available, all tested by arduous war service. One point that requires to be borne in mind in connection with these joints is that whereas with the Hooke's joint the floating shaft is located through the joint from the fixed shaft, with a C.V. joint this is generally not the case, and the free shaft requires supporting radially, and in some eases axially as well by a spherical casing surrounding the joint. Brakes.—When the drastic step of mounting the brakes inboard has been

taken, it is clear that the brake mechanism will require considerable Modification, particularly in the method to be adopted in getting rid of the heat generated. To give some idea of the heat flow involved, the amount of energy converted into heat in stopping a 1,800 lb. car from 150 m.p.h. is about 970 Centigrade heat units, and if the car stops with a negative acceleration of half "g," which is well under maximum deceleration, this heat is produced in only 14 seconds. Some of this energy will be used in overcoming air resistance, but only a small part. What is to be done with the rest ? If it is passed directly to air moving over the brakes and the air is heated 10° C. in doing so, then 400 lb. or 5,100 Cu. ft. of air is required—quite a large amount to have passing through the body, and it will warm the cockpit unduly, too. An alternative is to use some form of liquid cooling, connecting it up with the engine cooling system. A third possibility is to use an evaporative cooling system. The amount of heat mentioned above can be absorbed in evaporating one pound of water, and the steam so formed could be condensed in quite a small condenser in the airstream or, maybe, in a surface type condenser iorming part of the body panelling. The only mechanical complica tion needed is a very small pump to force the water to where it is to be evaporated. Having reached the conclusion that either liquid or evaporative cooling is to be used, it remains to decide what form the braking mechanism will take, as obviously the conventional drum and expanding shoes are ill adapted to any thing but air cooling. Some form of Eddy brake in which energy is converted into heat by magnetically-induced internal electric currents, or a hydraulic tarrangement in which energy is converted into

heat by churning oil, may provide the eventual solution, but as I see it, the most practical solution at present is a disc-brake similar to those used on Capt. Eyston.'s "Thunderbolt." The brake linings, circular in form like a clutch lining, are carried on a light rotating member not unlike a clutch plate. The brakes are applied by clamping these plates between two stationary iron plates cast with internal cooling passages, to which the coolant is led through flexible pipes. The relatively small amount of heat that is conducted through the heat-resistant liners can be dissipated by spacing the liners apart

somewhat and inducing a radial air flow by means of suitably-shaped vanes between them. For the actual engaging and disengaging of the plates I do not think that

a hydraulic mechanism can be bettered. Any required degree of self-servo action can be provided by arranging for the pressure plates to move together along helical guides instead of axially.

One brake unit per axle, working on the differential cage would give automatic compensation between the wheels, limited by the self-locking action of the differential. This would also give a considerable measure of braking on one wheel even if one half shaft parted. Front Suspension.—For taking the thrusts caused by driving, braking and cornering with a minimum of whip, there is no system to improve upon the twin transverse wishbone arrangement for the independent front suspension. At the same time, accurate steering geometry can be arranged without undue complication, and if equal wishbones are used with the universal joints in the same planes as the hinges, the use of heavily-loaded splined couplings in the front half shafts can be avoided. In order to reduce the number of joints involved, the stub axles can be ball-jointed to the wishbones, so

that motions due to steering and to suspension are confined to only two joints on each side. Wishbones may be steel forgings or pressings or light-alloy forgings.

Rear Suspension.—The most satisfactory system of rear suspension for racing cars so far discovered is undoubtedly the De Dion-type axle. I think, however, that instead of the usual arrangement of longitudinal radius arms which set up a steering effect at the back end when one wheel rises independently of the other, more satisfactory results may be obtained by shackling the axle by a very wide shackle, to a wide-ended transverse wishbone-link on each side. This arrangement would, however, not be capable of taking brake torque reactions when a normal brake layout is used. Whichever system of fore-and-aft location for the axle is used, a means of transverse location is also required, and this can be arranged in many ways. Mercedes-Benz racing cars used a ball pin engaging in a vertical slot, and their tourers had two wishbones balljointed together at their apices for their tourers, while other alternatives are a transverse radius rod or a transverse system of rods giving a straight-line motion.

Springing.—Springing may be by hydraulo-pneumatic struts combining the elastic media with a damping system, like the oleo legs of an aircraft, but for convenience of installation on the chassis, torsion bars connected to the suspension wishbones, with large piston-type hydraulic shock-absorbers built into the mounting of the same link, would be hard to improve upon. Driver control of shockabsorber adjustment is desirable.

Steering.—Modern sports cars have demonstrated that with a well-laid-out steering geometry completely reversible steering mechanism can be used, with a consequent gain in. lightness, and it is probable that the same can apply to our racing car. Otherwise, any of the modern proprietary steering gears will give excellent results. Owing to the mass of mechanism forward, the steering box will have to be mounted behind the cylinder block over the driver's feet, motion being transmitted by a long push-and-pull rod. Wheels and Tyres.—If the tyre designer will permit it, there are several advantages to be gained from a reduction in wheel rim diameter. There is a reduction in unsprung weight of the wheel and tyre and a slight reduction in weight throughout the transmission, due to the lower torque transmitted at a slightly higher rotational speed. The reduction at the final drive can also be rather less. Probably most important of all is the reduction in wind resistance resulting from the smaller wheels. All this is rendered possible by taking the large diameter (and very hot) brake drum away from the wheel rim. On the other hand, the same

weight of tread rubber Must be provided, and centrifugal effects on the tread, for the same peripheral speed, increase as the radius is decreased.

Frame.—While, in theory, a stressed skin or geodetic form of construction for the combined body/chassis may be ideal, I think that by the time the points of attachment of the suspension and power unit have been suitably strengthened and the holes for the insertion of the driver and removal of the power unit have been reinforced, the saving in weight or rigidity over a rectangular frame welded up from large-diameter thin-walled tubing, will be quite negligible. Certainly the latter mode of construction offers considerable advantages in the way of accessibility.

Body.—I am not too sure whether the all-enveloping type of body enclosing the wheels is suitable for a road-racing car, though it is true to say that this car, with fluid-cooled enclosed brakes, is as suitable as any for this type of body, which will give very considerable advantages in maximum speed and acceleration in th.e higher speed ranges. Probably the enveloping body will be used on very fast circuits and an alternative with unenclosed wheels on circuits where maximum speed is limited.

When the latter form of body is used, I hope that the squat appearance of the 1939 1 Iand 3-litre Mercedes will be avoided, as from photographs these cars seem very ugly, although the new E.R.A. looks to be one of the handsomest cars yet built. Maybe this is just racial prej Lid ice.

As regards the actual construction of the body, a light-alloy, tubular framework with panels held on by pop rivets and aircraft-type fasteners will combine lightness with accessibility.

Fuel Tankage.—On the car with the enclosed wheel form of body there will be adequate fuel accommodation in tanks in the space between the wheels, and even in the alternative body, a considerable proportion of the fuel can be carried in tanks in fairings between the front and rear suspensions, in the scuttle and over the transmission on the driver's left so that the quantity in the tail tank, where weight distribution varies as the tank empties, is kept to a minimum.

Performance and Dimensions.—The engine and transmission layout that I have described could be accommodated in a chassis with a wheelbase and track of about 8 ft. 8 in. by 4 ft. 6 in. Height to the top of the scuttle would be about 32 in. and frontal area some 8.5-9 sq. ft. for the body with unenclnsed wheels and about 20 per cent. more with wheels enclosed. Weight one would anticipate to be about 1,500 lb. bare and 1,850 lb. in racing trim with driver and full tanks.

As regards performance, I think that without being unduly optimistic one could expect about 180 M.p.h. in the less streamlined form and about 30 m.p.h. more with the more aerodynamic body. Acceleration from rest to 100 m.p.h. might take about 7/ sec. with either form of body.

It is, of course, on the type of circuit with slow corners connected by fast straights that one would expect such a car to perform brilliantly.