The Inside Story of the Irving-Napier "Golden Arrow"
In April 1976 we published intimate information about the unsuccessful Sunbeam “Silver Bullet” Land Speed Record car and as we have also looked at the 200 m.p.h. twin-engined “1,000 h.p.” Sunbeam of 1927, about which a contemporary has also published much data, and some years ago Motor Sport gave a “mechanics’ eye view” of the 400 m.p.h. Railton Special, it now seems opportune to consider the 1929 Irving-Napier “Golden Arrow,” one of the most successful LSR cars of the pre-war period, which earned Sir Henry Segrave his Knighthood (Campbell wasn’t so honoured until 1931), a car about which some interesting information has come to hand. Apart from the ease with which the “Golden Arrow” recovered the LSR from America, it is remarkable that it cannot have run more than about 40 miles in its entire life. It can be seen today, looking exactly as it did at Daytona 52 years ago, in the National Motor Museum at Beaulieu. – W.B.
Segrave set a significant landmark in LSR history when he bravely raised this prestigious record to over 200 m.p.h. (actually 203 .79 m.p.h.) in 1927 with the 44-litre chain-drive Sunbeam, using the Daytona course for the first time by a British contestant. However, more foolhardy than merely brave, Ray Keech regained this record from Sunbeam’s in 1928 with a speed of 207.552 m.p.h., in the crude three-Liberty-engined 81-litre White-Triplex, becoming the first holder of a new International Trophy presented by Sir Charles Cheers Wakefield of CastroI-oil fame. Before that Malcolm Campbell had also raised Segrave’s record, to 206.956 m.p.h., likewise running at Daytona (speeds around the pace at which the American Space Shuttle glider touched down after its 1981 orbiting of the Earth).
Consequently, Segrave had to do something about this. He appointed Capt. J.S. Irving, who had left the Sunbeam Motor Company after completing his task for Louis Coatalen of planning the successful 200 m.p.h. Sunbeam, while Segrave no longer drove for the Sunbeam Company. Finance was the first requisite and Segrave got the backing of the two senior Directors of Portland (Red Triangle) Cement (of which Company he was himself a Director and Sales Executive), Castrol, Lord Rootes (for whom he had instituted the side-valve Hillman Fourteen Segrave fabric coupe that cost £398 and was guaranteed to do 60 m.p.h,), BP and KLG, with many well-known engineering firms being persuaded to contribute free labour and materials. That settled, work on the new LSR car commenced at the KLG factory at Putney Vale, close to Segrave’s house at Coombe Warren, Kingston Hill.
Irving was able to base his design of the new LSR car on his experience with the big twin-engined Sunbeam. He would have aimed for 274 m.p.h. had not Dunlop’s told him that they could not guarantee their tyres safe at over 250m.p.h. So Irving settled for a mean speed, under good beach (Daytona) and weather conditions, of around 240 m.p.h. To attain the advance needed over the Sunbeam’s theoretical top speed of 210m.p.h. it would be necessary either to have some 60% more power or to cut frontal-area by nearly 50%, assuming almost identical aerodynamic efficiency. It was necessary to use a British engine to humour the sponsors and Segrave, so the offer of a 1,250 h.p. (guaranteed) foreign engine (LSR expert Posthumus thinks a Fiat) had to be refused. The only alternative was the 1927 Type 7A 23.9-litre (not 26.9-litres as Posthumus quotes) Napier Lion aero-engine, as used, in 7B geared-form, by Flt.-Lt. Webster in the Supermarine S5 to win that year’s Schneider Trophy seaplane race at 281.91 m.p.h. Incidentally, it is interesting to note the higher speed possible in the air, in spite of the drag of the enormous floats, necessitated because this was a water-‘plane contest and also because no aerodrome then had sufficient space in which to take-off and land such very fast machines.
Irving obtained the latest racing engine made by Napier’s, capable of developing 938 b.h.p. at 3,300 r.p.m., but which, after exhaustive tests, was run on BP alcohol-mixture instead of on tetraethyl-leaded petrol, again probably because of BP sponsorship, although better low-speed acceleration, and cleaner sparking-plugs, were given as outweighing the 2.04% power-loss that this fuel change entailed, to 912 b.h.p. , which compared to 900 total of the twin Sunbeam engines. The Napier engine required a shorter water pump, the coolant outlet changed to the propshaft end, etc., to get it into the sleek shape of car Irving was planning. As Reid Railton, another great designer of LSR cars, said later: “We should consider ourselves very lucky in that the present-day racing aero-engine is so admirably shaped for fitting into a motor car”. Yet I do not think he moulded bodywork so closely to an engine as Irving did in the “Golden Arrow”, which he based on the nose cowling of the S5 racing seaplane, a snug fit round the three cylinder blocks of the 12-cylinder engine, thus obtaining a frontal-area of only 11.1 sq. ft., compared to that of 20.8 sq. ft. of the 203 m.p.h. Sunbeam. At first it was intended to model the cockpit round Segrave, but a steering wheel with a minimum diameter of 16″ was deemed necessary for proper control of the car, so this, plus hand-grip clearance, determined the actual width of the cockpit. Irving had thought about using “one or other of several types of control”, one supposes the lying prone scheme, or the driver encircled by a steering-ring, but this idea was quickly abandoned, probably after the fatal accident to the driver of a racing Panhard-Levassor equipped with the latter arrangement. A 1/16th-scale model of the “Golden Arrow” was wind-tunnel tested but the results were not available until nearly three months after construction of the car had commenced; fortunately they tallied very closely with the full-size design, although minor modifications were made to improve accessibility, comfort and safety, which slightly increased head resistance. Nevertheless, the final shape showed a drag of 708lb. at 250 m.p.h., which the integral tail-fin increased by 41Ib., whereas the comparable drag of the Sunbeam had been 860 lb. and had the apparently-contemplated tail-fin been fitted, this would have been increased by 118 lb.
To his sleek shape Irving had to incorporate “necessary evils” such as clutch, gearbox, transmission, axles and steering (not to mention four wheels!). Although reduced to the minimum, this caused the original seaplane outline to become largely submerged. Even so, aerodynamic efficiency was so good that a maximum of 275 m.p.h. seemed possible. Irving had used fairings that completely enclosed the space between front and rear wheels, and this alone represented nearly 6% reduction in head resistance. Putting additional fairings in front of the wheels gave no appreciable improvement, probably because of the drag of their struts, and as they were difficult to devise and obstructed accessibility, they were abandoned.
To attain Arrow-like stability, even when the “Golden Arrow” was deflected by undulations at over 200 m.p.h., caused Irving more trouble than any two other factors. In the end he opted for a big tailfin which located the centre of gravity an inch in front of the centre of pressure. It must be remembered that rear wheel drive was employed, so that aft weight was needed. It was thought that, if the yaw-angle could be kept above 10 deg., risks resulting from a gust of wind striking the car or a minor skid deflecting it could be disregarded, allowing that a 10 m.p.h. cross-wind at 200 m.p.h. could equal a yaw-angle of 3 deg. and a gust of wind one of 6 deg. The critical yaw-angle for the Sunbeam had been reckoned as eight deg. above the danger point at 200 m.p.h. and the “Golden Arrow” had a safety factor of 13 deg., when going 50 m.p.h. faster.
It is very interesting, in view of today’s “ground effect” thinking, that Irving superimposed his streamline shell on a horizontal plane 23’0″ long and 15″ deep expressly to exert a downward air pressure, apart from this structure enclosing the transmission. It was intended to exert a down-thrust of 138 lb./sq. in. on the front axle and 410 lb./sq. on the rear axle at 250 m.p.h., Irving using the greater thrust at the rear to ensure that if the “Golden Arrow” left the ground the back end would land first, thus obviating a nose down return that would cause the car to “turn turtle”. He was worried about an end-over-end crash, obviously, but not with the car going over backwards, as Donald Campbell’s boat did. He noted that aerodynamic lift was probably 100% more vital than the mere reduction of weight for this purpose – one feels that Irving would fully appreciate what designers of winged and skirted F1 cars are about. .. .
Returning to the higher speeds possible in the air, it was calculated that whereas 462 b.h.p. (or 585 b.h.p. after safety factors had been incorporated) was required to overcome the Irving-Napier’s head resistance, another 165 b.h.p. was needed to overcome mechanical losses and a further 150 b.h.p. to cope with tractive resistance. Irving was hoping for up to 60 lb./ton tractive resistance at Daytona beach but, in fact, when Segrave went for the record, the poor conditions must have made this nearer 80 to 100 lb./ton, over most of the course. The weight on the rear axle (ready to run) was 4,149 lb. and no differential was used. Irving reckoned that the car’s weight distribution meant that while it was not possible to transmit full engine torque under any conditions in 1st gear, it was in 2nd gear if both wheels stayed on the ground, and possible under any conditions in top gear, which was thought sufficient to ensure safe control; to make certain, 260 lb. of lead ballast in the tail added to the aerodynamic download. The low weight of the Napier engine (850 lb.) was actually a disadvantage as a heavier unit would have got the c.-of-g. further forward.
The clutch used to get the “Golden Arrow” going caused Irving some anxiety. With the twin-engined Sunbeam a very high spring pressure had been required to prevent slip, and to enable Segrave to cope manually with this, considerable pedal travel had to be provided, which “give” in the linkage and wear on the clutch plates further increased . So for the “Golden Arrow” Irving wanted absolute minimum pedal travel (shades of the Austin 7!) allowing for clutch wear, and easy operation. He designed a multi-plate clutch with five driving and four driven plates, of 12″ dia., all lined with Ferodo, held together by 18 coil springs which exerted a total pressure of 2,340 lb. with the clutch in and 2,880 lb. when it was held out. A two-to-one leverage reduced the load on the withdrawal race to 1,440 lb. maximum. Pedal travel was only two inches and the strong toggle-levers and links eliminated any perceptible “give”. Full foot pressure was required to put the clutch out with the engine not running but a Dewandre vacuum-servo close to the withdrawal race reduced the effort so much that the pedal could be depressed with one hand, when suction was up. (Irving remarked that he believed that this servo system had since been adopted for a Continental firm’s 40 h.p, car. One wonders which make. He forgot, or had never heard of, Parry Thomas’ pioneer use of vacuum-servo on the 1920 Leyland Eight). In spite of the care devoted to the “Golden Arrow’s” clutch, the unit pressure on the friction surfaces was almost 70 lb./sq. in., or six times that recommended for clutches engaged at 3,00.0 r.p.m. But, of course, the Irving-Napier wouldn’t have much use, especially in traffic! Nevertheless, Parry Thomas had some trouble with overheating, I believe, with the small multi-plate clutch he used on “Babs”.
Variations between alignment of the clutch and gearbox were allowed for by connection via dogs, regarded as “very essential”, to ensure smooth clutch action – one might have thought Segrave was going in for a sprint from a standing-start, as in some ways he was, as he would need to accelerate hard into the measured mile, even with a four-mile run in. Irving designed a compact three-speed and reverse gearbox incorporating the gearing-up for the car’s twin propeller shafts. The ratios were 1:1, 1.54:1 and 3:1, giving speeds of 78 m.p.h., 156 1/2 m.p.h. and 241 m.p .h. at 3,400 r.p.m., the engine then having 200 r.p.m. in hand. The casing was of L.8 alloy, with aluminium-bronze end-plates reinforced with steel plates. The gears were of Vickers VNCG oil-hardened steel, hand-polished after hardening, the shafts of Vickers CHNC steel. Two parallel prop.-shafts conveyed the drive to the back axle, geared-up 49:33 , so that they turned at 5,040 r.p.m. at 3,400 engine-r.p.m. Both shafts turned inwards, to balance torque. A gear-type pump fed oil through nozzles into the teeth of the gearbox.
The prop.-shafts were considered safe up to 6,000 r.p.m. but were enclosed nevertheless in strong casings; a precaution omitted, I believe, on Cobb’s Brooklands Napier-Railton as first contrived, for its single prop.-shaft. The makers of the “Golden Arrow’s” were Hardy-Spicer. The back axle weighed 67.0 lb. with brakes and radius-arms but without wheels. It had two bevel-cases of L.8 alloy, the centre and end pieces being of Vickers VACS steel forgings, the shafts of Vibrac, and the bevel gears and pinions of BND. Worm and hypoid gearings were discarded, in favour of straight bevel gears.
The chassis frame was very stiff, the 19′ 6″ side-members deflecting less than 3/8″ under total load, a fine tribute to the makers, John Thompsons Motor Pressings. Channel-section and tubular cross-members were used. The front axle was a three-piece affair, of Vickers VACS steel forgings. It carried the front brake gear (the “Golden Arrow” was notable as an LSR car with 4WB) and the castor angle was 4 1/2 deg. The swivel pins were supported on fibre friction pads, each with an area of 7 sq. in., to damp the steering. The complete axle, less wheels, weighed 33.0 lb. Marles steering was used, geared 2 3/4 turns, lock-to-lock (28 deg.).
The stopping of the “Golden Arrow” was yet another problem. The big Sunbeam melted its aluminium brake shoes when going for the 5-km. record, which reduced the stopping area. If Segrave had had no brakes and engine to stop him after his record run in the Irving-Napier it would have run four miles in 124 sec. Using the brakes Irving gave him, 31 sec. and a mile would be required to come to rest, dlssipating 635 h.p. and producing 13,500 BThUs, sufficient.to raise 10 gallons of water from 70 deg. to boiling point in the process and bringing the brake drum temperature to some 2,000 deg.F. Ferodo-lined steel shoes were used for the “Golden Arrow”, in 18″ ribbed drums, the brake back-plates having eight 3″-dia. holes, for additional cooling. These brakes were adjusted once before the car was first run and never touched again. But although Trulay cables gave some degree of compensation, the action must have seemed odd to Segrave; the pedal went down about 3/4″ and then felt solid, although increased pressure on it, while still not moving it, did improve retardation.
The car was sprung on half-elliptic leaf springs all round, axle travel being 1 1/4″ front, 1 3/4″ rear. The rear springs were like two 1/4-elliptics joined at the axle. So stiff was the springing that one hundredweight caused a deflection of only 0.5 mm. Dunlop having made excellent wheels and tyres for the big Sunbeam, they naturally supplied those for Irving’s car. Ace wheel discs reduced rotational drag. The cotton-ply 37 x 7 tyres with 1/16″ treads and inflated to 125 lb./sq. in. were of very special construction, and heat/time charts were drawn up for their use. But as Irving said “No inflated tyre of rubber and cotton will ever be proof against some hidden sharp stone or hard seashell cutting it, and every would-be recordbreaker must needs take his life in his hands.”
So the car took shape, the body panels supplied by Thrupp & Maberly, who were more used to making coachwork for the carriage-trade, the gold-hued panels attached to strong hoops with countersunk screws. A 20-gallon fuel tank was installed, and Irving suppressed a desire to have a cockpit cover, which would have given a useful speed increase. Standard-size Ransome & Marles bearings were used throughout, as Irving remembered how much time and money had been wasted over Coatalen’s insistence on special-sized ball-races for the 203 m.p.h. Sunbeam. Built-in screw-jacks, in streamline fairings, were fitted, to enable the wheels to be rapidly changed; they functioned more quickly and easily than a touring car’s jack.
The “Golden Arrow” was duly shipped to America, arriving on Feb. 9th, 1929. Six mechanics, five of whom can be seen on the tow-truck in our photograph, and a Hillman Segrave coupe went with it – it would be interesting to know how many of these Hillmans were bought, perhaps out of respect for the famous British driver.
The “Golden Arrow” had quickly come to fruition, Capt. J.S. Irving being helped by his Chief Draughtsman W.U. Snell, his brother whom the Alvis Company had loaned to supervise the building of the car, and his daughter who dealt with all the clerical, accountancy and progress-chasing work. The “Golden Arrow” had been estimated at £10,000, plus £2,000 of free supplies. The final cost was £10,059 15s. 4d., with £1,500 supply of free parts. The radiators, slung between the wheels, had come out at £604 instead of an estimated £300, and wheel-balancing an astonishing £43 5s. It then cost about £2,000 to pack and ship the car to Daytona. Incidentally, the engine cost £3,078 10s. and design charges were £2,228 4s. 7d. The finished car was 27′ 6 1/2″ long, with a wheelbase of 14′ and it was 43″ high. It weighed a total of 7,814 lb.
The engine had been tested by Napier’s but neither it nor the car were so much as filled with fluids until the car’s arrival in America. On Feb. 26th 1929 this tedious experience was gone through and Segrave had one part-throttle run up and down the beach, finding the bulkhead so effective that at 182 m.p.h. he didn’t need to wear goggles. He drove off the beach, up two planks, and from the Promenade through the main street of Daytona to the garage! There followed a most trying wait for the weather, until March 11th, before the record bid could be made. After the first run new wheels and tyres were fitted within the requisite time allowance and the engine restarted using a trolley-mounted auxiliary engine. It is amusing that although the lorry carrying the trolley started off at the same time as Segrave, it needed nearly quarter of an hour to cover the distance the “Golden Arrow” had done in three minutes!
Segrave had found that with the 10 to 1 compression ratio it was not possible to use all the throttle opening until past 2,400 r.p.m. and some care on the accelerator was thus needed until the car was moving at over 55 m.p.h. in bottom gear, but at least this obviated possible tyre damage due to wheelspin. On the other hand, no adjustments of any sort had to be made, and at all times the Napier engine started immediately and ran without a mis-fire throughout. Irving had arranged for ice-cube cooling, with admission via RP thermostats, to supplement the 20 gallons of water in the side surface radiators if required. In the event, the ice cooling wasn’t needed, the temperature at the end of the first record-run being almost too cool, at 74 deg. C. The hot water was emptied and replaced by cold after each run.
Segrave took the World’s f.s. mile record at 231.362 m.p.h. (mean speed), and the kilometre record at 231.446 m.p.h., on just the two runs. -Never has a LSR been more easily achieved, in spite of the Triplex’s speed having been bettered by nearly 25 m.p.h. As I have said, it is unlikely that the ” Golden Arrow” has run more than 40 miles in its life, if that. (On his test run Segrave may not have needed the full 9 miles course.)
Irving had allowed for the absence of testing before the actual record-bid was made by asking that certain components, the gearbox, steering-box, brakes and transmission, be assembled with clearances to be expected after, say, 5,000 miles’ running, achieved where necessary by initial grinding to such clearances. The problem was that the British workmen regarded these clearances as a “sloppy fit” and Irving had personally to inspect every part to ensure they had been kept to! Incidentally, the use of side radiators partly fairing-in the wheels is interesting, because one theory suggests that Parry Thomas may have been killed by packed Pendine sand building up in the wheel fairings to a dangerous degree – but perhaps the Daytona sand was softer.
The car returned to this country after its fine performance and was put on show, after which it went into storage. In 1958 Wakefield Castrol presented it to the Montagu Motor Museum – the driver being commemorated on the car’s show-plaque as “Seagrave” – and this so very successful and little-used LSR car can still be seen in the National Motor Museum at Beaulieu.
I once suggested that Lord Montagu should use the “Golden Arrow” to attempt a new British speed record, or to establish an English one (at Southport sands?). I am sure he has the necessary skill and would not have shirked such an endeavour, but I was told that the Napier engine had never been properly inhibited and that its cylinder blocks are therefore now too porous for it to be run again. A pity! Unless, of course, someone with a Napier “Lion” motor to lend, and more persuasive than I am, tries again . . ..
Even if it was not written into his contract with Portland Cement, I believe the Directors thereof disliked Segrave going motor-racing, which is may be why he turned to trying for the Water Speed Record although he had resigned his Directorship before his fatal attempt on the record. One remembers Gilbert Frankau, the novelist, meeting Segrave at the send-off party for the “Golden Arrow” at Rootes’ Devonshire House showrooms in London and saying to him “I wouldn’t drive that car for all the money in the world”. To which de Hane Segrave replied ” It’s the boat that puts the breeze up me” . In 1930 Segrave lost his life when “Miss England II” broke up after setting a new Record . . .
As a sequel to this story, a year to the day of Segrave taking the LSR, Capt. Irving delivered his paper on the “Golden Arrow”, from which much of the foregoing has been culled, to the IAE, at the King’s Head Hotel in Coventry, and later in Wolverhampton and Manchester. The intention at Daytona had been to run again and try for 240 or more m.p.h. It was only the fatal accident to Lee Bible in the White-Triplex a few days after Segrave’s record had been set, that caused the second attempt to be abandoned, because the beach was closed. – W.B.