The automotive inventions of Ettore Bugatti
Extracts from a Paper read before the Newcomen Society by H. G. Conway, M.A., M.I.Mech.E.
Ettore Bugatti was perhaps the most famous manufacturer of racing and sports cars that the automobile world has so far known. In the years 1924 to 1926 his cars won no less than 1,045 victories. As Pomeroy says, "Bugatti cars have won more road races than any other make, quite possibly more than all other makes added together." Not until 1932 did Alfa Romeo produce a serious competitor, and not until 1935 or 1936 was Bugatti eclipsed by the German Mercedes and Auto-Union cars. Bugatti car production ceased in 1939 and no more than sporadic attempts were made at restarting after the war.
This paper however is more concerned with the man as an engineer than his racing successes. He was born in Milan on September 15th, 1881. His father, Carlo, was an artist, painter, sculptor and furniture designer; his brother Rembrandt became a gifted sculptor, but although he too was sent to an art academy for training, he soon, at the age of 17, was allowed to turn engineer.
Bugatti's first car design was in 1899-1900; it had a four-cylinder overhead valve engine, four-speed gearbox and chain rear drive. It was produced by the Bugatti-Gulinelli Motor Corporation at Milan. Soon after (1902) Bugatti was engaged by the De Dietrich Company at Niederbronn to supervise their car designs and moved to Alsace, where he remained on and off for the rest of his life. Bugatti designed cars for De Dietrich, Hermes at Graffenstaden, and in 1907 for Deutz at Cologne. In 1909 he moved to Molsheim and established his own Bugatti Company.
Ettore Bugatti -- mad inventor or mechanical genius? Viewed front the point of view of his car designs and productions, he was certainly a mechanical genius although few engineers would doubt that he had a streak of eccentricity to stubbornness which made him persist with bad design features long after their faults had become obvious to his customers.
But viewed after a study of his inventions it is not so easy to form an opinion. Many of his patents are unbelievably odd, and impracticable, examined thirty years after they were filed. Considered in total and attempting to put them in perspective, they are fascinating.
A search in the Patent Office in London shows that there are 340 listed in French files (together with another 32 in the name of his son Jean) and that the number in Britain is 176 (25 being in the name of Jean and one Roland -- this one filed after the father's death). No time has yet been available to study all the French specifications which, although available in London, are not always readily accessible, but the British documents have all been viewed and a few of the others, and the notes that follow are the result.
The history of Bugatti's Patent Specifications begins with the filing on June 26th, 1904, in Germany of a patent relating to a yoke end for steering gear; it is possible that he filed previously in Italy but a search does not show this. His last case was filed in Belgium on January 14th, 1944 -- during the war -- and the same case in France a year later on January 12th, 1945, for a means of mounting a trueing diamond on a cylindrical grinding machine. Thus he had almost forty years' active experience as a patentee, which by any standards, and taking into account the number of cases filed, marks him out as exceptional if not unique.
By permission of the Controller of the Patent Office it has been possible to reproduce some of the diagrams from the more interesting of his car and engine specifications; the author feels able to assess the value of these, but not of the many relating to machine tools and rail-cars. To understand the documents it has to be appreciated that French Patent Law is much different from British or American practice (although the author is not qualified as an expert on this subject). The habit of French inventors is to publish ideas on, and have a patent granted for, subjects which would not be accepted as patentable elsewhere. The French patent office does not appear to have a novelty search as we do in Britain, where the very granting of a patent implies that the Patent Office thinks you may have something and have not been able to find to the contrary. Thus many of Bugatti's French specifications have not been accepted over here due to lack of novelty -- for example the idea of building up a roller-bearing crank and holding it together in alignment with bicycle-type cotter pins.
These differences must account for much of what has been seen in the specifications. In Britain you must have a novel idea and must describe and illustrate it clearly. In France any relatively wild idea can be filed, with the most rudimentary sketch of it, not in any way appropriate to illustrate the practical realisation. Another difference which may be important is that under French practice the Patron "invents" everything including the ideas of his staff. In Britain the true inventor must be declared, except on foreign or "convention" cases; Bugatti may not have thought of all the idea; he himself, or at any rate others, may have contributed. Although most of the patents relate to motor vehicles a wide variety of miscellaneous ideas has been noted in the review of the inventions:
Needleholders (1905), cylindrical razor (1906), a diesel shunting locomotive with two engines so it can go in two directions (1911); a lathe with twin slides and feedscrews to increase production (1917); a multi-spindle drilling machine (1922) -- no doubt signs of an increasing production; several other aids to productions, toolholders and the like about this period; boiler oil-firing system (1936); Bren-gun carrier army vehicle (1938); a wonderful fishing reel with a power drive to make shark and tunney fishing easier (1938); speed-boat hulls (1938); venetian blinds (1938); several patents covering an aircraft design (1938) -- all filed in Luxembourg (to avoid difficulties with nationalised aviation in France ?); a particularly ingenious indicating micrometer (1941); a telescopic bicycle fork (1943); sail winding gear (1944); and numerous machine tool inventions filed during the war.
There is a series of designs relating to the Bugatti rail-cars, starting with B.P. 394877 (1933) which is for a bogie construction. There are many others on bogies, sprung railway wheels, car construction, window frames, and even swivelling, reversible railway seats -- a delightful construction as shown in B. Patent 468110.
As far as the author can determine, Bugatti's filing of patents began with the German specification 158066 of June 26th, 1904, which illustrated a yoke fitting for a steering gear (Bugatti's address is Strasbourg). The elegant and simple means of assembling a split yoke on a taper is shown in Fig. 1. An identical diagram appears on P. 925 of the Z.V.D.I., 1908, describing the Deutz car; it is clear that this was a detail from this car although it must have been used also on the Hermes designs, judging by the date.
The next patent (German 167618 of May 7th, 1905) illustrates an internally expanding rear brake assembly and again an identical diagram is reproduced in the same description of the Deutz car.
Patent 170691 (Germany 1905) follows. This illustrates (Fig. 2) a four-cylinder engine with overhead valves, the camshaft being below the crank and operating the valves by means of long tie-rods. It has not been established definitely if this construction was used on the Hermes engine, but the idea of a low-mounted camshaft appears again later.
Passing over the next three German cases for needleholders (!) we find No. 203453 of March 22nd, 1907, dealing with a toggle clutch (Fig. 3). This design was definitely used on the Deutz and indeed became a standard Bugatti design on all his cars up to about 1930, and incidentally figuring in later patents. The elegant feature is the ease with which the clutch engagement force can be adjusted by setting the toggles just to align under action of external spring operating on the thrust collar. Centrifugal force on the long outer levers facilitates engagement until the toggle goes into alignment, and in practice the clutches are delightfully smooth. A fault with the design as illustrated is that the thrust collar should be able to float laterally to minimise the load in the two side levers (see later).
Patent 21159/1911. This is the first British Specification filed from Bugatti's address, Hardt-Muhle, Molsheim, Germany. It relates to a rotary disc type of valve with a bronze collar bearing on the disc to seal the port. This has been re-invented since for hydraulic devices and a well-known slide selector made for aircraft by the Dowty Company uses the identical mechanism. Note that the piston rings seal the outer diameter of the collar and pressure (if there!) in the manifold will force the collar onto the disc to seal it. However the scheme obviously would not have worked in practice on an engine, certainly not in 1911.
21160/1911.This describes a chassis with reversed 1/4-elliptic springs at each end and evidently covers the classic rear suspension used on all Bugatti's cars. He claimed that for a given axle motion the chassis motion is reduced by the outboard hanging of the springs, as compared with the A.C., or Frazer Nash system of outward pointing 1/4-elliptics. It is difficult to believe that he was right about this; it is more probable that the advantages of the reversed rear suspension system are that (a) the spring is in tension during traction; (b) axle motion is forward when the spring deflects, which would appear to give a slight degree of understeer. However, no matter what he claimed, the system was one he retained for some 35 years.
3430/1913.This shows a very cunning means of holding a wheel to a hub so that it could be detached readily. A wedge was used for locking firmly, screwed up by a spanner, and another screwed ring could be used for breaking the taper on removal. No doubt Bugatti preferred to use the Rudge hub, which must have been available about then, as no use of the invention is known.
7131/1913. This, an improvement on 21160/1911, shows a 1/2-elliptic spring with a 1/4-elliptic between the middle of the 1/2-elliptic and the axle. This could hardly be called a 3/4-elliptic spring but is really its equivalent if arranged differently.
19655/1913. This describes a magneto driven by the end of the crankshaft, the magneto in turn driving the water pump which was immersed in the radiator. This does not seem to have much virtue.
19656/1913. A means of making an axle end into a yoke (i.e., to carry a T-shaped "Elliott " sub axle) by splitting and forging the tube ends. A more interesting method was to follow a few years later.
19657/1913. This disclosed the idea of splitting longitudinally the main leaf of a 1/4-elliptic spring so that it could twist more easily, as it has to do. In fact, this is quite unnecessary.
20385/1913. This specification shows a weird engine design with the camshaft underneath the crankshaft (and well oiled !), operating overhead valves with very long tie-rods, and having centrifugal pump rotors to throw oil onto the main bearings of the crankshaft. It is probably a development of the earlier German case 170691 of 1905.
20386/1913. This shows a pair of transverse shock-absorbers similar to the system used on the front axle of the original Austin Seven, but with the starting handle passing through the centre of the shock-absorber discs.
20387/1913. This again relates to shock-absorber mountings but uses a resilient link between the shock-absorber arm and the axle. Later cars actually used this idea with leather links but the specification envisaged rubber to allow small axle motions without actual motion of the friction damper. This would not have been found effective.
20388/1913. This relates to a radius-rod layout for use with a transverse front spring, and was probably not used; the novelty lies in the means for allowing the ends of the radius-rods to twist when one wheel only lifts.
27047/1913. Now we see Bugatti becoming interested in roller-bearing crankshafts. A cage construction is described where the rollers are inserted through holes drilled in the side of the cage flanges and then sealed by screwed plugs. It would be a heavy construction, the only reason for the side flanges being to retain the cage on the connecting-rod end.
1357/1914. Bugatti continues to be interested in transverse springs, and describes one of the spring eyes mounted in a sliding spring-loaded, or a rubber, shackle in place of the normal hinged link as used by Ford. One end of the spring is pinned to the axle, the other having the new shackle.
1358/1914. This shows the British beginnings of the famous Bugatti clutch and describes the mounting of the centre shaft square driving section in line with the clutch axis to avoid offsets; the clutch itself has already been shown in the.German case 203453/1907.
2038/1914. This relates to the mounting of friction shock-absorbers, protected from dirt, inside the brake drum; they can be run lubricated with back axle oil.
5407/1914. A worm-and-nut steering gear. An interesting trick, probably not original to Bugatti but still not well known, is the splitting of the nut so that wear can be taken up.
6530/1914. Another (and what an) idea ! To avoid rattle of the camshaft gears, drive it by means of two gear trains of slightly different ratios and interpose a friction clutch between the two gears (Fig. 8.) This sounds all right if one is prepared to overlook the generation of heat and the waste of energy in the clutch.
8128/1914. More transverse springs with other small transverse springs at the ends to join them to the axle. As the Ford Model-T was well established by then, this invention does not seem very sensible.
10160/1914 and 11688/1914. These are listed but were never published. Probably the war and Bugatti's status as an enemy prevented completion, so they were allowed to lapse.
12523/1914. This specification was not accepted until 1921; it relates to hydraulically lifted valve gear and probably stemmed from 20385/1913. The saving of weight in the valve mechanism would be a factor in Bugatti's mind but he could not have appreciated the difficulty with shock waves and leakage which the construction would have entailed. One wonders, all the same, if he ever tried the idea.
10549/1915. This is a patent of addition to the previous one and was published in 1916 -- five years before the master patent. This seems strange! The end of the cylinder is a fixed plug attached to the bottom of the engine, so that the adjustable cylinder is in hydrostatic balance.
(We now note the change in the Patent Office numbering system, the numbers being numerical and not related to a particular year. Dates of publication will be given for interest, filing dates generally being some two years earlier. Bugatti is now in Paris.)
100266/1916 (Fig. II) describes an interlaced valve rocker system obviously suitable or engines with a single overhead camshaft. It was probably for an aero-engine.
100298, 101390/1916 and 105218. Cylinder construction with integral head and sheet metal water jacketing is described in the master patent with two improvements. It is possible that Bugatti had in mind a construction particularly suitable for the new aero-engine on which he was working at the time; he is pre-occupied with sealing the jacket to avoid brazing or welding as others had done. He uses an O-ring of modern type; this was patented in the U.S.A. and elsewhere in the 1930s and much royalty was paid to the inventor. Bugatti's diagram might have helped to dispute the novelty of the later invention but he spoils it by referring in the text to "a compressible band or fillet-capable of being squeezed when required by tightening ..." The essence of the O-ring is that it is elastic, not compressible (i.e., rubber) and that its pressure tightness comes from the groove dimensions rather than by being tightened up as with a gasket.
101534/1916. A simple and effective double crank layout (again for an aero-engine) where the cranks are phased at 90 deg. to allow closer spacing.
103115/1917. This is an improvement to 100266, with rollers on the rockers.
106795/1917. We now have the first specification for a machine tool -- many are to follow. It describes a lathe with two leadscrews and two saddles to increase production.
126672 (Filed 1916--published 1919). The first cooled valve: air or oil can be sucked, on the valve opening stroke, into the annular chamber around the stem, and expelled up the centre of the valve. Bugatti used a similar idea on his aero-engine and later on racing cars. In practice the passages became choked with burnt oil and contributed little.
131652 (Filed 1917 -- published 1919): This is a most intriguing specification, which was not granted. Bugatti, now at 86, Rue Chaptal, Levallois Perret, Seine, describes an aircraft cannon mounted to shoot through the propeller axis by off-setting the engine crankshaft from the propeller by a gear train. Bugatti filed on June 11th, 1917, and the specification, as printed, made reference (probably under pressure) to a certain Marc Birkigt who had filed a similar idea six months earlier on January 27th, 1917. He of course was the designer of the Hispano engine and the famous "moteur-cannon" of that make. Did Bugatti knowingly copy the idea? --surely not: probably knowing the need to fire through the propeller many clever engineers had the same idea.
141693/1920 (Filed April 1919). Back at Molsheim in Alsace, now incorporated into France, and claiming the mounting of accessories inside the dash to keep them clean, dry and warm. This was probably the genesis of the dashboard mounted magneto which persisted in most of his racing cars. Driven by the end of the camshaft, a magneto makes a good torsion damper but the design has little else to commend it.
146866/1920. An ingenious ball-operated clutch mechanism evidently for a machine tool but possibly intended for cars as well.
159482/1921. This describes the mounting of a dynamo or dynamotor around the shaft between gearbox and clutch, which seems interesting, and then spoils it by suggesting an alternative drive by bevel gears at this point. A belt drive was a better realisation.
159486/1921. The illustration (Fig. 16) is an exact drawing of the Bugatti clutch which was so successful for many years. Bugatti obviously had trouble with his claims as the specification admits two other similar designs and only claims the point that the yoke has lateral freedom to equalise the forces in the link members. This feature was in fact used in the post-war racing cars, the yoke being nicely centred as it slides along. This specification is of course very similar to the much earlier case 203453 of 1907.
159822/1921. The specification describes a large-diameter valve guide so arranged that when it is drawn out, the hole left is large enough to allow machining of the valve seat from the valve end. However, the valve itself was larger than the hole and had to be fitted front below. It is not clear why Bugatti patented this idea in 1921 as he had used it in 1907 on the Deutz engine.
160147/1921 (Filed March 20th, 1920). It is astonishing that Bugatti was allowed to file this specification. It illustrates his prewar and post-war valve gear as used since 1907 (Fig. 17). Indeed The Autocar of June 20th, 1917, had described the same gear in some detail. The essence of a patent is an absence of prior disclosure. One can only assume that the Patent Office-Examiner did not read The Autocar, or the Z.V.D.I. of 1908.
(To be continued).