Tuning the 4 1/2-Litre Lagonda and Invicta
By L. S. Michael, O.B.E. (Continued from the January issue)
In calculating compression-ratios use the method advocated in popular books on tuning, usually involving the pouring of liquid into the cylinder head from a burette, thus discovering the combustion space at t.d.c. Then, by calculation, arrive at the combustion space required for the new compression-ratio, measure that volume of liquid into the burette, pour it into the cylinder head and scribe a line at the liquid level, making due allowance for the meniscus; thus obtaining the level to machine to. This is to be checked several times, and anyone who has tried it will recall the pretty pattern on the cylinder head where he has scribed lines at the different levels after each re-check!
Most of these pistons are not flat topped nor absolutely flush with the block, therefore such operations cannot give useful information if done on the head alone. If it is attempted with the head in situ, on the assumption that the piston rings will prevent a downward escape of liquid, it will be found that owing to the disposition of the sparking plugs it is impossible to get accurate results.
I suggest a different approach. It is easy to calculate the compression space required for various ratios as the formula
CR = V + CS
gives this, where CR = compression ratio; V = swept volume; CS = compression space. A table showing the answers for the various compression ratios is given below. With a cylinder of 88.5-mm. bore the amount of machining required to reduce the volume by 1 c.c. is 0.0063 in. These cylinder heads do not reduce in bore until after the plug holes penetrate the head, so one is not involved in problems occasioned by the cylinder-head shape.
In order to do the calculations required, it is necessary to know the accurate swept volume of one cylinder. This, one would imagine, was easily determined by dividing the capacity of the engine by six. Unfortunately a large number of different figures have been published for this engine, 4,467 c.c. and 4,453 c.c. being the most popular. In all the Invicta publications, and the Lagonda Motor Show catalogue of October 1933 describing the 1934 models, the capacity is said to be 4,467 c.c. This agrees with the figures published by Henry Meadows Ltd., who made the engine; one might think that they should know ? Nevertheless, for the 1935 T.T. and on all subsequent Lagonda race entry forms, still available for inspection, the capacity is declared as 4,453 c.c., and from 1935 onwards this figure is repeated in Lagonda catalogues, while Meadows continue to use the figures 4,467 c.c. The bore and stroke given by both is 88.5 x 120.64 mm. and it must be emphasised that no other measurements were put out by the makers of the engine. Since Meadows have taken the trouble to give the stroke to two decimal places of millimetres it is reasonable to suppose that the bore is given with equal accuracy and that they represent the designed dimensions. The mathematically minded may care to work it out for themselves. The answer is 4,453 c.c. The weight of evidence seems to favour 4,453 c.c. and for the purpose of the following calculations this is accepted as correct. Even if one is working on an engine bored to the maximum safe oversize, the effect will not be enough to cause any trouble as far as the degree of machining contemplated here is concerned.
By far the majority of 4 1/2-litre Lagondas are fitted with 6.68-to-1 pistons. These were always fitted as standard after 1935, and were supplied as replacements even when earlier engines were reconditioned. Most Rapides had 7.0-to-1 pistons and these can be recognised as compared with the standard ones because they project distinctly higher above the block face. These 7.0-to-1 pistons were extremely difficult to obtain after 1947 and the chances are that any engine overhauled after that date had 6.68 pistons in the rebuild. Sometimes an individual had special pistons made, but this is usually known to the owner of the car concerned.
The table below sets out the amount of machining required to raise the compression of a 6.68-to-1 engine to any desired figure up to 8.5 to 1. If your engine is known to have a higher compression, a subtraction sum will show what is needed to alter it to any desired extent.
The early M45 Lagondas and most Invictas were fitted with 6.0-to-1 pistons. The 6.0-to.1 pistons are usually flat-topped and are distinctly below the top of the bore at t.d.c.
Most 6.68-to-1 pistons are slightly domed and the top of the dome is nearly flush with the top of the block at t.d.c. The ex-W.D. engines which were available immediately after the war had 6.0-to-1 pistons, and a large number were fitted to both Lagondes and Invictas. The ex-W.D. engine can be readily recognised because it has a wide flat sump instead of the narrow deeper sump of the standard engine. In case of doubt it is not difficult to withdraw a piston and compare the height above the gudgeon-pin with that of a standard 6.68-to-1 piston. A large number of those latter are about, and no doubt any of the garages specialising in Lagonda work would let one make this comparison, as indeed would the Lagonda Club spares registrar.
To bring an engine fitted with 6.0-to-1 pistons tip to 6.68-to-1 it is necessary to machine 0.097 in. off the block, then the table shown below can be applied.
Usually it is advisable to machine the block if more than 0.096 in. is to be taken off, because of the danger of running into the sparking plug holes. The block of course can be machined on the head and crankcase surfaces. Although there is no likelihood of the valves hitting the pistons I have not given figures above 8.5 to 1. The works never went higher than 8 to 1 in their racing engines, though they may have been influenced by the fuel laid down for the races in which they took part. Nevertheless, they did not choose to go higher for the 1936 500-Mile Race, for which any type of fuel was permitted.
Colin Campbell gives some interesting figures relating to engine speeds in the second edition of "The Sports Car," due to be published shortly. His suggested formula for safe continuous cruising speeds gives something over 4,200 r.p.m. for the Lagonda. When this was raised with him, he pointed out that his work related to modern engines, where at the present state of knowledge the behaviour of bearing material had become the critical factor. He considered that the con.-rod and torsional vibration of the crankshaft were the limiting factors in the Meadows engine. He also suggested that the big-end bearings tend to crack under compression loads when pulling hard at low speeds. Since he competed with an Invicta which he owned for some time, I have no doubt he was referring to that version of the engine. Such trouble will doubtless occur with the M45R and LG engines but they will stand higher stressing than the earlier ones, and have a much more robust con.-rod without separate bearing shells. Apropos of the foregoing it is a good idea to have thebearings cast in "racing metal" when such attention becomes due, for although this will slightly increase the rate of wear of the crankshaft it is not serious, and the bearings will stand higher compression better. Racing metal, being harder than ordinary white metal, will hasten the scoring of the crankshaft if used with dirty oil because abrasive paticles take longer to become buried in the bearing material. Therefore, slightly more frequent oil and, when fitted, filter changes are advisable.
It is worth recalling that in 1931 a special engine was built by Meadows, for Raymond Mays' famuous white lnvicta. According to Mays, alterations were made to the valve gear, valves and pistons, and special four-bolt con.-rods of greater strength and rigidity than standard were designed by Murray Jamieson and Peter Berthon. As the normal Meadows head was retained, there was no room for larger valves. No doubt superior material was used, and the rockers machined all over and lightened as much as possible and lighter cam-followers and push-rods used, as in the Lagonda Rapides of later date.
In spite of continuing to use the standard crankshaft a considerable increase in compression was risked (some put it as high as 10 to 1 but Mr. Crump is unable to confirm this; in fact, he thought it an exaggeration). The engine, on alcohol fuel (Shell M.C.3), ran for an hour on test at 3,900 r.p.m., giving 158 b.h.p. The unit was then stripped for examination, found to be satisfactory, and after reassembly was installed in the Invicta, where it gave long and trouble-free service in sprints, hill-climbs and Brooklands' races. The power produced is by far the highest recorded by one of these engines. Even so, it is to be noted at what relatively low engine revolutions this was realised.
Because many modern bread-and-butter cars are built to run with compression-ratios substantially over 8.0 to 1, it does not follow that it is safe to do so with a large pre-war power unit. In any case, the higher ratios are usually associated with small cylinders; those vehicles with a cylinder capacity approaching the 4 1/2's, at present, seldom go much above 8.0 to 1. This ratio in the Meadows engine makes the use of 100-octane fuel desirable, though provided the ignition control and gearbox are used intelligently, 50 per cent. 100-octane and 50 per cent. premium spirit gives only a little audible pinking.
I feel that 8.0 to 1 is going to the limit for most purposes, and durability may suffer if the full performance thus made available is exploited continuously. An increase in compression to 7.5 to1 gives a really worthwhile improvement which the owner will find most gratifying. This ratio was offered as an option on the pre-war production Rapides without affecting the guarantee, when fuels available were much more prone to give trouble. It has been used satisfactorily for many thousands of miles by quite a number of Lagondas on the road today, and at least three LG45s are in daily use with substantially higher compressions.
One of these cars did throw a con.-rod but that was attributed to the owner's practice of regularly exceeding 4,000 r.p.m. in the gears when in a hurry, in addition to which it had had a very hard competition life in rallies, sprints and races before the breakdown happened; this same car has recently completed a successful tour to Spain and back.
It is difficult to assess exactly what improvement in terms of b.h.p. is achieved by the various changes in compression-ratio. The brake-test figures available do not give this information clearly; it must be deduced from a mass of notes and correspondence. Furthermore, it is very rare for a compression-ratio change not to be accompanied by other modifications affecting power output.
It does appear that raising the compression from 6.68 to 1 to 7.0 to 1, combined with 6.5 deg. greater ignition advance and fuel to eliminate pinking, gave an extra 5 b.h.p., and although power fell above 3,400 r.p.m. in both cases, the higher-compression engine gave superior power right up to 3,800 r.p.m., at which crankshaft speed it exceeded the power of the 6.68-to-1 unit by 7 b.h.p., in both cases power being well below the peak which remained at 3,400.
In 1934 an engine (M45/271) was specially prepared for racing: polishing ports, etc., gave it a better performance than standard. During the course of tests its compression was raised from 7.0 to 1 to 7.25 to 1. This increased the power by 2 b.h.p., the engine peaking at 3,600 r.p.m. Thus it seems that an increase of compression-ratio from 6.68 to 7.25 to 1 will produce an extra 9 b.h.p., which is quite appreciable.
When it comes to greater increases in compression-ratio one is on less sure ground. Higher ratios, used with suitable fuel, must produce more power. It is exactly how much more that is difficult to establish. The only significant bench tests that can be identified without doubt relate to an LG45 engine in 1936. This engine had a compression-ratio of 7.96 to 1 and, in conjunction with 2-in. carburetters, polished head and ports, and a six-branch exhaust manifold, gave approximately 35 b.h.p. more than standard.
It is evident from the correspondence between the works and Fox & Nicholl in the period 1934 to 1937, that originally the compression-ratio of those engines was limited by the fuel, which produced severe pinking if high (by the standard of time) ratios were used. The works recommended the use of 30 per cent. Benzol with premium fuel, even for normal touring.
Today, 100-octane spirit will permit the use of very high ratios without pinking and mechanical considerations alone limit what can be used.
Some interesting calculations were published by E. G. Higham in The Automobile Engineer of August 1955. For a typical long-stroke engine with a bore and stroke ratios of 1.35 to 1 peaking at 4,000 r.p.m., the maximum bearing load is on the compression stroke at t.d.c., and this amounted to 3,379 lb./sq. in., assuming a compression-ratio of 7.5 to 1. These figures must be close to those arising in the Meadows engine, which has a bore/stroke ratio of 1.36 to 1. Calculations for many other engines were made. For comparison an engine peaking at 6,000 r.p.m. with a bore/stroke ratio of 1.25 to 1 had a maximum bearing load of 7,860 lb./sq. in. The XK150 engine has a bore/stroke ratio of 1.28 to 1, the figure for the Coventry-Climax FPF 1,100-c.c. unit were bore/stroke 0.88 to 1. maximum load 4,580 lb./sq. in. Thus, even allowing for the inferior bearing material and the date of the design, the loads in the Meadows. engine cannot be regarded as excessively high, being about 4,000 lb./sq. in. less than the Jaguar and 1,000 lb./sq. in. less than the Climax, which was considered one of the most lightly stressed modern high-performance designs, having a lower bearing load than all the other 6,000-r.p,m. engines considered by Higham. Nevertheless, when a Meadows engine is stripped down after a season's hard competition it is common to see signs of cracking of the bearing metal, unaccompanied by any suggestion of incipient seizure or oil starvation. This is more noticeable in the earlier engines with smaller rods, smaller crank pins, and separate brass bearing shells instead of the metal being run direct into the eye of the rod itself. Therefore, it must be accepted that since the bearing loads are not higher than is theoretically acceptable by the bearing material, the trouble is due to other factors. These are mainly: con.-rod design, lack of rigidity of the crankcase (a matter that always worried Fox & Nicholl, who repeatedly tried to get the design of this component modified), and torsional vibration of the crankshaft. The vibration damper limits the latter to some extent, and the stronger rods of the LG-type engines, together with their more massive main bearing caps, help to overcome the former trouble, but both factors become more critical as engine speed rises.
A conclusion can be drawn that, provided engine revs are kept below 3,800, a compression-ratio of 7.5 to 1 with modern fuel will not sensibly impair engine life, and will give a much better performance than standard. If the same engine speed limit is observed, and racing metal used for the bearings, 7.75 to 1 is equally satisfactory, while if reduced bearing life is acceptable 8.0 to I1 can be employed without disaster. It should be remembered that with the two higher compression-ratios these engines will readily exceed 3,800 r.p.m. in the gears, so attention should be paid to that point if a reasonably long engine life is desired.