Arab Valve Gear

A Reid Railton Cameo

When I wrote briefly of Brian Demaus’ rare 1926 Arab sports car in the December issue I was in error over it having a single cam to operate both the inlet and exhaust valves in each cylinder, as Ken Fantom, who made a new camshaft for Demaus, points out. It is a cardinal sin for any historian to jump to conclusions and that is what I did in this instance. Knowing that the Arab had valve-gear almost identical to that pioneered by the great automobile engineer and racing motorist J.G. Parry Thomas for his Leyland Eight and other cars, in which a single cam served both inlet and exhaust valves, I concluded that the same must be the case beneath the valve cover of Reid Railton’s later Arab engine. Not so!

The Parry Thomas valve-gear was a classic of neat and ingenious design, at a time when overhead-camshaft production engines were few in number. Not only was the admittedly short-lived luxury Leyland Eight of 1920 the first British production straight-eight-cylinder car but it was one of the first in this category to have hemispherical combustion chambers, the tulip-headed overhead valves inclined to this end.

The whole layout was clever in the extreme. The camshaft ran along the cylinder head in bearing blocks that also served as bearings for the rocker spindles, one spindle for each pair of rockers and the leaf valve springs, which were a notable feature of all Thomas’ engines, ran across the head below the camshaft, one spring closing both inlet and exhaust valves. Maybe Thomas felt the clatter from eight cams was better than from 16…

The main reason why the famous Welsh engineer used leaf valve springs was probably to obviate the chance of loss of tension or even eventual breakage from which the coil springs of the time were apt to suffer if they became excessively hot, as they were apt to do in overhead-valve engines. But there were additional advantages. Leaf springs no doubt exerted more consistent friction damping than coil springs, thus helping to quell valve float, and as Thomas had them pivoted about their central anchorage-points on the cylinder head, the small amount of movement they were permitted helped to pull each valve more firmly into its seat as the other valve opened. More than that, shorter valve stems were possible than would have been the case had coil springs had to surround these, which saved weight, and with the widely-inclined valve angle, short valve stems meant that, more importantly, the rockers could be considerably reduced in length, and therefore weight also. This enabled Thomas to use quite substantial rockers without exceeding the desirable weight norm, although these girder-like rockers were drilled to further reduce weight. The ends of the valve stems protruded through slots in the rocker extremities and were retained by cross collets. (I have one such leaf valve-spring in my possession, a much appreciated gift from a MOTOR SPORT reader. It measures 6.3 in long by 0.8 in wide at the leaves and has twelve leaves, naturally of varying lengths, the shortest only 1.1 in long, but the spring size would clearly have been different for Thomas engines that varied from 1 ½-litres to 8-litres.)

The speed at which the big engine of the Leyland Eight revolved was hardly such that much valve bounce or thrash would have been expected, but Thomas may have been anxious to obviate any trace of clatter in a power unit which he intended to compete, on his own admission, with that of the ghost-quiet 40/50 hp Rolls-Royce. Or was he already looking ahead to the racing potential of his new engine?

I have never understood how he obtained the power he did from his later racing Leyland Thomases with the “soft” valve-timing the solitary cam for both valves must surely have produced, sufficient though this may have been for “town-carriage” purposes. Cost considerations very obviously had no bearing on any part of the Leyland Eight’s design, so we can only conclude that Thomas was satisfied with a total of eight cams on his eight-cylinder engine, and it has to be remembered that even so the power output seems to have exceeded 20 bhp per litre in standard form and approaching 30 bhp per litre in racing form, at maximum crankshaft speeds of 2,800 to 3,000 rpm.

My mistake in describing the valve-gear of the Arab caused me to ponder on whether Parry Thomas, like Reid Railton whose “baby” the Arab was, had departed from the single-cam configuration on any of his other engines. I decided to do as much research as seems possible at this late date. Fortunately for me, the BL Motor Industry Heritage Trust had the engine of their late-model Leyland Eight dismantled at the time, and their managing director Peter Mitchell kindly agreed to have his engineers look at it for me, and Faud Majzub also happened to have the 1.8-litre Hooker Thomas engine of his Scriven Special “No No Nanette” down, and kindly got his son Julian to photograph its valve arrangements.

From this evidence, although the position relating to the Leyland-Thomas racing engines is still surmise, it seems definite that Parry Thomas used virtually identical valve-gear for Leyland Eight, Hooker-Thomas, and 1 ½-litre racing Thomas-Special engines, that is to say, a single cam serving each inlet and exhaust valve, with a common leaf spring per cylinder beneath the girder rockers. Only in the method of driving the single oh-camshaft were there differences. The Leyland and Hooker engines had Thomas’ ingenuous eccentric-drive for the camshaft, no double used initially because he needed a quiet means of operating the oh-camshaft in what was to be a silent super-luxury car, which … would have been impossible with a gear train or a vertical-shaft drive, and perhaps not even a chain-drive would have met the Welshman’s high standards. It was unusual that he put this drive at the back of the engine, contrary to normal practice. The Leyland Eight had a triple-eccentric mechanism (copied five years later by W. O. Bentley for his 6½ litre Big Six Bentley and retained for the 8-litre Bentley). This hushed form of camshaft drive appears to have been adequate for the racing versions of the Leyland Eight and the racing Bentley engines, but even the latter did not run at much above 3,000 to 3,500 rpm, although Walter Hassan, OBE, got 3,640 rpm from the Barnato-Hassan special, with no problems in this department, although the original H-section con-rods used to break.

When it came to the Hooker-Thomas four-cylinder engine, I believe Thomas used a camshaft drive with two eccentrics, again at the rear of the engine (I have heard that this was sometimes apt to turn the camshaft the wrong way, but maybe only after wear had occurred!) For his racing 1½-litre straight-eight Thomas Special engine, which had a stroke of only 88 mm (compared to that of the 8-litre Leyland Eight’s 146 mm stroke and a stroke of 120 mm for the 1.8-litre Hooker-Thomas engine). Thomas used a train of gears to drive the oh-camshaft, possibly because there would not have been space for the eccentric rods, and anyway noise suppression was of no moment in a racing power-unit, nor, perhaps, would this form of drive been suitable for an engine intended to run up to 6,500 rpm.
Be this as it may, Walter Hassan, who knows about these things, tells me that in his opinion there would be no difficulty about running an eccentric-drive at his speeds, as all the motion is circular, with no real reciprocating movement. To digress slightly, Parry Thomas overcame any difficulties over varying heat expansion between crankcase and head in his eccentric-drive by pivoting the top eccentrics to one side and using an Oldham coupling to provide sufficient movement. There was also a means of micro-adjustment of the timing, through a bolted-up coupling, on the Leyland Eight engine. Walter Hassan thinks that, with his vertical-pillar to retain the true centres, Parry Thomas’ solution was probably a more purely engineering design than W. O. Bentley’s eccentric camshaft-drive, but he reminds me that “W.O.” knew all about eccentrics, from his railway-locomotive apprenticeship. Incidentally, whereas Thomas used drilled, H-section eccentric rods, W. O. Bentley preferred tubular eccentric rods, in three pairs, and again the drive was at the back of the engine.

“W.O.” foresaw the problem of heat expansion, which he tried to solve by incorporating helical springs above and below the upper bearings of his eccentric rods, possible as the rods pulled under load, but never “pushed”. After these springs had broken occasionally on production Bentleys, he replaced the springs with packings formed from shim-washers of pen-steel compression because the air gaps filled up with oil, thus providing a degree of damping. The six-cylinder Bentley camshaft was long and of relatively small diameter and it used to suffer from some torsional vibration, until a Lanchester-type damper was fitted to the rear end of it.

Parry Thomas used a very small-diameter iron coupling between eccentrics and the camshaft of the Leyland Eight, no bigger than that used to drive the water pump on the Bentley Six engines. He presumably “got away” with this due to an eight-cylinder engine having a smooth valve-train. (The Leyland Eight was the first British production straight-eight). The only other car engine I can recall with eccentric oh-camshaft drive was the 2-litre twin-cam Maudslay, of which only two were made; its plate-type eccentrics had no provision for combating heat expansion, the designer, J. A. Kemp, maintaining that the shape of the driving member, and the temperature-stabilising effect of the oil mist it encountered, were sufficient. I am much indebted to Walter Hassan for a lot of the foregoing information. Incidentally, Thomas was intending eventually, it is thought, to use desmodromic valve-closing on his 1½-litre straight-8 racing engines, and I understand there was provision for this, as built.

For the Arab engine, Railton put the oh-camshaft drive conventionally at the front and used a rather elaborate form of single roller-chain, with which he drove water and oil pump, and dynamo and magneto, incorporating an ingenious means for adjusting chain tension by means of an external lever, at all events in the earlier Arab engines. With a stroke of 127 mm, it would have been possible to have used the Thomas eccentric camshaft-drive, but perhaps this was regarded as too costly, or there may have been patents covering it.

However, it is the change in the rocker gear to which this article is really directed. Presumably to give his Arab engine the benefit of valve overlap, which was hardly possible with the Thomas single-cam arrangement, the Arab engine had normal rockers, ie, one for each valve, operated by two cams per cylinder. This was neatly achieved without much alteration to the Thomas layout by off-setting the rockers so that each was engaged by its own cam. This caused abandonment of girder rockers for conventional slender ones, and single rocker-spindle housing and shorter spindles were used, replacing two housing per spindle, a legacy, I believe, from the 1½-litre straight-8 Thomas engine.

Incidentally, Ken Fantom, who made a new camshaft for Mr Demaus’ Arab after it was seen that the profiles of the cams of the old one had worn completely away, points out that the small oil holes drilled in each cam of the Arab engine and intended to lubricate directly these and the end of the rockers they bore on, would no doubt soon have become blocked by sludge, contributing to the early demise of the profiles, although it is appreciated that the desire was to lubricate the ohc gear sparingly, to obviate too much oil getting down the valve guides.
The main difference, between the Thomas and Arab engines, however, was in the use of two instead of one cam for each pair of valves, to give the Arab the benefit of overlap. Another feather, albeit it only a small one it would seem, in the cap of that truly talented engineer, Reid A. Railton.