Chapter 22 - The Terrible Twins

I have never liked the air cooled vertical twin, but they looked neat, were simple, had an even firing, smooth exhaust note, fitting neatly into the same frame as their predecessors, the singles. The vibration was terrible, it seemed worse than the single, at least it was more noticeable. To my mind the head joint on all vertical twins with one-piece heads and cylinder block was marginal and subject to distortion. At BSA we must have done a fantastic amount of work on twins, judging by my notes. BSA had made up their minds that the twin was the motor cycle engine of the future. Every time I rode one it gave me the willies - the vibration was so noticeable. We tried all sorts of cures to get rid of the shakes and I still think our best ever effort in that direction was a short stroke 500 we made in 1953. Charlie Salt & I ran and ran it until we wore it out but no one seemed interested in it. This short stroke engine was a 500 A7 using a 650 head, barrel and pistons slightly modified. The crank was machined from the solid. This arrangement gave a 70 x 64.5 bore and stroke and instead of having a bolt on flywheel it had a triangular bob weight machined in the centre of the crank and discs next to the journal bearings. It was a very smooth running engine, both on the dyno and on the road. Power output was similar to the Star Twin, using 7.25:1 compression ratio.

With further tuning and lighter valve gear it gave 36bhp at 7000 a specially developed two into one exhaust system, we got as much as 39 horsepower at 7000. It was a most exciting machine to ride, as you accelerated it went on a normal power curve, then suddenly the exhaust note would change and it would 'yowl' right on up to maximum about 107mph Back in 1953 on low octane petrol this was quite an exciting performance on the road.  Charlie and I enjoyed that engine.   

The so-called Star Twin camshaft was a sporty one designed I think by Jack Amott although the design office never gave him credit for it. It was quite useable on the road with silencers fitted and also useful with open pipes. The fellows in my shop told me Jack did all the work on cams and I have no reason to doubt their words. Later on we went to the Daytona camshaft which was really a race cam and not very good at low revs but good at high speeds.  Amott did design a racing cam for the twin, but we were unable to make it work properly, it was too radical and hard on valve springs. Reg Wilkes mentioned that if the drawing office had stuck to Jack’s original design they would have been all right. I asked to see Amott's original design and Reg got his camshaft out of a cabinet and showed it to me.  It had a much larger base circle than the drawing office version, although to be quite honest I could not see what effect this would have.  Wilkes suggested I try it sometime, he was sure it would work better and it did. It was still pretty radical. I never could understand why the drawing office decided to change the base circle diameter. One of the disadvantages of a larger base circle is a higher rubbing speed on the other hand you can get a smoother easier ???  

Twin camshaft problems.
Whilst working at BSA Dennis Lashmar carried on racing my old 'Pike BSA' with the alloy twin engine. I think he had one good win on a wet day at Snetterton, when it ran cool enough not to blow up. Latterly whenever we went to Silverstone or anywhere to watch him, he finished up sliding along on his backside at about 100mph with a broken crankshaft.  This was too dangerous I considered. Once just after the massed start at Silverstone, with the pack all bunched up and approaching Woodcote corner at over 100mph the Beesa suddenly seized and he was sliding along with bikes all round him! I told Mr Hopwood that if we could not do something about the crankshaft breakages we should drop the racing twin. He pointed out that as far as Lashmar was concerned it was his own bike and he could do what he liked with it, but the policy of the factory was to make a twin. I felt we should make a new crank for it to which he agreed and wanted to know if I had any suggestions. By this time Group Research were in the picture at BSA and they demonstrated to me in ten minutes what I had suspected for years. They came up with some very good ideas. One was that the crank needed larger crankpin journals .which would make it stiffer and get away from the frequency at which it now vibrated. Mr Hopwood agreed that their suggestions were fine for 1955 but at that time we had to use what we had.  He was interested to know if any other suggestions and Group Research said to put a rolled radius around the ends of each crankpin. They demonstrated by putting a standard crankshaft on vee blocks over the Back to top of the electro-magnetic vibrator they had built.  

Then started some sort of motor generator. The noise was like standing next to a jet engine at the airport, it went into a scream and from a scream to an outer pitch sound, onto another phase even higher, then they brought in the electro-magnetic vibrator and the whole place began to buzz.  The frequency of the vibrator was adjusted to a multiple of 6,600, the crank vibrated with a high speed buzz, inside ten minutes there was a loud bang, the crank fell in two parts, broken at the usual point, showing the same sort of fracture that we had experienced after three hours running at 6600 RPM. It was a much quicker way of testing a crankshaft and without wrecking an engine. Next they set up one of their special cranks, with the rolled filet radius on the crankpin, it was subjected to the same test, buzzing away on the vee blocks, ten minutes passed, thirty minutes and it was still in one piece. I got tired of waiting and asked to be informed when it broke. It did not break and after one hour it was still good. I was very impressed and so was Mr Hopwood but the factory again did not seem interested, perhaps they could not believe it.   This rolling process consisted of applying a ball ended tool to the radius under high pressure in a big lathe.  

A few cranks were made, some for research, some for use on the dyno. It seemed fantastic that this simple process could make such a difference. Many years later when I was working for Volkwagen in the USA. I discovered that they cured a rash of broken crankshafts on the 1965 truck engine by the same procedure, except that they did not use a ball, but a small roller to form the radius. It cured the problem for Volkswagen. 

Twin engines for Formula 3
We made a number of A7 race engines for this small race class in 1953. They were set up to run on alcohol fuel but were not very reliable and although we got close on 50bhp were repeatedly beaten by Norton singles doing far less. We actually 'borrowed' a 500 Norton engine designed for a formula 3 car, it pulled 42bhp but had 39 ft lbs of torgue against our 36 ft lbs. Part of the trouble with our engines was the installation, they did not get enough cooling air and in one case they had a restricted oil supply, using a long small bore hose from the tank to the engine, the sticky Castor oil simply would not flow fast enough. Another problem was at the head to cylinder block joint, there was a tendency when very hot for a slight blow across to the oil drain holes thus pressurising the crankcase, this blew oil out of the crankcase at every possible point. The cure was to fit small hollow dowels into the head and barrel joint, thus rigidly locating the head, and keeping gas pressure out of the crankcase.

The single overhead camshaft A7 In 1952 they were working on a single overhead camshaft 500 twin. It was a handsome engine with alloy head and barrel, fine pitch finning, exhaust pipe held to the head by nuts. The single overhead cam was driven by bevel gears and a shaft running up what would normally be the push rod tunnel, the valves were operated by rockers. Arthur Bridgewood was working on it in great secrecy in a little shop at the end of the test shop, he had to put up with the noise and fumes of engines being tested. At this point in time it gave the same power as the A7 push rod engine and broke crankshafts with equal regularity. The camshaft was lubricated by the rocker feed which was completely inadequate and of course it wore out cams and nd rockers. Doug Hele suggested using a single wide cam as they did on some Ariels. This cured the rapid wear but the performance suffered due to the geometry of the rockers to cams, one rocker being a trailing rocker, the other a leading one this giving different opening diagram for each valve.  I personally felt that the engine could have been made to go if I had been given more freedom as I had with the Gold Star and MC4. One problem was excessive oil consumption due to an accumulation of oil in the rocker box which ran down the valve guides. This we cured by using a C11 oil scavenger pump to pump surplus oil direct to the tank. The feed side of the little oil pump was used to squirt oil through 1/16th holes on to the cam lobes, this cured the rocker and cam wear. At this point we needed a stronger crankshaft and a different design of cams to take advantage off the overhead cam arrangement. By this time Mr Hopwood had lost interest in the engine and said he did not think it would ever be any good, so scrap it. I think that’s what finally happened to it. 

The A10 was if anything worse than the A7 when it came to vibration and crankshaft breakages because of its longer stroke.  Actually most complaints were about the A10.  Although the A7 was the one we raced so that we were more intimately involved in its development. BSA brought about their own demise by their attitude to progress. A favourite saying of Mr Leakes was "Don’t let us be pioneers". Back in the dim and distant past they had lost money on a few experimental ideas that had not worked and that had never been forgotten.   Nevertheless when you are the biggest motor cycle manufacturer s in the world you have got to do some pioneering if you want to stay on Back to top. You cannot leave it to small firms like Velocette, who introduced foot gear-change, a successful spring frame, eccentric rocker spindles, a production overhead camshaft engine and many other innovations.

Alloy heads for the twins
About the beginning of 1954 it was decided to make a production version of the twin with an alloy head, we had made experimental ones before for the racing people. The head had shrunk in valve seats which gave very little trouble but the spark plugs fitted into bronze inserts screwed and pinned into the head, these persisted in coming out when the plug was removed. I persuaded Mr Hopwood to use long reach plugs screwed directly into the aluminium head, this was very satisfactory. At one stage we could not make a decision on which was best, a single inlet port and one carburettor or twin ports and two carburettors, on the dyno there was very little in it. The first of the Star twins had twin port heads with detachable manifold, it could be used with either one or two carburettors. In point of fact the difference in power is only worth while when running on open pipes. Twin carbs and their cables have to be carefully adjusted to keep both cylinders pulling equally at small throttle openings and just off idling, which is a nuisance to the rider at best, assuming he can cope with the bit of DIY mechanical skill necessary.   

After a year we had improved the port configuration and it ran so much better with one big carb that the twin carb option was dropped. We had found it useful to measure the capacity of the inlet ports and check on performance, about 142 to 150cc's gave optimum results. If an engine was down for power we often found the ports undersize.  Like all aluminium cylinder heads, these expanded a lot with heads and at the outset we experienced stretched or broken head bolts. These bolts went downwards through the head into the iron cylinder block and would usually break at the root of the last thread, which was the weakest point because it took all the stretching. To overcome this we quite simply put the bolts in a lathe and reducing the diameter of the pIain portion to 10% less than the diameter of the root of the thread. This meant that the thread was no longer the weakest point and that the plain portion could stretch without exceeding its elastic limit.

This was completely successful on the first attempt and no more trouble was experienced with the bolts. They could stretch when the head expanded and return to their original length as the head cooled own. When the new twins were going into production however, Alan Jones who was Works Manager at that time phoned me to say they were unable to make the head bolt as needed, despite my pointing out that we bad found them necessary, he just continued to say they were unable to make them. I went on with my work and forgot about the matter, but I did not have to wait long, within a few hours the motor cycle test shop foreman was on the phone to me complaining the head bolts of the new twins were breaking right and left, so I referred him to Alan Jones.  Mr Jones reiterated they were unable to make the bolts we had designed. Prior to this last call I had taken the precaution of calling the drawing office to say that Jones would not follow their drawings of the bolt thus securing an ally. Alan Jones got no sympathy when he had to pull all the bikes concerned back, dismantle the engines and use the bolt we had specified.