The early frames were made in the frame shop on a fairly simple jig and welded. They had to do a bit of setting and straightening at first to allow for the distortion, but once this was overcome and production jigs built they had few problems. The only serious one I can remember, was the twin front down tubes breaking at the engine attachment lugs which were welded on to the tubes, this was overcome by using a longer scarfed lug slipped over the tube and welded. As well as road testing the frame was mounted on a special test rig in the experimental shop with eccentric rollers under the wheels, these were driven by a powerful electric motor which gave the frame a good workout. The frames completed their tests on this torture machine alright  but we had reports of daily breakages on the Belgium pave at MIRA by the experimental shop testers. It was beginning to become worrying to the design dept as the go ahead for production could not be given until these breakages were cured. I remember Mr Hopwood was particularly worried at this time, as apart from the little Bantam, welded frames were new to BSA Mr Hopwood sent for me and pointed out I had had a lot of experience welding frames, and what did I think the problem was.  

All I could suggest was that we found out the way they were testing them. He decided he would go to MIRA and see for himself. That night I was working later than usual, I heard a bike pull up outside the experimental dept and went out to see who it was, there were two young apprentices, one with a sidecar outfit which was carrying a new spring frame, on enquiring where they had been they said MIRA. testing the latest new spring frame and were quite pleased with themselves that they had broken another. I was not so pleased with their results and told them so, but they said it was what they had set out to do. On enquiring how it handled before, they admitted it was fine, and that it handled better then the old frame, on which if you had gone as fast you would have killed yourself!

 Next day I met Hopwood and told him of my previous nights conversation, and he agreed I was right, he had been out that morning to MIRA. and had found them tearing around on the pave. He wondered what the best plan would be, so I suggested that they sent one of the old style spring frames and one new and make the apprentices ride them together, the old plunger frame setting the pace. Hopwood thought this a good idea and said I was a bloody genius. Next day they did this, young Povey wound up in hospital, he was riding the plunger frame bike, which broke when they started speeding he fell off and the other lad hit him. That settled the question of the better frame and production went ahead. The only criticism I had of the new frame was flexing when braking hard, the front mudguard used to hit the frame. We converted a standard frame to a single oval and tapered down tube, it definitely put a stop to the flexing, but the head angle was not quite right as we did not have the benefit of any sort of jig. Handling was not quite as good as the standard frame. 

Low temperature bronze welding was used by Bill Nicholson and myself on our one off frames especially when using the aircraft quality high tensile chrome-moly steel tubing which is not suitable for normal fusion welding, it causes embrittlement. We used nickel bronze welding rod  which was expensive and rather difficult to use, but which gave a very strong welded joint. Tests at the National Physics laboratory tore the steel tube before it broke the weld.  

Ignition & Carburation.
Our method of getting the optimum ignition timing was purely practical, we deliberately timed the engine on assembly about three degrees over advanced from normal, we then ran the engine on the dyno wide open throttle, full load at about 6000 RPM, then slowly retarded the ignition lever until the dyno showed maximum torque, then continue retarding the ignition timing until power started to drop off, then advance it back to a point where it gave maximum power. Stopping the engine we checked the timing on a degree disc and recorded it for that particular engine, then set the timing at this point with ignition lever at full advance. This method of finding the best ignition timing can be used on the road, the mountain mile in the Isle of Man was a good place to do this.

I am reminded of a mysterious malady that affected all our Daytona twin engines in 1955. Cyril Halliburn who was out there that year said they were in terrible trouble with points burning on the Lucas racing magnetos. Lucas were consulted and suggested fitting an extra condenser externally, connected in parallel with the internal condenser, utilising the contact breaker earthing brush terminal, normally used as a "kill" button. Everyone was of the opinion it was condenser trouble as the points rapidly became blackened and then misfiring set in. They could not do better than 7th and 9th place. Back at the works we set up a similar engine on the dyno and got Ken Norton from Lucas Experimental to come round and help with our tests. It did not take long to show results, oil in the contact breaker which in turn caused the burnt points. We stripped the magneto expecting to see oil coming through from the engine but on examination it was found to be clean and dry. After running the engine again with a manometer hooked up to the timing case we found that at 6000 RPM we had a pressure 4" to 7" of water below atmosphere in the timing case. The magneto had a single lip seal to keep oil out of the magneto but the low pressure was enough to lift the lip seal causing a drop in pressure in the magneto.

These racing magnetos had a breather pipe on the contact breaker cover which for some reason was bent in a loop ending right opposite the hole for the crankcase breather, so of course the faint oil mist coming from the breather was sucked into the contact breaker. All that was necessary was to bend the breather pipe on the magneto away from the vicinity of the crankcase breather hole. Since then I have experienced similar problems with Volkswagens with high mileage worn engines blowing oil up the distributor shaft and on to the points. On the TT & GP. carburettors the so-called air slide in the side of the carburettors admitted air to the air correction jet, it was only effective from fully closed to about 1/3rd open when the area of the air slide equalled the air correction jet, this range was equivalent to about three sizes of main jet. I think a tapered needle would have given more sensitive results and suggested this to CRB Smith of Amal but he was not keen to try it although he agreed that it would probably work.  

One of the strangest occurrences to do with carburettors was when one day Reg Wilkes sent an apprentice from the test shop into our, main shop for a 1 3/32 GP carburettor and he misread the size and gave Reg: a 1 3/16th carburettor and they immediately got quite a jump in power. They reported this to me and I went to the shop and the test was repeated and there was no doubt it was quite a gain in power. I suggested to Reg to open the port to match, immediately we lost what we had gained plus a bit more, so we made a thin sleeve and pressed it in and the power was back. We repeated this on other engines always with the same results. A number of private owners of DBD Gold Star noticed the carb being bigger than the port and opened up the port thinking to gain power, but not having a dyno were not aware of the results. Later on we fitted a venturi behind the carb and picked up even more power. The smallest diameter of the venturi could be 80% of area of the carburettor. We tested this on several engines and it always worked and seemed to improve carburation. The venturi had to have the classic included angles of 22 degree in and 7 degree out.

We did some testing with SU constant vacuum carburettors on the Ariel 500 twin and on the 650 Gold Flash. At this time Triumph Thunderbird 650 was equipped as standard with an SU carburettor. We got good results on test but not so superior to the Amal that it did not seem to warrant a change. Also in the carburettor field we did some test runs on the Wal Phillips injector carburettor, not conclusive and in my opinion only suitable for grass track or cinder track racing. We were also convinced of the necessity of having racing carburettor flexibility mounted and insulated from engine vibration. We had a striking illustration of this one when Reg Wilkes was running one of our 500 Gold Stars, he called me into the test house to observe this engine and pointed out that it exhibited all the symptoms of weakness, fitting bigger main jets did not cure it even very big jets. So we ran the engine up at about 5000 rpm, the exhaust pip e glowed red and the exhaust note sounded flat and harsh. I thought I could see some movement of the flexibly mounted carburettor, so I grabbed hold of it, I experienced a strong tingling sensation in my hand, but held on and a remarkable change came over the engine. The note became crisp and the red glowing pipe literally went out and the mixture became rich. The float bowl on the dyno was always remote and insulated, but on this particular engine the carburettor was mounted on rubber canvas hose, to cure the problem we had to stretch rubber bands from the carburettor to the bed of the dyne etc in several directions thus damping out the high frequency movements of the carb. 

Connecting Rods and BSA. 
I am not an expert on forging steel but I know the BSA smithy had a good reputation in Birmingham going back many years. They did a lot of work for outside firms as well as for the various divisions of BSA The Australian riders have told me that they Lave used BSA connecting rods in Manx Nortons with great success and a considerable saving in cost. I have been of the opinion that one of the secrets of BSA connecting rod forgings was in the tumbling they received. The forgings left the smithy all rough and covered in scale, a whole batch would be tumbled in a large wire basket, tipped from end to end over and over for a day or so and by the time they came out they were clean and descaled, almost polished. I believe the residual stresses in them from the forging operation were neutralised just as if they had been heat treated. The stresses had been literally knocked out of them and they were all ready to machine.

There was a learned article in the S.A.E. journal I had read describing how General Motors or someone had found that by shot peening connecting rods they became more reliable, with far less chance of surface cracks spreading and causing failure. I drew Mr Hopwoods attention to this article and he agreed that we should try it, by having three new connecting rods shot peened with the appropriate lead shot. (Mr Lupton laid this on and it was not long before I had three of the current Gold Star rods ready for test. After installing in an engine a series of power curves were taken trying out some other parts not connected with rods at this time. The first one broke after quite a short run so it was decided to go no further and the remaining rods were scrapped. 

And we concluded that shot peening was not successful for us. Sometime later I wondered whether those three rods had been tumbled, on making enquiries months later no one was sure whether they had been or not, but thought probably not Anyway one wrecked engine was enough. When a rod lets go it usually does quite a bit of damage.  Gold Star connecting rods were always polished as a final precaution against surface cracks, it could be that shot peening in place of the relatively expensive polishing might be equally effective, but it was never explored. The length of the connecting rod in relation to piston stroke has an important bearing on performance. For many years a ratio of stroke to -con rod length of 2: 1 was thought to be a good compromise. Some exhaustive testing along these lines was done and we found that for the standard 88 mm stroke 1.86:1 seemed best all round ratio. However, when it came to ultra short stroke engines a ratio of 2.2:1 became desirable to reduce piston acceleration at around TDC. 

Pistons and Piston Rings.
When I was first at BSA in 1952 I was able to borrow the 1949 500 Gold Star engine that I had used in the 1951 TT. It was of particular interest as I had on the last lap three times passed Dennis Lashmar who was riding a brand new featherbed Norton belonging to Harold Daniel, the same Norton that I later took on the Continent and finished 2nd at Hockenheim against some pretty good opposition. So this engine was no slouch and was pulling 6600 RPM. with the highest Back to top gear we could use, approx 125mph This old 1949 Gold Star had been specially built for Ray Hallett of Canterbury by Jack Amott, it was the old style non detachable rocker box, small fin engine. It had 8.1:1 ratio,  1 3/16" RN Carb , ??? inlet cam, Z exhaust,  4.375:1 Back to top gear and pulled 6200 on Sulby straight. When we got it on the dyno we got 37.5bhp which was what Nortons claimed for the long stroke Manx engine of 1951. We stripped the engine to find out what it was like internally, it seemed pretty good except for a cracked piston boss. So we replaced the piston and rings with the latest type, with narrow rings etc. On testing on the dyno with the new parts it gave 33bhp so we ran it a bit more and fiddled with jets and timing with no appreciable improvement, so we replaced the new style piston with one of the earlier type and immediately got over 36bhp. On examination we found the piston rings to be a little wider and the second ring had a groove machined in the inside diameter.

We contacted Brico and asked them why and was told it caused the ring to twist and thereby sealing better. To make a difference of over 3bhp it must have sealed a lot better. We then put a new 1952 500 Gold Star on the dyno it did not do as well as the old 49 job. The valve springs surged and it blew oil out of the breather and gave less than 35bhp. I spoke to Mr Hopwood about it but he did not seem to worry saying the 500 Gold Star was obsolete anyway and not to worry ourselves about it. The twin is the thing now. So I did not get any encouragement upstairs. We were however having so much trouble with the twins at that time that I felt we should do something with the 500 at least along the same lines as we had the 350. 

For 1952 the 350 Gold Star had a new head with a little more cooling fins than the previous years, more downdraft to the inlet port and the inlet valve axis was changed, moving the valves apart a little which allowed a larger inlet valve. Mr Hopwood at this time was interested in giving the scramblers bike more torque at low speeds, so we had some 350 heads cast with smaller ports and used smaller valves as well. One of these was prepared and fitted to a factory scrambler machine, they reported that they could not tell any difference, so I took another of these castings and fitted it with l/16” smaller exhaust seat and smaller valve and at the same time made the inlet valve bigger, the results were very encouraging, we got more power and even more important perhaps a cooler running exhaust valve. The production Gold Star at this time used Jessups G2 for exhaust valves and every now and then one would break and wreck the upper part of the engine.

We had to go to a lot of trouble with these valves getting the shape and finish just right making sure the valve guides were not slack. A different type of split cotter was used and a quietening ramp added to the cams in an endeavour to let the valve down on its seat a little more gently. We also reduced the valve spring seated pressure, but still they would strip looking very hot, but with the 1/16" smaller exhaust valve after some hard test running imagine our joy when we stripped the head and found that the exhaust valve looked in better condition and it appeared to be running cooler. The bigger inlet valve did not lose any torque, in fact it was better.

High Camshafts.
After the success of the MOV Velocette and Vincent HRD using these high camshafts, people have wondered why we did net try such an arrangement, in effect we did, the ultra short stroke 350 82 x 66 was very short overall, the push rods being about half the length of the normal B31 push rods. I would never have advocated making a high camshaft BSA Gold Star, if you are going to go to the trouble of redesigning, why not make the camshaft really 'high' and wake an overhead cam engine. A modern design could probably use a cogged belt, as used in many cars. People often associate overhead cams with high revving race engines, this is not necessarily so. Overhead cams will allow quicker lift and closing to valves which means more conservative timing and lighter valve springs (This in turn means better flexibility and quietness.

What about a Four? 
Why didn't we build a four cylinder, similar to the Italian MV 's and Gilera and later on the Japanese? In the first place the Italian fours were almost purely racers financed by other parts of the same company. The Japanese four cylinders were raced it is true but their production fours were only possible due to the conditions in Japan in the 60's and continued successfully by huge production figures using modern production methods. The British motor cycle was what people describe as designed around a centre lathe and at the end of the war we had a lot of worn machinery. In an endeavour to be less dependent on the tool room and model room I asked for and received two Colchester centre lathes and a Cincinatti miller, they all needed overhauling but once this was done were quite useful and I added own small 5" centre Little John lathe. On requesting a vertical Miller we got a veteran WW1, you could move the vertical shaft nearly 1/16" in any direction. We also managed to obtain an old surface grinder which was useful. I hoped this wasn’t typical of their reserve tools. The three cylinder BSA Triumph was typical of this attitude to production and must have been an expensive engine to produce. 

Water Cooling.  Not much was thought about water cooling until Tony Vandervell got Joe Craig to make a water cooled Manx Norton with the idea of using four of them in a row to make the Vanwall race car. When we heard about the gain in horsepower and its ability to maintain full power, I think we were shaken up a bit, but we loved all those cooling fins, why worry with anti-freeze or leaky radiators. However, with the aid of hindsight I think water-cooling might have been the answer to many of our problems with the vertical twins and also the MC 1.

Cast Iron Crankshafts.
Harry Taft of Idoson made us or had made six crankshafts for A10 twin in 'nodular' or spherical graphite iron. We tested two of these to destruction, running at 6600 for 6 hours, there was nothing special about these cranks, they were made as close as possible to the standard steel crank and balanced to the same factor. After the first one broke the other was stripped for examination and then passed to the experimental dept. for testing on the road. I don't remember if I ever got a report on that one. With six hours at that high speed I thought that was pretty good especially since it did not have the milled fillet radii. Mr Hopwood however was not impressed, I guess 6 hours at 6600 did not seem much, so I ran a standard steel crank at the same speed, it broke at just around two hours, this of course changed his opinion.

The factory said they were not practical as they would have had to have 100% X-ray inspection. I do not believe this would have been necessary as they did not X-ray other castings. In my opinion a spot check from time to time would have sufficed. One unexpected bonus with the iron cranks was an increase in power, apparently the iron made a better bearing surface, it was also easier to cast the metal for balance purposes just where it was wanted. 

Nodular iron is as strong as steel, Buick and Pontiac in the U.S.A. made connecting rods from this material. The name nodular iron comes from the formation of graphite molecules in the iron being round, whereas in plain cast iron the graphite molecules are flat, this makes a difference to the strength of the material. I designed a crankshaft assembly in this material for the Gold Star and got the research people to give an opinion on its feasibility, but it made no difference the factory apparently did not want to save money. 

Herbert Hopwood and other personalities at BSA.
  Herbert Hopwood came from Nortons to BSA. Prior to that he had worked several years at Triumphs. At BSA he was first a draughtsman but very quickly was promoted to chief designer. At Nortons he had designed the Norton vertical twin, in my opinion the best of the many twins. He had imtimated to Mr Leake,  Managing Director at BSA that he wanted a free hand in design, also a development shop which they had never had before. Previously whatever development had been done was in a corner of the engine test shop and by the experimental department, he got his way and had the competition shop made separate from the experimental. Mr Hopwood redesigned the A7 500 twin followed by the A10 650, both of which were badly needed and were an immediate success with the public. The original post-war A7 twin was a very different design and not too dependable. It had a very novel and expensive built up crankshaft designed I believe by Herbert Perkins. In those early days Mr Hopwood was busy turning out new designs, the MC1 250 racer and the MC4 which was half of an A7. The war department 500 side salve twin, the 500 overhead camshaft twin and a new four speed gearbox known as the 'A' box. 

The best BSA's in the immediate post war years were probably the B31 and B33. The M31 and M 33 were the same engine in a stronger heavier frame, more suitable for pulling a sidecar. In those days the only BSA gearbox was the 'B', a big heavy gearbox that had occasional selector problems. It was this gearbox that was fitted to the Gold Star I rode in the 1952 Senior TT.  The 'A' box was a more modern design with some resemblance to a Triumph box and it was very much easier to change gears. Since Mr Hopwood had worked so long at Triumphs and had a hand in most of their designs it was not surprising that there were similarities in the two designs. The only serious problem I can remember in the 'A' box was a failure under racing conditions of the layshaft bushes. I suggested fitting Torrington needle bearings which cured the problem. 

I had already met Mr Hopwood once when I contacted BSA, about getting two engines for the 'Pike' BSA I intended to build for 1952. We did not have much to say on that occasion. I do not think he really was interested in someone with my Rudge and racing background but as he said after hiring me he did not know anybody else. He wanted someone he could trust to carry out his ideas and not oppose them, that was one of the problems with Jack Amott who if he did not like an idea would not use it. During my time I found this fault was not peculiar to Amott, the same problem I had at one time or another with all the fellows in the development shop, particularly with the best ones, the more creative they were the more stubborn they were likely to be.

 To overcome the problem in the case of Reg Wilkes I tried a bit of applied psychology, but it back fired on me as I found that if you let Reg think he had thought of an idea himself he would work on it with enthusiasm. After about three year of this he was convinced that practically all the ideas in the Gold Star were his own. He told me so one day saying all I had done was use his ideas, what could I say? Perhaps I should have asked him to build a Gold Star 500 without the general shop improvements and see how it ran. Part of the problems at Nortons that Mr Hopwood had were the opposition of Joe Craig to the twin. Craig was an outstanding development engineer, the Manx Norton racing engine was probably the most highly developed single cylinder engine of all time. Unfortunately he completely ignored the rest of the Norton range.

You can imagine his humiliation when the Norton International which they used in the Clubmans TT was beaten handsomely by a Gold Star. We heard later that he had borrowed a350 Gold Star engine and tried it on their dyno and found out for himself that it put out 31-32bhp as against the 28-29 of the overhead camshaft International. The competition shop at BSA in 1952 was the responsibility of Bert Perrigo who was one of the most famous pre-war trials riders, always on a BSA.  He was incidentally also helpful in getting me my job at BSA There were three mechanics and Bill Nicholson who was in and out all the time building his special frames. In those days Gold Stars had their own assembly line at the back of the office block with Sam Jones an old TT rider himself in charge. They dropped this arrangement when demand outstripped the facilities but that was when the best bikes were built.  Every one was assembled by a man who what he was doing. When later they ran the Gold Star down the normal assembly line, they had to be virtually rebuilt at the end of it! 

When Bert Perrigo was made sales manager to Ariels Dennis Hardwicke was brought in to take his place. Dennis Hardwicke came to BSA from Temple Press where he was on the staff of Motor Cycling and had written very flattering articles on BSA which led to his being given the job of competition manager. When he first came to BSA Mr Hopwood asked me to try and help Dennis until he finds his feet. I do not think Hardwicke appreciated this as before long he was trying to get my best mechanics to work in the competition shop and did eventually have a dyno installed so they could test their own engines, which was a waste of effort and unnecessary expense. He did a pretty good job as competition manager, too good in a way, he got all the best riders from other makes on BSA's which reduced public interest.

There were about nine draughtsmen in the drawing office over which Mr Hopwood presided with Mr Perkins in charge, later Ernie Webster - when Mr Perkins left due to ill health. Beside Mr Perkins office was an office where all records and specifications were maintained, several girls were employed as tracers and then another sub-office where secret designs were made by Doug Hele and Charlie Salt. These people were kept busy all the year round, what with yearly model changes, modifications as a result of service department complaints, and entirely new designs. Many of the ideas we made and tried in the development shop had to be drawn and have job number allocated, sometimes it happened the other way round, some small component would be made and tried and if successful have a drawing made afterwards. An example of how a major component, like the big fin cylinder head for the 350 Gold Star was planned would be a meeting like we had in1953 in Mr Hopwood’s office, attended by Mr H, Perrigo, Ernie Webster, Capewell and myself.

Mr Hopwood and I had briefly discussed the objects of the meeting beforehand. Once the meeting got under way Mr H turned to me and suggested as I had had some experience of these Gold Star engines, what did I think we should have for next year. I then listed benefits such as a new cylinder head and barrel with more fins, if possible a little more down draft in the inlet port. These ideas would be noted by Mr. Webster who then wanted to know what about the piston? I suggested a high dome design to use with the shorter connecting rod. Also I would like to make some changes to the crankpin design to ensure the cage runs on the outside flanges and does not touch the bearing surface, some small holes drilled to feed oil to the cage.I explained we had done this experimentally with very good results. Also, how about eliminating the small woodruff key on the timing side, as the key tends to weaken the flywheel. After this Ernie Webster wanted to know how they would ensure the oil feed holes would be lined up on assembly? Someone suggested a mark on the end of the crankpin. After the meeting Webster and I would discuss the details of this or that idea, before giving the job to one of the draughtsmen.

You will notice up to now there had been no mention of a 500 Gold Star. The company’s policy was that the 500 should be a twin, although of course they continued to produce the B33 and M33. Mr Hopwood managed to persuade the Board that having agreed to spend the money on a new 350 Gold Star it would not cost much more to make a 500 version.  

The alloy twins made in very small numbers in 1951-2-3 were intended to be the 500 Gold Star of the future. The 1954 Gold Star engine was almost a new engine, it had the same bore and stroke as before but nothing else.  ???  Jones was given the job of drawing up all the new parts and a very good job he made of it. 

We eventually got a new crankshaft for the twins and connecting rods to suit the larger diameter crankpin journals. On first testing it we were 2bhp down on the older crankshaft, presumably due to the extra friction. When Mr Hopwood heard of this he wanted to know why and what were we doing about it, I mentioned the extra friction and he said to get it down to Halliburns shop who was in charge of all Gold Star engine testing. I was reluctant to do this as I felt one dyno is not going to be any different from another, but he insisted I do so. So I sent the engine to Halliburn for a test and he came up with the same results as we did. About two weeks later Arthur Butler by fiddling with the balance, drilling holes here and filling them there smoothed out the running and recovered the lost 2bhp, we never got more power from that engine, but we did get rid of the breakage problem. Mr Hopwood adopted a rather "I told you so" attitude which rather irritated me as it appeared that he did not trust me. But to a man like Mr Hopwood who had been opposed in every move he made, and who had tricks played on him, like the deliberate attempts to break the new welded spring frames, already described in a previous chapter, I suppose being suspicious becomes a habit and I should have expected it.

 Doug Hele who also had come from Nortons had also worked at Douglas where he was concerned with their opposed cylinder 350 twin, with torsion bar suspension. Whilst at BSA he acquired a degree in engineering of which Mr Hopwood was envious but he soon overcame this. At BSA Doug Hele was concerned with the design of all the new designs the MC1 250 racer, MC4, the overhead camshaft twin, the scooter and the Dandy. He was a very good draughtsman and a fine mathematician, he was also a keen motor cyclist and used to like to ride motor cycles with some of our experimental engines. Later he returned to Nortons where he did the design work on their little Jubilee twins which unfortunately were not very successful, he also became involved in the testing of the desmodromic valve gear for the racing Norton.