A second Roarer roars
Readers were introduced to the remarkable Dan Smith in issue 75 with the feature story on his V-4 AJS. Now it’s time to look at another creation from the modest Smith workshop in Vancouver BC.
As noted in the AJS story, Dan considers the 1930s to be the zenith of the British motorcycle industry, considering the post-war products and their sameness. So after completing the herculean task of building his AJS V-4 from a single photograph, Dan set about making his own version of another of his favourite designs, the 500cc supercharged twin Velocette known as The Roarer.
Now, for those not familiar with the saga of the Roarer, it’s worth a brief rundown here. In the ‘thirties, thanks mainly to the rapid development of supercharging that began in WW1, cramming increasing volumes of air and fuel mixture into cylinders was proceeding apace. Military engineers had long grappled with the problem faced by their aircraft, which had to cope with cooler and less dense air the higher they flew. ‚
The higher the altitude, the lower the volumetric efficiency, with the resultant drop in power. The best way to overcome this was to artificially increase the amount of mixture going in, and the favourite method was by supercharging. The same principle applied to ground-based engines. Bentley had done much work in this field, as had Mercedes Benz in car racing, and BMW had successfully supercharged their record breaker, the WR750. Significantly, those in charge of the rules for international motorcycle racing made no distinction between ‘atmospheric’ (non-supercharged), and ‘boosted’ (supercharged or turbocharged) engines. As the thirties progressed, several German and Italian factories tried supercharging, not just in the 500cc class, but also in smaller displacement categories. Velocette had dabbled in boosted engines as far back as 1930, with their blown KTT (using a Foxwell blower) which was quaintly known as “Whiffling Clara” that ran in the Junior TT in 1931 and 1932, and of course, so had AJS with their water-cooled V-4 500cc racer. Both of these engines had begun life as normally-aspirated designs, with supercharging later grafted on, but what became the twin-cylinder Roarer was designed by Harold Willis and Charles Udall at Velocette specifically for supercharging. To overcome the BMW’s inherent torque reaction and the gyroscopic precession effect from the horizontally-opposed layout with its large flywheel (which became even more marked and violent when boosted), Udall’s design placed twin crankshafts, geared together along the frame axis, with the cranks rotating in opposite directions, thus eliminating both the torque reaction and the gyroscopic effect which in BMW’s case, affected the cornering characteristics. One crankshaft drove the supercharger, which sat behind the crankcase, while the other drove the transmission components, including the shaft to the rear wheel. A skew gear from the front of the left side crankshaft drove the oil pump which ran at one-third engine speed. The choice of shaft drive was interesting, but Udall’s reason was that it kept the rear of the motorcycle free from the oil that was necessary to lubricate the usual drive chains. Indeed, Udall’s insistence on shaft final drive virtually dictated that the engine would need to use longitudinal crankshafts. The top end of the Roarer engine was more or less conventional, with single overhead camshafts for each cylinder. One unusual aspect of the design was that the exhausts ran from the rear of the cylinder head, with a long, curved inlet manifold running over the top of the engine and into the leading side of the cylinder head. This allowed the incoming mixture to shed some of the heat generated in the blower, and to reduce the inherent pulses also created by the vanes. To assist in cooling the exhaust area of the cylinder head, a pair of ducts was bolted to the sides of the heads.
By 1939, Gilera was rapidly coming to grips with its supercharged four-cylinder 500 which even by the mid ‘thirties was pumping out around 85hp when in trim for record breaking, and in 1937 set a new world record for the flying kilometre at 170.15 mph. Despite the inherent peculiarities of the opposed twin design, the supercharged BMW had taken an emphatic 1-2 in the 1939 Senior TT, while the Gilera won the season’s final Grand Prix in Ulster and rider Darino Serafini became that year’s European Champion. The Roarer, on the other hand, completed a solitary lap of the TT course in 1939, during a practice session in the hands of TT maestro Stanley Woods. The factory claimed the lap was not timed, which seems highly unlikely, and Woods reported that the machine handled extremely well and was incredibly smooth, but there were concerns with overheating and the Roarer was withdrawn from the Senior TT. A factor in the decision to withdraw the machine may have been the absence from the TT of Harold Willis, who had entered hospital suffering from meningitis. He passed away on June 11th, just two days after that single practice lap.
Udall originally planned to use glycol cooling for the Roarer, but this was vetoed on cost grounds. Then when peace returned and racing struggled back onto its feet, the FIM intervened and banned supercharging outright in 1947. Thereafter the Roarer gathered dust in the workshop, until in 1956 the engine and
gearbox were stripped to permit journalist Vic Willoughby to inspect the design and interview Charles Udall. At the same time the artist Frank Beak created exploded drawings of the engine – something that would prove invaluable in the future. The Roarer was externally refurbished and reassembled, minus its engine components, and was trotted out for various motorcycle shows during the ‘fifties. It is believed the internals were placed in a box under a bench in the Velocette workshop, from where they gradually disappeared. When Velocette finally went under in 1970, the Roarer carcass (an the prototype Model O) was acquired by journalist John Griffith who exhibited it in Stanford Hall Museum in Leicestershire, which he had co-founded. Following Griffith’s death in a motor accident in 1975, ownership of the Roarer passed to John’s son Stephen, and it was subsequently sold in 1983 to Ivan Rhodes in Derby, who began a long and painstaking task to rebuild the bike – a task that involved recreating the missing engine internals which fell to his son Grahame. Luckily, Charles Udall was living in retirement in UK and still had his notebooks from 1938 and 1939. Since its completion, the original Roarer has appeared around the world, including Australia and New Zealand, much to the delight of Velocette aficionados and those who appreciate this immensely significant piece of racing history.
Meanwhile, in Canada…
Just as he did when creating his V-4 AJS, Dan Smith began at his drawing board, in ‘The Basement’ under his modest house in Vancouver BC, where he worked out the dimensions from photographs and cutaway drawings taken from magazines which needed to be scaled up in isometric projection. It was a long and arduous process that involved drawing up every component individually, and often more than once, before he was confident that everything would fit together as it should. However unlike the AJS project, Dan was able to get some assistance along the way, as Ivan Rhodes happily volunteered to check Dan’s drawings against his own. Still, it was no cakewalk, and getting the crankcases right was the first task. “The cases are made up of five separate castings,” says Dan. It had been documented that the balance factor of the cranks was 100%, a figure confirmed by Ivan Rhodes, whereby the piston movements are exactly matched to those of the crankshaft bob weights, giving perfect primary balance. After making the wooden patterns for the crankcases sections, Smith had them cast by a Vancouver company in aluminium alloy, although the original Roarer castings were mainly in heattreated magnesium. He machined the castings in his basement workshop and then moved on to the top end of the engine, first tackling the complex system of bevel shafts and gears used to drive the camshafts. Inside the engine he opted for off-the shelf parts in several areas (“Why make what you don’t have to”, is Dan’s motto) such as Suzuki pistons, a BSA oil pump and NSU Max valves and hairpin valve springs. The lubrication system was another aspect that occupied countless hours, with myriad oilways that needed to be drilled and matched up. Then there was the gearbox to deal with! Unlike conventional British designs with a countershaft, the Roarer gearbox has just two shafts, for input and output, with the former sitting below the latter and connected by large gears. The input shaft is splined into the clutch centre, while the output shaft is splined into a constant-velocity universal joint on the axis of the swinging arm
pivot. The final drive shaft runs inside the left side of the swinging arm fork. A suitable CV joint could not be found in Britain, so Velocette Managing Director Percy Goodman used his contacts in the aero industry to obtain a couple from the Bendix Corporation in USA. One was installed in the Roarer, the other lost. Many years later, new owner Ivan Rhodes located it, still in its wrapping, under a bench where it had been for over 40 years. Next came the most unique feature (for a Velocette) – the supercharger. Dan’s research convinced him that the original Shorrock blower had been specifically made for the Roarer, so there was no chance of finding one of those. What to do? Make his own of course. “I made the whole supercharger but I had no idea when I was starting. When I made the vanes for the supercharger I ended up with a bit more clearance than I would have liked but I wasn’t about to remake them. Velo used a Shorrock and you can still buy those, but there’s none that would fit in, and you’ve only got so much room. To me they had to have gone to Shorrock and said we need a blower that will fit in this cavity, so they made it, and the pictures show six vanes, but I couldn’t put six vanes in, it was too short. There isn’t room for the whole stack to go on the shaft, so I made four vanes which doesn’t really matter because you don’t get any more pumping, it’s displacement per rev as it goes around. If I was going to pursue it I would have to cut the clearance down on the vanes but I’m not racing it so I didn’t. The cams were shaped wrong for the blower – you should have zero overlap with a blower. When they (Velocette) were running it on the bench they talked about having trouble getting the blower pressure up. I’m getting 6 pounds boost on it at 6,000 rpm but I haven’t been able to spin it up further, but I think I’ll get 8 pounds if I get it to 7,000.” With the power unit and transmission basically complete, it was time to turn to the chassis, which was naturally built in the Basement workshop by Dan – a process he describes as being, “Just a whole bunch of straight tubes, fairly straightforward.” Straightforward it may have been, but it was still a case of trial and error, and immensely time consuming, combining as it did the swinging arm with the integral drive shaft tube. The engine is clamped into place at four points, acting as a stressed member and contributing to the rigidity of the chassis. Thanks to the inherent vibration-free nature, no head-steady is necessary. “I got a set of Webb forks out of India. They were re-popping Webb forks for the 16H Norton so I got them back here and converted them to the Webb TT forks but they must have made the patterns off an old set and it was just about as much work to fix them as it would have been to start fresh. I was a month or more building the fuel tank by hand out of steel.” Dan also made the hubs and brake drums, exact replicas of the originals, but using Suzuki brakes shoes.
Three years after the project began, the replica Roarer burst into life. It was time to put his work to the acid test, so man and machine ventured south to Seattle International Raceway. “When I was on the track at Seattle, which was two years ago, I was a wreck when I finally came off and I said, ‘That’s it, I’ve had enough’. You don’t know what to expect, down the straightaway and I’m watching blower pressures and mixtures and I haven’t really gotten enthusiastic enough to go back and do it again. I haven’t put many miles on it, and then I’ve come back and I’ve made some changes but I haven’t been back out again. I’m going to be 80 next year so going 100 mph is not for me. The gearing they (Velocette) used for the one lap they did at the TT was very high, and I’ve got the gearing lower in the final drive. I couldn’t fit in the lower gear (in the box) because you don’t go through four gears you just cut through one gear, up and out so you have to do all your ratios just between two, they’re not compounded. So in the final drive you get a choice of 3:1, or 4:1. I couldn’t get full revs out of it because they shortened the straight, so I couldn’t really get it wound out. It was funny; I’m in third gear and it won’t rev and after many times around the track I’m getting used to it so I thought I’d try fourth gear and see what that’s like, and it went down into third! I’d made my cam plate so it went first, second, fourth, third! I remember having to change my gear sets in order to fit everything in there and I swapped third and fourth around and I guess when I made the cam plate my drawings and sketches got swapped around, you forget these things. The cam plate is correct now but I haven’t ridden it, apart from in Montano on a little short track but I couldn’t get it out of second gear. It won’t pull up any sort of hill without a good start; it needs a lot of clutch.” With the Roarer now a complete and running
item, Dan has moved onto his next project, a Norton-framed café racer using one of his own HRD Series A Rapide engines, and he has decided to put the Roarer up for sale, to a buyer “who has deep pockets”. “I’m slowing down, but I’m still in the Basement eight hours a day.” You get the impression that while ever the sun rises, Dan will not be able to stay away from the drawing board and the machine shop for long.