Riddles: Starting with a line drawing that showed nothing more than engine internals, Vancouver builder Dan Smith began the process of reconstructing the sporting but obscure AJS V-4, of which perhaps only 30 were ever built. Many of the parts are hand-turned while some have been modified to fit as donor parts. However, the last component to be cast was the primary cover. Smith had been unsure about how the drive to the forward-mounted generator would work and had even contemplated an auxiliary shaft. When the rest of the pieces were in place, the solution was a simple chain drive.Dan Smith was a man obsessed with a 70-year-old black- and-white photo of a motorcycle few even knew had ever existed. This machine he was determined to build, from the ground up, with nothing more than his own remarkable skills as a guide. Robert Smith tells the story.
Most of us can turn a wrench to do basic bike maintenance. Some are capable of restoring a rusty barn find to its original state, even making parts where the originals no longer exist. But few are those who can build a complete motorcycle engine from bare metal. Even then, a capable machinist could probably shape the parts from billet with the aid of drawings, a CAD-equipped computer and a CNC mill. But what if there were no drawings, or even a model to work from? 
All this makes Dan Smith’s re-creation of the groundbreaking air-cooled 500cc AJS V-4 even more remarkable. Not a restoration or a rebuild, and without components to copy, Smith designed, cast, machined and assembled his V-4 from a black-and-white photograph and a cutaway sketch of the 1936 prototype.
And if you didn’t know AJS built an air-cooled V-4, you’re not alone.
THE WOLVERHAMPTON FIRM OF A.J. STEVENS EARNED A SOLID sporting reputation during the 1920s with their racing “Big Port” 350cc single, while also building sturdy side-valve V-Twins for sidecar use.
Like many other enterprises, it was the financial collapse of 1929 that scuttled AJS. In spite of diversifying into car and truck bodies, furniture and even radio cabinets, the company continued losing money and by 1931 was unable to pay its creditors. Jumping at the chance of acquiring the AJS name, the Collier Brothers of Plumstead, London snapped up the company, merging it into their own business. The famous grouping of AJS and Matchless was the result.
Though they had a sporting reputation of their own (a Matchless won the single-cylinder class in the first Isle of Man TT race in 1907), the Colliers certainly intended to exploit AJS’s racing heritage. Hence the motorcycle that was the sensation of the 1935 Earls Court show in London: an air-cooled V-4 designed by Bert Collier but dressed in AJS livery. It was shown with lights and dynamo as a fast road model, but exposed “hairpin” valve springs and space for a supercharger suggested it might be raced, too.
The four cylinders were arranged in a 50-degree V with a single camshaft on each head. A central crankshaft sprocket drove the single timing chain, which was tensioned by an idler between the cylinders. Two carburetors, one on either side, fed mixture to the 50mm bore by 63mm stroke cylinders, with the exhaust exiting through four separate pipes. Two bevel-drive magnetos hung on the “timing” side of the engine providing the sparks, while a front-mounted DC generator—right where, some speculated, a supercharger might fit—fed the battery.
Though it never went into production, contemporary reports suggest that parts for as many as 30 air-cooled V-4s were produced. It was certainly raced in 1936, the factory entering two machines in the Isle of Man TT, with Harold Daniell and long-time AJS works pilot George Rowley as riders, though neither machine completed the course. The racer appeared again in 1938 with a supercharger, but apparently suffered overheating problems.
In 1939, AJS entered a liquid-cooled V-4 in the TT and the Ulster grand Prix: in the latter race it led for three laps before rider Walter Rusk retired with a broken fork link. The “wet” V-4 was also raced briefly after the break for hostilities, until the ban on superchargers terminated its development, and the parallel Twin “porcupine” took its place.
VANCOUVER’S DAN SMITH IS A LEGEND IN CANADA’S VINTAGE motorcycle scene. A long-time Vincent owner and a guru of the Stevenage machinery, he owns two complete examples: a fully tricked out series C Shadow with alloy wheels, modern carbs and his own modified “short rod” engine; and a Series B Rapide that he’s ridden not only to Tierra del Fuego and most of the way back, but also north to the Yukon territory.
And it was his Vincent tuning skills that got him involved with the VIBRAC team, which mounted an attempt on the blown fuel world speed record at Bonneville Salt Flats in 2003 with the ex-Max Lambky streamliner. As crew chief, Smith coordinated rebuilding and tuning the two Vincent engines in the streamliner. With Dave Campos as pilot, the streamliner clocked close to 200 mph before a backfire broke one of the crankshafts.
His last two restoration projects, a 1934 BSA Blue Star and a 1933 Matchless Silver Hawk, were rescued from almost total obliteration, with Smith making many of the missing parts himself. How did this man get into motorcycle restoration?
“The motorcycles came first,” he says. “I was knee high when I realized if a guy learned how to be a machinist, he could make motorcycle parts. The fascination with the whole thing is tools. You use the tools in connection with the motorcycles.”Why the AJS V-4: and how does the Ajay V-4 compare with the Silver Hawk?
“This engine is just fascinating,” he says. “The Silver Hawk is pretty mundane. You can’t spin that thing fast, you couldn’t get any horsepower out of it. It’s not made for what the Ajay was made for, for going fast. This is quite a design.”
Smith admits, though, that the V-4’s power is always going to be restricted by the sharp turn the mixture has to make to get from the carburetors into the cylinder heads.
His interest in building the V-4 goes back 20 years, though preparation really started seriously in the mid 1990s. And not surprisingly, re-creating a motorcycle that was never much more than a prototype and was last seen nearly 70 years ago took a while.
He started with a solitary black-and-white photograph and a cutaway drawing—not an engineering drawing but a line drawing showing the general layout of the engine internals. Both Vic Willoughby and Bruce Main-Smith in the UK provided some useful insight into the V-4 project, and Dan Smith also visited Sammy Miller’s museum in New Milton, Hampshire to collect some dimensions from the supercharged, liquid-cooled 1939 race bike.
“I did some sketching and scaling,” he says. “I had these dimensions, and I couldn’t fit anything around to make it work. After quite a bit of time, I concluded that the water-cooled one was completely new. I was trying to duplicate the crankcase: water cooling would have allowed them to make the engine more compact with shorter rods.
“I couldn’t make the dimensions work for the air-cooled engine.”
Smith next attempted to scale many of the components, interpolating angles and dimensions by projecting the axes shown in the cutaway drawing. It seems there may have been some artistic licence used to make the drawing look right to the eye, and the angles shown are deceptive.
“There are, I think, six different diminishing points,” he says, “ so it’s impossible to scale.
“The cambox was very difficult to decipher because of the view. I tried to mimic it as much as I could and keep it to scale. It’s very difficult to make something exactly the same.”
Starting from the known two-inch bore diameter and the 50-degree vee (and allowing a wide enough flat between the bores to accommodate the cylinders) governs the deck heights for the cylinders on the crankcase mouth. Then, by working out the rod length based on the pistons’ dimensions, “I got to an accommodation of numbers I could use,” Smith says.
Designing the bevel gear drives for the twin magnetos was a challenge. With dimensions taken from the liquid-cooled engine, it proved impossible to fit gears with the right number of teeth with the 50-degree angle between the two magnetos. He settled on spacing the magnetos at 60 degrees, which gave room for the bevel gears. The half-time speed for the magneto drive is achieved through two reduction gears.
Next Smith made a full-scale model of the engine in wood and fitted it with a period Burman gearbox into a “Denly” AJS frame. The frame was developed by Norton/AJS racer Bert Denly, who added a third chainstay running from the seat tube to the rear axle mount to the standard AJS frame. Denly and his co-pilot, one Mr Baker, took the three-hour 350cc world speed record at Montlhery in 1930 at exactly 100 mph on a Denly-framed AJS R7 racer. The frame was acquired through Smith’s network of contacts on Vancouver Island.
So, with a wooden mock-up of the engine for final sizing and for fitting into the frame, work could start on making the major castings. Smith produced the patterns from which the components were cast. Then with the crankcase dimensions fixed, he was able to put together the crankshaft, flywheels and camshaft drive. Though the original V-4 was said to have used six main bearings for the crankshaft, Smith only saw the need for five.
“I don’t know why they’d need six,” he says. “In order to keep it narrow, you don’t want to put two bearings on the drive side. It’s only maybe 40 horsepower at 6,000 rpm and perhaps 40 pounds of torque.”
He used two inboard main bearings in the crankcase, two outboard of the flywheels and one in the timing chest to prevent end float. Next came the connecting rods.
“It’s a knife-and-fork rod like a Harley,” says Smith. “I had to make those, work out the cranking ratio and establish the rod lengths.”
Then the top end. Smith used pistons from a Suzuki DR100, “so I consequently ended up copying the combustion chamber, and I also used their cam profile,” he says, “but the cams are individually verniered with pins, like a Manx.” The valves themselves are turned down items of Vincent heritage, something that would surprise no one who knows Smith’s reputation as a tamer of the Stevenage monsters.
“I was going to make valves, then it dawned on me I have old Vincent valves by the bucketload around here,” he says, “so I re-machined them.
“I think the splay of the valves might have been slightly wider than I have, but I was restricted by the squish band.”
Smith has incorporated eccentric rocker adjustment. “The V-4 used alloy rockers, so that’s what the replica has,” he says. There are also small pumps to scavenge oil from the camboxes back to the tank. “I had to make those too.”
The single timing chain runs from the central crankshaft sprocket, over one camshaft, under an idler sprocket between the banks of cylinders, then over the second camshaft. The chain itself is from a Suzuki GS 400—or rather, from two GS400 timing chains riveted together. Like the other Suzuki parts used in creating the V-4 (and the hairpin valve springs from an NSU Max!), the chains were supplied by ex-racer Murray Neibel of Vancouver dealership Modern Motorcycling.
For the magneto drive, Smith not only cut the bevel gears himself, but also made the gear cutters. For practice, he’d just finished cutting three sets of planetary gears for a 1915 Harley-Davidson two-speed hub. And though it took him a month to finish them, “they made a beautiful fit,” he says. “I’m really happy with how they came out. It took a month to complete just this job, but cutting bevel gears is always fun.”
The magnetos themselves are both by British Thompson-Houston: the rear mag is a BT-H KDTT donated by local ex-motocross racer, Denis Mitchell. The forward mag was an eBay find and came from Tasmania. Remarkably, the name plate is stamped “KDV 50 AJ4,” though it’s unlikely to be a reference to the V-4.
As well as the magnetos, the timing drive spins the gear oil pump borrowed from a BSA A10, with worm drive taken from a B-series BSA engine. An anti wet-sumping valve and pressure relief valve complete the lubrication system.
Smith chose two float bowls for each carburetor: contemporary reports suggest the V-4 may have suffered fuel starvation problems; if so, he hopes this will prevent it.
The last component to be cast was the primary cover. Smith had been unsure about how the drive to the forward-mounted generator would work and had even contemplated an auxiliary shaft, but when the rest of the components were in place, the solution was a simple chain drive. Also chain driven is the standard Burman four-speed gearbox mounted behind the engine.
“I didn’t have that until (the engine) was in the frame. I thought, ‘Okay, the crankshaft’s here and the transmission’s here, so now I can draw the primary case.’ You look at the photograph and it’s really just joining the lines.”
By April 2006, Smith’s recreation was essentially complete and “dry” assembled, though the cams still needed to be indexed. Before that, he had to dismantle for painting and plating. Oh, and work out the carburetor settings, ignition timing …
ON JULY 8, 2006, JUST OVER 70 YEARS SINCE ITS PROGENITOR HAD run in the Isle of Man TT, Dan Smith’s V-4 burst into life with a throaty rasp from the four exhausts—not smooth like an in-line four, but with a ragged beat not unlike a Laverda triple.A couple of weeks later, I followed him to Spanish Banks on Vancouver’s Burrard Inlet for some photographs, reveling in the sound and sight of the V-4. It’s a magnificent machine, and a wonderful realization of Bert Collier’s creative vision. The thirties were a time of great engineering advances, and the AJS V-4 represents perhaps the pinnacle of that decade’s motorcycle design.
Very few machinists have the foresight, ingenuity and skill to pull together such a project, and I know I’m not alone in appreciating Smith’s efforts to recreate one of the era’s most significant motorcycles.
Robert Smith Canadian Biker May 2007 Issue
RELATED CONTENT : Resurrecting a 1946 A.J.S.
BACK TO MORE VINTAGE CONTENT
