Rocket Science: "Tul-Aris" Prototype Racebike | Cycle World

CLASSICS: Rocket Science

Professor Robin Tuluie’s divining rod for better handling.

Tul-Aris prototype studio view

From the March 2001 issue of Cycle World.

Every one of us has fantasies of building wonderful new motorcycles. In the imagination, engines, wheels and chassis morph into fresh and uniquely appealing combinations. The reality of making a living pulls most of us back down to earth, and our prototypes are never built.

Never is too strong a word. The beautiful machine you see in these pictures exists. Its twin-cylinder two-stroke snowmobile engine produces 165 crankshaft horsepower, and dry weight is 264 pounds. That’s Grand Prix territory.

The project is the work of a Minneapolis-area team inspired by Robin Tuluie. There are no plans to produce this motorcycle, but it is just too delicious a creation to ignore.

Tuluie himself is just as eclectic a prototype as his machine. When I first met him at AMA roadraces in the 1980s, I thought him a charming eccentric, a Norton nostalgia nut of the kind that collected in Battle of the Twins races. Later, I encountered his accurate, well-written comments on an Internet two-stroke forum, but didn’t associate the two. When the connection was made, I learned that he was a post-doctorate in astrophysics, with a serious minor in motorcycle racing. Interesting combination. When he replaced the Norton with a well-executed homebuilt, based on a Honda CR500 engine, I was impressed by the flexibility of his enthusiasm. This wasn’t a man who would still be trying to set a land-speed record with a Henderson at age 70.

Next he switched worlds. Astrophysics was great fun, but there was no prospect of ever knowing everything–the universe is just too complicated. Besides, what university would give tenure to a motorcycle racer? Tuluie looked for a motorcycle-industry job.

Tul-Aris prototype engine close-up

Powering the Tul-Aris is a highly modified Polaris snowmobile engine. The two-stroke Twin has been enlarged to 780cc with hop-up components intended for snowmobile competition, though the straight-shot expansion chambers actually allow it to make more power than in a sled–165 crankshaft horsepower. The engine breathes through twin carburetors and reed valves at the front.

When I next heard from him, he was part of Polaris’ Victory motorcycle-engineering team as chassis-dynamics specialist. The people who gravitate to astrophysics find math a natural language, so I wasn’t surprised to learn that instead of looking up the various expressions describing the motions of a vehicle, Tuluie found it easier just to derive them as needed. I thought of air racer Darryl Greenameyer. When he switched to Top Fuel drag racing, he didn’t have to waste time getting used to the speed. Flying 500 mph at 25 feet altitude was adequate preparation.

While at Polaris, Tuluie found out how much serious horsepower was being stuffed into snowmobile Twins that you could carry under one arm. Who could resist building a motorcycle around such an engine?

The engineering workplace of today replaces drawing boards with large computer monitors. Objects are modeled in code and can be manipulated on screen, their parts analyzed for operating stresses, components shuffled around to find the best packaging. Data then goes to rapid-prototyping machines that in a few minutes make three-dimensional paper or plastic objects you can play with, pass around and refine. Or, if you’re ready, data can go to CNC machine tools to cut metal. These are not only great tools for work, they’re the best playthings since erector sets. Surrounded by equipment like this for making the imaginary become real, what would you do? Stay late and play! Or, if your company has a stiff-necked no-play policy, somebody from the department is sure to have bootleg copies of all the software at home.

Tuluie’s practical education advanced. Once the Victory V92C reached production, the task group that had created it fissioned in all directions–call it “project burnout.” Tuluie landed at MTS, a giant company that makes specialized vehicle testing and simulation machines used all over the world.

Tul-Aris prototype studio radiator close-up

The radiator mounts horizontally along the base of the rear subframe. The short engine allows an efficient air duct to be formed between the top of the engine and the bottom of the gas tank, while the low-pressure area at the rear of the bike promotes airflow.

In fact, their use is spreading outward from auto manufacturing. Formula One is an example. Rules now forbid race teams from unlimited testing on actual tracks, but by using these exotic simulation machines they can do even better. Teams have therefore lined up to buy “seven-posters”–big, hydraulic-powered, computer-controlled devices on which full-scale cars can be placed. These systems accurately replicate suspension action on any track, complete with simulated aerodynamic downforce and can do it 24 hours a day if so desired. Users were claiming significant reductions in lap times from such testing. Why not apply similar methods to motorcycles?

From these diverse origins, the 165-bhp, 264-pound “Tul-Aris” special came into being. The design and manufacture of the physical hardware was a big project, even for Tuluie and his large staff of volunteers. And there is more. This is not just a beautiful and potent motorcycle. It is also the sharp point of a much larger and more ambitious projectile–an attempt to raise the understanding of motorcycle dynamics as has been done for F-1 cars. There are well-funded motorcycle racing teams today that have spent “years in the wilderness” with new bikes, searching for the elusive setup that would make their machines handle. At present, the only way to a good setup is experienced guessing, followed by endless testing.

Back to hardware practicalities. How do you connect a snowmobile engine to a motorcycle’s rear wheel? When Rokon built its sled-engined roadracer in the mid-1970s, it used the snowmobile’s variable belt-drive. I was present at a Loudon, New Hampshire, test and it was weird to hear the engine speed remain constant as the vehicle accelerated, the transmission shifting ratios continuously.

Tul-Aris prototype studio stripped view

The Tul-Aris is a professional-looking package and while small, is not tiny. It carries its wheels slightly less than 54 inches apart, and runs 23.5 degrees of steering rake and 3.9 inches of trail. The main chassis structure consists of aluminum plates that carry the engine rigidly. The chassis forms a small tank behind the steering head that carries almost a gallon of gas. That combines with a 3.5-gallon tank to provide plenty of range. The rear subframe is fabricated from honeycomb sheets with kevlar and carbon-fiber faces, much like current Formula One car construction.

But the pulleys of these systems are too bulky for motorcycles. Tuluie laid plans to build an aluminum-and-carbon-fiber-case for a Ducati gearset onto the back of his snowmobile engine. Its clutch would be driven by a toothed belt. Reed valves and carburetors would face forward, while the engine would be integrated into a minimal chassis–a machined-and-welded “steeple” that joins the engine/transmission unit to the steering head. Exhaust ports would face to the rear. The ducted radiator would lie near-horizontal, above the engine and under the seat. Cooling air would reach it via a “window” through the steeple, above the engine. The components fit together so neatly they seemed made for each other.

A great deal of analysis was done to ensure that some stack-up of stresses did not occur at an unforeseen point. This meant lots of computer time, creating models of all the parts and assemblies, then creating realistic stresses to apply to them with dynamic finite-element analysis.

What is real? The track is real, but not always available. The test rider is also real, but subject to human variability. That makes lesser degrees of reality more attractive for testing–as now with F-1. Testing machines are always available, and they are repeatable. Enhancing their usefulness are dynamic models–mathematical constructs that accurately represent the way a real vehicle reacts to road and aerodynamic forces. Before you dismiss the words “dynamic model” as more computer-nerd jargon, think of what such a model could do. To run 10,000 setup combinations on the track would take the crew the rest of their lives. To run 10,000 setup combinations on the computer would take mere hours. That difference makes it appealing to develop and prove a dynamic model. Such models are now routine for car and aircraft design.

Tul-Aris prototype studio gearbox close-up

One problem with using a snowmobile engine is the lack of appropriate transmissions. Tuluie made his own, using shafts and gearsets from the latest Ducati Superbike. By pulling the clutch, the gearbox can be accessed in position, allowing individual gear ratios to be matched to particular tracks. The carbon-fiber housing for this box is almost weightless.

For these reasons, Tuluie has developed and is refining a dynamic model for his new machine. In comparison with real-world testing, it has already produced some specific results within a few percent, which is impressive. Which is the real motorcycle? It can be hoped that over time, this kind of work can tell us what riders are really talking about when they discuss handling–which has lagged behind the other motorcycle arts because it couldn’t be easily defined or measured like power, braking force or mass properties. Now it is coming within reach of analysis. Aircraft were first with this because military aviation has the support of entire governments. Cars are using simulation and dynamic modeling now in durability testing because it saves money and it works. Now that the methods exist, they are being applied to F-1. Why not motorcycles next?

I have seen small-scale motorcycle engineering projects perish for two reasons: Their creators made every design detail a revolution, and they amateurishly insisted on riding the machines themselves.

No single human being is that powerful–not even the late John Britten. This is why Tuluie’s machine is a conventional design, and will be tested and raced in the 2001 season by a professional racer, Mike Ciccotto. When someone asked Tuluie why he didn’t design an alternative front end, he replied reasonably, “Why would I?” Telescopic forks are very good now, and no challenger has shown overall superiority.

Is there even a racing class for this bike in our four-stroke world? Fortunately, WERA has Open classes that fit. Even so, racing success is only part of the reason for this machine’s existence. It is inspiring that a beautiful, complex project like this one can still be conceived and carried out by a few determined people, using limited money and borrowed facilities. The project is not just the physical object that you see. It is also a tool for analysis, an element in a larger, longer-term project to understand all motorcycles better. Tuluie was an astrophysicist because he wanted to understand everything. Now, he’s a motorcycle constructor for basically the same reason. “I detest empirical methods,” he said recently, referring to blind cut-and-try. “It’s almost unimportant, what the ‘answer’ is.”

Tul-Aris prototype studio suspension close-up

An Öhlins spring/damper is stroked by a rocker linkage to provide 120mm of rear suspension travel. The entire mechanism operates on preloaded needle bearings to eliminate any undamped travel, and the bottom pull link is designed to operate as a flexure in two different dimensions to minimize binding loads.

Finding a good setup without knowing how and why it works just means that when a new tire construction, or a new chassis, or a new rider changes the conditions, the same blind process will have to be repeated. This should be unnecessary.

What is handling? I hope Tuluie and his fellow engineers and craftsmen will begin to answer this question formally and in language we can all understand, using the dynamic model they have developed for this machine. If major manufacturers already know all this stuff, they show little sign of it. Otherwise, a good race setup would be a lot more common than it is. Also, manufacturers have no motivation to share what they know. Tuluie has already presented one SAE paper on what he has learned with this project, and there will be others.

In mid-December, the bike was sent to California to be test-ridden at Willow Springs Raceway by Ciccotto. After two laps, he came back into the pits, causing the crew to fear they’d made some awful mistake. Ciccotto removed his helmet and announced, “Man, this thing handles awesome!”

After another five laps, he came in again, saying, “I just went into Turn 2 faster than I’ve ever gone–even on my Superbike!”

This is an encouraging beginning, but there is hard work ahead. Even the most refined dynamic models–those used in aviation–show only where the best combinations are likely to be. Digging them out and perfecting them still requires hardware testing. Strange, isn’t it, that as you arrive at Willow Springs for a motorcycle test day, Edwards Air Force Base–a major test center for military aircraft–is right next door?

Photo #1

Tul-Aris studio right-side view.

Brian Blades

Photo #2

Tul-Aris studio left-side view.

Brian Blades

Photo #3

Tul-Aris studio stripped view.

Brian Blades

Photo #4

Powering the Tul-Aris is a highly modified Polaris snowmobile engine.

Brian Blades

Photo #5

The radiator mounts horizontally along the base of the rear subframe.

Brian Blades

Photo #6

Tul-Aris studio details view.

Brian Blades

Photo #7

An Öhlins spring/damper is stroked by a rocker linkage to provide 120mm of rear suspension travel.

Brian Blades

Photo #8

Tul-Aris studio details view.

Brian Blades

Photo #9

Tul-Aris studio details view.

Brian Blades

Photo #10

Tul-Aris studio details view.

Brian Blades