The Kawasaki H2-R
An Inside Look
By Kevin Cameron
Kawasakiís modern dynamometer facility at Akashi is just like you imagined it. The serious, slight man who is your guide pushes open a steel door and motions for you to enter. A fluorescent-lit corridor stretches down our pathway, one wall punctuated by the control consoles and soundproof double glass of the test cells. Two gray-uniformed men emerge from a one foot-thick door, pushing a lab cart. On that cart is an engine that is completely unfamiliar to you, but your business is farther down the hall.
What is that high-frequency buzz? You walk to a lighted window and look in. This is a Kawasaki 750 H2-R road-racing engine under test. Three heat-blue pipes jut crazily from the big-finned cylinders. Magnesium carburetors gleam dull gold. Power is passing down a trembling chain to a massive Meidensha electric dynamometer. The test is supervised by two technicians: one monitors the instruments, the other records data. A large faced tachometer in the panel shows 9300 rpm. And now your guide nods abruptly to show thatís all you will see here. You have been lucky to get this far.
What is the rest of the story? What is this engine that has won eight National AMA Road Race Titles in two years? How did it come to be?
There is no successful technology that doesnít sit squarely atop a mountain of development work. There are reasons why this engine has the strengths and weaknesses it does. It is based on a road machine, but the Kawasaki performance reputation determined that the H-2 street bike would have the highest power-to-weight ration in the business, yet sell for less than a kingís ransom . . .a good starting point to say the least. From the introduction of the advanced A1 and A7 rotary-valve twins in 1966, Kawasaki has worked to create a performance image for itself.
To get part of the action that Yamaha was monopolizing with its TD1-C 250cc racer, Kawasaki readied a race-version of the 53 x 56mm A1 twin. The A1-Rís, as they were called, had many of the problems youíd expect from a first effort. In 1968, special versions were built incorporating much-needed improvements into the carburetion and ignition systems, and moderate success encouraged the company to revise the bike from top to bottom for 1969, and engage several top-line riders.
The 1969 A1-RAís had the first of the now familiar Mikuni center float carbs on vibration-absorbing rubber mounts. These bigger 30mm carbs, with oversized rotary valves and new pipes, upped the power enough to put the old oil-bath clutch out of business. An extremely neat and powerful dry clutch took its place, with the straight-cut primary gears in their own magnesium case, and the clutch basket and release outside.
This machine in the 250cc and 350cc models carried the works team while racing prototypes of the newly introduced 500cc HI Triple were undergoing testing in Japan and on the West Coast. Not surprising, the three-cylinder racer drew upon the twinís design in important respects. The 75 bhp 500 needed a strong clutch, so the proven idea was extended from the A1-RA, but this time with sintered copper friction material in place of the former asbestos-based plate facings. Center-float carburetion, now out of the prototype stage, had already appeared on 1969 Yamaha production racers. The 500 mounted three of these die-cast 35mm instruments in rubber sleeves.
Owners of the original A1-Rís found that prolonged operation over 10,000 rpm would crack the cast pistons. The factory A-models, therefore, had forged pistons that lasted, and this type of piston became the standard for the Triple.
Next to pistons on the two-stroke critical parts list is the crank, especially the connecting rod big-end bearing. A1-Rís had lubrication pumps, supplying oil first to the main bearing, then by collector ring to a drilling in the crankpin, finally to dribble 25-50cc per hour through the big-end. The first 250 racer had a big-end cage of street design, first developed by the German Durkopp firm in the Ď50ís and now made under license by NTN. The racing cages were identical with the street unit, save for being plated with silver instead of copper. These would last well at production-racer speeds, but gave up quickly in the A-models. A much stronger cage was designed, placing two rollers in each of nine slots. This allowed the separator bars to be much thicker than in the earlier cage, which had fourteen single rollers. It worked, and was retained for use in the 500 racer.
Another part of the A1-R-to-A1-RA metamorphosis had been the abandonment of the street frame and adoption of an enlarged version of the double-loop "featherbed-type" frame from Kawasakiís 125cc GP bike. This new frame was a marvelous handler, so when the 500 HI-R came along, the design was simply scaled up again.
When committed to battle in 1970, the 500 wasnít as successful as had been hoped. That was the year of big-buck racing for Honda and BSA-Triumph, who introduced the full 750ís that were then the pinnacle of four-stroke Class-C design. While only one expert National was won on the 500 in two years, an immense amount was learned, Junior and non-AMA races were won, and it did lead many laps of the Nationals it didnít win.
So there were problems. But from their solutions grew the 750 H2-R. Crankcases that might run forever on the street cracked right up under racing conditions. This led to rubber engine mounts, now standard on all big Kawasaki Triples, racing or street versions. Although the first 750ís had extra reinforcing ribs in their cases, for racing they had to be built up with weld before final machining. Finally, all-new cases were introduced which are used in racer and street machines alike.
The 500ís close-ratio transmission was near-perfect when properly shimmed and run-in, needing only the lightest dab for a positive shift, but gear-tooth and dog life wasnít so hot. Better material and surface finish would have solved this problem; in the H2-R, it got both. A more expensive alloy is used, and the teeth are finished by grinding, hot hobbing.
The H1-R cranks had run as long as 1500 miles in 1970, but in 1971, mysterious troubles ruined crank after crank. Another approach to strengthening the cages was tried, reducing the number of rollers from eighteen to thirteen, with one per slot. This was stronger than the street design, with fourteen rollers, and it didnít have two rollers rubbing each other the wrong way. It was very successful, and the design was enlarged for the H2-R.
Early 500ís had some main-shaft failures originating in the fillet where the shaft joins its flywheel. Later cranks had larger fillets, and when the 750 was designed, it had immense main shafts with extra-large polished fillets. Breakage, then, became non-existent.
While street 500ís had the radical new battery/inverter CDI (Capacitive Discharge Ignition), the racer used a points system more familiar to mechanics at that time. In 1971, the H1-RAís pioneered a magneto-type CDI of clever design, incorporating a complex advance curve. The direct descendant of this system fires the H2-R, and street 750ís are equipped with CDI, though of a different type.
The rubber-mounted carbs of the 500 showed that fuel-bowl frothing could be overcome, and now the street 750 has such mounts. The big race motor shakes more, so its carbs are joined by a plate and bar which makes them vibrate in unison, if at all, equalizing mixture to all three cylinders.
Each companyís designers have their own particular styles and special policies. Kawasaki likes to use the same part in several models, and to use the same production equipment to size several different parts because itís good economics. The 750 received a wristpin of the same diameter as that of the 500 and 250 before it. Many gears from the 250, 350, 500, and 750 models are interchangeable, or have only slight differences, being made on the same center-distance and often to the same pitch module.
Performance and reliability are only useful together. Thereís no point in providing enough port area to peak at 10,000 rpm if the mechanical parts canít handle the load. Thereís also no use in entering racing with a sure finisher that canít get into the top ten no matter who rides it. Kawasaki favors high engine speeds over sophisticated breathing and high brake mean effective pressure in this design. Why? The central idea of Class-C is to race developed versions of street motorcycles. The H-2 was built to a price, so that it could reach a large market. The least expensive way to build the motor was with the ignition on one end of the crank and the primary drive on the other. Thus, eliminating the expense of idler shafts and extra gearing. The power section was caught in between, and had to shrink so they wouldnít have to put "cornering wheels" on the engine. The only elements that could be reduced were the transfers, so they are rather small. Based on considerations of time-area, they are just right for its street peak 6800 rpm, but for the racing target of 9000 rpm, they are tiny indeed.
One possibility was to keep the transfer ports open longer, thus extending the timing. This pushed the exhaust port up as well, because at that high rpm, there must be approximately 35 degrees between the opening of the exhaust and the opening of the transfers. Otherwise, pressure in the cylinder will be high enough to blow down into the crankcase when the transfers open.
The earlier the exhaust open, the higher the exhaust temperature and the greater the heat transferred to the exhaust side of the piston and cylinder. This is the situation that made the H1-RA, with its exhaust at 79 degrees ATDC, so prone to sticking.
Okay, make the port wider. That will give more time-area to the blow-down, or pre-exhaust phase, without raising the port. AMA rules required stock cylinder castings, so they couldnít put a divider in the port to prevent ring snagging. Some widening was possible, without a divider, but not muchójust a few percent.
So, the limits on exhaust port width and height limited the transfer time-area increase that could be achieved by raising the transfers. Widening the latter created a situation in which the port was bigger at its outlet in the cylinder than down in the crankcase, where itís limited by the narrow cylinder spacing. This wouldnít work either.
What else could they do? Stuck with the small transfers, the next reasonable approach might be to raise the scavenge pressure by reducing crankcase volume. Again, they were up against a wall of physical law. In order to flow a crankcase full of mixture in the short timing allotted, the velocity through the small transfer windows had to become very large, so much that the mixture streams would curl up through the cylinder and out the exhaust well before the piston covered the opening. You might say, "Yes, but I thought the expansion chamber was supposed to send a wave that blocks the exhaust port, recovering some of the lost charge." The blocking wave has a duration and strength that are determined by, among other things, what size pipes will fit under the bike, and it canít make up for such an extreme situation. The secret crank is not a full-circle design, and it surprises people to see that the racer is the same. Special full-circle cranks were in fact built for the higher scavenge velocity only reduced power, so they were put on the shelf.
Hemmed in by these limitations and the AMA rule on stock head and cylinder castings, the only out was to raise rpm, and thatís what they did. The standard dyno curve for the race motor shows 100 bhp as the peak reached at 9000 rpm, while in service, engines frequently peak as high as 9500 rpm, and sometimes see 10,000 rpm in the lower gears. Instead of fewer deep breaths through large, fair transfers, it must take 9500 rpm smaller huffs through its little ones.
If you look at the exhaust pipes of an H2-R, you will see that they arenít like Kawasakiís past efforts. Gone is the very large diameter dwell section and the strikingly long tailpipe. Gone, too, is the characteristic short, steep convergent cone, the baffle that generates the blocking wave. For the H1-R, the big pipes were considered important enough to sacrifice handling to make room for them under the frame. Why the change? Because since the 750 already has too much pressure drop across its transfers, it doesnít need the strong suction wave that a large diffuser generates. With the smaller diffuser, the dwell section diameter becomes smaller. Since the bigger motor could be more powerful than the 500 and still afford a desirable wider powerband, there was no need for the steep, short final come of the narrow-band engine. The H2-Rís longer cone used up length that would have appeared as tailpipe, or stinger, so that part has shrunk several inches.
This wider powerband was welcome to all the riders, since the hardest thing in riding the 500 was to watch a 750 BSA jet away from you off a slow corner. With the 750ís milder tune, it didnít need the largest carb that its displacement would indicate, so the 35mm size was retained from the 500. Three types of carburetors are used on the various versions. In the first year, special smooth-bore Mikunis, much like Amal GPís, were created on the theory that more air could flow at the same pressure drop than with the standard type, which has a bore enlargement to guide the throttle slide. Velocity across the fuel orifice is what atomizes the fuel, it was thought, but it turned out that the turbulence created by the bore enlargement of the standard type more than compensated for any local velocity drop. Midrange was better with the old 500 carbs, so for 1973, new specials were produced, this time magnesium replicas of the standard carb, using the same tuning parts, and having a bell-mouth extension of about 40mm. These gold-colored wonders reputedly cost as much as some motorcycles, but were worth it because they were very progressive and extended the usable power range.
The new ignition, developed from that of the H1-RA, had a four-pole N-S-N-S permanent magnet rotor supplying magnetic flux to a stator of six poles. Alternate poles were wound with fine-wire coils for charging the capacitors to 160-200 volts, while the others carried lower voltage coils that functioned in triggering the sparks. A clever system was incorporated by which a special circuit added the voltage of one charging coil and that of the adjacent "trigger" coil. As the magnet pole swept by, the voltage in one coil would fall as that in the next would rise. When the sum reached a specified level, triggering occurred. By varying the relationship of the coil inductance, quite complex advance curves could be obtained. The engine receives six sparks per revolution, of which three are used for ignition.
Back from the engine, a narrow, finely finished primary drive with a rubber shock hub transfers power to an improved H1-R clutch with twelve (instead of the former ten) friction discs, compressed by a 40% more powerful spring pack, makes this clutch virtually bullet-proof.
There are three transmissions made for the 750, each suited to a particular track requirement, and they parallel the latest trend in auto racing transmissions in having coarse-pitch teeth. Hard-faced forks that ride on optional standard or reverse-pattern shift drums shift the clusters. While there have been troubles with selector mechanism springs, there has never been a gear failure with this transmission.
"What frame is this?"
"Oh, this is the Mark 46."
Lines like these are provoked by the fact that Kawasaki has done more frame and handling experimentation than just about any other racing team, thus earning the 500ís frame an evil reputation. But the fact was that when it was lowered a bit, and equipped with better rear suspension units and shod with Dunlopís "soft-wall" 3.50 M tire, it became just like any other motorcycle, only faster. The problem with providing the necessary rigidity was eventually tackled in the US, and the frame that Yvon DuHamel rides is a very stable product of that work. The frame that Gary Nixon rode to three National wins this year is factory-standard. Spectators and riders alike saw this unit wiggle violently on more than one occasion, but as long as the center of mass is pointed down the track and the thing is on its wheels, all the oscillations can do is try to frighten the rider. They didnít.
The fork presently used is based on the street fork, and shares many parts with it. Some versions were made with extensive use of magnesium, but all have light-metal triple clamps.
The first of the double-disc Kawasaki front brakes was seen in a practice session at Mosport in 1971. Its moving mass was very large the, over 50 pounds complete, but work soon started on lighter alternatives. After a brief romance with titanium, development centered on the aluminum disc coated by Kawasakiís patented tungsten process, but their use was by no means universal until showing up at Laguna Seca in 1973. In the meantime, the very distinctive drilled discs appeared, in cast iron rather than the rust-free stainless steel used on the street version, which, by the way, exhibited far less adhesion. The present brake, which incorporates road calipers and pucks with dished aluminum discs, is certainly the best brake to use in AMA road racing from the standpoint of overall stopping power and low weight.
At the other end of the machine, where braking requirements arenít so severe, there is more variety. Some machines use the old magnesium drum brake with a wire-spoked wheel while others use wire wheels with Kawasaki disc brakes. But, the lightest and most-trouble-free combination is the Morris magnesium cast wheel with a Lockheed disc.
Swing arms vary, too. Every imaginable combination of rear shock length, type, angle, and position has been tried, together with swing arms from Japan or from local sources such as Harry Hunt or Brown. The conclusion is always the same as for frames: there is no one combination that pleases every rider on the circuit.
The aircraft-type refueling system that Kawasaki introduced has become something of an institution. Gas from an overhead gravity tank feeds through a large hose to a fitting that quickly couples to the gas tank. As gas goes ink, air comes out at the top of the tank through another smaller coupler. Both couplers are so constructed that no fuel can be spilled during a gas stop. The exact mechanics of how this is done in the minimum time are matters for practice by the team and correct construction of the fuel tank. Despite a lot of money being spent on imitation, no other team has made its system work as well as Kawasakiís.
The fairing of the latest H2-Rís represent current thinking in motorcycle streamlining. The area directly behind the front wheel is almost a flat surface, perdendicular to the direction of travel, and there is considerable width at the level of the front axle. It is possible that this flat surface pushes a plug of lower velocity air ahead of it, thereby reducing the drag of the front wheel assembly, which, because of AMA rules, could not be streamlined. A similar scheme was employed on WW- II aircraft to reduce aerodynamic drag of "dirty" radar dipole arrays. The fairing recently seen on Honda, Norton, and Triumph seem to show general agreement on this concept. In addition to other benefits it may have, this fairing lowers cylinder head temperature by routing a considerable amount of air through, rather than around, the motorcycle.
Itís true that there have been problems with these machines, and that piston and ignition failures have cost them more than one race. Itís also true that the much heavier Suzuki 750 racer has more horsepower, and that the 350 Yamaha water-cooled racers are nimbler, but the all-around performance of the Kawasaki has earned it eight National Titles in two years, twice as many as the next most successful contender. To counter claims that Kawasaki competed against smaller Yamahas and outdated four-strokes, I can only say that the other manufacturers knew the rules too. The thing that counts is being first when the checkered flag falls across the finish line.
Kawasaki H2-R Road Racer
Two-stroke, three cylinders at 120-degree piston port.
Bore & Stroke
71 x 63 mm (2.795 x 2.480 inches)
100 at 9000 RPM
Piston Speed at Rated Power
(3) Mikuni 35mm center-float type.
Five-speed, various ratios available.
Dry, multi-plate w/metallic friction materials.
Variable according to track (E.O. 81 Ė 86 degrees ATDC)
Variable according to track (T.O. 58 Ė 65 degrees BBDC)
Variable according to track (I.O. 75 Ė 84 degrees BTDC)
Variable according to track (2.5 Ė 3.0mm BTDC)
Exhaust Port Width
Variable from 42.5 Ė 47mm
Inlet/Transfer Port Width
Castrol R-30 at 30:1 in fuel plus pump at 55 cc/hr/cyl.
100 octane (by motor method).
Crank Main Bearings
Roller or ball, NTN EC type.
Connecting Rod Big-end Bearing
Special NTN high-speed steel needle type in silver-plated cage.