Bishop Rotary Valve Engine

Advanced Engine Technology The Bishop Rotary Valve by Tony Wallis, Bishop Innovation

To demonstrate the credibility of Bishop Innovation’s new rotary valve technology it joined forces with Mercedes-Ilmor to develop the technology for use on their V10 Formula One engine only to have its strategy destroyed by a change in engine regulations.

10% power By the early 1990s Bishop Inno- offered by the rotary valve. Fur- ney Australia. Bishop was respon- advantage vation had completed the initial ther, a successful public demon- sible for the cylinder head design, and improved development of its promising new stration of this technology in the development and demonstration durability rotary valve concept for IC en- extreme operating conditions of of the required durability and per- gines and was looking for a way F1 provided a mechanism to ad- formance. By late 2000 back to to further develop the technology. dress the industry’s prejudice. back testing with the poppet valve The automotive industry, having In 1997 Bishop started work- single cylinder engine demon- observed a succession of failed at- ing with Ilmor Engineering (later strated a 10% power advantage tempts spanning the last century, Mercedes-Ilmor) to develop their and improved durability. In 2002 no longer believed the rotary rotary valve technology for F1 the first V10 engines using this valve concept was mechanically engines. The initial development technology were built and tested viable. It chose Formula One (F1), was carried out on 300cc single exhaustively. A completely new as the criteria for success was well cylinder bottom ends supplied by V10 engine was designed and matched to inherent advantages Ilmor at Bishop’s premises in Syd- manufactured in 2003. Testing of these engines was prematurely terminated when the FIA an- nounced changes to Article 5.1.5 Figure 1: of the engine regulations late in Valve drive train 2004 with the specific purpose of and inlet tract banning this rotary valve tech- arrangement on nology. V10 engine. The Technology The F1 rotary valve assembly is shown in Figure 2. The Bishop Rotary Valve (BRV) is an axial flow rotating valve incorporating both the inlet and exhaust port in the same valve. There is one valve per cylinder positioned with its axis perpendicular to that of the crankshaft. The steel valve is sup- ported in two shell type needle roller bearings that ensure the valve’s stepped centre portion al- ways runs with a small radial clearance to the housing. The outside diameter of the valve’s centre portion and the bearings are similar, allowing the valve assembly to be housed in a stepless bore. Face seals located at each end of the valve’s centre portion prevent cooling and lubrication oil enter- ing the centre portion and the cylinder. At the exhaust end a carbon face seal keeps the oil and

2 Motorsport Technology Advanced Engine Technology

Figure 2: F1 rotary to accommodate local distortion, valve assembly. resulted in high friction and seizure. In the BRV arrangement the small radial clearance between the valve’s periphery and its housing is designed to ensure that any thermal or mechanically in- exhaust gas sep- duced distortion of the valve is arate while at the in- accommodated without the let end a lip seal separates valve’s periphery ever touching the inlet air and oil. The valve is its housing. Provided the radial driven by a gear at the inlet end. exposed to the clearance is kept small it provides The outside diameter of the combustion process sufficient flow resistance to pre- valve generally lies in the range it is inevitable that thermal and vent significant flow of gases be- 0.67 - 0.74 that of the cylinder mechanical distortion of the valve tween the exhaust and inlet port. bore diameter allowing it to be will occur. Bishop focused on In applications where fitted to engines with convention- finding a solution that would this is not the case al cylinder bore spacings. allow this small but inevitable The valve rotates at half en- distortion to be accommo- gine speed and eliminates the in- dated. In arrangements Figure 3: ertia induced forces that have where both the inlet and Gas sealing plagued the development of re- exhaust port are in the arrangement. ciprocating poppet valve mecha- same valve, a satisfacto- nisms since the invention of the ry solution is complicat- IC engine. It is this feature that ed by the requirement has inspired numerous inventors to prevent leakage be- over the last century to chance tween these ports. A their hand at developing rotary typical previous ap- valve engines. These develop- proach was to use a ments have generally failed due stationary split sleeve locat- to a combination of problems in- ed around and lightly preloaded volving gas sealing, oil sealing, against the valve’s periphery. In excessive friction and seizure such arrangements it was very caused by thermal and mechani- difficult to create an even distri- Bishop has developed additional cal distortion of the valve. bution of oil between the valve technology to control these flows. As a portion of the rotary surface and the sleeve. This, com- This approach was immediately valve’s periphery is periodically bined with the sleeves poor ability successful and allowed the early development to proceed without seizure, friction or lubrication problems. The gas sealing arrangement is shown in Figure 3. It consists of 2 axial seals and 2 circumferential seals located adjacent the window, housed in slots in the cylinder head and preloaded against the periphery of the valve. These seals function in a similar manner to piston rings. Unlike the piston ring they are subjected to a con- stant sliding velocity with no re- versal of direction throughout the cycle. This arrangement allows the use of very long windows, an essential requirement if the breathing potential of this concept is to be achieved. Window length determines the opening and closing rate of the valve and the fully open flow area. The opening and closing rates on modern F1 engines are very high Figure 4: CFD of production valve showing oblique flow through window and the and can only be matched by ro- start of tumble flow. tary valves which have window

Special AutoTechnology 3 Advanced Engine Technology

peak flow capacity of these valves increases linearly with valve diameter. By selecting a suitable valve diameter, peak VE can be achieved at what ever speed is required. CFD simulations suggest that valves could be produced that enable peak VE to be main- tained at speeds up to 25,000 rpm. Further increases are unlikely as it is difficult to make the inlet tract short enough to achieve the cor- rect wave action. Whilst the rotary valve engine has demonstrated breathing capacity that is at least equivalent to the best F1 poppet valve engines, it has the huge advantage that it does this without the dra- matic reduction in life that occurs with F1 poppet valve heads. As inertia induced loads in the valve Figure 5: Single Cylinder BRV engine on Bishop’s 18,500 rpm dynamometer. train are absent in the rotary valve, the forces that destroy the poppet valve heads are also ab- In axial flow ro- lengths greater than 0.77 that of tumble flow. The tumble ratio on sent. Further the mechanical and tary valves the air the cylinder bore diameter. engines with near square gas loads seen by the valve are flow through most In axial flow rotary valves the bore/stroke ratios is typically essentially independent of speed. of the inlet tract is air flow through most of the inlet twice that reported for similar 4 The only issue affecting durabili- parallel to the tract is parallel to the window and valve engines. Unlike the poppet ty that changes with engine speed window and per- perpendicular to the cylinder axis. valve this high tumble flow is is the peripheral speed of the seal- pendicular to the This is quite unlike poppet valve generated without any loss of vol- ing elements and the bearings. As cylinder axis arrangements and often leads their umetric efficiency (VE) and is re- the peripheral speed of the valve designers to incorrectly conclude sponsible for very fast burn rates over the sealing elements is ap- that the rotary valve cannot observed. Production based en- proximately 80% that of the max- breathe as well as the poppet valve. gines built in the early 1990’s had imum F1 piston ring velocity, this In the Bishop design much of ignition timing of 15°, or less than is of little concern. In production the air passes obliquely through half that of the best four valve en- poppet valve engines, engine life the window. This is well illustrated gines. considerations require changes in the CFD image shown in Figure that greatly curtail their breathing 4. The air flow adjacent the port The Testing capacity from the level achieved floor is turned through approxi- in F1. This is clearly not the case mately 35° before it passes Extensive back to back testing with the rotary valve and Bishop’s through the window while the air demonstrated that the peak volu- research suggests a production adjacent the port roof is turned metric efficiency was the same for rotary valve has a breathing ca- through approximately 90°. As all both poppet and rotary valve en- pacity up to 45% greater than the flow must at some stage be gines. The difference, if any, oc- that observed on current 4 valve turned parallel to the cylinder ax- curs in the engine speed at which production engines. is the remaining rotation takes the VE first starts to fall when op- The numerous BRV engines place in the cylinder itself. While timum length tracts are fitted. The tested over the last 18 years have this oblique flow suppresses the 10% power advantage observed all demonstrated remarkable re- discharge coefficient the unob- on the single cylinder BRV en- sistance to engine knock. In the structed flow path through the gines in 2000, was in part a result early 1990s engines with conven- window tends to compensate. of the rotary valve being able to tional bore/stroke ratios ran com- While the discharge coefficient maintain peak VE to higher speeds pression ratios as high as 15:1 on (based on the window area at the than the poppet valve. By 2004 unleaded 93 octane pump petrol. combustion chamber surface) of the poppet valve and rotary valve The F1 single cylinder engine ran early valve designs was only 0.54 engines were both able to main- compression ratios as high as (fully open valve) steady develop- tain peak VE to engine speeds 17:1 using standard F1 fuels be- ment has seen this improved to greater than 18,000 rpm. fore settling on 15.3:1 as opti- 0.72 on current F1 heads. During the course of the BRV mum. No evidence of knock has This oblique flow through the development valve diameters be- ever been observed and this is window is responsible for one of tween 58 and 70 mm have been thought to arise from an absence the rotary valves most useful at- manufactured and tested. Flow of any hot surfaces in the com- tributes - its strong in-cylinder testing has demonstrated that the bustion chamber (the valve sur-

4 Motorsport Technology Advanced Engine Technology

face moves continuously through On the power front both the This technology the combustion chamber) and the rotary and poppet valve engines has potential in very fast combustion rates. Bishop were designed for the same maxi- production anticipates rotary valve produc- mum speed and both engines pro- applications tion engines could run compres- duced near identical power out- sion ratios as high as 15:1. puts despite far greater resources being devoted to the development of the poppet valve engine. From Formula One Rotary Valve a F1 perspective the rotary valve Engine had the potential to increase power faster as it could increase its A schematic showing details of the breathing capacity by merely in- valve drive train and positioning creasing the valve diameter and of inlet tracts and injectors for the unlike the poppet valve it had no V10 engine is shown in Figure 1. inertia issues preventing its oper- A section through a single Figure 6: Section through Rotary Valve ation at higher speeds. cylinder version of the F1 cylinder Cylinder Head Assembly. With advantages in height, head assembly is shown in Figure weight, C of G, elimination of in- 6. The cylinder liner is cast inte- ertia loads, breathing and greater gral with the cylinder head thus stiffness of the head. durability Bishop is confident eliminating the need for a head As most of the weight saving that, had the technology not been gasket (one of the weakest ele- came from the top half of the en- banned, it would have become the ments on a poppet valve F1 engine there was also a considerable technology of choice for racing gine) and for cored water passages reduction in the engine’s centre of engines. in the cast cylinder blocks. In the gravity (C of G). The cylinder This technology has consider- cylinder head there is no require- head height was reduced by ap- able potential in production ap- ment for the complex water proximately 50mm producing a plications where it has the addi- gallery cores typically found in strikingly compact engine. This tional advantages of fast the poppet valve head. Apart from was a significant achievement as combustion, low valve drive some simple water transfer gal- F1 engines are already remark- torque and high compression ra- leries the water cooling galleries ably compact compared to their tio capability. Bishop anticipated are all drilled. production cousins. Figure 7 that demonstration in F1 would The 3 litre V10 F1 rotary valve shows a comparison between a provide the impetus for the con- engine weighed in at less than Mercedes 230 SLK cylinder head siderable further development re- 80kg, making it easily the lightest and a single cylinder version of a quired to bring this technology to V10 F1 engine ever built. It rotary valve cylinder head suit- production. Regrettably F1 has weighed approximately 16kg less able for use on this size engine. abandoned its long standing rai- than the equivalent poppet valve The height saving is 150mm. son d’être to “improve the breed” engine with which it shared a sim- Bishop estimates that, compared and replaced it with “protection ilar bottom end. A significant con- to current 4 valve engines, weight of the status quo”. tributor to this weight reduction is savings up to 4kg/cylinder could the integral liner and head. The be achieved on production BRV presence of a head gasket requires engines. very rigid heads and blocks and large stud loads to maintain ade- quate pressure on the gasket. Not only can smaller studs and less stiff housings be used but the geometry of the integral liner and head itself greatly increases the

Figure 7: Comparison of height between a production 4 valve cylinder head and a single cylinder version of a production rotary valve cylinder head.

Special AutoTechnology 5