(Originally published Monday, December 17, 2007)
One thing every engineer in the paddock is agreed on is that big-bang engines work. They know this because the stopwatch says so. What they don’t know is why they work. Conjecture has centred around the effect of torque pulses on the rear tire’s contact patch, leading to a widely accepted theory that closely grouped firing pulses allow the tire to slip then recover and grip over the extended interval before the next power pulse. The origins of this theory come from US dirt-tracking where firing two big Harley pistons close together within a few degrees of crank revolution gave what was called ‘the big sneeze’. Personally, I never found this argument convincing when applied to a MotoGP bike, the time frame for events to happen seems far too short. Can a construction like a tire carcass really react that rapidly?
Still, it was the best model anyone could come up with. When MotoGP engines moved away from big-bang towards long-bang by rephrasing crankshafts to spread the power pulses out slightly there was no change in traction. You would have expected some significant changes if the ‘big-sneeze’ analysis were correct. Maybe it wasn’t the tire at all. In the absence of hard data, it was all conjecture.
Or it was until Saturday night at Valencia when Yamaha actually gave us some numbers to crunch for the first time. Given that the M1 didn’t have the best of seasons, it was quite brave of Yamaha to open up like this. They started with a few comparisons between the performance of the new for ’07 800s and the dear departed 990s. These figures from Mugello are probably in line with what you would have come up with after a few minutes’ thought: the new bikes were over 6mph slower down straights but up to 5mph faster in corners. Throttles were fully open for 25% of a lap compared to just over 15% previously. Data from Jerez showed the 800s braking up to 30 metres later for corners and getting on the throttle up to 15 metres sooner, which helps to explain why lap times stayed static or came down. Incidentally, Yamaha say they improved the M1’s power output by 4% over the season, raised the rev ceiling by 1000rpm, and improved fuel economy by 3%, going through three versions of the motor.
The first only did the first two GPs with the final version appearing at Brno after the Summer break. Yamaha started the year with a top speed deficit of 4% to the Ducatis and had more than halved the gap by the end of the year. The engine management electronics could switch between three different ignition maps. Previously, the bike could be set to use whichever map was required in each gear. In 2007, it selected the right map not just for each gear but for each corner. This means the bike ‘knew’ where it was, just as the Ducati must have done when they tried out their semi-automatic gearbox in 2004. Corner counting would be too simple and would need resetting after a visit to the pits or an off-track excursion. According to one eminent engineer (not a Yamaha man), GPS is perfectly accurate enough for this. Wouldn’t want to trust it at, say, Brands Hatch to differentiate between Paddock and Bottom Bend, or Turns 1 and 3 at Laguna Seca. And Degner’s and the 130R are almost on top of one another at Suzuka!
Interesting as this performance data was, it was just an hors d’oeuvre. The main course was presented by Masao Furusawa under the title ‘What is Big Bang?’ Fursawa’s area of expertise is harmonics, so perhaps it was no surprise that he chose to use the analogy of signal-to-noise ratio to explain his theory. You understand that best from tuning your radio every day. Accurately setting your radio to the desired station means the signal comes in strongly and overpowers any background noise.
Noise is always present, what you want is a strong enough signal to render it irrelevant. So what is signal and what is noise in the context of a motorcycle engine? This is best explained by thinking about that word ‘connection’ you keep on hearing riders use in testing. This is shorthand for the connection between the throttle and the rear tire. In an ideal world, opening the throttle by 10% would deliver 10% of available power (actually torque, but never mind) to the rear tire. Life is rarely this convenient or simple, and racing engines certainly aren’t.
Modern electronics should be able to provide the linear throttle response riders crave; in Furusawa’s model a high signal-to-noise ratio. And what his research suggests is that that is what you do get—up to a critical rev level where the signal is severely distorted by ‘noise’. The question is, what is this interference? Furusawa says it is ‘inertia torque’, that is the torque due to the motion of the heavy moving parts in the engine—crankshaft, con rods and pistons. This is totally separate from the torque generated by the combustion process. At low revs, the level of interference from the rotating mass is insignificant, but around 12,000rpm it starts to become greater than combustion torque and by around 16,000 is double. This is counter-intuitive because you would assume, with a conventional 180-degree crank, that everything would balance out. Not so, as you discover when you look more deeply at the direction in which torque is exerted at different points of a crank’s rotation.
Combustion torque is easy to understand: it’s produced by ignition of the fuel/air mixture. Inertia torque is much trickier to define and understand. Let’s try. Forget combustion and just consider the piston and con rod travelling up the bore. At BDC the piston, con rod and crank pin are in line and no torque can be applied to the crankshaft (in fact at top and bottom dead centres, the con rod is momentarily stationary and vertical). Now move through 90 degrees. The big end of the con rod together with the piston is moving quickly with lots of energy and is about to decelerate to a halt at TDC. That energy of motion (kinetic energy) has to go somewhere, and the only place it can go is into the crankshaft. So inertia torque is positive in that it is applied in the direction of rotation of the crank. On the down stroke, the converse is true. The lower part of the con rod together with the piston has to be rapidly accelerated from rest at TDC to a high velocity, which requires an input of energy. That removes energy from the crankshaft so here inertia torque acts against the direction of rotation.
Without doing the math, you can see how this variation of torque over each revolution might produce some small variations in the torque seen by the tire contact patch. On your 180-crank, four-cylinder road bike, you won’t notice the effect because you don’t use high enough revs, but as this inertia torque is proportional to rpm squared, you can see how a 17,000rpm MotoGP engine might have problems. At those sort of engine speeds, the ‘noise’ of the inertia torque is ‘louder’ than the ‘signal’ of the combustion torque. The rider’s connection with what’s happening at the rear tire’s contact patch is lost both with the throttle open and with it closed.
The cure is equally counter-intuitive; an irregular firing pattern, 90- degree crankshaft. The conventional 180 crank has its two outer pistons at TDC while the centre pair are at BDC. Leave cylinders number one and three as they are then move two and four through 90 degrees in opposite directions and you have the 90-degree crank with one piston coming to TDC every 90 degrees of crank rotation. Yamaha tried firing all four cylinders in one revolution and compared the result to the more conventional firing order of two cylinder firing at a 270-degree interval in the first revolution of the crank and the other two firing just 90-degrees apart in the middle of the next revolution. The first surprise is that they sounded the same, the second is that there was no difference in traction. That effectively killed off the ‘big sneeze’ theory.
The mathematics say that inertia torque is reduced to almost zero before 10,000rpm and—crucially—to only about 3% of the 180-crank’s value at 15,000rpm. The experimental test to confirm the theory involved measuring rotational fluctuation of the rear wheel, a consequence of uneven torque delivery. With the 180 crank there are big torque spikes at all throttle openings, but with the highest peaks just as the rider gets on or off the throttle. The 90-degree crank shows no such behaviour, suggesting it would make getting into and out of corners a lot easier for the rider. Inertia torque (noise) is still there, it’s just at such a low level it doesn’t have a significant effect. Of course the first law of engineering says you never get something for nothing and an irregular firing order means vibration that may require a balance shaft or heavier components to tame, thus losing you part of what you’ve just gained.
These findings are of course all for in-line four-cylinder motors, but it’s easy to see how the 90-degree crankshaft can effectively mimic a V4—the back tire doesn’t know what direction the cylinders are pointing in! Is this an inherent advantage of the 90-degree V4 engine? Yamaha think not, and will continue with the in-line engine which they regard as enabling them to build a shorter and therefore more nimble machine. But they will have to use an irregular firing order crank.
Furusawa’s work is significant in that it is the first coherent explanation of why bigbang engines actually work despite the fact that in design terms they look like horrible out-of-balance lash-ups. And there wasn’t one mention of tire contact patches slipping and gripping. What the Yamaha team have done is define what that nebulous term ‘connection’ means when applied to motorcycle racing: it means the ratio of combustion torque to inertia torque, with a high ratio being the good connection of a contemporary four-stroke MotoGP engine and a low ratio being the distinctly dubious connection of a 500cc two-stroke being aimed out of a corner. It feels like our knowledge of how motorcycles behave has moved on another step.