Throttle by Wire and Virtual Powerband

Every ambitious dirt-track racer learns the limits of engine tuning. In olden times, the Webco catalog provided ample winter reading, listing all the many cam grinds promising to boost engine power. Peak-power cams worked by keeping the intake valves open longer after bottom center. This allowed the incoming fuel-air mixture, flowing at high speeds, to continue to coast into the cylinder longer. Excellent! Trouble was, as you tried to accelerate off turns with such power, you went slower instead of faster—less than excellent.

Why? At the lower engine rpm of throttle-up, intake velocity wasn’t fast enough to keep the mixture flowing into the cylinder after bottom center. Instead, the rising piston stopped the inflow and reversed it, pushing some of the mixture out. The result was weak—a flat spot in your torque curve, followed by a steep rise as higher revs and increased intake velocity began to work as planned, greatly boosting engine torque and doing so quickly. This sudden rise and swooping peak produced a powerband that was impossible to ride: first it left you hanging, throttle open and going nowhere as rivals passed (flat spot), then kicked your back tire loose, destroying your chance of traction and acceleration (steep rise to peak torque).

The long valve overlap usually found on “full race” cam grinds made this even worse. Overlap is the period around top center at the end of the exhaust stroke when the exhaust valves are still in their closing process, but the intakes have already begun to lift. Why do this? Contrary to popular belief, valves can’t just snap shut or pop open―those motions take time, and so the exhaust valves normally close a few degrees after top center. As for the intakes, it’s beneficial to have them open at their fullest as the piston is travelling at its greatest speed, which occurs about 76 degrees after top center (the crank throw and con-rod are at right angles to each other at that point). This motivates our cam grinder to start lifting the intakes some degrees before top center. The result is a period during which all the valves are slightly open together―the exhaust and intake valve events “overlap” each other slightly.

Related: TECH UPDATE: Why Variable Valve Timing?

Overlap boosts power in the rpm band in which the exhaust pipe is tuned to work. As the exhaust valves open, a wave of leftover combustion pressure rushes out into the header pipe. The wave expands as the header ends at the larger collector pipe, sending a pulse back toward the engine. This expansion or negative wave arrives at the exhaust ports during valve overlap. Traveling into the small space between piston (now very close to TDC) and cylinder head, this negative wave “sucks out” the inert exhaust gas remaining there and then travels out through the slightly open intake valves to start fresh charge flowing into the cylinder―before the piston has even begun its downward intake stroke. These effects boost torque because:

Sadly, that’s not the end of the story. At some lower rpm, instead of boosting torque the exhaust wave action creates a flat spot of reduced torque. Waves are constantly reverberating in the header pipes, alternating between positive and negative. At some lower rpm, it will be a positive wave that returns to the exhaust ports during valve overlap. It will push more exhaust gas back into the cylinder, out through the intake valves, and into the intake system. Then, when the piston begins its descent into its intake stroke, the first thing it draws into the cylinder will be this exhaust gas, which cannot burn. Diluting the fresh charge results in weak torque.

The traditional way past all this toil and trouble is to shorten the overall time that the valves are open (valve duration) and try to recover the loss by lifting the valves farther (increased valve lift). To get rid of the overlap flat spot, just grind your cams with less overlap! Yes, you will lose some peak, but the solution fills in the flat spot enough to make the torque rise as the engine revs up, and makes the increase gradual enough that a human rider can handle it.

Now, when you feed the engine some throttle, there is prompt torque rather than a flat spot―you accelerate with the others. And as torque rises, it won’t rise fast enough to kick your back tire out from under you. You’re in the game.

What does this have to do with throttle by wire? All the riders who have been through the above experiences have had the same thought: “Man, if I could just somehow learn to ride this thing, nobody would see which way I went!”

Ah, but there is a way, and it’s called the virtual powerband. The actual powerband of a highly tuned engine is an alpine landscape of steep peaks and valleys. So instead of connecting the rider’s twistgrip to the throttle slides or butterflies with a steel cable, we connect it to a TPS or throttle position sensor. This generates a signal that reports throttle-grip angle.

Related: Are Locomotives Doing It Right?

What the rider wants is not a certain amount of throttle opening, but rather a certain amount of torque. But if torque is rapidly changing as a function of engine rpm, the same throttle angle at 7,600 gets you a very different amount of torque from what that same throttle angle gives at 8,000. Things can in fact happen so fast during acceleration that a rider finds it necessary to turn the throttle rapidly backward during acceleration just to keep the rear tire from snapping loose.

People familiar with electronic controls―sensors, computers, digital stepping motors―got interested in this problem. “What if we interpret the rider’s throttle angle not as a demand for a specific throttle opening of carburetors or fuel-injection throttle bodies, but as a demand for what the rider really needs to control―engine torque?

On a dyno, we map the engine’s torque output as a function of rpm and actual throttle opening, storing that information in a torque map. Now, when the rider rotates the throttle grip a certain amount the computer interprets that as a demand for that much torque. It looks up in its map how much slide lift or butterfly angle is required to give that torque, and the stepper motor moves the throttles to achieve that. If the rider holds a constant throttle-grip angle as the engine accelerates, the computer continuously adjusts the engine throttles to give that constant torque―if we could look at the throttle butterflies as the engine accelerated, we’d see them constantly in motion.

This changes everything, because now if there is a dip in the engine’s torque curve, the computer opens the throttles enough to fill it in, keeping torque output constant. If “natural” engine torque is rising very rapidly in a certain rpm band, the computer rapidly moves the throttles to hold torque at the value chosen by the rider.

If you’d like to see a system like this in operation, check out some films of aircraft making arrested landings on carriers. The pilot, by movements of stick and rudder pedals, chooses a path for the airplane as it approaches from the stern of the ship, but we see the control surfaces making large motions as the flight control electronics of the airplane do whatever is necessary to stay on that chosen path. The pilot chooses the path and the flight control makes the control movements necessary to remain on that path.

Related: Special Feature: Throttle by Wire

“Virtual powerband” is the term chosen by Formula 1 engineers to describe the process by which control electronics move the engine’s throttles to deliver smooth torque to the wheels rather than a series of torque peaks and valleys. This is even more desirable for motorcycles: The grip from their small tire footprints is more easily upset than is that of the giant rubber rollers of racing cars.

With the virtual powerband turned off, the engine’s power curve can be all peaks and valleys, but with it turned on, the valleys are mostly filled in and the traction-destroying peaks are trimmed down. The result is much closer to the ideal―the horizontal flat line of constant torque at all rpm.

Within limits, this makes it possible to tune engines more radically, yet still have a package that is comfortably usable by a human rider. But if you take away one thought from this story, let it be this: Virtual powerband cannot increase power―it merely makes it possible for riders to use the power of more highly tuned engines.