by David Vizard
An engine’s BSFC is almost certainly one of the most misunderstood numbers, even by professionals, in the engine performance field.
In my experience, novice enthusiasts most commonly confuse and misinterpret two numbers that are virtually joined at the hip: torque and horsepower. And it is surprising that even big-time race engine builders with big-time race wins still fail to fully understand what I perceive as the third most important number: brake specific fuel consumption.
Let’s start with what it is not. BSFC is not a measure of the mixture ratio. I sometimes hear engine builders say they like to see the BSFC in the high 0.3s to low 0.4s, inferring that this is better than, say, 0.28.
While lecturing at UNC Charlotte some years ago, I asked a highly successful NASCAR Cup Car engine builder what he thought were optimum BSFC figures. He was under the impression there was an optimum somewhere in the 0.3 to 0.45 range, so I posed the question to the entire audience.
This Chevy 350 small-block was built for the magazine story “Son of Sledgehammer,” which appeared in the September 2007 issue of Popular Hot Rodding. Brake specific fuel consumption was good because the numbers were low. However, if the low numbers had been interpreted as a lean mixture (about 15 hp of the 470 hp), the engine produced would have been jetted out!
Of the 115 attendees (standing room only in a 100-seat auditorium), about 100 were engine builders, one was a member of the press, and the rest were students. Of all the answers given only the press guy, Johnny Hunkins, editor of Popular Hot Rodding, got the answer right.
The optimum BSFC is zero. This means you have an engine that is producing power and using no fuel to do it. Let’s be honest here, you can’t get more efficient than that.
Let’s consider the individual words. First it is called “brake” because it is measured on the brake. This is an old term for a piece of hardware now more commonly called a dyno. It is “specific” in as much as it applies specifically to the weight in pounds (lb) of fuel needed to generate one horsepower (hp) for one hour (hr). This is commonly written as lb/hp/hr and is a direct measure of the amount of fuel, in pounds per hour, consumed by each individual horsepower of the engine being tested. That makes the number useful across the board as a means of comparing the fuel efficiency of engines regardless of their displacement and specifications.
The truth is that the BSFC figure can change while the mixture is unchanged. If that is the case, the BSFC figure is hardly applicable or reliable as a gauge of the mixture ratio.
Let’s look at an example: Assume we have a test engine that makes 500 hp at 6,000 rpm, and it loses 100 hp (in pumping losses and internal friction). Attached to this engine is also a means to simulate added internal friction, achieved by means of a disk brake on the front of the engine. This simulates the loss of another 100 hp in internal friction.
The engine is dyno’d and, with a 13:1 air/fuel ratio, it uses 200 pounds of fuel per hour. This means the BSFC is 200 pounds per hour of fuel divided by 500 hp, which equals 0.40 lb/hp/hour. So for every 1 horsepower generated, the engine uses (in one hour) 0.4 pound of fuel.
If the brake at the front of the engine is applied to simulate a 100-hp loss then, at the same 6,000 rpm, we see only 400 hp, as measured at the flywheel by the dyno. The induction system has no idea that power is being absorbed by something other than the absorber on the end of the flywheel so it continues to feed an identical amount of fuel (200 pounds per hour) and mix it with an identical amount of air. Since neither the flow of fuel nor air has changed, the air/fuel ratio is still exactly 13:1.
I built this hydraulic-roller valvetrain, Chevy 383 small-block test engine in 2007. The way to get the mixture right was to use a wide-band oxygen unit. The best option was to have one in each header pipe, but failing that an excellent job can still be done with one in each collector (as shown here). This 10:1 engine made more than 530 hp and easily eclipsed the 500 ft-lbs mark on a hydraulic roller and out-of-the-box AFR 200-cc port heads. Additionally, the engine turned flawlessly to 6,700 rpm.
The same cannot be said for the BSFC, however. It is 200 pounds of fuel per hour divided by 400 hp. That comes out to 0.50 lbs/hp/hour. We have made no change in the mixture but the BSFC has changed by 20 percent. So we can, with total certainty, say that the BSFC is not a measure of the mixture ratio!
Now some old-timers who have been looking at BSFC to determine if the engine is rich might be pleased to know that one of the symptoms of poor (i.e., big number) BSFC is an overly rich mixture. However, poor BSFC could also be the result of incorrect ignition timing or a number of other reasons.
A really good spark-ignition piston engine can get down to the low 0.3s. An example is the 18-cylinder Wright Cyclone compound turbo engines used in the last derivatives of the Lockheed Constellation airliners of the late 1950s. They had a BSFC of 0.32. Getting much below that seems hard to do, but there are developments in the pipeline that could crank the barrier down to an estimated 0.28 (see page 68 for a further discussion).
Today, the best way to measure mixture with the minimum of unaccountable error is by a wide-band oxygen system. I have extensively used the system by Innovate Motorsports on the dyno.
This chart illustrates the torque and horsepower curves of an engine equipped with a 4500 Series Super Victor and a 4150 Series Super Victor manifold.