Engine dyno testing has become something of a black art. But if you don't understand it, how can you believe in it? More importantly, how can you trust the results? One way is by knowing how to analyze the data and spot "anomalies." Today's computerized SuperFlow engine dynos have become the industry standard and are capable of generating staggering amounts of data.,. but as with any tool, they're only as good as the operator who uses them. With the help of SuperFlow, Kevin McClelland at Flowmaster, and John Baechtal at Wes-tech Performance, we've developed this sheet-reading primer to help the average car crafter understand the test parameters that go into dyno testing and how to spot trends that could reveal whether you're getting faulty data. While some of the methodology is unique to the SuperFlow dyno, the general principles apply across the board. 1. Test TypeIn an acceleration test, the engine is accelerated or decelerated at a preselected rate (300 rpm/second is common). While acceleration testing saves time and more closely mimics how your engine behaves under actual operating conditions, the engine's recorded torque output is reduced due to the inertia of both the engine and the dyno system. The higher the acceleration rate, the greater the amount of change in torque output.In a step test, the engine is stabilized at each selected rpm data point (typically 250- or 500-rpm increments). The dyno-control software runs the engine at each test step and only records data when the speed is within 20 rpm of the set speed for at least 1 second. Data is then taken, and the test moves to the next step. Step testing is harder on the engine but generally yields more repeatable and accurate data, usually produces higher numbers than acceleration testing, and is extremely useful in tracking the effect of small changes.Whatever the chosen test method, one test type, test increment, or test rate cannot be directly compared to the data derived from a different test regimen. In fact, since every dyno facility is slightly different, results from one dyno facility should not be directly compared against results obtained at a different facility.2. Vapor PressureHumidity or vapor pressure is an important factor in calculating the corrected horsepower and torque values. An abnormally high vapor-pressure figure will inflate the torque and power numbers. Cool days typically produce 0.3-0.4 vapor-pressure readings. You'll likely see 0.6-0.7 vapor-pressure numbers only on hot, humid days; 0.9 or higher would occur only if it's raining cats and dogs.3. Engine TypeUnless you're testing a dirt bike, the listed engine type always should be "4-Cycle Spark." This affects the friction-horsepower portion of the correction factor.4. Fuel Specific GravityA hydrometer is included with every SuperFlow dyno. Knowing the fuel's specific gravity is important because it affects the accuracy of the BSFC readings, not to mention optimum carburetor calibration. Gasoline specific gravity typically ranges between 0.695 and 0.750. Today's lighter oxygenated pump-gas would be at the low end of the scale.5. Barometric PressureImputing the right barometric pressure is one of the most important factors in obtaining accurate corrected torque and power numbers. Traditionalists still rely on a wall-mounted mercury barometer, but its reading must be corrected for temperature and gravity. The latest SuperFlow 901 dynos come with an internal self-correcting barometer. Beware of abnormally low barometric-pressure readings. For example, if the barometric pressure is set to 27.00 inches Hg, it's probably bogus--unless you're testing during a hurricane.6. Engine DisplacementThis value should match the actual displacement of the test engine. Smaller engines will show less frictional horsepower loss than larger-displacement engines. On a SuperFlow dyno, frictional power loss also factors into the VE and corrected-torque and horsepower algorithms.7. Air SensorThe number in this data block equates to the diameter of the turbine fan in the "bell" above the carburetor. Imputing the correct value is important for obtaining accurate airflow information, which in turn forms part of the scfm and A/F calculations. When it comes to turbine-fan diameter, bigger is not necessarily better. A 6.5-inch turbine fan is ideal for a typical 4150-type standard-flange Holley. Dominators require a higher-capacity 9-inch turbine fan.8. RatioAs long as the engine is coupled directly to the dyno, this number should always be "1.00 to 1." Imputing a bogus number can wildly skew the corrected power and torque numbers.9. StrokeThis value should match the actual stroke of the test engine. Shorter-stroke engines show less frictional-horsepower loss than longer-stroke engines. Although not shown on this sheet, the latest software also includes input values for cylinder bore size and connecting-rod center-to-center length; this factors into the correct VE and horsepower and torque numbers.10. SpeedThe engine speed, in revolutions per minute (rpm), at which all the readings in the same row were taken is shown in the first data column.11. Torque (Uncorrected)The engine's uncorrected torque output is usually expressed in pound-feet (lb-ft). A comprehensive dyno report should show the uncorrected (raw) torque and horsepower numbers, not just the corrected numbers. If the uncorrected to corrected figures differ by more than 50 numbers, something's not right.12. Power (Uncorrected)The engine's uncorrected power output is expressed as horsepower (hp) in the U.S. Dynos measure torque, then horsepower is computed using the classic equation:Horsepower = rpm x torque divided by 525213. Mass Fuel FlowThe mass fuel flow through the engine at each rpm increment is expressed in pounds/hour (lb/hr). Making more power with less fuel is a sign of an efficient engine. The general rule of thumb is that you need 0.5 lb/hr fuel flow per 1 hp.14. Mass AirflowSuperFlow dynos monitor the airflow volume and transmit it to the central processing unit where the data are processed and calculated with the local barometer, vapor pressure, and carburetor air temperature (CAT) to provide mass airflow data in standard cubic feet per minute (scfm). The scfm measurement is what the airflow would be if the atmospheric conditions were measured at a barometric pressure of 29.92 inches Hg, 60 degrees F, and with no water in the air (dry vapor pressure). Generally, an engine will consume approximately 1.25 scfm of air per horsepower at peak torque while using approximately 1.4 scfm at peak power. The sheet notation "A1" indicates only one air turbine was in use for this test.15. Air/Fuel RatioMass fuel flow divided by mass airflow yields the engine's air/fuel ratio. When dialing in the carb for best power, the tuner should strive to keep the air/fuel ratio between 12.8:1 and 13.2:1. Slightly exceeding these parameters at several data-collection points is not uncommon.16. Brake Specific Fuel ConsumptionBrake specific fuel consumption (BSFC) is a measure of how many pounds of fuel it takes to make 1 hp for 1 hour (lb/hp-hr). A 0.5 BSFC number is considered the norm for an average, normally aspirated, carbureted engine. The lower the BSFC number, the greater the engine's efficiency (but the number should not be so low that the engine leans out). Typical well-tuned race engines get down in the "low 4s." On the other hand, supercharged and nitrous engines usually are run rich to play it safe, so it's not unusual to see "high 5s" or "low 6s" in these applications. In the absence of custom air-bleed mods, it's not unusual for Holley carbs to go rich on the top end.17. Brake Specific Air ConsumptionBrake specific air consumption (BSAC) is a measure of how many pounds of air it takes to make 1 hp for 1 hour (lb/hp-hr). The lower the number, the better: It should be in the 6s or lower if the carb is set up right; a 5.55 BSAC number is really good. However, because BSAC varies per engine displacement and power level, there isn't any hard and fast "correct" value.If both the VE exceeds 100 percent and the BSAC is above 6.0, it's a good indication that some of the air and fuel that could have helped make more power went right through the engine and was wasted.,. check for poor valve sealing or a faulty camshaft design.18. Manifold PressureExpressed in inches of mercury (in Hg), this is the classic manifold vacuum measurement. For best power, this value should be as close to zero as possible on the top end. A negative number (one that's less than zero) indicates the carburetor is pulling vacuum and has become a restriction (the carb does not flow enough air for the application). For good street part-throttle drivability, you might want to tolerate some top-end restriction.19. Engine Oil PressureThis one's self-explanatory: If you want the engine to live, you need oil pressure. There should be at least 15 psi at idle and 10 psi for every 1,000 rpm. Carefully monitoring run-to-run oil-pressure fluctuations can provide advance warning of an engine going sour.20. Carburetor Air TemperatureAn accurate CAT is a critical part of the torque and horsepower correction factor; higher temperatures inflate the numbers. In terms of getting repeatable results when making small incremental changes, try to keep the CAT within 15 degrees for each test.,. the correction factor is not perfect.21. Fuel TemperatureIf the fuel gets too hot, the lighter hydrocarbon elements in the chain evaporate, changing the fuel's specific gravity (and causing vaporlock and percolation problems). As we've seen, if the fuel's specific gravity changes, the various fuel-flow-dependent data calculations will be invalid. Optimum carb calibration is also adversely affected. Keep fuel temperature between 60 and 90 degrees ***22. Engine Oil-Out TemperatureOil temperature has a significant effect on engine power. The engine should be heated or cooled as needed to maintain oil-out temperature within 3-5 degrees F at the start of each test. For longevity and test repeatability, most engines like oil temperatures around 190 degrees F. Assuming the engine does not detonate, hotter oil and water temps will add some power, but at the expense of ultimate longevity.23. Engine Water-Out TemperatureMaintaining consistent water temperatures is important for obtaining accurate and repeatable results. The engine should be heated or cooled as needed to maintain water-out temperature within 3-5 degrees F at the start of each test. McClelland likes to start each test with the engine coolant temperature at around 170 degrees F. The water-in temp (as well as the oil-in temp) should be within 10 degrees F of the "out" numbers.24. Corrected Brake TorqueBased on formulas developed by SAE, this number is derived from plugging in the uncorrected (measured) torque numbers to an algorithm that takes into account barometric pressure, vapor pressure, and carburetor air temperature. SuperFlow dynos also factor in a friction horsepower-loss factor. Most race dynos correct the measured data to 29.92 inches of mercury, standard dry air, and 60 degrees F air temperature. SAE net-power correction factors, as used by major automotive manufacturers, use a less favorable correction factor and on average yield 5 percent lower torque and power numbers.It is important to remember that the correction procedure is really only intended to compensate for small pressure and temperature differences. In reality, it merely approximates what the power would be if you actually tested under those conditions. The correction procedure only compensates for changes in air density. It ignores coolant-and oil-temperature differences. Humidity also has a variable effect on power because it affects the point at which detonation occurs. And the theoretical correction factor can't compensate for real-world anomalies such as the engine having the same cylinder-block temperature regardless of the ambient air temperature. If the block temperature is constant, air entering the cylinder head is heated more on a cold day than it would be on a hot day!There's no doubt that corrected power yields more valid numbers than uncorrected power, but there's no real substitute for testing under the same conditions when you have to make close comparisons. For best results, try to maintain constant air temperature, fuel temperature, oil temperature, coolant temperature, humidity, and barometric pressure. For example, if the test facility is in a high-humidity area, try to test early in the morning.25. Corrected Brake PowerCorrected brake horsepower is derived from corrected brake torque, using the previously stated standard torque-to-horsepower formula. The correction-factor limitations noted above for corrected torque obviously also apply to corrected horsepower.26. Friction PowerThe estimated average losses due to internal engine friction are based on torque/cubic inch multiplied by engine displacement at each engine's piston speed. Larger displacement and/or longer-stroke engines theoretically suffer more frictional losses than do smaller, shorter-stroke engines. SuperFlow dynos obtain friction horsepower from a preprogrammed "lookup chart;" subtract that figure from the observed torque before applying the SAE correction factor, then add the loss back in to come up with the total corrected torque data as displayed on the printout.27. Volumetric EfficiencyThis is the ratio, expressed in percent, of actual air drawn into the cylinder to the maximum possible amount of air that could enter into the cylinder. If VE equals 100 percent, the cylinder is completely filled with air on the intake stroke. Supercharged or highly refined racing engines can exceed 100 percent VE. An abnormally high VE (120 percent and up) indicates the dyno operator "forgot" to input the correct engine displacement, bore, and stroke. Maximum VE should occur at or near the torque peak; this indicates the exhaust and intake tract are correctly sized and tuned for the application.28. Mechanical EfficiencyThis is the ratio, expressed in percent, of the engine's torque output to torque output plus friction torque. Run-of-the-mill engines typically have around 70-80 percent mechanical efficiency. This means that 20-30 percent of the engine's torque is lost to friction. Anything over 80 percent indicates an efficient engine. This is an assumed, not an actually measured, value. To actually calculate frictional losses, you'd have to spin the engine over using a large electric motor. The amount of electrical energy required to rotate the engine can be converted to horsepower and would represent a "real-world" friction number.
When The Flag Drops.,. The Bull ***t Stops.,. P. Engineer, Engine Builder