Home
Hydraulic dynamometer
Engine testing / data interpretation
Read an engine dyno test report in this order: confirm the engine configuration and test conditions, check the measured channels and units, then evaluate torque, power, fuel consumption and thermal behavior at matching operating points. A peak-power figure alone cannot establish whether an engine passed its test or whether two runs are comparable.
This guide is for development engineers, engine rebuilders and quality teams reviewing shaft-output tests. It explains the relationships behind a dynamometer graph and the questions to ask before accepting the data. For the underlying measurement boundary, start with our guide to what an engine dynamometer measures.
01 / ENGINE TESTING GUIDECore measurements and engineering units
Start with the report header
Before reading the curves, identify the engine, its configuration and the test method. Record the engine identifier, calibration or software version, fitted accessories, fuel specification, intake and exhaust arrangement, and whether the test used stabilized load points or a speed sweep. These details determine what the numbers describe.
An engine tested without its installed cooling fan is not automatically comparable with one tested with that fan operating. Likewise, a repaired engine evaluated against an OEM acceptance schedule should not be judged against an unrelated development sweep. The required configuration belongs in the test specification, not in an assumption made after the run.
Check the date, run identifier, operator record and instrument configuration. A summary exported without its raw-data reference is difficult to investigate when a value looks unusual. The report should let a reviewer trace the displayed result back to the recorded test, including any excluded intervals or annotations.
Separate measured and calculated channels
| Report field | Typical unit or format | What to establish before interpretation |
|---|---|---|
| Shaft torque | N·m or lbf·ft | Sensor location, zero reference and calibration record |
| Rotational speed | rpm | Actual measured speed, not only the commanded setpoint |
| Shaft power | kW or explicitly defined hp | Torque-speed calculation, averaging and correction basis |
| Fuel consumption | kg/h, g/s or volume per time | Mass or volume basis, fuel density and return-flow treatment |
| BSFC | g/kWh | Fuel mass divided by brake output on a consistent basis |
| Temperature and pressure | °C or K; kPa, bar or other stated units | Measurement location and absolute or gauge pressure |
| Test conditions | Time, ambient conditions and configuration | Matching conditions, acquisition settings and test sequence |
Torque and speed provide the basis for calculated shaft power; fuel and other instrumentation add context. The Hance control and data acquisition system integrates configured measurement equipment through HD-3000 and Dyn Pro. A particular channel still requires the appropriate instrument, interface and configuration; a field name in a report does not prove that it was physically measured.
If a channel comes from the engine controller, label that origin. An ECM estimate and an independently instrumented measurement may answer different questions. Keep their names distinct rather than presenting both as interchangeable readings of the same sensor.
02 / ENGINE TESTING GUIDETorque-power-speed relationships
Use the equation as a consistency check
Power is torque multiplied by angular speed. HORIBA's technical article on engine testing describes shaft torque and rotational-speed measurement as the basis of engine power calculation.
Power (kW) = Torque (N·m) × Speed (rpm) ÷ 9549.3
Power (mechanical hp) = Torque (lbf·ft) × Speed (rpm) ÷ 5252.1
These are unit relationships, not performance specifications. Use values from the same operating point and processing basis. Multiplying maximum torque by maximum speed is not valid when those values occurred at different points. Also distinguish mechanical horsepower from metric horsepower; an unlabeled “hp” field leaves a unit ambiguity.
For averaged data, ask how power was calculated. Averaging the instantaneous torque-speed product is not necessarily identical to multiplying separately averaged torque and speed when the signals vary. This distinction matters more during changing conditions than at a well-stabilized point.
Do torque and power curves have to cross?
With mechanical horsepower and torque in lbf·ft, their numerical values are equal at approximately 5,252 rpm. A plotted intersection there follows only when both axes use matching numerical scales and the data share the same basis. Different axis ranges can move the visible intersection or remove it altogether.
For kW and N·m, numerical equality instead follows at approximately 9,549 rpm under the same mathematical conditions. Neither crossing is a special combustion event. A graph outside that speed range is not defective simply because the curves never meet.
Read each axis label before interpreting the picture. If one trace is corrected and another is uncorrected, or the speed axis represents a different signal, visual comparison may be misleading even when the drawing looks tidy.
Compare the useful operating region
Peak torque and peak power can occur at different speeds because power depends on both torque and speed. For an engine used at a defined working speed, the relevant question is often its output and stability at that point, not the largest value anywhere on a sweep.
Compare runs at matched speed, load and configuration. Review the complete region required by the test plan, including any low-speed or part-load points. Do not convert engine shaft output into wheel output with an assumed fixed loss percentage; our engine dyno vs chassis dyno comparison explains why their measurement boundaries differ.
03 / ENGINE TESTING GUIDEFuel consumption and BSFC interpretation
Confirm what the fuel channel measures
Brake-specific fuel consumption, or BSFC, expresses fuel mass consumed per unit of brake output energy. With fuel mass flow in kg/h and shaft power in kW, the relationship is:
BSFC (g/kWh) = Fuel mass flow (kg/h) × 1000 ÷ Brake power (kW)
The word “brake” refers to useful shaft output measured at the test boundary. It does not mean the quantity is a braking-system measurement. Fuel flow alone is not BSFC: it must be paired with the corresponding brake power.
For engines with a fuel return, establish how the instrument determines net consumption. If the result is based on volume, identify the density and temperature basis used to convert it to mass. A supply-flow reading that ignores returned fuel cannot simply be labeled engine fuel consumption.
Lower is meaningful only on a comparable basis
Lower BSFC means less fuel mass per unit of shaft output energy on the stated basis. It does not, by itself, establish overall efficiency across different fuels with different energy contents. Fuel identity remains part of the comparison.
Compare like-for-like speed, load, fuel and thermal conditions. A full-load sweep does not describe the engine's entire part-load fuel-consumption map. Nor should an apparently favorable value override a missing fuel-calibration record, a changing operating point or an incomplete return-flow arrangement.
Near zero positive shaft power, dividing by power makes BSFC highly sensitive to small changes. At zero power it is undefined; negative or motoring intervals require separate treatment. Do not treat a dramatic number in those regions as a normal efficiency result.
Align fuel and power in time
Fuel instrumentation and torque-speed acquisition may have different response times, averaging windows and timestamps. During a sweep, pairing delayed fuel data with current power can create a misleading BSFC feature. Check the channel timing and use the averaging method specified for the test.
For an agreed interval, fuel mass divided by positive shaft energy provides an interval-specific measure. An arithmetic average of instantaneous BSFC values is not automatically equivalent. State the calculation method, excluded intervals and denominator so that another reviewer can reproduce the result.
04 / ENGINE TESTING GUIDETemperature and pressure channels
Channel location is part of the result
“Coolant temperature” is incomplete without a location. Engine inlet, engine outlet and facility-water supply readings are different channels. The same applies to oil temperature, intake-air temperature and exhaust temperature: the report must identify where each sensor was fitted.
Pressure channels need an equally clear reference. Gauge pressure and absolute pressure cannot be compared as if they are the same number. For differential pressure, identify both measurement connections. A boost-pressure trend means little if the reader cannot establish the sensor location and pressure basis.
The engine cooling water column addresses engine coolant control, while a dynamometer water recirculation system supports the process-water loop. Their readings should remain distinguishable in the report rather than being merged into a generic “water temperature.”
Read trends alongside the load history
Review thermal and pressure trends with actual speed and torque, not as isolated peaks. For a stabilized-point test, use the agreed stabilization criteria and accepted recording window. A short instant at the commanded load does not prove that the relevant engine conditions were stable.
If a value drifts, examine the load sequence, supply conditions, sensor behavior and engine status before assigning a cause. A temperature rise is an observation, not a diagnosis of a failed component. There is no single coolant, oil or exhaust-temperature threshold that this general guide can apply to every engine.
Ambient conditions also belong in the record when output correction is used. Dynojet's correction-factor documentation explains the role of reference atmospheric conditions. Identify the applicable method and distinguish observed from corrected output; a correction factor does not repair a sensor fault or a changed test configuration.
05 / ENGINE TESTING GUIDERecognizing invalid or non-repeatable data
Use a review checklist before acceptance
| Observation | Question to investigate | Appropriate review action |
|---|---|---|
| Power does not agree with torque and speed | Are units, timestamps and processing consistent? | Recalculate matching points and inspect the configuration |
| Sudden discontinuity in one channel | Did the engine change, or did the channel drop out? | Check raw signals, events and instrument status |
| Low BSFC during a fast change | Were fuel and power aligned in time? | Review response delays and the agreed calculation window |
| Repeats drift as the test continues | Did thermal or facility conditions change? | Compare the conditions behind each accepted run |
| Curves look unusually smooth | What filtering or interpolation was applied? | Retain raw data and document processing settings |
| Only peak figures are provided | Were required operating points actually covered? | Request the test schedule and complete result set |
These are investigation prompts, not automatic rejection rules. A real engine response can resemble an instrumentation problem. Preserve the original data and document the explanation rather than silently deleting an inconvenient point.
Make comparisons traceable
For each comparison, keep the engine configuration, test profile, actual operating points, stabilization conditions, instrument records and processing settings together. Power Test's WinDyn5i documentation illustrates why channel configuration, calibration and test templates belong to the measurement context. It does not establish the functions included in any particular Hance installation.
A repeatable engine dyno testing procedure provides the conditions that make a report interpretable; a test-cell system checklist identifies the equipment and interfaces needed to produce it. Review these as connected tasks, rather than treating the report as a standalone graph.
The final acceptance decision should reference the agreed engine specification and explain missing, failed or excluded points. If the report cannot support that decision, the next step is to identify the missing evidence, not to infer a pass from the best-looking curve.
Share your channel list, test method and required report fields with Hance to discuss a suitable measurement configuration.
Discuss your test-data requirements.
Questions and answers
Why do torque and horsepower curves cross at 5,252 rpm?
That relationship uses mechanical horsepower and lbf·ft, with matching numerical axis scales. It comes from unit conversion. Other units, different axis scales or a graph that excludes that speed need not show the same crossing.
Which channels are needed for BSFC?
Net fuel mass consumption and corresponding brake power are required. Power normally comes from torque and speed. Record the fuel measurement basis, channel timing and averaging method; volume-based measurements also need an appropriate density conversion.
How can sensor lag distort a report?
A slower channel can represent an earlier operating condition than a faster one. During changing load, combining them without accounting for their response can produce misleading calculated values. Review timestamps, instrument response and the approved processing method.
Should I compare corrected or uncorrected power?
Use the basis required by the test specification and label it. Preserve observed output and the correction method where applicable. Compare equivalent bases; never assume a correction makes different engine configurations interchangeable.
Is a repeatable report necessarily accurate?
No. Repeated runs can share a calibration or configuration error. Repeatability checks and measurement-validity checks are both needed, along with any uncertainty evaluation required by the test program.
Technical references
HORIBA — Engine Testing, Readout No. 13, February 2010: shaft measurement, stable conditions and engine-test principles.
Dynojet — To Apply Correction Factors: atmospheric correction context.
Power Test — WinDyn5i: configuration, calibration and test-template context.
This is an interpretation framework, not a customer test report or a substitute for the engine manufacturer's acceptance specification. No illustrated measured results or universal pass limits are implied.