Machine Testing and Efficiency Determination For students.

The efficiency of an electrical machine is the ratio of useful output power to input power, expressed as a percentage: η = (Pout / Pin) × 100%. A 90% efficient motor wastes 10% of its input as heat, noise and vibration. Machine testing is the set of practical procedures used to measure this efficiency honestly, because nameplate numbers mean nothing until they are verified on a real machine under real load.

Testing matters for money and safety. A factory running a 100 kW motor at 85% instead of 93% efficiency burns roughly an extra 8 kW continuously — over a year of round-the-clock operation that is about 70,000 extra units of electricity. Tests also reveal overheating windings, unbalanced phases and deteriorating insulation before they cause a breakdown, which is why utilities and industries specify acceptance tests before any large machine is put into service.

Key Concepts

• Direct loading test: the machine is loaded to its rated output with a brake or dynamometer while input and output power are measured directly. Accurate but needs full-load power and a loading device.
• Swinburne's test: an indirect method for DC machines run at no load to find the constant losses (core loss plus mechanical loss); copper loss is computed as I²R. Simple, but assumes constant losses stay constant under load.
• Hopkinson's test: two identical machines are coupled back-to-back, one motoring and one generating, so only the losses need to be supplied. It gives full-load temperature and efficiency without a full-load power source.
• Retardation test: a running machine is switched off and its slowdown rate measured to find mechanical losses and moment of inertia.
• Constant vs variable losses: core and friction losses barely change with load; copper losses grow with the square of current. Efficiency peaks where constant and variable losses are equal, usually near 75–100% load.

Formulas

• Efficiency: η = (Pout / Pin) × 100% = (Pin − losses) / Pin × 100%.
• Mechanical output: Pout = 2πNT / 60, where N is speed in revolutions per minute and T is shaft torque in newton-metres, giving Pout in watts.
• Copper loss: Pcu = I²R per winding, with I in amperes and R in ohms.
• Swinburne's (motor): η = Pout / (Pout + Pc + I²R) × 100%, where Pc is the constant loss found at no load.

Worked Example

A DC shunt motor is tested by direct loading. Supply: 220 V, line current 15 A, so Pin = 220 × 15 = 3,300 W. The brake measures shaft torque 12 N·m at 1,500 rpm, so Pout = 2 × 3.1416 × 1500 × 12 / 60 = 1,885 W. Efficiency η = 1885 / 3300 × 100% = 57.1%. The missing 1,415 W went into copper loss, iron loss, friction and windage. If the same test at half torque gave much lower efficiency, that would confirm this small motor is inefficient at light load — exactly why direct testing beats nameplate claims.

Common Misconceptions

• "A machine is equally efficient at all loads." Efficiency peaks near rated load and falls sharply at light load, because constant losses stay the same while output shrinks.
• "The no-load test alone gives efficiency." It gives only the constant losses; without load current you cannot compute copper loss, so you cannot compute efficiency.
• "Bigger machines are less efficient." The opposite is usual: large machines reach 95–98% efficiency while small motors struggle to cross 80%, because losses do not scale up as fast as output.

Real-World Applications

• Acceptance tests on new transformers and motors verify the manufacturer met the contracted efficiency before payment is released.
• Energy audits use field efficiency tests to decide whether rewinding an old motor or replacing it saves money.
• Electric-vehicle makers map motor efficiency over the whole torque-speed range to predict real driving range.
• Utilities test large generators with Hopkinson-style back-to-back methods because no test lab can absorb hundreds of megawatts directly.

Efficiency is never a guess — it is measured, directly with brakes and dynamometers or indirectly with no-load and back-to-back tests. Learn to separate constant losses from copper losses, measure at several load points, and you can judge any machine's real performance instead of trusting its nameplate.