YARDPUMP · Technical Article

Why Can a Small Torque Offset Distort a Pump Shaft-Power Test?

Convert torque zero offset into watts at actual speed. Separate instrument zero from rig drag and real shaft load before correcting the result.

Suppose two small submersible pumps deliver similar flow and head on the same bench, but the reported shaft powers differ by several watts. Ask what that gap means in torque and whether the observed zero drift could account for it. Shaft power at the measurement section is torque times angular speed. The torque is small on a small pump; a modest offset multiplied by running speed can be large enough to reverse the apparent comparison.

Keep zero, rig drag and real load separate

A torque transducer may produce a nonzero signal when installed in the specified no-torque condition. That is a zero check. During rotation, couplings, bearings or the load device can consume power. Which losses cross the measuring section depends on where the sensor is mounted. A motor running unloaded, an impeller churning water or a rotating test fixture can all apply real torque. Setting an arbitrary running indication to zero assumes that all of it is instrument error. That assumption may be wrong.

Draw the train from the motor to the load before reviewing results. Put the transducer, couplings and supporting bearings on the sketch. Look up the required zero condition, warm-up and permitted reset procedure. Where the test method allows characterization of rig drag, record it separately at comparable speed, temperature and assembly. Say how it enters the calculation. An undocumented correction that combines initial offset, thermal drift and actual drag can make reported efficiency look more precise while concealing what was measured.

What does 0.03 newton metre mean?

Suppose measured speed is 2,850 revolutions per minute, about 298 radians per second. A positive zero offset of 0.03 newton metre adds about 8.9 watts to calculated shaft power. If two pumps differ by six watts and the zero uncertainty is comparable, the test does not yet support ranking them. Measure speed in the same stable interval as torque. Substituting synchronous speed or a nameplate number adds a separate error. The figures illustrate how uncertainty propagates; they are not product data.

Check range and usable resolution at the low load you care about. Confirm that calibration covers the torque range and direction being used. Keep the pre-run and post-run zeros, original torque readings and actual speed as well as the software's watts. When the apparent difference is close to uncertainty, repeat the relevant duty or use a suitable range. More digits in the exported result do not make the transducer more accurate.

It helps to agree on the reporting boundary before comparison. If the stated result is motor shaft output, identify which load-side losses are included. If the stated result is power entering the impeller, the measurement and any correction need to support that different boundary. Keep the hydraulic and electrical quantities in separate columns. An input power meter can flag a gross inconsistency without establishing an independent shaft result. On a wet rig, have qualified personnel use guarded, isolated equipment and the specified loading procedure.

Verify the sign convention before running. A sensor orientation or software sign change can turn a correction into an addition. Use the prescribed check to confirm loading direction and retain the coupling and support layout with the exported data. For thermal drift, compare samples under similar rig conditions and interleave a known reference where appropriate. Testing all of one type on a cold rig and all of the other after warm-up can turn time drift into a supposed product difference. If accuracy is specified relative to full scale, translate it into the uncertainty at the small torque actually measured. The headline accuracy percentage is not the relative accuracy of every operating point. Use that calculation to decide whether the setup can answer the buying or design question.

Frequently asked questions

Does re-zeroing before every run fix this?

Zero only under the condition prescribed by the instrument. It can address a defined offset, but cannot correct range, alignment, calibration or temperature effects.

Can I check it by applying an assumed efficiency to input power?

That produces an estimate. An assumed efficiency is not independent shaft-power evidence. The measured torque and speed are needed for this comparison.

What if the zero changes substantially by the end?

Record the size and timing of the change. Identify which duties may be affected and repeat them if needed. Do not subtract the final offset indiscriminately from the entire run.

Related reading