A low-level pump delivers steadily on a test bench. The flowmeter display looks normal, but the external record gives less water than a separate collection by weight. Do not adjust the pump or multiply displayed flow by a correction factor. First compare the meter's own total, the external count and the independently collected quantity over the same interval. A meter can measure flow correctly while an external system uses the wrong volume per pulse, misses pulses or integrates over a different period.
Establish what one pulse represents
Check the instrument configuration for a pulse, frequency or other output. Verify units, volume per pulse, output limits and the receiving input requirements. A pulse per litre and a litre per pulse happen to give the same number. Half a litre per pulse and half a pulse per litre differ by a factor of four. Do not compare settings by their number alone. The receiver can also have minimum pulse width, filtering and count rate limits. A pulse generated by the meter is not necessarily accepted at the next stage.
Record the local total, raw external count, converted external value and independently collected quantity separately. Use the same start and end, note any reset or uncounted fraction, and record the density and completeness of collection when weighing. A pulse or two can represent a substantial fraction of a short test with a coarse pulse value. Use a longer safe interval before calling the discrepancy a missed count. Tare the collecting vessel and account for the method's temperature or entrained air limits rather than assuming weight validates every situation.
Compare error patterns over more than one flow
For illustration, suppose the output is one pulse per litre. A flow of 20 litres per minute gives about one-third of a pulse per second. At 120 litres per minute it gives two pulses per second. If the receiving arrangement in this example accepts only one pulse per second reliably, low flow can agree while the high-flow total falls short. These figures illustrate frequency and scale, not the specifications or adjustment instructions for a real instrument. A similar proportional error at low and high flow suggests a unit or pulse value check. An error that grows at higher flow calls for a check of receiver limits, pulse width, filtering or the meter's output capability. More than one fault can be present.
If the instrument total and independent collection agree while the external total does not, examine the chain from output to software first. If both instrument and external totals are low, go back to full pipe, air, installation and the independent collection method. Change one controlled variable at a time and retain the original settings and readings. Qualified personnel should handle signal checks or adjustments under approved isolation and test procedures. Do not work on live terminals to reproduce the error.
Recheck the range used for decisions
After correction, verify representative low, middle and high points across the actual range. Check the resolution over the shortest reportable interval too. Link the instrument total and external report with units, pulse value, configuration and independent quantity. Where records support lot release, determine which earlier tests may have used the affected setting. A normal flow display is the beginning of a measurement chain, not proof that every part of it is sound.
If the local rate is correct, an independent counter and the controller can observe the same pulse train at the same time. Check whether the logging window misses the first discharge after startup or whether the controller merely sends totals late. A missing count and a delayed report need different corrections.
Frequently asked questions
Does the local display prove the external total is right?
No. Measurement, output, receiving and conversion are separate stages.
What if the same factor appears at low and high flow?
Check units and pulse scaling first, then the interval and output. Do not start with a compensating multiplier.
Does one missing pulse prove a faulty counter?
No. Check resolution, test boundaries and reset conditions. Repeat over a longer controlled window if needed.
