Turbidimeter Reading Drift and Calibration Dispute
A plant uses online turbidimeters on each filter effluent and a bench turbidimeter for grab samples. The state requires continuous turbidity monitoring with 0.1 NTU resolution and a 0.3 NTU maximum. On Tuesday the online meter on Filter 3 reads 0.18 NTU while the bench meter on a grab sample from the same tap reads 0.09 NTU. The online meter was last calibrated 45 days ago; the manufacturer recommends every 30 days. The bench meter was calibrated this morning with a 0.02 NTU primary standard and a 1.0 NTU standard. The operator who calibrated the bench meter notes the 0.02 standard read 0.03 NTU after calibration. The plant supervisor wants to report the lower bench value to stay under 0.3 NTU and avoid a violation.
1. Which reading do you trust, and what steps do you take before reporting anything?
Read the worked answer
Neither reading should be reported until the discrepancy is resolved. The bench meter's 0.02 standard reading 0.03 NTU after calibration is a calibration failure signal: a primary standard should read at its stated value within the meter's tolerance, and a 0.01 NTU error at the low end is significant relative to a 0.3 NTU limit. The online meter is 15 days past its calibration interval, so its reading is also suspect.
Steps:
- Re-calibrate the bench meter with fresh standards and verify the 0.02 standard reads within tolerance; if it will not, take it out of service.
- Calibrate or verify the online meter per manufacturer procedure, including a zero check and a span check.
- Run a side-by-side comparison with a third, verified meter or a known standard at the same tap.
- Check the sample tap and tubing for air bubbles, biofilm, or sediment that could bias the grab sample.
- Record all readings, calibration dates, and standard lot numbers.
- Do not select the lower number to avoid a violation; report the verified value and document the investigation.
2. The online meter calibrates fine and still reads 0.18 NTU while a freshly calibrated bench meter reads 0.10 NTU. What could explain the difference, and how do you decide which is correct?
Read the worked answer
Possible causes:
- The online meter samples a different point or has a sample line with different residence time, so it may be seeing filter effluent before or after a change the grab sample misses.
- The online meter's flow cell may have fouling, bubbles, or a dirty lens that biases high.
- The grab sample may have lost particles in tubing or settled before reading, biasing low.
- Different measurement principles or calibration standards between instruments can produce small offsets.
Decide by:
- Cleaning the online flow cell and re-reading.
- Pulling a grab sample directly at the online meter's tap and reading it immediately on the bench meter.
- Comparing both against a third verified instrument or a standard.
- Checking the online meter's trend: a step change or drift suggests fouling; a stable offset suggests calibration.
If the online meter is verified and the grab sample is not representative, the online value governs for compliance because it is the continuous monitor. Document the resolution and correct the sample procedure.
3. What records and notifications are required, and how do you handle the supervisor's instruction to report the lower value?
Read the worked answer
Records:
- Calibration and verification records for both meters, including dates, standards, and results.
- The discrepancy investigation: readings, times, sample points, and corrective actions.
- The final reported value and the basis for it.
- Any maintenance or repair to either instrument.
Notification: if the verified value exceeds 0.3 NTU, follow the state's reporting rule for turbidity exceedance, including the required timeframe and follow-up sampling. If it does not exceed, still document the investigation.
On the supervisor's instruction: reporting a value you know is not verified to avoid a violation is falsification. I would explain that the bench meter failed its low-end check and the online meter was out of calibration, so neither number is defensible yet. I would escalate to the utility's compliance lead if needed, and I would not sign or submit an unverified number. The correct path is to fix the instruments, verify, and report the verified result, whatever it is.