Two recorders in one container may show different values. First check whether the records are comparable: device identity, clock, logging interval and sensor position are relevant context.
Calculate the difference
Match values with the same timestamp and record one device minus the other. Exclude unmatched times and report their count; do not create artificial matches by rounding timestamps.
Record the context
Note both serial numbers, sensor positions, logging intervals and displayed clock settings. Download time is a separate event; do not use it in place of measurement time.
Interpret the gap. Direction and change over time describe the comparison, but do not prove a cause by themselves. The model documentation or a reference check can help determine what verification is needed.
Which record goes in the report?. Choose the record for reporting, with its measurement point, under the applicable product specification and authorized site quality process. Changing devices from run to run makes comparisons unclear.
Verification record. If a check or calibration was performed, link its date, method, reference device and result to the device ID. Mark any later device or position change in subsequent comparisons.
Monthly summary
Report the number of matched readings, direction and range of differences, unmatched count and control records reviewed. The summary makes the gap between curves visible; it is not by itself a decision on device accuracy.
