.png)
Pulse Totaliser Dropped? Distinguish Counter Rollover from a Reset
A cumulative pulse count falls between two collections. That observation alone cannot tell you whether the counter wrapped, a technician reset it, the device restarted or the wrong source was mapped. Preserve the discontinuity and establish its cause before converting the difference into consumption. Otherwise one ambiguous interval can become a large, plausible-looking usage figure.
Original conceptual illustration; not a specific product, live software screen or customer installation.
Key takeaways
A negative difference does not prove rollover.
Use the documented counter cycle and independent continuity evidence.
Keep unresolved intervals identifiable rather than manufacturing consumption.
1. Define the counter you actually collect
Write down the source device, register or signal, integer representation, count direction and pulse scale. Establish whether you are collecting a lifetime counter, a resettable batch total, a signed position counter or an accumulated value maintained by a gateway. Those are different records even when the display calls all of them a total.
Obtain the exact model and firmware documentation for rollover, reset and persistence. Counter width alone does not establish its operating cycle: a configurable rollover limit may be smaller than the storage range. A restart may preserve a count on one device and clear it on another.
Rockwell Automation’s ring-counter documentation is one example of configurable rollover and rollunder boundaries. It supports the need to inspect the configured cycle; it is not evidence that a different instrument follows that particular implementation. Do not copy another product’s register assumptions into the project.
2. Collect evidence at the discontinuity
Retain the last valid count before the fall and the first valid count after it, with source timestamps, receipt times and quality. Preserve intervening communication gaps. Record approved maintenance, configuration changes and replacement devices that could explain the boundary.
Use supported restart indicators, reset records or rollover flags where the selected equipment provides them. Confirm what each indicator means and whether it persists long enough to be collected. Absence of an uncollected event flag is weak evidence; a short-lived flag can disappear between polls.
Check the mapping before interpreting the process. An address change, source replacement or signed interpretation error can create a discontinuity without any physical change in consumption. Keep the original raw values alongside the decoded engineering record.
3. A hypothetical single-wrap calculation
Suppose an invented, forward-only counter cycles through 0 to 9,999 before returning to 0. Its documented modulus is therefore 10,000. A count of 9,980 followed by 25 represents 45 increments if independent evidence establishes exactly one rollover and no reset: (10,000 − 9,980) + 25 = 45.
If the hypothetical pulse scale is two litres per increment, the corresponding interval quantity is 90 litres. This arithmetic describes the stated example only. It does not prove that the observed instrument wrapped, establish a real product’s scale or authorise a billing adjustment.
Ask whether more than one complete cycle could occur during a collection gap. A maximum credible pulse rate and elapsed interval can reveal that ambiguity. If multiple wraps are possible and no reliable extended counter exists, two endpoint values cannot recover the missing number of complete cycles.
Original explanatory diagram. Examples are hypothetical; verify the actual equipment, interface contract and project requirements.
4. Agree a discontinuity policy before handover
Define separate outcomes for an ordinary increment, an evidenced wrap, a documented reset, a source replacement and an unresolved fall. State which intervals can be calculated automatically and which require review. A reset boundary should remain visible even when subsequent collection is healthy.
For a documented reset, collect suitable before-and-after evidence where the operational method allows it. Decide whether an interval can be reconstructed from independent measurements. Do not assume that resetting to zero means no consumption occurred between the last poll and the reset.
Test these cases using synthetic records or an isolated approved test arrangement. Keep control and protective functions outside the exercise. The acceptance question is whether the reporting system classifies each case correctly, not whether a chart can be made continuous.
Identify the source, cycle limits, scale and persistence behaviour.
Capture the discontinuity and independent maintenance or event evidence.
Check whether one wrap, several wraps or a reset is possible.
Retain uncertain intervals and validate the agreed classification policy.
5. Keep the derived record traceable
Store any corrected interval as a derived result with the source values, reason, formula, pulse scale and review status. Preserve raw history. A future investigator should be able to reproduce the result without relying on the memory of the person who fixed the chart.
Avoid applying a modulo formula to every falling value, assuming the largest representable number is the configured rollover limit, or treating an unavailable sample as zero. Each shortcut can produce an attractive trend that answers the wrong question.
For a ProSense Solutions discussion, provide the exact counter documentation, configured limits, pulse scale, sample interval and a sanitised discontinuity record. Define whether the result supports maintenance, internal allocation or regulated measurement so that the appropriate validation scope can be established.
Related guide and project resources
Need help with your application?
Send the relevant equipment documentation, field definitions and a sanitised example of the behaviour you need to verify. Confirm the available functions and integration requirements for the exact equipment and deployment.
