When a system accumulates a quantity over time, it usually needs to answer two different questions at once: how much has passed since the last reset, and how much has ever passed. A resettable totalizer answers the first, clearing to zero on demand so it can measure a shift, a batch, or a day. A non-resettable totalizer answers the second, accumulating without ever being cleared so it forms a permanent lifetime record. This guide explains why well-designed metering runs both accumulators side by side, why custody and audit rely on the one that never resets, and how to reset the first without losing counts.
Resettable vs non-resettable totalizer in one line: A resettable totalizer is an accumulator that can be cleared to zero on demand so it measures a defined period such as a shift, batch, or day, while a non-resettable totalizer accumulates continuously and is never cleared, forming a permanent grand total. Good metering practice runs both from the same measurement: the resettable one gives operators period figures they can start fresh, and the non-resettable grand total gives an audit-proof lifetime record that custody and reconciliation depend on.
The core idea is that a single stream of flow or pulses feeds more than one accumulator. Each increment of measured volume is added to both a resettable register and a permanent grand total at the same instant, so the two never diverge except that one gets cleared periodically and the other does not. This mirrors the way a domestic utility meter works: the odometer-style dial that only counts up and is never reset is the grand total the utility bills against, while a separate figure someone jots down at the start of a period is the resettable reading used to work out a period's usage.
The resettable totalizer exists to serve operations. A plant runs in shifts, produces in batches, and reports by the day, so operators need a figure they can zero at the start of each period and read at the end to see exactly what happened during it. Loading a truck, filling a tank to a target, or measuring a well test all call for a total that starts at zero and stops at a known quantity, and only a resettable accumulator can provide that cleanly without the operator having to subtract one large number from another.
The non-resettable grand total exists to serve accountability. Because it is never cleared, it is a continuous, monotonic record of everything the meter has ever registered, which makes it the trustworthy backstop for reconciliation. If there is ever a dispute about how much passed a point over any span of time, the difference between the grand total at the start and end of that span settles it, independent of how many times anyone zeroed the resettable figure in between. The grand total's whole value comes from the fact that nobody can reset it.
In custody transfer, where measured volume determines who pays whom, the non-resettable grand total is the anchor. A resettable total is convenient but inherently editable, because the act of resetting it is a deliberate operator action, and anything an operator can zero is something that could be zeroed at the wrong moment, by accident or otherwise. The grand total removes that risk: since it only ever climbs and is never cleared, any period's volume can be recovered as the difference of two grand-total snapshots, and no single reset can erase history.
This is why audit trails favor recording grand-total readings at period boundaries rather than trusting the period figure alone. If the historian captures the grand total at the top of each day, then every daily volume can be reconstructed from those snapshots even if a resettable daily register was cleared early, late, or twice. The two views cross-check each other: the sum of the resettable period totals should equal the change in the grand total over the same span, and a mismatch is an immediate flag that something was reset when it should not have been or that data was lost.
For batch and shift work the same logic applies at a smaller scale. A batch total is a resettable figure zeroed at the start of the batch, but the record that survives an audit is the pair of grand-total readings that bracket the batch. Treating the grand total as the source of truth and the resettable totals as operational conveniences, rather than the other way around, is what keeps period reporting robust even when resets happen frequently and under time pressure.
The hazard in any resettable totalizer is losing counts during the reset itself. If flow is still passing at the instant the register is zeroed, the increments that arrive during the reset operation can fall into a gap and never be added anywhere, so the period totals silently under-report. Safe reset logic avoids this by capturing the value first and then subtracting it, rather than blindly writing zero: the system reads the current period total, records it as the completed period's result, and clears the register by subtracting that captured value, so any increments that arrive mid-reset are preserved and simply belong to the next period. Because the grand total is never touched, it acts as an independent check that no counts vanished during the reset.
Timing matters as much as method. Resets are best triggered at defined, logged boundaries, such as a shift change or midnight, and recorded as events so there is a clear before-and-after. A reset that happens at an undocumented moment makes the resulting period total hard to interpret, while a logged reset with a captured value and a grand-total snapshot on each side leaves no ambiguity about which volume belongs to which period. The event log turns each reset from a potential source of confusion into a documented boundary.
A cloud SCADA platform such as Merobix is well suited to this two-accumulator architecture because it can hold the permanent grand total centrally and independently of any field reset. Even if a resettable register at a remote site is cleared by a local operator, the platform keeps accumulating its own historian-based grand total from the same rate, so daily and monthly figures for allocation and reconciliation come from a record that no field action can zero. Resets performed in the field are received as logged events with their captured values, so the central system can rebuild any period from grand-total snapshots and immediately flag a case where the resettable totals and the grand total no longer agree. That combination gives operators the convenient period figures they need at the site and gives accounting the tamper-resistant lifetime record it needs in the cloud, from one measurement.
A resettable totalizer can be cleared to zero on demand, so it measures a defined period such as a shift, batch, or day. A non-resettable totalizer accumulates continuously and is never cleared, forming a permanent grand total. Good practice runs both from the same measurement so operators get period figures while accounting keeps an audit-proof lifetime record.
Because a figure nobody can clear cannot be erased by a mistaken or ill-timed reset, so it is trustworthy as the basis for who pays whom. Any period's volume can be recovered as the difference between two grand-total snapshots, independent of how often the resettable figure was zeroed. The non-resettable grand total is therefore the anchor that reconciliation and audits rely on.
Capture the current period value first and record it as the completed period's result, then clear the register by subtracting that captured value rather than blindly writing zero, so any increments arriving during the reset are preserved for the next period. Trigger and log resets at defined boundaries such as shift change or midnight. Because the grand total is never touched, it provides an independent check that no counts were lost.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.