Knowledge baseConnect
Tags, virtual tags and deadband
How a device's values become tags with a history: system tags and mapping, what gets stored (deadband, report-by-exception, heartbeat), virtual tags, and correcting the history.
A tag is one measured quantity of one piece of equipment — a flow, a pressure, a pump’s running state. Each tag keeps its own history, and that history is what charts, alarms, reports and forecasts read.
System tags and mapping
System tags are the standard names, each with its unit, used across all equipment: Flow (m³/h), Pressure (bar). They are kept in Configuration → Tags.
On each equipment, the Tag Mapping tab links the tags the device sends to system tags:
- the equipment tag — the column name in a file, the register or node, the IOA, the MQTT topic;
- the system tag it feeds;
- a scale factor (and for live protocols an offset) to convert units;
- active or not.
Only active, mapped tags are stored and count towards the licence’s tag limit. A column a datalogger sends that is not mapped still appears in the list (so it can be mapped) and its current value shows on the Real Time tab, but its values are not kept.
Map first, rely later. Mapping a tag starts its history from that moment. Live protocols keep no copy of what was sent before; for file-based equipment, the files still in the intake folder can be read again.
Kinds of value
| Kind | What it is | How it is drawn |
|---|---|---|
| Instant | A reading: flow, level, temperature | A line |
| Totaliser | A counter that only grows: a volume or energy meter | Bars of the amount per period, not the counter reading |
| Digital | On / off, open / closed | Steps |
Alarm rules, reports and forecasts follow the kind: a totaliser alarm looks at the amount over a period, a digital alarm at the change of state.
What gets stored
New in 1.14.7 — coming soon
Storing every value of every poll fills disks without adding information. Three settings, per tag or per equipment, decide what is written:
- Deadband (per tag, on Modbus, OPC UA and IEC 104): a value is stored only once it moves more than the deadband from the last stored value. A slow drift is still recorded, because the comparison is with the last stored value, not the previous reading.
- Report-by-exception (per equipment): a value that has not changed is not written again.
- Heartbeat (seconds): a value is stored at least this often even when nothing changes, so a flat signal never looks like missing data.
With OPC UA subscriptions, the deadband is also sent to the server, so unchanged values never cross the network. Deadbands can be set one by one in Tag Mapping, or for many tags at once through the tag CSV import.
The history lives in PostgreSQL with TimescaleDB, split into time chunks; closed chunks are compressed. Every hour SyncTide picks the chunk size and how soon a chunk is compressed from how much data really arrives, so a small installation keeps weekly chunks while a large one compresses within hours. A file sent twice, or two polls that overlap, never store the same value twice.
Gaps on charts
Set the equipment’s expected data period (how often it normally produces a sample) and switch on Show data gaps on charts in Monitoring & Alarms. Two points further apart than one and a half times that period — or one and a half times the chart’s bucket, whichever is wider — then leave a blank space instead of a straight line across the hole. A coarse bucket never makes a healthy series look broken.
Virtual tags
A virtual tag is computed from a formula over other tags: a total of three meters, a difference of two pressures, a pump’s fault bit read from a status word. It is read like any other tag in charts, reports and alarms. Create one in Configuration → Tags by switching the tag to Virtual (computed).
Total flow = ${FT-101} + ${FT-102}
Pump 1 fault = bit(${Status word}, 3)
Head (m) = (${PT-201} - ${PT-200}) * 10.2
${Tag}refers to a tag of the same equipment,${Equipment:Tag}to a tag of another one.- Operators
+ − × ÷,%, powers, and the functionsabs,min,max,round,sqrt,pow,log,exp,bit(value, n)andbits(value, from, to). - It is recomputed when every input has a new value, when any input changes (the others carried forward), or on a fixed interval.
- Formulas run in a sandbox: no imports, no attributes, no other calls; a formula that refers back to itself is refused.
A site KPI is a virtual tag scoped to the whole site rather than one equipment. Give it a target and a tolerance (± %) and SyncTide colours it by its deviation and can raise an alarm when it leaves its band.
Virtual tags have no stored samples of their own, so they cannot be corrected directly: correct their inputs and the virtual values follow.
Correcting the history (Editor module)
Meters drift, a sensor is replaced, a link drops for a night. With the Editor module, administrators correct the stored history from the Explorer (the pencil in the range bar):
- Points: drag a point (in Raw mode, where one chart point is one stored value) or double-click to type a value. Nothing is written until you save.
- Periods: shift a whole window — instant tags by a percentage or an amount, totalisers by an amount only (a percentage would distort every consumption figure), digital tags forced on or off.
- Gaps from forecasts: with the Forecasting module too, a finished communication gap can be filled with the stored forecast, point for point.
Every change is a batch with who, when, the reason and the values before, recorded in the audit log and reversible (newest first). Corrected points carry a marker that every role can see. Reports, dashboards and exports use the corrected values; past alarm events are not rewritten — they record what happened at the time.