Calibration history
Synacl stores every reading raw and converts it to the displayed (engineering) value with the tag's factor and offset — see Device parameters (tags). Without calibration history, that conversion always uses the tag's current factor and offset, for every reading ever stored. Fix a wrong factor today and last year's chart changes with it.
Calibration history changes that. When it is on for your account:
- A new factor or offset applies from the moment you save it. Readings taken before keep the calibration they were taken under.
- Every change is recorded: the old and new values, when the change took effect, who made it, and an optional reason.
- If the old calibration was simply wrong, you can apply the current one to past data — an explicit, recorded step (see below).
Rollout. Calibration history is being rolled out account by account. Until it is on for your account, changing a tag's scaling re-scales all of its past readings, as described in Device parameters (tags), and the history and Apply to past data options below do not appear. Contact support if you need it.
Changing a tag's calibration
Open the device, choose Edit, find the tag and use Edit scaling. Set the factor (Multiply or Divide) and offset as before. There is also an optional Reason field — "replaced the probe", "datasheet says ÷10" — which is saved with the change. A reason costs a few seconds and is what answers "why does the line step here?" months later.
Saving records a new calibration only when the factor or offset actually changed. Saving the form with nothing changed, or changing only how the factor is shown (Multiply ↔ Divide for the same value), records nothing.
The first change to a tag also records the calibration it had before — so the history always starts from what your data was actually read with.
Where the history is applied
Every place Synacl shows stored readings scales each one with the calibration in force when it was taken:
- Reports, their exports, scheduled report emails, and job reports.
- Dashboard charts and history seeds, the device page chart and records table, and the time-axis line chart.
- Tag statistics shown while you set an alarm limit (min, max, mean, percentiles, the histogram and would have alarmed N times).
- Fleet views and the AI agent tools (
get_telemetry,get_latest_values).
When a report's range spans a change, an average over the bucket that contains the change weights each reading by its own calibration, and the report says that a tag was re-calibrated during the range. A percentile across a calibration change is not shown — a percentile of readings taken under two different calibrations does not describe either — and the report says why.
Live values, alarm limits, rules and the gateway always use the current calibration. That is correct: a live reading is taken now.
Seeing the history
In Edit scaling, History lists every calibration the tag has had, newest first: the factor and offset, when it took effect, who changed it, and the reason. Changes made by Synacl Support show as Synacl Support; changes made through an AI agent are marked as such.
Apply to past data
Sometimes the old calibration was never right — the tag was set up with ×1 when the register always held tenths. Then you want the corrected factor to apply to the readings already stored, too.
In Edit scaling, choose Apply to past data…. You choose:
- All past readings, or from a date — readings before that date keep the calibration they had;
- a reason (required).
Confirming re-scales how those stored readings are shown everywhere listed above. Nothing is deleted: the earlier calibrations stay in the history, marked superseded, so you can always see what the data read as before. This does not change the tag's current calibration, so the gateway is not sent new configuration and does not restart.
Applying to past data needs permission to change parameters (owners, admins and operators). Viewers can see the history but not change it. AI agents can change a tag's scaling but cannot apply a calibration to past data — that is a decision for a person, in the app.
Good to know
- A tag can hold up to 100 calibrations. If a tag reaches that, apply its current calibration to past data to consolidate the history, then change it again.
- Deleting a tag and adding it again starts a new tag with a new history; the old tag's history is not carried over.
- Older versions of the mobile app show past readings with the tag's current calibration until the app is updated.