Data sources

Every way data gets in

PLCs and RTUs read directly, MQTT, datalogger files, formulas and forecasts — and, from 1.14.8, field visits. Whatever the source, the values land in the same history per tag. Below is what each one supports, for the engineer who connects it.

PLCs & RTUs

Modbus TCP

For PLCs, power meters, drives and gateways. SyncTide polls each device on its own interval and merges neighbouring registers into as few requests as it can.

Runs on your server. In our cloud on request, through a secure tunnel or a small on-site server.

Role
SyncTide is the Modbus TCP client (master). It reads; it does not write to the device.
Function codes
FC1 coils, FC2 discrete inputs, FC3 holding registers, FC4 input registers.
Addresses
Classic 40001 / 30001 / 10001 / 00001 numbering, or explicit hr:100, ir:0, di:5, coil:12 (0-based). An address offset per device for devices that count from 1.
Data types
int16, uint16 and bool on one register; int32, uint32 and float on two; int64, uint64 and double on four; or a single bit of a register.
Byte order
ABCD, CDAB, BADC or DCBA, per device, for 32- and 64-bit values.
Scaling
value × scale + offset, per tag.
Batched reads
Neighbouring tags share one request — up to 125 registers or 2 000 bits — so hundreds of tags take a handful of requests.
Connection
Port 502 by default, unit ID and timeout per device. Polled once a minute or faster, the connection stays open between polls; an optional keep-alive read holds it for PLCs that close idle sessions. New in 1.14.7 — coming soon
Poll interval
Per device, down to 1 second.
Timestamps
The poll time: Modbus carries no time of its own.
Storage
Deadband per tag, report-by-exception and heartbeat. New in 1.14.7 — coming soon
Example Tag addresses
FT-101   hr:100   float   byte order CDAB
PT-201   40011    int16   scale 0.01
P-01     coil:12  bool

PLCs & RTUs

OPC UA

For PLCs with a built-in server, SCADA systems and gateways. Browse the server from the platform, then read on an interval or subscribe and let the server push changes.

Runs on your server. In our cloud on request, through a secure tunnel or a small on-site server.

Role
SyncTide is the OPC UA client, on an opc.tcp:// endpoint.
Sign-in
Anonymous, or user name and password.
Browse
Browse the server's namespaces and node tree from the platform and pick the tags; each tag's type is filled in from its node.
Node IDs
ns=2;s=Flow, ns=2;i=1234, i=2258 or ns=3;g=<GUID>.
Read mode
All of a device's tags in one Read request, split every 1 000 nodes; a bad node is skipped without stopping the others.
Subscriptions
Monitored items with a publishing interval (250 ms by default), a queue size, a sampling interval per tag and an absolute deadband filter on the server. If a server refuses the filter for a node, that tag is subscribed without it and keeps flowing.
Session slots
Browsing and reading take the server's sessions in turn, so a browse never fights the reader for the few sessions a PLC allows.
Timestamps
SourceTimestamp, else ServerTimestamp, else the time of receipt.
Values
Numbers and booleans (true = 1), with scale and offset per tag.
Storage
Deadband per tag, report-by-exception and heartbeat. New in 1.14.7 — coming soon
Example Endpoint and nodes
endpoint  opc.tcp://192.168.1.50:4840
FT-101    ns=2;s=Station1.FT101
PT-201    ns=2;i=1234

PLCs & RTUs

IEC 60870-5-104

For utility RTUs. SyncTide keeps the connection open, takes the RTU's spontaneous events as they come, and keeps the time the RTU put on each one.

Runs on your server. In our cloud on request, through a secure tunnel or a small on-site server.

Role
SyncTide is the controlling station: it connects to the RTU (port 2404 by default) and keeps one connection open per device.
Addresses
A common address (ASDU) per device — it must match the RTU's — an originator address, and one IOA per tag (1 to 16 777 215).
Types
Short float M_ME_NC_1 / M_ME_TF_1, scaled M_ME_NB_1 / M_ME_TE_1, normalised M_ME_NA_1 / M_ME_TD_1, single-point M_SP_NA_1 / M_SP_TB_1 — or any monitoring type named exactly, such as double-point, bitstring M_BO_NA_1 or integrated totals M_IT_NA_1.
Data flow
A general interrogation on connect for the current state, then the spontaneous events as the RTU sends them.
Timestamps
CP56Time2a time tags are kept: events an RTU buffered during an outage arrive with their own time when the link returns.
RTU clock
CP56Time2a carries no time zone, so the zone of the RTU's clock is set per device.
Unmapped points
IOAs the RTU sends that are not mapped are logged once, with their type, so they can be mapped without another tool.
Silent link
An optional reconnect when nothing has arrived for a set number of seconds.
Limits
No 64-bit types: the standard has none.
Storage
Deadband per tag, report-by-exception and heartbeat. New in 1.14.7 — coming soon
Example Points
CA 3 · IOA 1001   M_ME_NC_1   FT-101 flow
CA 3 · IOA 2001   M_SP_NA_1   P-01 running

Messages

MQTT and Sparkplug B

For devices and edge gateways that publish. SyncTide subscribes and stores what arrives, with the publisher's own timestamps when it sends them.

Role
SyncTide subscribes to an MQTT broker — your own, or the one in our cloud.
Connection
User name and password, TLS, keep-alive; it reconnects on its own, with back-off.
Topics
Each equipment has a topic prefix and SyncTide subscribes to prefix/#. The longest matching prefix wins.
QoS
0, 1 or 2, per equipment.
JSON
One value per tag topic {"value": 12.3}; several tags in one message {"flow": 12, "pressure": 5}; a list {"values": [{"name", "value", "ts"}]}; or a bare number.
Timestamps
ts or timestamp as ISO 8601 or Unix seconds or milliseconds; otherwise the time of receipt.
Sparkplug B
NBIRTH, DBIRTH, NDATA and DDATA: metric names and aliases from the birth messages, metric timestamps kept, numbers and booleans.
Buffering
Decoding and writing are separated by a buffer, so a slow database never stalls the broker connection.
Scaling
value × scale + offset, per tag.
Example A JSON message
topic    plant/station1/FT-101
payload  {"value": 12.3, "ts": "2026-11-20T10:00:00Z"}

Files

Datalogger files

For dataloggers and systems that export files, often over a cellular link and often late. SyncTide reads each file once it is complete, at the times written in it.

Formats
CSV with the delimiter detected (comma, semicolon, tab or pipe), and Excel .xlsx (first sheet).
Layouts
One column per tag, or long rows of time, tag and value. A units row is recognised; files without a header row and headers over several lines are read too.
Numbers
Decimal comma or point, thousands separators; on/off, true/false and yes/no read as 1/0.
Timestamps
2026-11-20 14:30:00 (also with a T), 2026/11/20 14:30, 20/11/2026 14:30, 11/20/2026 2:30 PM; anything else is read day first.
Time zone
Per equipment, e.g. Europe/Lisbon. Times that carry an offset keep it; summer-time changes are handled; a logger whose clock never changes uses a fixed zone such as Etc/GMT-1 (UTC+1).
In our cloud
Dataloggers send by FTP, FTPS (explicit TLS) or SFTP into the equipment's folder.
On your server
SyncTide watches a folder per equipment: your FTP or SFTP server, or a network share, drops the files there — or upload them from the browser.
Partial uploads
A file is read only once it has stopped changing, so a slow cellular upload is never read half-way.
Duplicates
A file sent again is recognised by its content and skipped; overlapping files never store a value twice.
Late files
Files that arrive hours late are filed at their own times, and alarms still fire for the crossings inside them.
Failed files
A file that cannot be read is parked with the reason and listed in the platform. It is tried again when the file, the equipment's settings or SyncTide change, or when you retry it.
New columns
New columns are found automatically and listed for mapping; mapped tags keep their history.
Example A wide CSV with a units row (decimal comma)
Date;FT-101;LT-07
;m3/h;m
20/11/2026 14:30;12,4;3,10
20/11/2026 14:40;12,9;3,08

People

Field work (Interventions)

New in 1.14.8 — coming soon

What technicians do on site is data too. With the Interventions module, visits are recorded on a phone or tablet, against the site's own zones and assets, in the same platform as the readings.

Visits
Preventive or corrective, at a site, with one sheet per piece of work on a zone or asset.
Site structure
Zones and assets up to 8 levels deep, with make, model, serial number and fields per asset type; every change keeps who, when and the old value.
Forms
Checklists, questions with text, numbers, choices and dates, notes and signatures — published as versions, and a visit keeps the version it started with.
Evidence
Photos with descriptions, and signatures, kept with the visit.
QR labels
Printable labels for every zone and asset; scanning one opens it in the platform.
After closing
Changes to a closed visit need a permission and keep the previous values in its history; a signature never changes.
Licence
Interventions is an optional module.

See Interventions

Computed

Formulas and forecasts

Some values are not measured anywhere: a total of three meters, a pump's fault bit, tomorrow's flow. Virtual tags compute them from other tags; forecasts come from the prediction module.

Virtual tags
A formula over tags of the same equipment, ${Tag}, or of another one, ${Equipment:Tag}; read like any other tag in trends and reports.
Functions
+ − × ÷, %, powers, abs, min, max, round, sqrt, pow, log, exp, bit(value, n) and bits(value, from, to).
When it computes
When every input has a new value, when any input changes (the others carried forward), or on a fixed interval.
Site KPIs
A site-level virtual tag with a target and a tolerance can raise an alarm when it leaves its band.
Safety
Evaluated in a sandbox: no imports, no attributes, no other calls; a formula that refers back to itself is refused.
Forecasts
With the prediction module: a forecast per tag and horizon, up to seven days, with its band — used by forecast alarms and to fill gaps in the editor.
Example Virtual tags
Total flow    = ${FT-101} + ${FT-102}
Pump 1 fault  = bit(${Status word}, 3)
Head (m)      = (${PT-201} - ${PT-200}) * 10.2

What happens next

Once it is in, every value is stored, watched, analysed, reported and, when needed, corrected — the same way, whatever the source.