Skip to content
SYNGETICTECHNOLOGIES

Our methodology

We Model the Manufacturing Process Before We Move the Data.

A structured method inspired by ISA-88 principles: understand the process, model it, then connect and deliver the data.

Seven stages

From Process Understanding to Reusable Connectors

Pick a stage to see what we do and what you get.

Stage 01 of 07

Understand

Learn the equipment, process and quality requirements.

We learn how the product is actually made — and where its data lives — before writing any code.

Key question · Which decisions should this data support?

Engineering activities

  • Process walkthroughs with engineers
  • Inventory of systems and files
  • Data required vs. data available

Artifacts

  • System inventory
  • Constraints log
  • Gap analysis
01 / 07

Interactive · Data architecture

A Four-Layer Manufacturing Data Architecture

Sources → pipelines → contextual model → applications. Select a source to trace its path.

Layer 01

Raw Data Sources

Select a source to trace it

Controller states, setpoints, and process parameters read from the equipment.Continuous signals such as temperature, pressure, flow, or velocity.Programs, spindle and axis data, tool information, and alarms.Recipe managers, line software, and equipment-specific applications.Material and coating tests, often exported as reports or result files.CMM, metrology, and SPC results stored by the inspection software.Work orders, operations, serial numbers, and material lots.Older databases, Access files, and proprietary exports.

Layer 02

Collection & Transformation

Connectors and pipeline stages

OPC UA clientsBrowse and subscribe to server nodes; map node IDs to equipment signals.REST API adaptersPaginated, authenticated extraction from application APIs.SQL extractionRead-only incremental queries using watermarks or change tracking.File ingestionVersioned parsers for CSV, Excel, XML, and proprietary exports.Data validationType, range, completeness, and duplicate checks with rejection reports.Timestamp normalizationConvert local times and device clocks to UTC with known precision.Unit conversionDeclare source units explicitly and convert to agreed engineering units.Schema mappingMap source fields to the contextual model, preserving original names.

Layer 03

Contextual Manufacturing Data Model

Syngetic implementation categories

Process Definition

What should happen: the designed process and its requirements.

  • Equipment
  • Recipes
  • Specifications
  • Process steps
  • Operation definitions

Material

What the process consumes and transforms, with lot genealogy.

  • Workpieces
  • Powder lots
  • Consumables
  • Material identifiers

Execution

What actually happened: runs, events, and actual parameter values.

  • Manufacturing runs
  • Timestamps
  • Actual parameters
  • Machine events
  • Operator actions

Observations

What was measured about the process and the product.

  • Sensor measurements
  • Inspection results
  • Laboratory tests
  • Quality assessments

Layer 04

Applications & Analytics

Where contextual data is used

OEM software platformsDelivered through the platform's own ingestion API or data model.Manufacturing dashboardsContext-aware views by line, machine, order, or part.Process monitoringActual parameters compared with recipe and specification limits.Quality analyticsInspection and lab results joined to the process that produced them.TraceabilityPart and lot genealogy across operations and equipment.Statistical analysisSPC, capability, and correlation on consistently structured data.Predictive modelsFeatures with known context and provenance, ready for validation.

Interactive · Process context

Definition Plane vs. Execution Plane

What should happen vs. what did happen — linked by identifiers.
A temperature reading is never just a temperature. It belongs to an operation, a machine, a recipe and a run. That context makes it useful.
Synthetic example data
  1. Definition

    Part Family

    PF-SEAL-RING

    ↕ part_number

    Execution

    Serial Number

    SN-1042

  2. Definition

    Manufacturing Operation

    OP-40 · Plasma coating

    ↕ operation_id

    Execution

    Production Run

    RUN-1042 · 14:00–14:38Z

  3. Definition

    Equipment

    SPRAY-01 · APS cell

    ↕ equipment_id

    Execution

    Actual Parameters

    600 A · Ar 45 slpm

  4. Definition

    Recipe

    RECIPE-07 · v3

    ↕ recipe_id @ version

    Execution

    Sensor Measurements

    substrate_temp 142 °C

  5. Definition

    Process Requirements

    Substrate ≤ 150 °C

    ↕ characteristic + limits

    Execution

    Inspection Results

    thickness 248 µm

Context walk · 1/7

A reading arrives

A sensor reports 142 °C. On its own, this is only a number with a unit and a timestamp.

substrate_temp: 142 °C

Identifiers

Shared keys are designed up front, not discovered later.

Time windows

No shared key? We match by validated time windows.

Confidence

Uncertain matches are flagged — never guessed.

Interactive · Data lineage

Follow a Manufacturing Data Point

One synthetic plasma-coating measurement, from sensor to platform.
All values synthetic

Stage 1 — Source. A spray-process diagnostic sensor measures in-flight particle velocity during plasma coating on SPRAY-01.

SHOP FLOORINTEGRATIONSPRAY-01plasma torchSN-1042couponPV SENSOR185 m/sCoating labCPN-1042-A248 µmConnectorextractPipelinenormalizeContext modelrelateCustomerplatform
sensor-frame.txt
Synthetic
// raw frame as emitted by the diagnostic sensor
"PV_AVG=185.0;UNIT=m/s;T=04/12 14:02:11.348"
// device clock · local time · no part or run reference
1 / 6

All identifiers, values, and endpoints shown are synthetic and for illustration only. Use ← → to move between stages.

Process data that doesn't add up?

Bring us the equipment list and the question you need answered.