Industry 4.0: Connect Your Factory Without Rebuilding Your Network

· 10 min read
10 min read
Lucas Moreau
Lucas Moreau
OT/IT Network Engineer

Industry 4.0 doesn't require you to rebuild your factory from scratch. It requires you to intelligently connect what you already have. Here's how to turn 15-year-old Modbus machines into IoT assets—in a matter of hours, not months.

The Real Challenge of Industry 4.0: Field Connectivity​

Industry 4.0 projects rarely fail because of the analysis or visualization phases. They fail because of issues with collecting field data.

The reality in French factories:

  • 70% of PLCs still communicate via Modbus RTU over RS-485
  • The production floors lack an industrial Ethernet network—or the network is fragmented and aging
  • The electrical cabinets are mounted on DIN rails in areas with temperatures of 60°C and subject to vibration
  • The CIO refuses to expose the OT network to the IT LAN

The result: we have machines that can generate data but are unable to feed it back into the analytics systems.

The solution is not to replace the PLCs. It is to add a layer of intelligent connectivity between the field and the cloud.

Before:

After:

An industrial IoT gateway like the Eziwan Gateway collects Modbus data from your machines and transmits it to your monitoring platform via LTE—without affecting your existing network.

What an Industrial Gateway Actually Does​

On-siteCloud-based
Reads Modbus RTU (RS-485) registersPublishes to your MQTT broker
Reads Modbus TCP registersExposes a REST API
Reads 4-20 mA analog inputsPushes data to InfluxDB / Grafana
Reads digital inputsTriggers email/SMS alerts
Buffers data in case of disconnectionResends data upon reconnection

Deploy Without Interrupting Production​

Step 1: Inventory of Modbus Slaves​

Before connecting anything, map your Modbus slaves:

Machine 1 : variateur Schneider ATV320 — Modbus RTU, ID 1, 9600 baud
Machine 2: Siemens S7-200 PLC — Modbus RTU, ID 2, 9600 baud
Machine 3: Endress+Hauser flow sensor — Modbus RTU, ID 3, 19,200 baud
Machine 4 : IHM Weintek — Modbus TCP, 192.168.1.50:502

Identify the registry keys you're interested in: energy consumption, temperatures, room meters, and alarm statuses.

Step 2: RS-485 Wiring (30 minutes per cabinet)​

The RS-485 Modbus bus supports up to 32 slaves over a distance of 1,200 meters. In most workshops, all the PLCs on a line already share an RS-485 bus—you simply need to connect the gateway in parallel.

Important: The gateway is the Modbus master. It initiates all requests. Your existing PLCs remain masters on their own loops—there is no conflict.

Step 3: Configuring the Registers to Collect​

In the Eziwan console, configure which registers to read and how often:

# Configuration collecte Modbus — Ligne d'assemblage Atelier 3
modbus:
  port: rs485
  baud_rate: 9600
  parity: none
  
  slaves:
    - id: 1
      name: "Variateur ATV320"
      polling: 10s
      registers:
        - address: 8602    # Output Frequency (Hz × 10)
          name: frequence_sortie
          scale: 0.1
        - address: 16421   # Motor current (A × 10)
          name: courant_moteur
          scale: 0.1
        - address: 16420   # État défaut
          name: etat_defaut
          
    - id: 2
      name: "Compteur pièces"
      polling: 1s
      registers:
        - address: 0       # Compteur 32 bits (haut)
          name: pieces_total_hi
        - address: 1       # Compteur 32 bits (bas)
          name: pieces_total_lo

Step 4: Validation of the Data Collection​

Once configured, the Eziwan dashboard displays the collected values in real time. Verify that:

  • All machines are reporting data
  • The values are consistent (no 65535 = register missing)
  • Polling does not generate any Modbus errors (timeout, CRC)

Industry 4.0 Use Cases That Deliver Immediate ROI​

Real-Time Production Tracking​

The simplest and fastest-to-deploy option. A single Modbus register (piece counter) is all that’s needed to obtain:

  • Current production vs. target
  • Instantaneous rate vs. nominal rate
  • Automatically calculated downtime (when the counter stops incrementing)
  • OEE (Overall Equipment Effectiveness) calculated automatically

Energy Monitoring​

By reading the energy logs from variable-speed drives and power analyzers, you can obtain:

  • Energy consumption in kWh per machine, per workstation, per shift
  • Detection of machines that consume energy outside of production hours
  • Identification of consumption peaks (billing for power draw)

Anomaly Detection via Modbus​

Engine Bearing Temperature Log:
  Valeur normale : 45–65°C
  Alert threshold: > 75°C → email alert
  Critical threshold: > 85°C → SMS alert + recommended shutdown

These thresholds can be configured in Eziwan with just a few clicks. There's no need to modify the PLC program.

Integration with Your Existing Tools​

An Industry 4.0 gateway should integrate into your ecosystem, not impose itself on it:

Existing ToolEziwan Integration
GrafanaMQTT publishing → InfluxDB → Grafana
Ignition SCADAMQTT via Sparkplug B or OPC-UA
SAP ERPREST API webhook → SAP connector
OSIsoft PI HistorianMQTT → PI Connector
Excel / Power BICSV export or REST API

See available integrations → · Download the Industry 4.0 deployment guide →

Calculate OEE in real time using Modbus data​

The TRS (Total Rate of Return) is the key indicator of a machine's productivity. It is broken down into three factors:

TRS = Availability × Performance × Quality

Availability = Actual production time / Planned production time
Performance   = (Parts produced × Nominal cycle time) / Actual production time
Quality = Conforming Parts / Total Parts Produced

Using the Modbus data transmitted by the Eziwan gateway, the TRS is calculated automatically:

# Modbus registers required for TRS calculation
machines:
  - name: "Ligne d'assemblage L3"
    slave_id: 1
    registers:
      # Disponibilité
      - address: 0x0010    # Machine en cycle (1=oui, 0=non)
        name: en_cycle
      - address: 0x0011    # Machine shutdown (codes: 1=malfunction, 2=batch change, 3=material shortage)
        name: code_arret
      
      # Performance
      - address: 0x0020    # Total Count Counter (int32 over 2 words)
        name: pieces_total
      - address: 0x0021    # MSW Piece Counter
        name: pieces_total_hi
      - address: 0x0022    # Current output rate (pieces per hour)
        name: cadence_actuelle
      - address: 0x0023    # Nominal rate (constant, e.g., 120 pieces/hour)
        name: cadence_nominale
      
      # Qualité
      - address: 0x0030    # Compteur rebuts
        name: rebuts_total

The Eziwan platform calculates these metrics in real time and displays them on the dashboard—no on-site MES server is required.

Industry benchmarks for industrial OEE:

  • OEE > 85%: excellent (World Class Manufacturing)
  • OEE 60–85%: acceptable, targeted improvements possible
  • OEE < 60%: significant hidden losses exist

A real-time OEE display, visible to operators at the end of the line, typically reduces unplanned downtime by 15 to 25 percent within three months—simply by increasing transparency.

OPC-UA and Sparkplug B for Modern PLCs​

For machines equipped with modern PLCs (Siemens S7-1500, Schneider M580, Beckhoff TwinCAT), there are two alternatives to Modbus RTU that offer more advanced integration capabilities:

# Example of OPC-UA data collection from an S7-1500 via Eziwan
# Access via the Eziwan VPN tunnel (OPC-UA endpoint exposed via a proxy gateway)

from asyncua import Client

async with Client(url="opc.tcp://[IP_VPN_EZIWAN]:4840") as client:
    # Self-published variables without modifying the PLC program
    temp_node = client.get_node("ns=3;s=Ligne3.Temperature_Four")
    pieces_node = client.get_node("ns=3;s=Ligne3.Compteur_Pieces")
    
    temperature = await temp_node.read_value()  # → 245.6
    pieces = await pieces_node.read_value()      # → 18432

Advantage of OPC-UA over Modbus RTU: No need to manually map registers; the semantics (name, unit, description) are included in the OPC-UA node. No risk of scaling errors.

MQTT Sparkplug B: Standardization of Industrial IoT Topics​

Sparkplug B (Eclipse Foundation specification) standardizes the structure of MQTT topics for the IIoT. Whereas a standard MQTT topic is free-form (usine/ligne3/temperature), Sparkplug defines a universal structure:

spBv1.0/{Group_ID}/{Message_Type}/{Edge_Node_ID}/{Device_ID}

SCADA systems that support Sparkplug (Ignition MQTT Engine, HiveMQ Sparkplug, EMQX) receive data from the gateway with automatic decoding—without the need for manual mapping configuration.

The Eziwan Gateway supports Sparkplug B for native publishing in Ignition integrations.

Conclusion: Industry 4.0 is a project that takes months, not years​

Connecting an existing production line to the cloud via an industrial LTE gateway typically takes 1 to 3 days of technician work per line:

  • Modbus inventory: 2–4 hours
  • Wiring and installation: 2–4 hours
  • Configuration and validation: 2–4 hours
  • Team training: 2 hours

The first ROI (production count, real-time OEE) is visible as early as Week 1.

Industry 4.0 doesn't require an infrastructure budget. It requires the right tools.

FAQ​

Do I need to replace all my machines to transition to Industry 4.0? No. The "retrofit" approach involves connecting existing machines to IoT sensors and gateways without replacing them. 80% of industrial equipment already has an RS-485 Modbus port or an Ethernet interface—all you need to do is connect a gateway to it. Replacing machines is optional and can be planned gradually over 5 to 10 years.

What is the minimum budget for an Industry 4.0 pilot project? For a pilot involving 3 to 5 machines: 1 to 2 gateways (€800 to €1,600), 1 cloud subscription (€40 to €80/month), and 1 to 2 days of technician installation. Total: €2,000 to €5,000 in capital expenditures (CapEx) and €50 to €100 per month in operating expenses (OpEx). This budget allows you to obtain initial real-time production data and validate the ROI before a full-scale deployment.

Is Industry 4.0 only for large-scale manufacturing? No. Industrial SMEs (20 to 200 employees) are often even more agile when it comes to adopting these technologies—fewer legacy systems, faster decision-making. Cloud solutions like Eziwan are designed for SMEs: no server infrastructure to maintain, deployment in just a few hours, and a monthly subscription with no long-term commitment.

Further Reading​

Calculate your Industry 4.0 ROI → · Start a free pilot →

Additional Resources​