Article

Architecture decides your integration hours

A digital sensor is not automatically an OPC-UA sensor. Where the wiring, protocol conversion and commissioning sit decides how much work you take on.

QB Systems

QB Systems

  • Modeled analysis · engineering hours
  • 3 architectures compared
  • 14 min read
6–14 h
native PoE, one vessel of 6 sensors (modeled)
10–30 h
transmitter / bus alternatives (modeled)
60–65%
peak effort reduction vs custom-hub builds

Most labs now buy “digital” sensors as a matter of course — Hamilton Arc, Mettler-Toledo ISM, Endress+Hauser Memosens. But a digital sensor is not automatically an OPC-UA or MQTT sensor. In nearly every architecture the probe still reaches the data layer through a chain of intermediaries — a transmitter, a converter, a gateway, a PLC, a SCADA system. Building and validating that chain is where the integration hours quietly add up.

01 What we measured

The metric: engineering hours to a live namespace

For a single vessel of six sensors, taken to one OPC-UA + MQTT namespace with alarms set and commissioning signed off.

6–14 h

Native PoE multi-sensor platform

QB Multisensor + QB Edge + QB Control

10–30 h

Transmitter-centric & bus/converter

per-box config and protocol bridging repeat

20–65%

Modeled reduction with native PoE

midpoint-to-midpoint across scenarios

02 The journey

Effort by architecture

One fixed scenario — six probes, one vessel — modeled midpoint hours.

Modeled integration effort (engineering hours)

Source: Modeled estimate — midpoints of the ranges in the analysis
  • Native PoE. Six probes terminate at one module; power and data share a single cable; the platform publishes OPC-UA / MQTT directly.
  • Transmitter-centric. Probes terminate at a multi-channel transmitter, then get bridged upstream through a PLC / SCADA / gateway.
  • Bus / converter. Probes reach a converter (~4 sensors each) or share an RS-485 Modbus bus; MQTT is added through a bridge.

03 Putting it together

What the totals say across use cases

Use caseQB native PoE (modeled h)Other architectures (modeled h)Indicative reduction
A. 1 vessel, 6 sensors6–14 h10–30 h~20–55%
B. 2 vessels, 12 sensors10–22 h14–56 h~20–60%
C. Mixed-vendor, 1 vessel~12–20 h~30–60 h~55–65%
D. Probe sharing across 2 vesselsSoftware reassignmentRewiring / tag change per swapRecurring effort removed

04 What it means for you

Consolidation is the lever

If integration effort matters, the architecture that removes repetitive per-box configuration wins before a single sensor is chosen. Sensor accuracy and product quality are a separate, out-of-scope question.

Methodology & assumptions

The three third-party architectures follow standard, publicly documented integration patterns — RS-485 Modbus RTU wiring, transmitter channel counts, converter capacities. Vendors also offer products and configurations that can shorten or reshape these steps. Mechanical port fabrication, SIP/CIP validation, GMP IQ/OQ paperwork and calibration SOPs sit outside these estimates and add similar effort whichever option you pick. Percentages are illustrative, derived midpoint-to-midpoint from the engineering-hour ranges above — not measured benchmarks or quotations, and not comments on measurement accuracy.

Curious to see how it works?

Schedule a demo. See how to accelerate and automate your bioprocesses.

Contact us