Solar DAS Network Design: Fiber vs Copper vs Radio

Why Network Medium Decides Whether Your Solar Data Acquisition System (DAS) Data Is Defensible

Sixty percent of solar site data gap investigations trace back to the physical layer, not sensors, not software, not the historian. The network medium determines whether every measurement reaches the solar data acquisition system (DAS) intact. Choose the wrong medium for the site conditions, and no amount of calibration or software configuration fixes the silent packet loss that follows.

When the Cable Run Length Decision Goes Wrong

A 50 MW site in the Southwest commissioned in 2023 had all sensors calibrated, all tag scaling verified, and all time sync confirmed. The commissioning team signed off. Operations took over.

Six weeks later, the asset manager flagged 18–22% missing data from the solar data acquisition system (DAS) historian during afternoon peak hours. Not every day, but on the hottest days when inverters ran hardest. The investigation took three weeks to isolate.

The cause was a 140-meter Cat6 cable run from the combiner box aggregator to the substation switch, routed alongside the string inverter AC output cable tray. Specifically, inverter switching transients were inducing CRC errors on the Ethernet run. The link showed green. Notably, no alarms fired. As a result, the solar data acquisition system (DAS) discarded corrupted packets silently.

That rework involved replacing 140 meters of Cat6 with shielded fiber, adding new conduit routing away from the cable tray, re-testing the full run, and updating as-builts. In total, it added 11 days to the project schedule and significant labor cost. Notably, the physical layer decision (made months earlier during design) had never been reviewed against the inverter layout.

In short, the solar data acquisition system (DAS) network carries every measurement from sensor to historian. The choice of physical medium (copper, fiber, or radio) determines latency, data completeness, failure modes, and maintenance burden for the asset’s 30-year life.

This post defines when each medium works, when it fails, and how to build a hybrid topology that meets IEC 61724-1:2021 Class A requirements from day one. It is written for project managers, commissioning leads, SCADA engineers, and O&M teams.

For a complete picture of how sensor data flows before it enters the network, see our guide to solar DAS sensor mapping and data flow.

Copper Ethernet in Solar Data Acquisition System (DAS) Design

Copper Ethernet’s fundamental constraint is the 100-meter segment length defined by IEEE 802.3-2022. In practice, on a utility-scale solar data acquisition system (DAS) site, string inverter rows are commonly 150–300 meters from the nearest aggregation switch. Consequently, every run over 100 meters requires an additional managed switch. That means another power supply, another enclosure, another configuration, and another failure point. For example, a 200 MW site with inverter rows on a half-kilometer grid could require dozens of extra switches under an all-copper design.

EMI Risk on Copper Runs Near Inverters

PV string and central inverters generate switching noise during DC-to-AC conversion. Specifically, the primary noise range is 2 kHz to 150 kHz. Still, modern inverter designs reduce conducted EMI at the output terminals. However, radiated EMI and cable coupling remain risks when Ethernet runs share cable trays with AC output cables.

For context, Cat6 UTP (unshielded) provides minimal protection against common-mode EMI. By contrast, Cat7 S/FTP adds foil-shielded pairs and an overall braid shield. As a result, it cuts EMI pickup by 20-30 dB. That said, Cat7 costs 3–4× more than Cat6 and requires proper grounding of the shield at both ends, adding installation complexity.

The failure mode is insidious: CRC errors cause packet drops, but the Ethernet link remains up. No fault alarm fires. As a result, the solar data acquisition system (DAS) historian shows missing data, which often looks like sensor failure, data logger failure, or software bugs. In practice, root-causing it to the cable run requires systematic CRC error tracking in the managed switch, a step most field validation protocols do not include by default.

When Copper Ethernet Is Appropriate

Copper is the right choice in three specific situations. First, use it for device-to-device connections within a NEMA 4 enclosure where runs are fixed and short (under 5 meters). Second, it works well for substation-local connections between rack-mounted equipment in a controlled room. Third, it is appropriate for last-hop connections to field devices within 60–80 meters of the aggregation switch, short enough to stay well under the 100-meter limit after accounting for cable routing overhead.

For any solar data acquisition system (DAS) network run that crosses inverter row cable trays or exceeds 80 meters in planned routed length, fiber is the engineering-defensible choice.

Bar chart comparing data completeness rates across six DAS communication media: single-mode fiber ring, multi-mode fiber star, Cat7 shielded copper, Cat6 UTP near inverters, licensed 900 MHz radio, and unlicensed ISM radio.

0%
25%
50%
75%
100%

Solar DAS Network: Data Completeness Rate by Medium


Single-mode fiber (ring)
99.8%


Multi-mode fiber (star)
99.5%


Cat7 shielded copper
96.2%


Cat6 UTP (near inverters)
82.5%


Licensed 900 MHz radio
97.1%


Unlicensed ISM radio
91.8%

Source: REIG Solar field commissioning data (n=52 sites)

Fiber Optic: The Backbone of Any Solar Data Acquisition System (DAS)

Single-mode OS2 fiber has a specified attenuation below 0.4 dB/km per fiber optics in utility-scale solar installations (Fluke). In practice, on a 100 MW site with field instruments at 2–4 km from the main SCADA server, fiber delivers full-signal-strength data with sub-millisecond propagation delay. Indeed, distance is not a design variable for fiber. A 5 km run, a 15 km run, and a 40 km run all behave the same from a data quality standpoint.

More important: fiber carries light, not electrical current. As a result, inverter switching transients cannot couple onto optical fiber. In addition, ground potential differences can damage copper cable runs that connect equipment at different earthing points across large outdoor sites. Fiber is immune to both. In most utility-scale projects, those two properties alone justify the higher installation cost over copper for all backbone runs.

Ring Topology vs. Star Topology

Most utility-scale solar data acquisition system (DAS) deployments use a fiber ring rather than a star. Specifically, a ring provides path redundancy. If one segment is cut by excavation equipment, a cable pull gone wrong, or rodent damage, traffic reroutes the other direction around the ring. As a result, SCADA and telemetry data continue with minimal interruption. Managed switches with RSTP configured typically reroute in under one second.

A star topology has no redundancy. Specifically, a single fiber cut between the aggregation switch and a field device kills all data from every device on that branch. For example, on a DAS where a substation fiber trunk serves 20 combiner boxes and 4 inverter skids, a star topology means one backhoe can silently take out a quarter of the site’s telemetry. For this reason, ring topology is the standard for utility-scale sites and is the architecture REIG Solar specifies on all projects above 5 MW.

In practice, two fiber types matter for solar DAS design (see FOA fiber optic reference guide for detailed specifications):

• Single-mode (OS1/OS2): Runs up to 40–80 km depending on link budget, used for backbone connections and long field runs. Required for any run exceeding 500 meters.
• Multi-mode (OM3/OM4): Runs up to 300–550 meters, supports higher bandwidth for short building interconnections. Use within enclosures or substation buildings, not for outdoor field runs.

Radio in Solar Data Acquisition System (DAS) Networks

900 MHz ISM radio tops out at 115.2 Kbps. That is enough for a handful of RTUs polled at 15-minute intervals, and nothing more. Still, radio has a narrow but legitimate role in DAS network design: bridging over physical obstacles where running fiber is genuinely not feasible: river crossings, lease-line crossings, and temporary monitoring during construction. Radio also provides backup uplinks for remote RTUs where cellular coverage is too unreliable. However, it is not a substitute for fiber. The bandwidth ceiling and interference susceptibility introduce failure modes that a wired backbone avoids entirely.

When Radio Is a Valid Choice

Three radio types appear in solar DAS deployments:

The 900 MHz ISM band runs at 115.2 Kbps. A site with 50 RTU polling points at 2-second intervals generates roughly 3.75 Kbps of steady-state traffic. That sounds manageable. However, add alarms, event logs, SCADA command traffic, and FHSS retransmission overhead, and the link saturates. As a result, a solar data acquisition system (DAS) using unlicensed 900 MHz as its primary backbone is one congestion event away from data gaps.

NIST SP 800-82 Rev. 3 also recommends licensed spectrum for all OT control network communications and classifies ISM-band radio as elevated risk on isolated OT networks where interference cannot be controlled or predicted.

Measurement, Meaning, Control: What Each Medium Delivers to Your Solar Data Acquisition System (DAS)

Fiber meets all three MMC criteria unconditionally across any site scale. By contrast, copper meets Measurement and Meaning only when runs stay under 80 meters and away from inverter cable trays. ISM-band radio meets none of the three reliably under continuous SCADA load. The sections below show exactly where each medium falls short.

Measurement: Raw Data Delivery

Can the medium deliver raw sensor readings with acceptable latency and completeness? IEC 61724-1:2021 Class A monitoring requires continuous data capture with 1-minute averaging intervals. Fiber delivers this reliably across any site scale. Similarly, shielded copper within 80 meters delivers it in low-interference environments. However, ISM-band radio delivers it conditionally. Under 2–5% packet loss during interference events, the 1-minute sample cadence produces compounding gaps. Consequently, those gaps invalidate performance ratio calculations for affected periods. See our guide to DAS commissioning targets for completeness, accuracy, and latency for the numerical thresholds used in REIG witness testing.

Meaning: Data Integrity Across the Link

Does the medium preserve data integrity: correct values, correct timestamps, correct sequence? Fiber CRC error rates in field conditions are below 0.001%. By contrast, copper within 80 meters of inverter rows, using Cat6 UTP, can produce CRC error rates of 3–18% under peak EMI loading.

Those errors produce packet drops, not data corruption. Specifically, a dropped 1-minute record means a gap, not a bad value. As a result, operations analysts sometimes misread those gaps as sensor failures. Similarly, radio under congestion produces gaps, with the addition of potential out-of-order delivery.

For any solar data acquisition system (DAS) feeding a financial-grade historian, data integrity is not academic. In fact, investors and O&M contracts specify minimum data completeness percentages for performance ratio validation. For a discussion of how data tagging preserves meaning at the application layer, see solar DAS tagging standards for units, scaling, and QC.

Control: SCADA Command Reliability

Can SCADA send command messages reliably over this medium? Fiber and shielded copper support deterministic, low-jitter delivery of DNP3 control commands. By contrast, unlicensed radio requires acknowledgment and retry logic, adding latency on each retransmission event. For safety-critical functions (inverter trip, reactive power setpoint commands, and curtailment), a solar data acquisition system (DAS) control path that relies solely on unlicensed radio is not commissioning-ready. Therefore, licensed spectrum or fiber should always be on the control path for commands that affect grid interconnection.

Failure Modes in Solar Data Acquisition System (DAS) Networks

Sixty percent of the data gap investigations REIG Solar conducts on sites not designed in-house trace back to two copper failure modes: EMI-induced CRC errors and corroded outdoor connectors. Importantly, both are preventable at the design stage. By contrast, a fiber cut announces itself: the link goes down, an alarm fires, and the technician dispatches with an OTDR. A copper EMI problem, however, announces nothing. It simply takes data while every status indicator stays green.

Prevention Is Cheaper Than Any Field Investigation

The two copper failure modes (EMI and connector degradation) account for approximately 60% of the data gap investigations REIG Solar conducts on sites not originally designed with REIG. Still, both are preventable at design, not at the troubleshooting stage.

Lollipop chart showing typical recovery time in hours for six DAS network failure types: fiber cable cut, fiber connector fault, copper EMI recabling, copper connector failure, radio path obstruction, and radio spectrum interference.

Solar DAS Network: Failure Recovery Time (Hours)

0h
2h
4h
6h
8h

Fiber: cable cut
6h

Fiber: connector fault
0.5h

Copper: EMI recabling
5h

Copper: connector fail
0.5h

Radio: path obstruction
2h

Radio: spectrum congestion
4h+

Source: REIG Solar field commissioning data

Choosing Your Solar Data Acquisition System (DAS) Network Topology

The right topology for a solar data acquisition system (DAS) network depends primarily on site scale, inverter layout, and available fiber installation routes. The decision framework below reflects REIG Solar’s current design standard, derived from deployment data across 52 utility-scale projects. In practice, most sites above 5 MW default to a fiber ring backbone with copper last-hops.

Four Rules for Any Solar Data Acquisition System (DAS) Design

Four rules of thumb apply to any DAS network design regardless of site scale:

1. Use fiber for all runs over 80 meters, or any run that crosses inverter rows, AC cable trays, or transformer pads
2. Use copper only for device-to-device connections within a single enclosure or temperature-controlled room where runs are fixed and short
3. Use radio only where physical cable is genuinely impossible, not merely inconvenient or slightly more expensive
4. Apply NIST SP 800-82 OT segmentation requirements regardless of medium: the solar data acquisition system (DAS) field device network should be on its own VLAN, isolated from corporate IT by a firewall or data diode; see our guide to solar plant SCADA network topology and OT zone design

REIG Solar deploys field-proven RenergyWare NEMA 4 / UL-listed solar data acquisition system (DAS) enclosures. Additionally, each unit ships with pre-terminated, factory-tested fiber and copper patch points. Specifically, they arrive built and verified to IEC 61724-1 Class A performance requirements. Yet across 50 utility-scale sites, the most consistent rework trigger is a copper run drawn at exactly 100 meters. In practice, the actual routed cable path through conduit and cable trays runs 20–30% longer. Consequently, that exceeds the IEEE 802.3 segment limit and produces intermittent CRC errors that no one caught at design review. The fix is straightforward: design to 80-meter maximums for copper and specify fiber for anything longer.

A solar data acquisition system (DAS) network that follows these principles will support not just the initial COD utility witness test but the full O&M data quality standard for the asset’s 30-year operating life. Contact REIG Solar to review your DAS network design before installation.

Further Reading

• Solar DAS Commissioning Checklist: Sensors, Scaling, and Timestamps
• Solar DAS Time Sync: NTP and PTP Without Drift
• Solar DAS Commissioning: Irradiance and Weather QA
• Solar SCADA ROI: How Controls and Data Increase Revenue

References

• IEEE 802.3-2022: IEEE Standard for Ethernet (distance and performance specifications for copper and fiber Ethernet)
• NIST SP 800-82 Rev. 3: Guide to Operational Technology (OT) Security (OT network segmentation, licensed vs. unlicensed radio guidance)
• IEC 61724-1:2021: Photovoltaic System Performance Monitoring (IEC Webstore) (Class A/B accuracy and data completeness requirements)
• Fiber Optics in Utility-Scale Solar Installations (Fluke) (commissioning and testing fiber networks in solar facilities)