String-Level Power Routing Unit: MOS Drivers, Fault Detection & Telemetry
← Back to: Renewable Energy / Solar & Wind
This article answers key questions about designing and optimizing string-level power routing units in PV systems. It covers topics like component sizing, fault handling, and system reliability in different environmental conditions.
What this page solves
Large PV plants often connect many strings into a shared combiner or DC bus. Once wired, the topology is essentially static: every string is bound to a fixed MPPT or boost input, and changing the routing usually means shutting down, visiting the site and physically rewiring or replacing fuses. When a few strings become weak due to PID, ageing or heavy soiling, they can drag down the operating point of the whole MPPT and waste available irradiance.
A shorted or heavily faulted string can also become a low-voltage sink for neighbouring strings. In a passive combiner, the only defense is fuse coordination, which may react slowly or operate in an uncomfortable region between “never blow” and “nuisance trips”. During that time, healthy strings may back-feed energy into the faulted path, overheating cables, connectors and module junction boxes and complicating fault location in the field.
Operations teams need a way to connect, disconnect and re-route individual strings under load; to isolate suspected strings without taking down an entire combiner; and to choose between hot bypass and hard isolation depending on the fault scenario. At the same time, upstream boost arrays and MPPT stages benefit from cleaner, more predictable inputs so that each power stage works with a cluster of healthy strings instead of being penalized by a few bad actors.
Modern SCADA and fleet analytics also expect basic telemetry and event logs at string level: which string was disconnected, why a protection event latched, and how often reverse-feed or overload conditions occurred. This visibility is difficult to obtain from inverter-side logs alone, especially when several combiners and dozens of strings share a common DC bus.
This page focuses on a string-level power routing unit built around multi-channel high-side MOSFETs. It explains how these switches, together with reverse-feed detection, fault bypass and basic telemetry, form a programmable routing layer that connects, isolates and reconfigures PV strings between the combiner and the boost or MPPT stages.
Functional scope and placement in the PV architecture
The string-level power routing unit sits between the PV module or string outputs and the front end of the boost or MPPT stages. It may be housed near a combiner or integrated into an advanced combiner enclosure, but its role is different from a passive fuse-and-terminal block. Instead of permanently wiring each string to a fixed input, it creates a programmable matrix of connections that can selectively connect, disconnect or re-route strings under control of a local controller or plant-level SCADA.
On the power path, the unit provides high-side switching for each PV string, supports controlled turn-on and turn-off under operating current, and enforces limits on reverse energy flow between strings. It does not replace DC-link energy storage or inverter ride-through functions; those remain in the boost converter, inverter and grid-protection stages. Its job is to ensure that each upstream power stage sees a clean set of strings that are electrically compatible and free from severe faults.
On the control and status side, the routing unit exposes per-string on/off commands, basic status bits and latched fault causes to a local MCU, PLC or gateway. High-level energy management and reconfiguration policies remain in the EMS or SCADA layer; the routing hardware simply guarantees that requested configurations are implemented safely, with appropriate interlocks and protection timing. Commands are typically exchanged over SPI, I²C, GPIO or a simple fieldbus, not over cloud-facing protocols.
Telemetry within the unit is deliberately modest: it concentrates on health and events rather than on precise energy billing. Typical signals include coarse per-string current regions, open or short detection, reverse-feed events, overcurrent or overtemperature trips and a small event log. High-accuracy IV measurement, revenue-class metering, arc-fault signatures and surge capture are delegated to dedicated measurement and protection front-ends, which interface with the same strings and DC bus at other points in the system.
Inside the functional boundary, the routing unit combines multi-channel high-side MOSFETs, suitable gate drivers, reverse-feed detection, basic overcurrent and temperature protection, a fault bypass path and a small telemetry and control core. Together, these blocks define a clear black box in the PV architecture: a string-level switch matrix with protection and status, positioned between the PV strings, the combiner and the boost or inverter stages.
Core architecture
The string-level power routing unit can be viewed as a protected switch matrix placed between multiple PV strings and the DC bus feeding boost or MPPT stages. On the power side, each string terminates at an input port and then passes through a dedicated high-side MOSFET channel before joining a shared routed output. Around this path, reverse-current detection, fault bypass or isolation channels and overcurrent and temperature protection circuits supervise how energy flows and when a string must be disconnected or bypassed.
On the control and telemetry side, gate drivers translate controller commands into controlled turn-on and turn-off of each MOSFET channel. A small analogue front-end monitors voltage or current differences to detect back-feed conditions and other fault signatures, while local protection logic enforces safe operating limits, independent of higher-level software decisions. A microcontroller, FPGA or PMIC core coordinates this logic, sequences switching operations and collects status bits and event flags.
Telemetry blocks aggregate per-string states, coarse current regions and fault causes into a compact view for SCADA and fleet analytics. Precise IV curves, revenue-class metering and advanced fault signatures are delegated to dedicated measurement and protection front-ends elsewhere in the PV system. Within this architecture, the routing unit remains a clearly defined black box: a set of string input ports, a multi-channel high-side MOSFET matrix, reverse-feed and protection functions, a fault bypass path, a modest telemetry core and a controller that binds everything together and exposes a clean interface to the rest of the plant.
This section describes how these blocks connect from left to right: from PV string inputs, through the MOSFET matrix, reverse-feed detection and fault bypass channels, into the routed DC output feeding boost arrays, while protection and telemetry circuits supervise and report the behaviour of each path.
High-side MOS driver and MOSFET selection
Selecting the MOSFETs and gate drivers for a string-level routing unit is not only about holding off the PV string voltage. Devices must carry the expected string current with acceptable conduction loss, survive worst-case inrush and fault conditions within their safe operating area and share current in parallel where needed. The choice of RDS(on), package, die size and the way several MOSFETs are arranged on copper determine both efficiency and long-term thermal stress in the field.
A lower RDS(on) reduces voltage drop and conduction losses but usually implies a larger die, higher gate charge and a higher device cost. At string level, the optimal point is often a compromise between marginal efficiency gains and the impact on driver sizing and layout. Very fast hard turn-on can excite large inrush currents into downstream capacitances, while overly slow switching leaves the MOSFET in its linear region and increases heating. The gate driver and external gate network therefore shape the dV/dt and dI/dt to respect both inrush limits and MOSFET safe operating area.
Safe operating area constraints are particularly important in PV routing. Source impedance varies with irradiance, temperature and the number of parallel strings, and fault currents can be far above the average operating current. Candidate MOSFETs should be checked against realistic worst-case waveforms so that pulse currents and durations remain inside their SOA curves across the full temperature range. Thermal design must consider high ambient temperature, direct solar loading and the cumulative heating of multiple channels in the same enclosure, not only average dissipation at standard test conditions.
When a single device cannot meet current or thermal requirements, several MOSFETs may be placed in parallel. In that case, current sharing depends on more than data-sheet R DS(on). Symmetric copper geometry, similar trace lengths and, where needed, small source resistors help avoid one device running consistently hotter than its neighbours. Temperature feedback and per-channel protection further reduce the risk of one MOSFET drifting into an unsafe region while others remain lightly loaded.
Reverse conduction is another critical aspect. The intrinsic body diode of each MOSFET can easily form an unintended back-feed path if the routed DC bus rises above a weak or shaded string. In some architectures, additional blocking MOSFETs or carefully controlled synchronous devices are added to prevent this reverse energy flow. The reverse-current detection described earlier works together with these devices so that suspected back-feed is detected and the relevant channels are turned off before diodes or MOSFETs are overstressed.
On the driver side, high-side or floating gate drivers must be able to deliver enough peak current to charge and discharge the combined gate capacitance at the chosen slew rate. Simple bootstrap drivers are rarely ideal at string level, because routing channels may stay on or off for long periods with no regular switching activity. Instead, isolated or auxiliary supplies for each driver cluster provide stable bias regardless of switching duty cycle. Many modern drivers also integrate desaturation or current-sense protection and fault outputs, which can be tied into the routing unit’s protection logic. These features help clamp fault energy and report abnormal conditions without relying solely on slower, software-based decisions.
Overall, the high-side MOSFET and driver set must be chosen as a pair, based on conduction loss, dynamic behaviour, safe operating area, thermal headroom, reverse conduction risk and the ability of the driver to handle gate charge and protection functions in the string-level routing environment.
Reverse-feed and backfeed detection
When several PV strings share a common DC bus, a weak or faulted string can sink current from healthier neighbours. Shading, PID, contact degradation or bypass diode conduction may pull one string’s voltage below the bus, turning it into a low-voltage sink. If not detected quickly, this backfeed causes unwanted heating in connectors and junction boxes and can mask the true origin of the fault when maintenance crews investigate in the field.
The string-level routing unit therefore incorporates a reverse-feed detection path for each protected string or group of strings. A small analogue front-end senses the voltage and current around the high-side MOSFET channel, looking for conditions where current is flowing from the routed DC bus back into a string. This is typically implemented with a low differential voltage measurement and a fast comparator, optionally supported by a shunt-based current sense amplifier to improve direction discrimination and noise immunity.
Thresholds and time-over-threshold behaviour are tuned so that genuine backfeed events are caught promptly while avoiding nuisance trips during normal MPPT dynamics or switching transients. A reverse-current threshold defines how much negative current or ΔV is tolerated before a fault is declared, and a short time filter or one-shot timer requires the condition to persist for a minimum duration. This prevents the MOSFET from chattering around the trip point and reduces thermal cycling on the power stage.
Once the backfeed comparator asserts and the time-over-threshold condition is met, a latch captures the event. The associated MOSFET channel is then forced off or transferred into a safe state regardless of higher-level routing commands, and the event is recorded for telemetry. Clearing this latched condition typically requires a deliberate reset from a local controller or SCADA workflow, ensuring that a suspected problematic string is not automatically reconnected without review.
In summary, reverse-feed detection acts as an edge-triggered protection path: it recognises backfeed conditions with dedicated analogue sensing and comparators, latches a fault decision to avoid oscillation, and immediately commands the routing hardware to disconnect the affected string while reporting a precise event for maintenance and analytics.
Fault bypass and routing logic
Detecting a problematic string is only the first step. The routing unit must also decide how the faulted path is treated on the power side: whether to fully isolate the string, provide a controlled bypass, activate a crowbar to provoke an upstream protection device or operate in a degraded mode that limits current without cutting the string entirely. These options are implemented as distinct hardware paths around the main high-side MOSFET channels and are controlled by a small but well-defined routing and protection logic core.
In the simplest case, a confirmed fault leads to hard isolation, where the main MOSFET is held off and the string is disconnected from the shared DC bus. For some fault classes, a soft bypass path implemented with dedicated bypass MOSFETs can route current around a damaged segment or local connection while enforcing a current limit. In severe or unsafe conditions, a crowbar path may deliberately short a node to drive an external fuse or breaker, ensuring that energy is removed quickly by upstream protection without allowing prolonged overload of the routing hardware.
Degraded modes sit between these extremes. When a string is marginal rather than completely failed, the routing logic can restrict its participation by limiting the time it remains connected, reducing its effective duty or constraining the maximum current allowed through its channel. These decisions are enforced by gate timing, current thresholds and temperature limits in the driver and protection circuits, so that a partially healthy string does not unduly stress the MOSFETs or drag down the performance of healthier strings on the same bus.
Because several strings and possibly several routed outputs share the same hardware, routing arbitration and load sharing rules are encoded in the control logic. Interlocks ensure that a given string is only connected to one output path at a time, that no output is overloaded with too many active strings, and that protection actions always override new connection requests. This arbitration is implemented with simple hardware priority and state machines, leaving higher-level optimisation algorithms to the EMS or SCADA layer.
Finally, safe routing changes require handshake with the downstream boost array or inverter. Before a group of strings is reconfigured, status lines or registers inform the power stage that inputs are about to change, allowing it to adjust duty cycle or enter a soft-update mode. Once bypass, isolation or re-routing actions are complete, a completion signal allows the power stage to resume normal tracking. This handshake reduces step-changes on the DC bus, avoids excessive inrush and helps MPPT algorithms converge rapidly on a new stable operating point.
The fault bypass and routing logic therefore binds protection decisions to concrete power-path actions: isolating or bypassing faulted strings, enforcing degraded operation where appropriate, arbitrating which strings connect to which outputs and coordinating transitions with the boost or inverter stages through simple hardware handshakes.
Telemetry & Event Reporting
This section covers the telemetry and event reporting of the string-level power routing unit, which includes low-resolution I/V sampling, MOSFET temperature monitoring, fault cause reporting, and communication with SCADA systems.
Telemetry data includes parameters such as MOSFET temperature and junction temperature estimates, string on/off status, fault cause codes (overcurrent, reverse feed, driver fault), and event logs. These parameters are essential for remote diagnostics and system maintenance.
The unit communicates via common interfaces such as RS-485, CAN, PWM telegraphy, or UART, transmitting telemetry data to SCADA systems and enabling both remote reset and local latch functionality.
This section also describes the role of event logs and fault codes in fault detection and reporting, facilitating quick response to system anomalies.
System-level integration & reliability
This section focuses on the integration of the string-level routing unit with other system components such as the combiner box, string-level fuses, and arc-fault units. It discusses how to avoid conflicts between protection systems and outlines the reliability features essential for safe, long-term operation.
Key reliability features include safe operating area (SOA) ratings, avalanche protection, thermal foldback to prevent overheating, redundancy for fault tolerance, and graceful degradation to ensure that the system continues to function at reduced performance in the event of partial failures.
This section also explains the importance of careful system integration to ensure that protection actions do not interfere with normal operation and that redundancies are in place to protect the system from single points of failure.
Recommended IC Role Mapping
This section outlines the key ICs involved in the string-level routing unit, detailing their roles and how they contribute to the overall functionality of the system. The recommended ICs include MOSFET drivers, current sense amplifiers, watchdogs, and communication transceivers, each fulfilling specific functions to ensure reliable and efficient operation.
High-side Gate Driver IC
The high-side gate driver IC is essential for controlling high-side MOSFETs, providing the necessary gate voltage to switch the MOSFET on and off, ensuring that the power path is controlled accurately.
Power MOSFET (low Rds(on))
A low Rds(on) power MOSFET is required to minimize conduction losses during operation, ensuring that power is efficiently transferred through the routing unit with minimal heat generation.
Reverse Current Detection Comparators
Comparators are used for detecting reverse current flow, enabling the system to take immediate action when backfeed from a faulty string occurs, protecting the overall system.
Sense-FET / Current-Sense Amplifiers (non-precision)
These components are used for low-precision current sensing. They provide valuable feedback on current flow, which is critical for fault detection and system monitoring.
Watchdog MCU
A watchdog MCU monitors the system’s operation, ensuring that it resets in the case of a fault or system hang, thus improving the overall system reliability.
Event Logging Flash
Event logging flash stores fault and event data, which can be retrieved for diagnostics and maintenance. It allows SCADA systems to track the history of the system and respond effectively.
RS-485 / CAN Transceivers
These transceivers enable communication with SCADA or remote systems, allowing telemetry data to be transmitted reliably over long distances.
Power-Good / Fault-Latch Supervisors
These ICs monitor the system’s power status and latch fault conditions. If a fault occurs, they ensure the system is safely shut down, preventing further damage.
Application Mini-Stories
20 MW PV Farm with 20 Strings, Two Severe Mismatches
In a 20 MW PV farm with 20 strings, two strings were severely mismatched, resulting in significant current imbalance. The string-level routing unit dynamically detected the mismatch and performed a **cut-off** on the faulty strings to prevent overall system inefficiency.
Desert PV Farm with Frequent Reverse Feed in High Temperatures
In a desert PV farm, reverse feed frequently occurred due to high temperature conditions. The solution was to integrate **fast comparators** and **latches**, which helped quickly detect reverse feed and immediately shut down the affected strings, preventing overheating and system failure.
Tracker PV Farm with Mismatched Strings at Dawn and Dusk
In a tracker farm, mismatches between strings were significant during dawn and dusk due to the low solar angle. The **routing unit** participated in optimizing the system by dynamically managing the string inputs to reduce mismatch and improve efficiency.
These stories highlight how the string-level routing unit is key to maintaining efficient system performance, even under challenging conditions. It dynamically responds to faults, manages mismatches, and ensures that the system remains stable and efficient.
Design Checklist
This checklist provides essential design verifications to ensure that the string-level routing unit meets all key performance, safety, and compatibility criteria. The items listed below are critical for validating system stability and functionality under extreme environmental conditions and in integration with other subsystems.
MOSFET SOA Meets Worst-Case Insolation
Ensure that the MOSFET’s **Safe Operating Area (SOA)** can handle the worst-case insolation conditions. The MOSFET should not exceed its rated limits under extreme environmental conditions, especially high solar irradiance.
Reverse-Feed Threshold Matches PV Measurement
The reverse-feed detection threshold should align with the **PV measurement** system to ensure correct detection of reverse current and prevent false triggers or missed detections.
Gate Driver Operation from -20°C to +70°C
Verify that the **gate driver IC** is capable of continuous operation in the temperature range of **-20°C to +70°C**, ensuring stable control of MOSFETs even under extreme temperature variations.
Fault Bypass Meets Cold/Hot Transients
Ensure that the **fault bypass mechanism** can withstand both **cold and hot transient conditions** without compromising system stability. This includes rapid transitions during faults and temperature changes.
Telemetry Compatibility with SCADA Protocol
The telemetry data collected should be compatible with the **SCADA protocol**, including correct data formatting, time-stamps, and event reporting to ensure smooth integration with SCADA systems.
Request a Quote
FAQs – String-Level Power Routing Unit
When do I need string-level power routing instead of a passive combiner?
String-level power routing is necessary when individual string monitoring, fault isolation, and dynamic power distribution are required. Unlike passive combiners, string-level routing allows for better control over each string, providing real-time fault detection and the ability to disconnect problematic strings without affecting the entire system.
How to size high-side MOSFETs for long PV strings?
Sizing high-side MOSFETs for long PV strings involves considering the maximum voltage and current ratings of the string, ensuring the MOSFET can handle the power dissipation and thermal load. The MOSFET should have a low Rds(on) to minimize conduction losses and be rated for the highest expected voltage from the PV array.
How fast must reverse-feed detection react?
Reverse-feed detection must react quickly to prevent reverse current from damaging the system. A typical response time is within milliseconds to ensure that reverse current is detected and the affected string is disconnected before significant damage occurs.
Should bypass be soft or hard isolation?
The choice between soft bypass and hard isolation depends on the fault severity. Soft bypass is useful for temporary issues like minor misalignments, while hard isolation is required for serious faults, providing a complete separation of the faulty string from the rest of the system to prevent further damage.
How to coordinate routing with boost array inrush control?
To coordinate routing with the boost array inrush control, the system must ensure that the input currents are carefully managed to avoid large current surges during startup. This can be achieved by using communication protocols to synchronize the routing unit and boost array, controlling the duty cycle and soft-start mechanisms.
What telemetry is essential at string level?
Essential telemetry at the string level includes parameters like string current, MOSFET temperature, fault cause codes, on/off status, and event logs. These parameters are critical for real-time monitoring, fault detection, and system diagnostics.
How to avoid MOSFET thermal runaway in desert conditions?
To avoid thermal runaway in desert conditions, the MOSFET should be properly heat-sinked, and the gate driver should support **temperature compensation**. Additionally, the system should incorporate **thermal foldback** to reduce power when the temperature exceeds safe limits.
When to use sense-FET vs shunt sensing for telemetry?
**Sense-FET** is preferred when high-precision current measurement is required, while **shunt sensing** is ideal for cost-effective, lower-precision monitoring. Sense-FET is more accurate but often more expensive than shunt-based methods.
How to design latch/reset logic to avoid oscillation?
To avoid oscillations in the latch/reset logic, incorporate **debouncing** and use **one-shot triggers**. The latch should only respond to clear, defined events to avoid false triggers and unwanted oscillations.
Can routing reduce mismatch losses in partially shaded arrays?
Yes, routing units can optimize power distribution in partially shaded arrays by dynamically isolating mismatched strings, preventing significant loss in system performance due to shading effects.
What gate driver protections are mandatory for PV routing?
Gate drivers in PV routing systems must include protections such as **over-voltage protection**, **over-current protection**, **short-circuit protection**, and **thermal protection** to ensure safe and reliable operation.
How to plan for derating under high-irradiance peaks?
To plan for derating under high-irradiance peaks, use **dynamic thermal management** techniques that adjust the power output based on real-time temperature and irradiance data. This ensures that the system can handle high-power peaks without exceeding thermal limits.
