What this page solves

Insulation in MV/LV systems rarely fails overnight. Moisture, age and pollution slowly drag resistance down until a sudden fault, nuisance trip or dangerous touch voltage appears. This section clarifies which problems insulation monitoring is supposed to solve and how it complements leakage relays and periodic megger tests.

In outdoor LV panels, roadside cabinets and charging-station boards, rain, condensation and salt fog drive the long-term decay of insulation. RCDs and ground-fault relays only react when enough current already flows to ground, leaving a wide zone where insulation is weak but fault current is still below trip thresholds.

On MV feeders and transformer circuits, long cable runs and buried joints mean that every unplanned outage is expensive. Periodic offline tests provide snapshots of insulation, but do not show how fast resistance is drifting between test campaigns or how weather events impact the trend.

DC control power, LV DC microgrids and UPS buses introduce another blind spot. AC-side leakage protection cannot see insulation degradation on isolated DC links, particularly behind converters with high dv/dt and filter capacitors. A dedicated insulation monitoring chain is required to watch these nodes continuously.

The insulation monitoring function therefore needs to answer three practical questions:

• Is a dedicated insulation monitoring device really required for a given MV/LV or DC segment, or can simple leakage protection cover the risk?
• Where should the monitored zone start and end so that a single device represents a meaningful electrical boundary instead of a vague portion of the network?
• At what insulation level should the system raise a warning, schedule maintenance or actually trigger a shutdown, and how should those thresholds align with IEC guidance?

The rest of the page builds on these questions, moving from system placement to the injection-and-measurement chain, and finally to concrete thresholds, diagnostics and IC selection.

Role in MV/LV system

Insulation monitoring does not live in isolation. It sits alongside ground-fault relays, RCDs, surge protection and conventional protection relays. Understanding where insulation monitors connect in the MV/LV stack is essential to avoid blind spots and duplicate coverage.

At the MV level, a monitor can supervise a feeder segment or transformer winding as a single electrical zone. The device sees the combined insulation of cables, joints and connected equipment, and tracks how this bulk resistance drifts over months and seasons. It does not pinpoint a specific joint or termination; detailed partial-discharge diagnostics are handled elsewhere.

In LV distribution boards and panels, insulation monitors usually watch the main busbar and outgoing feeders of that cabinet. The goal is to provide a clear health indicator for that local zone: outdoor LV panels, MCCs in industrial plants or auxiliary boards around a substation. Each panel can then report its own insulation status upwards, enabling targeted maintenance instead of treating the whole facility as one block.

On DC links and control power rails, insulation monitors cover the nodes that AC leakage protection cannot see. A DC 24 V or 48 V control bus, a UPS DC bus or a LV DC microgrid may be galvanically separated from the AC network. By injecting and measuring with respect to protective earth, the monitor provides a continuous view of DC-side insulation without interrupting service.

All of these monitors are most effective when their zones align with protection and automation boundaries. A feeder-level monitor should describe the same segment that a protection relay trips. A panel-level monitor should cover the circuits under a specific LV board. The result is a layered view: MV zone insulation, LV panel health and DC bus insulation, all feeding consistent alarms to IEDs, PLCs and SCADA.

Key requirements & thresholds

An insulation monitoring chain must be specified against realistic MV/LV and DC bus conditions. The requirements below focus on measurable ranges, thresholds and robustness levels that determine whether the monitoring function remains useful once installed in harsh environments.

The monitored segment can be modelled as a resistance to protective earth in parallel with parasitic capacitance. For MV feeders and transformers, healthy insulation typically sits in the multi-megohm region, while actionable thresholds are often discussed between tens of kilohms and a few hundred kilohms. LV panels and DC control buses work with slightly lower ranges but rely on the same concept of a bulk insulation value that slowly drifts over time.

In practice, an insulation monitor should support at least the following ranges and thresholds:

• Insulation resistance range: operation from roughly 10 kΩ up to at least 10 MΩ, with coarse resolution in the multi-megohm region and finer discrimination between 20–200 kΩ where most warning and trip levels are set.
• MV feeder and transformer zones: healthy insulation commonly above 1–10 MΩ, warning discussions often starting below a few hundred kilohms, and trip thresholds set in the tens of kilohms range based on local safety rules and IEC guidance.
• LV panels and outgoing feeders: focus on tracking drift from several megohms down into the few hundred kilohm region, with thresholds that distinguish healthy panels from those needing inspection after moisture, contamination or cable damage.
• DC links and control power buses: typical interest between 10 kΩ and 1 MΩ, where declining insulation can increase touch-voltage risk even though AC-side leakage protection still appears normal.
• Injection current level: usually within 5–500 µA so that the measurement signal stands out from leakage and noise but does not create additional risk or stress under fault conditions.
• Measurement bandwidth and injection frequency: DC or low-frequency injection for direct resistance estimation, or narrow-band AC injection at frequencies that avoid 50/60 Hz and their harmonics and fit around converter noise spectra.
• Response time for warnings and trips: measurement refresh and filtering in the 100–500 ms range to reject short transients, with warning and trip decisions delivered fast enough to integrate with protection and automation logic.
• Voltage and surge withstand: components in the injection and measurement path dimensioned for the maximum system voltage and relevant surge levels, such as 600–1500 V peak and several kilovolt surge events.
• CMTI of isolation devices: at least 50–100 kV/µs common-mode transient immunity when connected near drives, inverters or STATCOM equipment with high dv/dt on their outputs.
• Noise tolerance and EMC margin: analogue and digital filtering robust against switching noise, lightning surges and communication bursts so that thresholds are crossed only by real insulation degradation.

These ranges provide a starting point for selecting isolation amplifiers, ADCs, references and comparators that can survive MV/LV conditions while still resolving the slow drift of insulation resistance around planned warning and trip levels.

Architecture options

Several architectures can implement insulation monitoring. The most common types are active injection schemes, frequency-selective L-C or impedance-based schemes, and simplified voltage-divider approaches used in small systems. Each option trades complexity, accuracy and robustness in different ways.

Active injection architectures apply a controlled current or voltage between the monitored segment and protective earth and measure the response. The response is then processed to estimate equivalent resistance and, in some cases, useful information about parasitic capacitance. This family covers both DC injection and narrow-band AC injection at frequencies chosen to avoid strong interference.

L-C or impedance-oriented schemes remain in the injection family but operate more explicitly in the frequency domain. A drive source shapes a known excitation, often at one or a few frequencies, and a synchronous detection stage isolates the response. This approach suits long cables and noisy converter environments where separating resistive and capacitive behaviour helps interpret the data.

Simple divider or bias-monitoring schemes rely on high-value resistor chains or biasing networks to observe how the potential of a conductor shifts as insulation drifts. These architectures can provide a coarse indication of insulation quality in low-voltage, well-bounded systems, but they are rarely suitable as the main protection mechanism for MV/LV distribution networks.

The sections below contrast the three options at signal-chain level and indicate where each approach tends to be appropriate inside a smart grid or power-distribution project.

Injection & AFE chain deep dive

The insulation monitoring signal chain starts with a stable reference and injection block, passes through a measurement front-end exposed to MV/LV conditions, and then crosses an isolation barrier before digital processing and threshold logic. Each block must tolerate the system voltage, switching environment and ground-fault behaviour while still resolving slow insulation drift.

The reference and injection section generates a controlled current or voltage towards the monitored segment and protective earth. A precision reference, a well-defined current or voltage source and a high-value resistor network set the injection level in the microampere range. Protection components and clear creepage distances are needed so that the injection path does not become a weak point under surge or fault conditions.

The analogue front-end then measures the resulting voltage or current. Depending on the architecture, this can be a current-sense transimpedance amplifier, a differential input stage around a high-side or low-side shunt, or a voltage-sense path that tracks a node potential. Input common-mode range, overvoltage protection and the ability to work with the chosen injection frequency are key selection criteria for the AFE devices.

An isolation amplifier or sigma-delta modulator transfers the measurement across the isolation barrier into the low-voltage domain. This part of the chain must combine the required working voltage and basic or reinforced insulation rating with high common-mode transient immunity so that fast dv/dt events from drives, inverters or switching surges do not corrupt the reading or create false trips.

On the secondary side, an ADC or digital interface feeds the measurement into an MCU, SoC or dedicated digital core. Filtering, demodulation and resistance estimation turn the raw signal into an equivalent insulation value. Threshold comparators and latches then derive warning and trip outputs, often with hysteresis and defined response times, and present them as relay contacts or logic-level signals for protection relays, IEDs and PLCs.

A robust chain keeps the injection predictable, the analogue front-end survivable under surges and common-mode stress, the isolation stage immune to dv/dt, and the digital logic transparent enough that warning and trip levels can be justified to safety and standards teams.

Safety & standards perspective

Insulation monitoring is part of a broader protection and functional safety concept rather than a stand-alone safety measure. Standards such as IEC 61557-8 and IEC 60204-1 describe how insulation monitoring devices behave in IT and machine systems, while functional safety frameworks such as SIL, PL or ASIL determine how their outputs integrate into safety functions implemented by relays, IEDs or safety PLCs.

IEC 61557-8 focuses on insulation monitoring devices in IT systems whose neutral is not directly earthed. The standard defines aspects such as rated system voltage, test current limits, response to insulation faults and the ability to verify functionality by applying known test resistances. These concepts influence practical choices for injection levels, threshold ranges and response times, even when the monitored system is not a textbook IT network.

IEC 60204-1 addresses electrical equipment of machines and clarifies how IT systems can be used in machinery. In that context, insulation monitoring supports a controlled first fault strategy: the first insulation fault does not immediately create a dangerous touch voltage, but must be signalled, and the system must not be allowed to accumulate a second fault without protective action. Clear documentation of the monitoring function and operator response is required.

When insulation monitoring outputs feed safety-related functions, those outputs are treated as safety inputs to a safety PLC, safety relay or protection device. The implementation must consider fail-safe behaviour, diagnostic coverage and suitable redundancy. Typical interfaces include normally closed contacts that open on fault, dual-channel signals to safety PLC inputs, or separated warning and trip contacts for graded reactions.

In a smart grid or industrial distribution setting, insulation monitoring can support several safety strategies. It can block system energization when insulation is already below a minimum level, request a controlled power reduction when a warning band is reached, or contribute to shutdown decisions when combined with ground-fault and overcurrent protection. Functional safety analysis then determines the achieved SIL, PL or ASIL, based on the behaviour and diagnostics of all elements in the safety chain.

Design and documentation should therefore link insulation monitoring parameters to the relevant clauses of IEC 61557-8 and IEC 60204-1, specify default and adjustable thresholds, describe self-test capabilities and define how warning and trip contacts are connected into safety relays, IEDs and safety PLC programs.

IC mapping & vendor landscape

Insulation monitoring relies on a small but specialised set of analogue, isolation and power-management ICs. Instead of individual part numbers, it is more useful to map IC types to the main manufacturers that support these roles in medium-voltage, low-voltage and DC-bus projects.

Along the injection and measurement path, precision references and high-voltage-tolerant amplifiers generate and shape the test signal. Low-noise operational amplifiers, current measurement front-ends and instrumentation amplifiers then extract the response under high common-mode stress. Isolation devices and sigma-delta modulators transfer this information into the low-voltage domain for ADCs and digital cores.

On the low-voltage side, general-purpose microcontrollers or SoCs handle injection control, filtering and insulation estimation. Comparators, window detectors, watchdogs and relay drivers translate the digital result into robust warning and trip interfaces. Auxiliary isolated power supplies and surge-protection components complete the signal chain so that the insulation monitor can survive overvoltage events and comply with installation rules.

Functional blocks and typical IC types

• Reference & injection: precision voltage references and current-source drivers, often based on low-drift bandgap references and high-voltage operational amplifiers.
• Measurement AFE: low-noise op amps, transimpedance amplifiers, high-side or differential amplifiers with protection networks and anti-alias filters.
• Isolation & high-side interface: isolation amplifiers, isolated sigma-delta modulators and multi-channel digital isolators rated for reinforced insulation and high CMTI.
• ADC & digital core: 16–24 bit sigma-delta ADCs or ADC-equipped MCUs and SoCs that can implement injection control, demodulation and resistance estimation.
• Comparators, latches & drivers: window comparators, precision comparators, system supervisors and relay or high-side drivers that provide clear warning and trip contacts.
• Power & protection: auxiliary and isolated DC-DC modules, surge arresters, TVS diodes, GDTs and MOV devices that protect the injection and measurement path.

Vendor landscape by functional block

The table below focuses on IC categories that are directly relevant to insulation monitoring. Metering SoCs, full protection-relay platforms and communication PHYs are covered on other pages to avoid overlap.

Application stories

Insulation monitoring appears in several smart grid and power-distribution projects with different environments and expectations. The following application stories highlight how thresholds, architecture choices and IC types come together in real installations such as EV fast charging, substation low-voltage panels and DC microgrids.

EV fast charger DC link

A DC fast charger runs at several hundred volts up to around one kilovolt and exposes the insulation monitor to strong PWM noise, high dv/dt and connector contamination. Active injection with a carefully chosen frequency and a narrow-band measurement path helps separate the insulation response from inverter switching artefacts.

Typical requirements include coverage from roughly 10 kΩ to the megohm region, warning thresholds around the low hundreds of kilohms and trip thresholds in the tens of kilohms. Response times in the few-hundred-millisecond range allow the charger control to ramp down current before opening contactors. The IC mix emphasises high-CMTI isolation amplifiers, sigma-delta ADCs and MCU or SoC platforms that can run demodulation and diagnostic routines.

Substation LV panel

Inside a substation LV panel, insulation monitoring supports long-term reliability and maintenance planning. Moisture, dust and ageing are the dominant stress factors. Injection schemes tend to operate at lower frequencies with an emphasis on trend stability and predictable alarm behaviour rather than absolute speed.

Monitoring ranges extend towards 10 MΩ, with healthy operation in the multi-megohm band and warning thresholds in the few-hundred-kilohm band. Trip levels are set lower and can integrate with bay controllers, IEDs or safety relays. Signal chains typically reuse industrial-grade references, AFEs, isolation devices and ADCs, while relay drivers and digital interfaces connect directly into SCADA and protection systems.

DC microgrid or BESS DC bus

DC microgrids and battery energy storage systems combine long cable runs, high DC voltages and frequent reconfiguration of strings and branches. Insulation monitoring helps detect developing faults that might not immediately appear in AC-side protection, but still raise touch-voltage and arc-flash concerns on the DC side.

In this environment, multi-frequency or impedance-oriented injection can distinguish between changes in insulation resistance and variations in parasitic capacitance. ADC and MCU resources must support these algorithms while still delivering clear warning and trip contacts. Integration into energy-management and safety functions allows reduced-power operation, selective branch isolation or full shutdown when insulation crosses agreed limits.

These examples illustrate how the same insulation monitoring building blocks must be tuned differently for high-noise EV chargers, utility-grade LV panels and evolving DC microgrids, with vendor selection guided by voltage rating, CMTI needs, required diagnostics and integration with protection and safety logic.

Design checklist

This checklist groups the main questions that hardware, system and safety teams can use during insulation monitoring design and review. Items are organised by topic so each line can be checked explicitly during schematic, layout, prototype and field-validation phases.

System context & operating conditions

• Has the monitored system voltage range been defined, including normal operation, temporary overvoltage and surge levels?
• Is the monitored boundary clear: complete feeder, selected branches, DC bus segment or a specific IT section?
• Have the dominant environmental stresses been identified, such as humidity, condensation, pollution, salt-mist or temperature cycling?
• Has an estimate of cable length and parasitic capacitance to earth been prepared for use in the Riso model?
• Has potential interaction between injection signals and existing metering, RCDs, PLC or power-line communication been evaluated?

Standards & safety role

• Has the intended system type (for example IT system or equivalent arrangement) been confirmed against IEC 61557-8 and IEC 60204-1?
• Is the role of the insulation monitoring device in the overall protection concept clearly described, including first-fault and second-fault behaviour?
• Has any linkage to functional safety targets (SIL, PL or ASIL) been defined, including which safety function uses the IMD outputs?
• Are fault reactions for an internal IMD failure documented, including whether the system is allowed to continue, to degrade or to shut down?
• Are all relevant clauses and test requirements from applicable standards listed in the design documentation for traceability?

Thresholds, measurement range & response time

• Has the required insulation resistance measurement range (for example 10 kΩ to 10 MΩ) been defined with engineering justification?
• Are healthy, warning and trip bands specified, and are the values traceable to touch voltage, fault current or utility requirements?
• Are maximum response times for warning and trip defined and aligned with the protection and control strategy?
• Is the impact of digital and analogue filtering on response time and threshold accuracy understood and documented?
• Are on-site adjustable thresholds constrained to ranges that remain compatible with standards and safety analysis?

Injection concept & frequency planning

• Has the insulation monitoring architecture been selected (active injection, L-C based or resistive monitoring) with a clear rationale?
• Are injection current or voltage limits defined so that test levels remain within allowed values under all conditions?
• Are injection frequencies chosen to avoid 50/60 Hz, harmonics and major converter PWM bands present in the installation?
• Are resistor chains and coupling components specified for the highest continuous and surge voltages, including creepage and clearance?
• Has the risk of interference with metering, RCD or power-line communication been checked using frequency-domain analysis or testing?
• For multi-frequency or scanning schemes, are bandwidth and memory margins in analogue and digital stages sufficient for future algorithm changes?

Front-end analogue chain & protection

• Does the analogue front-end support the required input common-mode range, including temporary overvoltage and surge conditions?
• Are input protection elements (series resistors, clamp diodes, TVS) dimensioned so that the measurement remains usable after surge testing?
• Are amplifier noise, offset, drift and bandwidth aligned with the chosen injection frequency and target resolution?
• Has the combined effect of Riso and parasitic capacitance on the analogue transfer function been analysed?
• Are specified op amps or instrumentation amplifiers rated for the required temperature range and industrial or utility environment?
• Has the contribution of AFE tolerances to overall measurement error been quantified in an error budget?

Isolation devices & insulation coordination

• Do isolation devices meet or exceed working-voltage, surge and insulation-class requirements for the targeted installation?
• Is CMTI rated against the highest dv/dt expected near drives, inverters or converters that share the same environment?
• Are isolated power supplies for the high-side domain selected with appropriate isolation ratings and leakage-current limits?
• Are PCB creepage and clearance distances around isolation devices consistent with datasheet and standards for pollution degree and altitude?
• Has susceptibility to sporadic bit errors under high dv/dt been evaluated and mitigated by diagnostics or redundancy where needed?
• Is the choice between analogue isolation and sigma-delta plus digital isolation aligned with ADC resolution and latency requirements?

ADC, MCU & digital processing

• Does the selected ADC provide enough effective resolution to achieve the required insulation resistance accuracy over the full range?
• Are sampling rates, digital filters and decimation factors sized for the injection scheme and specified response times?
• Does the MCU or SoC have sufficient computing headroom and memory for injection control, signal processing, diagnostics and communication?
• Are fixed-point or floating-point formats chosen to avoid overflow and quantisation issues in resistance estimation?
• Are parameter storage, threshold tables and calibration data preserved across power cycles and protected against unintended changes?
• Is an interface defined for exporting raw or processed measurement data during commissioning and troubleshooting?

Warning and trip outputs & system interfaces

• Are physical forms of warning and trip outputs defined, including contact type, logic levels and load capability?
• Is output behaviour fail-safe, for example using normally closed contacts that open on fault or loss of supply?
• Are latching and auto-reset behaviours specified for warning and trip, including how manual reset is performed where required?
• Are mappings from IMD outputs to protection relays, safety PLC channels, IED inputs or communication messages clearly documented?
• Are additional elements such as terminal blocks and auxiliary relays considered in the safety loop analysis and wiring diagrams?
• For safety-related applications, are dual or redundant channels provided where the target PL or SIL requires them?

Self-test, diagnostics & failure behaviour

• Are built-in self-test mechanisms defined, such as known test resistors, internal test injections or simulated fault paths?
• Is the coverage of self-test stated, including which internal faults can be detected in the injection, AFE, isolation, ADC and output stages?
• Are rules defined for when self-tests run, such as at power-up only or periodically during operation, and how nuisance trips are avoided?
• Is event logging planned for insulation warnings, trips and internal IMD faults to support maintenance and audits?
• Are fail-safe behaviours defined for key failure modes, including loss of injection, saturated AFE, ADC timeout or stuck outputs?
• Is system-level behaviour clearly described for IMD malfunction, including any allowed grace period, de-rated operation or immediate shutdown?

Layout, grounding, testing & production

• Are high-voltage, low-voltage and protective-earth domains clearly partitioned on the PCB, with controlled routing of injection and measurement paths?
• Is the grounding concept defined, including connections between PE, shields and signal grounds to avoid unwanted loops?
• Are high-value resistor chains and sensitive nodes located in low-leakage areas and covered by conformal coating or other protection if required?
• Does the test plan include ESD, EFT, surge and conducted immunity cases that specifically exercise the injection and measurement path?
• Is there a defined method for verifying insulation resistance measurement accuracy and threshold behaviour during type testing and commissioning?
• Are end-of-line tests planned for injection amplitude, analogue gain, threshold comparison and basic self-test functionality?
• Is a calibration strategy defined, including number of points, storage of calibration data and measures against field tampering?

FAQs about insulation monitoring

These questions capture common decisions and doubts that arise when planning and reviewing insulation monitoring in MV, LV and DC-bus systems. Each answer stays compact so you can reuse it in design reviews, specifications and maintenance guidelines.

When are RCDs and ground-fault relays not enough, and you really need an insulation monitor?

You add insulation monitoring when residual current devices and ground-fault relays no longer give you enough early warning. RCDs react to fault current; an IMD watches the gradual drop in insulation resistance, especially in IT systems, long cable runs and DC links, so you see the first fault before people or equipment are exposed.

How do you choose suitable insulation thresholds for MV/LV and DC bus systems?

Start by defining your system voltage, grounding scheme and local rules, then choose a healthy band, a warning band and a trip band. In many LV and DC systems you keep normal operation in the megaohm range, raise warnings in the few-hundred-kilohm range and trip in the tens-of-kilohms range.

Can your insulation monitor still be accurate on very long cables with high capacitance to earth?

A long line with high capacitance to earth makes the insulation monitor see both resistance and capacitive leakage. You manage this by choosing an injection frequency and filter bandwidth that keep the resistive component visible, or by using multi-frequency methods and segmenting the network when a single monitor cannot cover everything cleanly.

When should you use DC injection and when is AC or multi-frequency injection a better choice?

DC or low-frequency injection works well when the network is relatively short, noise is modest and you mainly need a simple go/no-go indication. AC or multi-frequency injection makes more sense on long cables, noisy DC buses and microgrids where you care about separating resistance from capacitance and tracking trends over time.

How do you design insulation monitoring that stays reliable inside noisy VFD or inverter cabinets?

In a noisy drive or inverter cabinet you treat dv/dt and common-mode noise as first-class design inputs. Isolation devices with high CMTI, robust analogue front-ends, tight filtering around the injection frequency and careful PCB layout help the IMD ride through switching edges without mis-tripping or silently losing measurement accuracy.

What should you watch for when you connect an insulation monitor to a safety PLC or safety relay?

The insulation monitor output effectively becomes a safety input when you wire it into a safety PLC or safety relay. Use fail-safe contacts, dual-channel input where required and monitored wiring so open circuits and welded contacts are detected. Configure the safety logic with suitable delays, test pulses and clear diagnostic messages.

If EV chargers already rely on the vehicle BMS, why add a separate insulation monitor on the station side?

A traction battery or vehicle BMS usually monitors insulation around the pack and chassis, while the charging station IMD focuses on the station-side DC bus, cables and cabinet. Combining both gives you coverage on the vehicle and the infrastructure. Relying only on the in-vehicle BMS leaves the station-side insulation largely unobserved.

How should insulation monitor alarms in a substation LV panel interact with protection and maintenance?

In a substation LV panel, treat the insulation monitor as both a maintenance sensor and a protection trigger. Use the warning level to generate SCADA alarms and maintenance tickets, and reserve the trip level for coordinated actions with bay controllers or protection relays so the system can disconnect or reconfigure affected sections safely.

Can one insulation monitor safely supervise several branches in a DC microgrid or BESS installation?

A single IMD can be switched between segments, but every extra branch increases scanning time and makes fault location harder. In high-power or safety-critical DC microgrids you normally give key feeders or battery blocks their own monitors. If you must share one device, define maximum scan intervals and clear rules for unresolved faults.

How do you verify on site that your insulation monitor’s range and thresholds are configured correctly?

During commissioning you validate the IMD by inserting known resistances to earth and checking that readings, warnings and trips match expectations. Step through values around the thresholds and record response times and hysteresis. Repeat after wiring changes or firmware updates so you keep evidence that the device and the configuration still behave as specified.

Which datasheet parameters matter most when you select AFEs, isolation devices and ADCs for insulation monitoring?

When you review candidate devices, look first at isolation ratings, working voltage, surge withstand and CMTI because these define survivability. Then check analogue noise, offset, bandwidth and protection features, plus ADC effective resolution and reference stability. Only after those basics look solid do features like communications, packaging and extras become meaningful differentiators.

During design reviews, what are the biggest red flags that your insulation monitoring concept is not ready?

Red flags include thresholds that nobody can trace back to standards or calculations, no clear plan for surge and dv/dt testing, and outputs that are not wired in a fail-safe way. Missing self-test concepts, no end-of-line checks and no logging of IMD faults are further signs that the design is not ready for service.