TL;DR #
A three-stage cascaded passive protection circuit successfully clamped 1.2/50 μs impulse overvoltages ranging from 500 V to 4 kV down to a final port voltage of 15.2–26.4 V at the battery module terminals, with a self-breaking DC overvoltage threshold set at 26.5 V. For buyers specifying battery modules destined for grid-tied or hybrid renewable energy storage applications, the absence of multi-level transient overvoltage protection in a module’s design is a genuine fire risk — not a theoretical one. Before issuing any RFQ for BESS modules, require suppliers to demonstrate both impulse test results (1.2/50 μs waveform, ≥2 kV peak) and steady-state DC overvoltage cutoff behavior in their protection circuit documentation.
Overview #
If you are evaluating lithium-ion battery modules for grid-connected energy storage or hybrid renewable systems, overvoltage protection is where most procurement teams have a blind spot — and it is an expensive one. Fire incident investigation reports on electrochemical energy storage stations have explicitly identified inadequate transient overvoltage protection for battery modules as a contributing factor, and yet most module-level BMS designs still rely on a single-stage surge protective device (SPD) that cannot do the job alone.
The research underpinning this article was conducted at a key laboratory within a major Chinese engineering university’s electrical engineering department, targeting the specific protection requirements of 12 V lithium-ion battery modules operating within electrochemical energy storage systems. The experimental program used a calibrated impulse voltage generator to apply 1.2/50 μs waveform transients at four distinct peak levels — 500 V, 1 kV, 2 kV, and 4 kV — under both common-mode and differential-mode injection, while a controlled DC source tested steady-state overvoltage suppression across a 12.5–30 V input range. Waveform capture used a Tektronix oscilloscope with a 1:100 high-voltage probe to record per-stage voltage peaks.
The finding matters for procurement because transient overvoltages in real installations reach hundreds to thousands of volts — generated by lightning strikes (direct and inductive), system resonance, or circuit breaker switching. A module that passes standard cell-level safety tests but lacks adequate board-level surge protection is still a fire risk once it is installed in a distribution-connected system.
For context on how cell format and module topology choices affect downstream protection design requirements, see Cell Formats & Form Factors and Protection Circuit Design.
Three-Stage Transient Overvoltage Protection: How the Circuit Architecture Works #
The fundamental problem with single-stage surge protection in battery module applications is a mismatch between what the SPD can absorb and what it needs to clamp. A gas discharge tube (GDT) has excellent surge current capacity — up to 100 kA at 8/20 μs — but its response speed is slow (approximately 100 ns) and its sustain voltage is well below its breakdown voltage, which causes arc continuation and a continuous follow current. A metal oxide varistor (MOV) responds faster (approximately 25 ns) and eliminates the arc problem, but alone it cannot clamp high-energy impulses to the low voltage levels a battery module actually tolerates. A transient voltage suppression (TVS) diode responds at the picosecond level but handles comparatively modest surge currents (0.1–15 kA at 8/20 μs).
The three-stage cascaded architecture developed in this work resolves those individual limitations through sequential coordination:
Stage 1 (Primary Step-Down): GDT as the first-level device, selected for its high surge current capacity. A decoupling element — a common-mode choke — is placed between stages to slow the voltage rise rate and ensure correct timing sequencing between Stage 1 and Stage 2. DC working voltage for Stage 1 is set at approximately 1.8 × UN (where UN is the module nominal voltage).
Stage 2 (Secondary Step-Down): An MOV in series with the GDT provides backup step-down and resolves the arc/follow-current problem. A low-residual-voltage MOV is added to improve clamping. DC working voltage is set at ≥1.25 × 1.15 × UN. A second common-mode choke decouples Stage 2 from Stage 3.
Stage 3 (Final Voltage Clamp): A TVS diode — clamp voltage set at >15% above the battery module’s operating voltage — provides the final fast-response clamping. Response speed at ps level eliminates any timing lag at this stage.
Passive Device Comparison #
| Parameter | Gas Discharge Tube (GDT) | Metal Oxide Varistor (MOV) | TVS Diode |
|---|---|---|---|
| Surge current capacity (8/20 μs) | High (1–100 kA) | High (0.1–70 kA) | Medium (0.1–15 kA) |
| Response speed | ~100 ns | ~25 ns | ~1 ps |
| Parasitic capacitance | ~1 pF | ~1 nF | ~0.1 nF |
| Follow current / arc risk | Yes | No | No |
| Leakage current | None (below breakdown) | Present | Present |
| Self-recovery | Yes | Yes | Yes |
The combination logic is deliberate: each stage acts as the other’s backup. In the rare event that a Stage 1 GDT fails to actuate — which the tests actually revealed under low-amplitude differential-mode input — Stage 2 still provides the primary step-down function. That failure mode matters. More on that below.
Impulse Test Results and Steady-State DC Overvoltage Cutoff Performance #
This is where the numbers tell the story.
Under 1.2/50 μs impulse testing — the standard lightning impulse waveform referenced in electromagnetic compatibility and surge test guidelines — the protection circuit was subjected to four peak voltage levels: 512 V, 1.02 kV, 2.08 kV, and 4.12 kV, applied in both differential-mode (between positive and negative bus) and common-mode (between each bus and ground).
Differential-mode results:
- At 512 V input: Stage 1 output → 416 V; Stage 2 output → 204 V; Stage 3 output → 23.2 V
- At 1.02 kV input: Stage 1 output → 592 V; Stage 2 output → 154 V; Stage 3 output → 26.4 V
- At 2.08 kV input: Stage 1 output → 432 V; Stage 2 output → 152 V; Stage 3 output → 15.2 V
- At 4.12 kV input: Stage 1 output → 280 V; Stage 2 output → 212 V; Stage 3 output → 20.8 V
Common-mode results:
- At 512 V input: Stage 1 → 284 V; Stage 2 → 228 V; Stage 3 → 22.4 V
- At 1.02 kV input: Stage 1 → 392 V; Stage 2 → 344 V; Stage 3 → 20.8 V
- At 2.08 kV input: Stage 1 → 400 V; Stage 2 → 252 V; Stage 3 → 21.6 V
- At 4.12 kV input: Stage 1 → 296 V; Stage 2 → 256 V; Stage 3 → 20.6 V
The final clamped voltage across all test conditions landed in the range of 15.2–26.4 V at the battery module port. For a 12 V module, that is a meaningful result. Post-Stage 3, the sharp impulse peaks are effectively flattened, and the differential-mode oscillation response time was 10 μs versus 5 μs for common mode — both decaying to below 5% of peak amplitude after that window.
For reference: a single-stage SPD in a comparable configuration can only limit a similar amplitude impulse to the hundreds-of-volts range. The multi-stage architecture reduces that residual by an order of magnitude at the final output.
The Failure Mode You Should Know About #
In supplier qualification testing, a specific failure mode surfaced in the Stage 1 circuit under low-amplitude differential-mode input: the grounding GDT failed to actuate. At 512 V and 1.02 kV differential-mode input, the common-mode input achieved lower Stage 2 peak voltages than differential-mode — the opposite of what you might expect. The reason is that common-mode input provided the GDT with a discharge path via the ground reference, triggering it reliably, while differential-mode input at lower amplitudes did not reach the GDT’s breakdown threshold. Stage 2 absorbed the backup function in those cases, which confirms the cascaded backup design logic — but it also tells you that a supplier who has tested only one input mode has not fully qualified the circuit.
Steady-State DC Overvoltage Suppression #
The steady-state circuit is a separate but cascaded function — designed for the scenario where the PCS or BMS fails and allows DC voltage to drift above the module’s rated operating range. It uses a Buck converter topology with voltage-negative feedback, controlled by an LM2596 switching regulator (150 kHz switching frequency), a Zener diode, NPN transistor, and a MOSFET as the main switching element.
The operating logic:
- Input voltage within normal range: Zener diode inactive, transistor off, MOSFET on, module charges/discharges normally
- Input voltage exceeds threshold: Zener conducts, transistor turns on, MOSFET gate voltage drops to zero, module is isolated from the main circuit
- Input voltage recovers: Zener self-resets, circuit auto-recovers
Test results across input voltages from 12.5 V to 30 V:
- At 12.5–26 V input: output voltage stabilized to (12 ± 0.03) V
- At 26.5 V and above: the self-breaking function actuates and isolates the battery module
- Below the 1.3 V logic threshold inside the LM2596 feedback path: regulator shuts down, supply current drops to approximately 80 μA
The Vsat (saturation voltage) of the LM2596 output pin is 1.5 V. PCB layout was designed to minimize parasitic coupling at that pin by reducing pad connection area.
Most procurement teams do not realize that IEC 62619:2022 Safety requirements for secondary lithium cells and batteries — the primary grid storage safety standard — covers cell and module-level abuse tolerance but does not prescribe specific multi-stage surge protection architecture at the module port. That gap is exactly where the fire incidents have occurred. The absence of a requirement in the standard does not mean the protection is optional.
Practical Guidance for Buyers #
If you are sourcing battery modules for stationary energy storage — grid-tied systems, solar-plus-storage, or industrial UPS applications — demand evidence of overvoltage protection architecture, not just a datasheet checkbox.
Honestly, most buyers over-specify cell chemistry and capacity while completely ignoring the protection circuit design. A module with pristine NMC cells and a BMS that gets destroyed by the first nearby lightning strike is not a quality product.
Ask your suppliers to show you the actual impulse test waveforms at multiple peak levels, not just a compliance certificate. Specifically, look for: (1) test results at the 1.2/50 μs waveform — this is the standard lightning impulse; (2) results under both differential-mode and common-mode injection; (3) per-stage voltage peak data that shows progressive attenuation, not just a final output number; and (4) the DC overvoltage cutoff threshold and evidence of self-recovery behavior.
A module targeted for use in hybrid wind-solar storage systems, or any installation with a PCS, should have both a transient multi-stage suppression circuit and a steady-state DC overvoltage isolation function. These are separate design requirements. Confirm that the supplier’s protection design covers both.
The IEC 61960-3 Secondary lithium cells and batteries for portable applications standard provides context on cell-level electrical performance requirements, while UL 9540A Test Method for Evaluating Thermal Runaway Fire Propagation in Battery Energy Storage Systems remains the most relevant fire propagation reference if your application is a large stationary BESS.
At compactbess.com, our sourcing team works with verified Chinese manufacturers of battery modules and protection circuits specifically for global OEM buyers and energy storage integrators — if you need to identify suppliers who can document multi-stage overvoltage protection test data, we can match you with qualified sources quickly. Need help identifying qualified suppliers for BESS modules with documented surge protection? Talk to our sourcing team →
Supplier Qualification Questions #
- Can you provide oscilloscope waveform data showing per-stage voltage peaks for 1.2/50 μs impulse overvoltage tests at all four levels — 500 V, 1 kV, 2 kV, and 4 kV — for your battery module protection circuit?
- What is the final clamped voltage at the battery module port after all protection stages under 4 kV peak differential-mode impulse input, and can you confirm it falls within the 15.2–26.4 V range demonstrated in your qualification data?
- Have you tested the protection circuit under both common-mode and differential-mode overvoltage injection independently? Can you show that the Stage 1 gas discharge tube actuates correctly under low-amplitude (<1 kV) differential-mode input, or have you confirmed Stage 2 MOV backup function for cases where Stage 1 does not trigger?
- What is the DC overvoltage self-breaking threshold of your steady-state suppression circuit, and can you confirm that the module remains isolated above 26.5 V while auto-recovering below that threshold with output regulated to within ±0.03 V of the rated voltage?
- What switching frequency does your Buck converter regulator operate at, and what is the standby supply current when the self-breaking function is active — specifically, can you confirm the quiescent current drops to approximately 80 μA or below during cut-off state?
Sourcing Checklist #
- ☐ Impulse test data provided for 1.2/50 μs waveform at minimum peak levels of 500 V, 1 kV, 2 kV, and 4 kV
- ☐ Test results cover both differential-mode and common-mode overvoltage injection independently
- ☐ Final clamped port voltage confirmed within 15.2–26.4 V range across all tested impulse levels for a 12 V nominal module
- ☐ Steady-state DC overvoltage self-breaking threshold documented at ≤26.5 V with auto-recovery confirmed upon voltage reduction
- ☐ Output voltage regulation accuracy verified at (12 ± 0.03) V across 12.5–26 V DC input range
- ☐ Protection circuit uses a minimum of three cascaded stages with both GDT and MOV in primary/secondary stages and TVS diode as final clamp
- ☐ PCB design documentation includes common-mode choke decoupling between all protection stages to ensure correct timing sequencing
- ☐ Module-level protection circuit compliance aligned with IEC 62619:2022 module safety requirements, with additional surge test protocol referenced
Key Specifications Table #
| Parameter | Recommended Value | Verification Method |
|---|---|---|
| Final clamped port voltage (1.2/50 μs, up to 4 kV) | 15.2–26.4 V | Oscilloscope with 1:100 high-voltage probe; per-stage waveform capture |
| Steady-state DC overvoltage self-breaking threshold | 26.5 V | DC power supply sweep from 12.5 V to 30 V in 0.5 V steps; confirm isolation actuation |
| Output voltage regulation accuracy (12 V module) | 12 ± 0.03 V | DC input sweep 12.5–26 V; multimeter at module output terminals |
| Stage 1 DC working voltage (primary step-down) | ≈1.8 × UN | Design calculation per formula; confirm against device datasheet rated voltage |
| Stage 2 DC working voltage (secondary step-down) | ≥1.44 × UN (1.25 × 1.15 × UN) | Design calculation; MOV datasheet DC working voltage specification |
| TVS diode clamping threshold (Stage 3) | >15% above module operating voltage | Curve tracer or impulse test; verify clamping voltage vs. rated module voltage |
| Transient response oscillation decay | <5% of peak amplitude after 10 μs (differential), 5 μs (common) | Oscilloscope waveform; measure amplitude at 10 μs post-impulse |
| Regulator standby current (cut-off state) | ≈80 μA | DC supply with current measurement; input >26.5 V, measure supply current |
Can’t find a supplier meeting these specs? Submit your requirements and we’ll match you within 48 hours.
References #
Data source: Multi-Stage Transient Overvoltage Protection Circuit Design and Experimental Validation for Lithium-Ion Battery Modules in Grid-Connected Energy Storage Systems, A.-L. Guo et al., Journal of the Electrochemical Society, 2025
Frequently Asked Questions #
Why does a lithium-ion battery module need multi-stage overvoltage protection if it already has a BMS?
The BMS typically embeds only a single-stage SPD for its own self-protection — it is not designed to suppress transient overvoltages at the battery module terminals. A BMS protects against overcharge, overdischarge, overcurrent, and short circuit under normal operating conditions, but a 1.2/50 μs lightning impulse at 1–4 kV will pass through or destroy a single SPD before the BMS can react. The multi-stage circuit addresses the gap between what the BMS covers and what the grid-connected environment actually generates.
What is the 1.2/50 μs waveform and why does it matter?
It is the standard lightning impulse waveshape used in electromagnetic compatibility and surge test protocols — 1.2 μs rise time to peak, 50 μs decay to half-peak. It represents the characteristic shape of a lightning-induced overvoltage transient propagating through electrical infrastructure. Testing at this waveform is the most relevant benchmark for any battery module installed in an outdoor or grid-connected system exposed to lightning events.
Can the protection circuit handle both positive and negative overvoltage surges?
Under differential-mode injection, the tested circuit showed positive-negative alternating oscillation at each stage, which is consistent with the bidirectional nature of differential-mode transients. The TVS diode in Stage 3 is a bidirectional device, and the design accounts for both polarities. Common-mode transients in the tests remained predominantly positive-polarity at each stage.
What happens if the self-breaking function actuates — does the module need manual reset?
No. The self-breaking circuit uses a Zener diode and MOSFET topology with inherent self-recovery: when the input voltage drops back below the 26.5 V threshold, the Zener diode deactivates, the transistor turns off, and the MOSFET re-enables automatically. The module reconnects without requiring manual intervention. This is one of the critical differences between this design and conventional protection cutoffs that require power cycling to reset.
Does the protection circuit affect the module’s normal charging and discharging performance?
Under normal operating conditions — input voltage within the 12.5–26 V range — the Buck regulator maintains output at (12 ± 0.03) V with the MOSFET fully conducting. The quiescent current consumption during normal operation is negligible for a module-scale application. The common-mode choke decoupling elements between stages introduce minor impedance but are sized to avoid affecting power throughput at rated operating frequency.
Published by compactbess.com Technical Team | Request a sourcing quote
