In automotive BOMs, the failure that takes down a design is often not the big MCU or SiC — it's a few-cent TVS or ESD diode that didn't hold. JJW, DOWO and Leiditech can now second-source these parts, but only if you stop thinking "cheaper TVS" and start sourcing by interface and risk. Clamping voltage and junction capacitance — not price — decide whether the swap is safe.

A TVS suppresses transient voltage; an ESD device shunts electrostatic discharge. In plain terms: when the board gets kicked by a voltage spike, these parts take the hit so the chip behind them doesn't. Connector hot-plug, chassis ESD, window-motor start/stop, harness coupling, a spike on the 12 V / 24 V rail — any of these can send a downstream IC into early retirement.
So second-sourcing a TVS/ESD part is not finding "same package, similar voltage, lower price." What actually matters: reverse standoff voltage, breakdown voltage, clamping voltage, peak pulse power, junction capacitance, leakage, and the automotive standards — AEC-Q101, IEC 61000-4-2, ISO 7637-2, ISO 16750-2.
Two parameters trip people up:
A protection part is not "the more aggressive the better." On a power port it must survive the surge; on automotive Ethernet, SerDes or USB it must be light enough not to distort the signal.
The domestic supply base is no longer blank; it segments cleanly by application layer:
| Supplier | Where it fits | Representative capability (per public data) |
|---|---|---|
| JJW | Power-input, motor, body actuators | AEC-Q101 Rev-E auto TVS/ESD; DO-218 for 12 V/24 V, ISO 16750-2 load-dump P5a/5b, 4.6–8.0 kW pulse; >1.0 kW in SMB |
| DOWO | Solution-level: LIN, CAN, T-BOX, lighting, BMS | 48 V LIN protection (SD48C-Q, DWC4840NS-Q); LIN-bus TVS (TPSMBJ15A/24A, SMB/DO-214AA, 600 W, AEC-Q101) |
| Leiditech | Interface protection: CAN, auto-Ethernet, lighting | SMC24 for CAN ESD (24 V bidir, 25 pF, SOT-23, IEC 61000-4-2 L4, ±30 kV, AEC-Q101) |
The signal here: China-based protection is no longer just "a part number exists" — it now arrives organized by LIN / CAN / T-BOX / lighting / BMS use case, which is exactly how a Western buyer needs to source it.
The established Western suppliers give the model to copy: they don't sell "a TVS," they sell protection by interface.
The lesson for the buyer is blunt: don't second-source with the coarse bucket "a domestic TVS for an imported TVS." Source by interface and risk. Power input → power, clamping, load-dump. CAN/LIN → capacitance, EMC, bus waveform. Automotive Ethernet / SerDes → ultra-low capacitance, insertion loss, signal integrity. Motor / inductive loads → surge, freewheeling, thermal.
The cheaper the part, the less you can decide on price alone — because when it fails, what breaks is the expensive IC behind it and the reliability of the whole unit. A few cents saved that sends one ECU back for repair doesn't even cover the tape on the return shipment.
Engineering must confirm three things on any swap: (1) the test conditions match, (2) the protected target matches, (3) unit-level ESD / surge / EMC has been re-verified. A 24 V ESD part on LIN is not the same part on CAN-FD; a TVS on a 12 V input is not the same on a high-speed Ethernet line.
A practical, layered rollout:
TVS / ESD / protection parts are low-price, high-volume — but not low-barrier. They don't decide how smart a system is; they decide whether it survives ESD, surge, hot-plug, short circuit and inductive kickback. The mature second source isn't "the main chip went domestic" — it's a protection-alternative table sitting next to every interface a surge can reach.
The supply movement behind this piece, as recorded in the data. Figures are point-in-time snapshots carrying the date they were captured — they may have moved since publication.
Part numbers named in the piece. China-based parts are marked — those are the ones a buyer is looking for when qualifying a second source.