Protection Coordination: TVS, eFuse and PPTC

Protection Coordination: TVS, eFuse and PPTC

Quick answer

TVS, eFuse and PPTC are not alternatives; they answer different faults on different timescales. A TVS clamps voltage in under a nanosecond, an eFuse limits current in microseconds, and a PPTC limits sustained current in seconds. Coordinate three things: the TVS clamp must stay below every downstream absolute maximum, the eFuse current limit must sit above the real load with margin, and a PPTC — if used at all — must hold more current at the hottest ambient than the eFuse will ever pass. Get those three numbers in order and the chain works; get them out of order and the parts fight each other.

Executive takeaways

  • Speed is not interchangeable. A TVS cannot limit current; an eFuse cannot clamp a nanosecond ESD edge; a PPTC cannot protect a sensitive IC from a fast short.
  • Set the clamp first. The TVS clamp at the real surge current must be below the eFuse's absolute maximum (28 V on a TPS25947) and the load's absolute maximum.
  • Do not stack redundant current limiters. An eFuse plus a PPTC in series usually means one of them trips first and the other never acts; choose one and coordinate the rest around it.
  • Use the timescale to assign roles. Nanoseconds: TVS. Microseconds: eFuse. Seconds: PPTC or fuse. Anything longer is a system-level decision.

1. Three faults, three devices

Protection failures are usually described as “a surge” or “a short”, but the physics differ. An ESD strike is a nanosecond, high-voltage event with modest energy. A hot-plug or load short is a microsecond-to-millisecond current event at the rail voltage. A stuck load or a failed converter is a sustained overcurrent that heats wiring and connectors for seconds or minutes.

Each device is optimised for one of those. The TVS is a voltage clamp with a fast junction; the eFuse is a controlled FET with current sensing and a timer; the PPTC is a thermal device whose resistance rises when it heats. Using one device for all three faults is how boards end up with a protected-looking schematic that fails in the field.

2. The timescale table

Device What it controls Typical response Energy handled
TVS diode Voltage (clamp) < 1 ns to clamp Low to medium (surge, ESD)
eFuse Current (limit or breaker) µs (current limit), 1–5 µs OVLO/OVC Medium (until it opens)
PPTC Sustained current (thermal) Seconds Low; it limits, not interrupts
Fuse Current (one-shot) ms to s depending on overload High (it melts)

The table also shows why “add a TVS” does not fix an overcurrent fault and why “add an eFuse” does not fix an ESD fault. The devices overlap at the edges, not in the middle.

Fault Primary device Backup Key check
ESD TVS eFuse OVLO Clamp below abs max
Surge TVS eFuse OVLO TVS energy if eFuse stays closed
Hot-plug inrush eFuse timer — Peak below 2 × ILIM
Load short eFuse PPTC or fuse ILIM above load, below backup trip
Sustained overload eFuse or PPTC Fuse Derating at maximum ambient
Reverse polarity Ideal diode / FET TVS FET rating above clamp

3. The three numbers to coordinate

1. TVS clamp at the real surge current. The datasheet clamp is quoted at the rated pulse current; at a lower current the clamp is lower, at a higher one it is higher. Use the actual surge current and keep the result below the eFuse's absolute maximum and the load's absolute maximum. On a 12 V rail a 15 V standoff part is typical; a 600 W SMBJ15A clamps around 24 V at its rated current, which fits under a 28 V eFuse limit.

2. eFuse current limit. Set it above the maximum load with margin, and use the device's blanking timer to allow inrush. The TPS25947 has an adjustable limit from 0.5 A to 6 A with ±10% accuracy above 1 A, and a timer that allows peaks up to twice the limit. A 2 A load therefore uses a 2.5 A limit and a 5 A inrush window, not a 2 A limit that trips at startup.

3. PPTC hold current at the hottest ambient. If a PPTC is in series, its hold current at the maximum ambient must exceed the eFuse's limit, otherwise the PPTC trips first and the eFuse never acts. Because PPTC hold current derates with temperature — a 0.50 A part holds 0.36 A at 50 °C — this rule eliminates most “eFuse plus PPTC” designs.

4. Worked example: 12 V, 2 A input

A 12 V input feeds a 2 A load through a connector. The load's absolute maximum is 30 V and the chosen eFuse is a TPS25947 (2.7–23 V, 28 V absolute maximum).

  • TVS: a 15 V standoff, 600 W part clamps around 24 V at its rated current — below 28 V and below 30 V, so both the eFuse and the load survive.
  • eFuse: current limit 2.5 A, inrush window 5 A for a few milliseconds, and overvoltage lockout set to the highest available option (13.8 V) so a sustained overvoltage disconnects the load in about 5 µs.
  • PPTC: none. A PPTC that holds 2.5 A at 50 °C would need a 3.75 A-class part, which is larger and slower than the eFuse and adds series resistance for no benefit.

The three thresholds are now ordered: 15 V standoff below the 24 V clamp, the clamp below the 28 V eFuse absolute maximum, and the 2.5 A limit above the 2 A load but below any branch protection.

The TVS still has a job during a surge: it holds the rail below 28 V for the few microseconds the eFuse needs to react. Its energy is small because the eFuse opens fast, which is exactly the coordination the chain is supposed to provide.

5. When a PPTC does make sense

A PPTC earns its place where the fault is sustained, the current is modest and cost matters: USB ports, small adapters, battery packs, sensor branches. In those designs there is often no eFuse, so the PPTC is the only current limiter — and then its derated hold current is the whole design.

A second valid case is a product that must recover without service: a PPTC resets itself, while a fuse or an eFuse with a latch needs a power cycle or a controller action. If the equipment is remote and a power cycle is expensive, that self-reset can be worth the slower response and the series resistance.

Where both are present, the usual reason is a branch architecture: an eFuse protects the main input, and a PPTC protects one small sub-branch that can be isolated without taking the whole board down. In that case the PPTC's hold current must still exceed that branch's normal current at temperature, but it does not need to coordinate with the main eFuse limit.

6. Hot-plug: who handles the inrush?

A board plugged into a live rail charges its input capacitance through the connector and cable. A 100 µF input at 12 V holds 7.2 mJ; if it charges in 100 µs the average current is 7.2 A, well above any steady-state limit. The eFuse must be configured to ride through that pulse — usually with a blanking timer and a current limit set for the peak, not the average.

The TVS does not help with inrush; it clamps voltage, and the inrush voltage is the rail. The PPTC, if present, is the wrong device for inrush because its thermal trip is slow and its repeated trips degrade it. If hot-plug is a routine event, use an eFuse with a timer, or add a controlled hot-swap controller, rather than relying on a PPTC to survive it.

7. Energy: what the TVS must absorb before the eFuse opens

During a surge, the TVS clamps and the eFuse needs a few microseconds to disconnect. The energy the TVS absorbs in that window is small: E ≈ Vclamp × I × t. A 24 V clamp at 5 A for 5 µs is 0.6 mJ — far below the joule rating of a 600 W part. The TVS energy problem appears when the eFuse does not open (for example, a surge below the OVLO threshold), and the TVS must absorb the full event instead.

That is the coordination check most often missed: the TVS rating must cover the case where the eFuse stays closed. If the surge current is below the eFuse limit and the voltage is below OVLO, the eFuse does nothing and the TVS does all the work.

8. Reverse polarity and the ideal diode

Reverse polarity is a different fault again. A TVS may conduct in reverse and protect the downstream circuit, but it will dissipate until the source is removed or a fuse opens. An ideal-diode controller or a back-to-back FET (as in the TPS25947) blocks reverse current actively and is the better answer for connectors that can be miswired. Coordinate it with the TVS: the TVS clamp must stay below the reverse-blocking FET's rating.

9. Ordering the chain on the board

Protection chain from connector through TVS, eFuse and PPTC to the load
Placement follows the fault: the TVS at the connector, the eFuse before the bulk capacitance, the PPTC only on branches that need it.

Put the TVS at the connector, before any series element, so surge current is diverted before it enters the board. Put the eFuse after the TVS but before the bulk input capacitance if the eFuse is meant to protect against a failed capacitor; if it is only protecting the load, it can sit after the bulk. The PPTC, when used, goes at the branch it protects.

The ground return matters as much as the order. The TVS current must return to the connector ground, not through the signal ground; the eFuse's sense resistor and the PPTC's thermal path should not share copper with the surge path.

9. The low-cost chain: TVS plus fuse

Where an eFuse is too expensive, a TVS plus a fuse is a legitimate chain if the fuse is rated so the TVS survives until the fuse opens. The TVS clamps the surge, the fuse clears the sustained fault, and the coordination number is the fuse's melting time versus the TVS's pulse rating. A slow-blow fuse may take hundreds of milliseconds at 2× its rating, during which the TVS carries the current.

The chain is weaker than an eFuse for two reasons: the fuse is one-shot, and the TVS must be sized for the fuse's clearing time, not for a few microseconds. It is still common in low-cost adapters and small boards, and it works when the numbers are checked.

10. Five coordination mistakes

1. TVS clamp above the eFuse absolute maximum. The TVS “protects” the connector while the eFuse dies behind it.

2. eFuse overvoltage lockout below the TVS clamp. Every surge becomes a disconnect; if that is not intended, the load drops for no reason.

3. PPTC hold current below the eFuse limit. The PPTC trips first, the eFuse never acts, and the PPTC's slow reset becomes the system's recovery time.

4. Stacking three current limiters. Each adds resistance and a failure mode; two are already hard to coordinate.

5. Ignoring the timescale. A TVS cannot save an IC from a 10 ms short, and a PPTC cannot save it from an ESD edge.

11. Checklist

  • List the faults: ESD, surge, hot-plug inrush, load short, sustained overload.
  • For each fault, name the device that acts and the time it takes.
  • TVS standoff above the maximum operating voltage.
  • TVS clamp below the eFuse and load absolute maximums at the real current.
  • eFuse current limit above maximum load, with a blanking timer for inrush.
  • eFuse overvoltage threshold consistent with the TVS clamp.
  • PPTC (if any) hold current above the eFuse limit at the hottest ambient.
  • Placement: TVS at connector, eFuse before the protected capacitance, PPTC on its branch.
  • Ground return for the surge path separate from signal ground.
  • Verification: surge and short tests at the real ambient, with the TVS current and eFuse response measured.

FAQ

How do I set the eFuse current limit for a motor load?

Above the stall-free running current with margin, and high enough with the blanking timer to cover startup. Measure the startup profile; a motor can draw several times its running current for tens of milliseconds.

Is a resettable fuse better than an eFuse for a USB port?

For cost and simplicity, yes, if the fault is sustained overcurrent and the port can tolerate a few seconds of reset. For fast faults, inrush control or accurate limits, an eFuse is better.

Should the TVS be before or after the eFuse?

At the connector, before the eFuse. The surge should be diverted before it reaches the eFuse and the board; the eFuse then handles current faults.

Can the eFuse's overvoltage protection replace the TVS?

No. The eFuse responds in microseconds and is not a nanosecond clamp; ESD and fast surge will pass through before it reacts.

Do I need a TVS if I already have an eFuse?

Yes, for ESD and fast surge. An eFuse responds in microseconds; a TVS clamps in under a nanosecond. They cover different faults.

Can an eFuse replace a fuse?

Often, if the eFuse is rated for the fault current and the application accepts active limiting. For safety-critical isolation, a fuse or breaker may still be required.

Why does my PPTC trip when the eFuse is current-limiting?

Because the PPTC's hold current at the actual ambient is below the eFuse limit. Either raise the PPTC rating or remove it and let the eFuse do the job.

What about a TVS plus a fuse?

A valid low-cost chain: the TVS clamps, the fuse clears a sustained fault, and the fuse must be slow enough not to open on a normal inrush but fast enough to protect the TVS. The fuse must be rated so the TVS survives until the fuse opens.

How do I test coordination?

Apply a surge and a short at the real ambient, and measure the TVS current, the eFuse response and the load voltage. A bench test with a short cable often misses the loop inductance that makes the real event worse.

Glossary

Clamp voltage
The voltage across a TVS at a specified pulse current.
Current limit
The maximum current an eFuse allows before regulating or breaking.
Hold current
The current a PPTC passes without tripping at a given temperature.
OVLO
Overvoltage lockout; an eFuse threshold that disconnects the load above a set voltage.
Blanking timer
A timer that ignores short current peaks so inrush does not trip the limiter.

Sources

Building a protection chain?

Browse our protection and diode lines, or send the rail, load and fault levels for a coordinated recommendation.

Browse In-Stock Components  ·  Submit Your BOM

Check stock by part number

Turn this insight into a sourcing decision.

Need pricing or a drop-in alternative? Submit a BOM / RFQ or talk to our team.