Inrush current is the surge a product draws in the first milliseconds after power-on. It blows fuses, trips breakers, welds relay contacts and stresses bridge rectifiers, yet it is often measured wrongly: with the bandwidth limit on, in Auto trigger mode, at a random switch-on angle, or with no clear idea of which part of the waveform the datasheet limit refers to. This guide shows engineers and technicians how to measure inrush current correctly, how to read the inrush specifications in datasheets, and how to use your measurement to decide pass or fail.
The illustrations in this guide are live: click the buttons under each one to change the setup and watch the result change.
- What Is Inrush Current?
- How to Read Inrush Specifications in Datasheets
- Equipment and Safety
- Scope and Probe Settings
- Capturing the Inrush Event
- What to Measure: Level and Time Frame
- Worst-Case Conditions
- Pass or Fail: Checking Against the Limits
- Step-by-Step Procedure and Report
- Common Mistakes and Fixes
- Inrush Calculator: Estimate and Pass/Fail
- FAQ
1. What Is Inrush Current?
When an AC-DC power supply is switched on, its bulk capacitor is empty. For a moment, the only things limiting the current are the line impedance, the bridge rectifier, the capacitor ESR and any inrush limiter such as an NTC thermistor:
Ipk ≈ Vpk / (RNTC + Rline + Rbridge + ESR)
A typical inrush event has up to three parts:
- EMI filter spike: the X-capacitors charge in a few microseconds. Very high peak, almost no energy.
- Bulk capacitor charging pulse: the main inrush, lasting from a fraction of a millisecond to a few milliseconds, followed by smaller pulses on the next half-cycles.
- Second events: an NTC bypass relay closing, a PFC stage starting, or a DC-DC converter charging its output capacitors, often tens to hundreds of milliseconds later.
Transformers (magnetizing inrush), motors, lamps and hot-plugged DC boards have their own inrush, but the measurement method is the same.
2. How to Read Inrush Specifications in Datasheets
An inrush number is meaningless without its conditions. Read the whole line, the footnotes and the test-condition section.
Power supply or module datasheet
| What the datasheet says | What to check |
|---|---|
| “Inrush current: 40 A max” | Is it max or typical? A typical value is not a limit. |
| “at 230 V AC” / “at 115 V AC” | Inrush scales with the input peak voltage. 264 V AC gives about 15 % more than 230 V AC. |
| “cold start, 25 °C” | A hot restart gives a higher peak with an NTC limiter. Ask the maker if hot restart is not specified. |
| “first half-cycle” / “excluding EMI filter charging current (under 0.2 ms)” | Tells you which part of the waveform the limit applies to and which spikes are excluded. |
| Measurement source impedance | Some specs assume a defined line impedance. A stiff lab source gives higher inrush than a weak one. |
| Duration or I²t | Some datasheets also give a duration (for example “under 10 ms”) or I²t. Use them for fuse and breaker selection. |
Fuse datasheet
- Melting I²t (A²s): the energy that opens the fuse. Compare it with your measured I²t, not the peak current.
- Pulse derating: repeated inrush pulses age the element. A widely used fuse-maker guideline allows about 38 % of melting I²t for 1,000 pulses, 29 % for 10,000 and 22 % for 100,000. Use your fuse maker’s own figures where available.
- Time-current curve: fast-acting (F) fuses tolerate much less inrush than time-lag (T) fuses.
Bridge rectifier datasheet
- IFSM (peak forward surge current) is specified for a single 8.3 ms (60 Hz) or 10 ms (50 Hz) half-sine pulse. Your inrush pulse is usually much shorter, so comparing your peak with IFSM is conservative.
- I²t rating (for t under 8.3 ms) is the better comparison for short pulses.
- Check whether the rating is at TJ = 25 °C or at operating temperature; a hot restart happens with a warm bridge.
NTC thermistor datasheet
- R25 sets the cold inrush. Resistance at operating current (often well under 1 Ω) sets the hot-restart inrush.
- Maximum capacitance at a given voltage (or maximum energy ½CV²): your bulk capacitor must be within it.
- Cool-down time: an NTC can take a minute or more to recover. Repeated on/off cycles give high inrush.
Relays, switches and breakers
- Relay and switch datasheets give an inrush or make current rating (for example UL “TV” ratings for lamp loads). Compare with your measured peak.
- Miniature circuit breakers trip instantaneously at 3 to 5 × In (type B), 5 to 10 × In (type C) or 10 to 20 × In (type D). Several units switched on together add their inrush.
- Some standards and customers define an inrush time-current mask (for example telecom -48 V DC equipment). The measured current must stay under the mask at every point in time.
3. Equipment and Safety
| Item | Notes |
|---|---|
| AC/DC current probe (Hall-effect clamp) | Peak current rating above your worst case (including hot restart), bandwidth of a few MHz or more. Degauss and zero before every capture. |
| Current shunt + differential probe | Alternative to a clamp. Low-inductance shunt rated for the pulse energy. Never use a grounded passive probe on mains. |
| High-voltage differential probe | Line voltage on a second channel, to confirm the switch-on angle. |
| AC source with phase-angle control | Switches on at exactly 90° (line peak). Otherwise use a zero-cross/peak-switching relay, or repeat many random switch-ons. |
| Oscilloscope with deep memory | Long single-shot captures (up to 1 s) at a high sample rate. |
4. Scope and Probe Settings
| Setting | Use | Why |
|---|---|---|
| Bandwidth limit | OFF | Narrow pulses and the EMI-filter spike lose amplitude with a filter. |
| Coupling | DC | AC coupling distorts a single pulse. |
| Probe ratio and units | Amps, ratio matching the probe range | For example 100 mV/A on the 30 A range. |
| Vertical scale | Expected peak in about 5 div | Clipping hides the true peak. Estimate it first with the calculator. |
| Sample rate | ≥ 1 MS/s for the main pulse; ≥ 50 MS/s to see the X-cap spike | Check the rate the scope actually uses at your timebase and record length. |
| Acquisition | Sample or peak detect | Never average a single event. |
5. Capturing the Inrush Event
- Single-shot mode. Inrush happens once; Auto mode re-arms and loses it.
- Trigger: edge on the current channel, either slope (the first pulse can be negative), at about 10 % of the expected peak, above the noise.
- Pre-trigger 10 to 20 %, so the moment of switch-on is visible.
- Switch on at the line peak (90°). At 0° the capacitor charges gently and the inrush is much lower.
- Capture several windows (or one long, deep-memory capture and zoom): see the next section.
6. What to Measure: Level and Time Frame
“The inrush” is not one number. Measure and report each of these, and state which one the datasheet limit refers to:
| Quantity | Definition | Timebase |
|---|---|---|
| X-cap spike peak | Highest current in the first microseconds (EMI filter charging). Report separately; many specs exclude it. | 10 to 50 µs/div |
| Inrush peak Ipk | Highest instantaneous |i| of the bulk-capacitor charging pulse. This is the value most specs mean. | 0.5 to 2 ms/div |
| Time to peak, pulse width | From switch-on to peak; width at 50 % of the peak. | 0.5 to 2 ms/div |
| Event duration | From switch-on until the line-current peaks settle within about 110 % of their steady-state value (or the % your spec defines). | 10 to 20 ms/div |
| I²t | ∫ i² dt over the event (scope math: square, then integrate). Use the first half-cycle for bridge I²t, the whole event for fuses. | 10 to 20 ms/div |
| Second events | Relay bypass, PFC start, DC-DC start-up, downstream loads. | 50 to 100 ms/div |
7. Worst-Case Conditions
- Maximum input voltage (for example 264 V AC): peak current scales with Vpk.
- Switch-on at 90° every time, or at least 20 to 30 random switch-ons and take the maximum.
- Hot restart: run at full load until warm, switch off for 1 to 2 s (bulk cap discharged by the load, NTC still hot), switch on.
- Cold start at minimum ambient if the bridge or capacitors change much with temperature.
- Bulk capacitor fully discharged before each cold test.
- Output load as in the application; some converters start into their full load.
8. Pass or Fail: Checking Against the Limits
Use the worst-case measurement (max Vin, 90°, hot restart) and compare each quantity with the limit that matches it:
| Limit | Compare with | Pass if |
|---|---|---|
| Product spec “max inrush” | Measured peak, in the spec’s own time frame and exclusions | Ipk ≤ limit, with margin (often 10 to 20 %) |
| Fuse | Measured I²t of the whole event | I²t ≤ melting I²t × pulse factor (for example 22 % for 100,000 cycles) |
| Bridge rectifier | First-pulse peak and I²t | Ipk ≤ IFSM (conservative) and I²t ≤ bridge I²t |
| NTC | Bulk capacitance and voltage | Within the max capacitance / energy rating |
| Relay / switch | Measured peak | ≤ inrush (make) rating |
| Breaker | Total inrush of all units on the circuit | Below the instantaneous trip level for the inrush duration (check the trip curve) |
| Time-current mask | Whole measured waveform | Every point under the mask |
9. Step-by-Step Procedure and Report
- Estimate the worst-case peak with the calculator and choose the probe range and scale.
- Degauss and zero the current probe; clamp the line conductor. Line voltage on CH2 via a differential probe.
- Bandwidth limit OFF, DC coupling, sample mode, scale for about 5 div at the expected peak.
- Single shot, edge trigger (either slope) at ~10 % of the peak, 10 to 20 % pre-trigger, long record (≥ 0.5 s) at ≥ 1 MS/s.
- Cold start: bulk capacitor discharged, NTC at ambient. Switch on at 90° at max input voltage.
- Zoom and measure: X-cap spike, Ipk, time to peak, width, event duration, I²t, second events.
- Repeat at least 5 to 10 times; keep the maximum.
- Hot restart: run at full load until warm, off 1 to 2 s, on at 90°. Repeat and keep the maximum.
- Compare with every limit (spec, fuse, bridge, NTC, relay, breaker, mask) and record the margin.
Report template: input voltage and frequency, source type, switch-on angle, cold/hot, ambient, probe and range, Ipk (with exclusions), time to peak, width, duration, I²t, second-event peak, each limit and its margin, PASS/FAIL.
10. Common Mistakes and Fixes
| Mistake | Result | Fix |
|---|---|---|
| 20 MHz limit or averaging on | Peak reads low | Bandwidth limit OFF, sample mode |
| Auto trigger | Event missed or overwritten | Single shot |
| Random switch-on angle, one try | Worst case missed | Switch at 90° or repeat many times |
| Only cold start | Hot-restart peak missed | Test a hot restart |
| Short capture window | Second inrush (relay, PFC) missed | Capture 0.5 to 1 s |
| Comparing peak with fuse rating | Wrong conclusion | Compare I²t with derated melting I²t |
| Including/excluding the X-cap spike inconsistently | Disagreement with datasheet | Follow the spec’s definition, report the spike separately |
| Probe range saturated | Flat-topped, low peak | Higher current range |
| Weak source or isolation transformer | Inrush reads low | Stiff source, or state the source impedance |
11. Inrush Calculator: Estimate and Pass/Fail
Tab 1 estimates the inrush and suggests scope settings. Tab 2 checks your measured peak and I²t against the product spec, fuse, bridge rectifier and breaker.
