A power-inductor datasheet may show Isat beside Irms, Itemp, IDC, or two columns simply labeled “rated current.” Those numbers do not express two versions of one universal current limit. Isat is an inductance-retention criterion under DC bias. Irms or Itemp is a temperature-rise criterion under a stated thermal setup. A converter design normally has to clear both gates, then pass waveform, loss, temperature, and control checks on the intended hardware.
Start with the footnote, not the symbol. A current number changes meaning with the L-drop threshold, ambient, board, signal, temperature, and evidence identity. For context, read the electronic-components overview, datasheet abbreviations guide, and AC-DC power-supply application note. None replaces the exact part’s conditions.
A datasheet ampere value is a measured boundary, not a complete current ceiling
Murata describes two ways of rating power-inductor current: one uses self-temperature rise as the index, and the other uses the rate of inductance change (B5). The first addresses heating. The second addresses loss of inductance and consequences such as increased ripple or unstable converter behavior. Because the observables differ, the larger number does not cancel the smaller one, and the smaller number does not by itself finish the design.
Manufacturers also use notation differently. Isat, Irms, Itemp, IDC, and generic rated-current labels are not cross-vendor definitions. Record the method, threshold, and qualifier before the amperes. If those fields are missing, keep the part unverified.
Read Isat as an L-versus-DC-bias result
Coilcraft Document 361 describes an Isat procedure that first measures a representative sample’s inductance at a specified frequency with no DC current, then increases DC current and repeats the measurement. Isat is the current at which inductance has fallen by a stated percentage from the no-bias value (B2). That definition requires more than a current number: full order code, reference inductance, drop percentage, L-test frequency and AC stimulus, temperature, sweep method, and typical/minimum/guaranteed-limit status all belong with it.
Core construction can produce a pronounced knee or a gradual roll-off. Changing the L-drop criterion moves Isat even if the underlying curve is unchanged. Isat is not a universal damage threshold. The application must decide how much inductance loss it can tolerate before ripple, current slope, operating mode, or control behavior becomes unacceptable.
Reconstruct the thermal fixture behind Irms or Itemp
A thermal current rating is also method-dependent. Coilcraft’s Document 361 uses 25 °C ambient, still air, no heat sinking, stabilization at each DC-current step, and a temperature measurement on top of the part. It says a 40 °C rise is typical for its power-inductor rating method. That is a Coilcraft method description, not a universal threshold for all vendors.
Würth Elektronik ANP096 shows why board trace width and copper thickness, solder pads, natural or forced convection, measurement location, and the stabilization rule can change a reported result (B6). In the product, ambient plus self-heating must remain compatible with the exact part’s maximum allowed temperature. A catalog Irms value cannot guarantee temperature in an arbitrary PCB, enclosure, airflow, or neighborhood of hot components.
Map the actual current waveform before comparing ratings
For the triangular ripple in its buck-converter tutorial, Murata expresses maximum inductor current as Iout + ΔIL/2 and screens Isat separately from the temperature-rise current (B4). Keep that expression inside its steady-state waveform assumptions. Other topologies, discontinuous operation, startup, load steps, current-limit tolerance, sensing delay, and fault response can create a different instantaneous envelope.
Use the worst credible instantaneous peak to assess inductance retention. Use the RMS value of the complete waveform, together with total real loss, for the thermal assessment. Total loss may include hot DCR copper loss, frequency-dependent AC winding loss, and core loss. Switching frequency, ripple amplitude and shape, duty cycle, temperature, and DC bias therefore belong in the record. A DC-only rating test provides a useful baseline but is not a simulation of every switching application.
Why choosing the lower rating is only a screening step
Taking the lower of Isat and Irms can remove an obviously unsuitable candidate. It does not establish remaining inductance, acceptable ripple, efficiency, hot resistance, transient behavior, or final part temperature. Nor can case size, nominal inductance, or one current value establish a substitute.
Qualifiers matter independently. A typical L-versus-current curve is not a guaranteed minimum. A minimum Isat does not guarantee a thermal result, and maximum DCR does not make the temperature-rise current a maximum limit. A comparison that drops these identities is not normalized.
A condition-locked example: two ratings on one XAL4020-102 row
Coilcraft’s 2017 Document 469 contains a teaching-table row for XAL4020-102 (B3). The values below are transcribed from that row only. No graph was digitized, and this older application note is not being presented as current purchasing data.
| Field | Object and value in Document 469 | Permitted interpretation |
|---|---|---|
| Full part number | XAL4020-102 | Every number remains attached to this teaching row |
| Nominal inductance | 1.0 µH ±20% | Not transferable to another L value in the same case |
| L measurement | 1 MHz, 0.1 Vrms | Bound to this row |
| DCR | 14.6 mΩ maximum | Maximum and typical are not interchangeable |
| Isat | 8.7 A at 30% typical L drop | A magnetic criterion, not a temperature rating |
| Irms | 9.6 A at 40 °C typical rise | A thermal criterion, not an L-retention guarantee |
| Table test temperature | 25 °C | Different environments require a new assessment |
The narrow lesson is that one component can receive two current numbers because two different endpoints were used. The 8.7 A value cannot stand in for the thermal meaning of 9.6 A, and 9.6 A does not establish compliance with the 30% L-drop criterion. Their ratio is not a rule for another family.
Normalize the method before ranking parts or vendors
The IEC catalog page identifies IEC 62024-2:2024 as the current edition for measuring rated DC limits of inductors used in DC-to-DC converters. Its public scope says the methods address saturation and temperature-rise limitations induced solely by DC (B1). That catalog summary establishes edition and scope; it is not the licensed normative text, and it does not prove that every datasheet follows one identical procedure.
The currently downloadable ANP096 document says Würth’s procedure is based on IEC 62024-2:2020 and describes implementation details such as replacing a thermocouple with an infrared camera. Since the IEC catalog now lists the 2024 edition, preserve the edition and any deviations instead of treating “IEC-based” as a universal part rating. A comparison worksheet should capture L-drop percentage, L stimulus and frequency, DC bias, temperature, allowed rise, ambient, board class and copper, airflow, thermometry, stabilization, DCR temperature, maximum part temperature, and typical/minimum/guaranteed-limit identity. Keep “maximum” only where the source explicitly applies it, such as DCR. Mark missing or unmatched rows “not normalized.”
Use an application-level acceptance workflow
- Lock the full order code, datasheet revision, nominal L and tolerance, DCR, and maximum permitted part or operating temperature.
- Build steady, RMS, peak, startup, load-step, current-limit, and fault envelopes across input, load, component tolerance, and control modes.
- Read the exact part’s L-versus-DC-bias data. Preserve test frequency, AC stimulus, temperature, drop criterion, and typical/minimum/guaranteed-limit identity. Decide how much L loss the converter can accept at the worst peak and temperature.
- Estimate hot DC copper, AC winding, and core losses from part-specific manufacturer data or tools. Record an unknown instead of borrowing a curve from a neighboring series.
- Run the intended PCB, copper pattern, enclosure, airflow, ambient, and neighboring heat sources to thermal equilibrium, then measure the defined hottest accessible location.
- Measure inductor-current slope, ripple, peak, overshoot, and protection action. Include cold and hot startup, line/load corners, and load transients.
- Accept magnetic behavior, temperature, efficiency, and system stability separately. “Below rated current” is not a substitute for those results.
- Save waveforms, thermometry location, conditions, source revision, assumptions, and pass criteria. If a test was not run, label it pending rather than reporting a pass.
Frequently asked questions
Is Isat an absolute-maximum current?
Not automatically. It is normally an L-drop point under stated conditions. The exact datasheet may impose other limits, and the converter must define how much inductance reduction is acceptable.
Should converter peak current be compared with Isat or Irms?
Use the worst credible instantaneous peak for the inductance-retention check. Use the full waveform’s RMS value and total loss for heating. Both checks also need their stated temperature and measurement conditions.
Does Irms include switching-frequency core loss?
Do not assume that it does. The cited catalog methods use DC or low-frequency conditions for a baseline; application AC winding and core losses need part-specific treatment.
Can two inductors with the same Isat be treated as equivalents?
No. The drop criterion, reference L, frequency, stimulus, temperature, qualifier, DCR, thermal setup, and full L/current curve may differ. Package and one current value cannot establish equivalence.
Can a design briefly exceed Irms or Itemp?
The temperature-rise rating is not a general pulse allowance. Any temporary exceedance needs exact-part manufacturer evidence, a transient thermal assessment, and validation in the intended system. There is no universal percentage allowance.
Decision rule and first-party references
Treat Isat as a condition-bound inductance-retention gate and Irms/Itemp as a condition-bound temperature-rise gate. Map peak and RMS quantities from the real waveform, preserve frequency, ripple, temperature, DC bias, board cooling, and evidence qualifiers, and close the decision with hardware results. A current value without its object and method is a screening clue, not a design guarantee.
