Este guia está disponível atualmente em inglês.
Technology / Technical guide
Parasitic Inductance, Current Slew & Switching Overshoot
An electrical interconnection can contain inductance even when no separate inductor is drawn in the schematic. When current changes, that inductance contributes a dynamic voltage. The resulting local overshoot depends on the complete circuit rather than only the nominal DC voltage or average converter loss.
An interconnection stores energy and responds to current change.
Inductive voltage follows a declared current path and time derivative; it is not a universal extra voltage added to every terminal.
01 / Distributed geometry
A current path can have an unintended inductive contribution.
Physical outgoing and returning current paths create magnetic fields.
Physical outgoing and returning current paths create magnetic fields. Their geometry therefore matters even if the intended circuit drawing treats the connections as ideal wires. Parasitic inductance is a model of this electromagnetic behaviour over an applicable range.
TI’s half-bridge reference identifies the switching-cell loop as one example of a relevant path. That example does not establish MARSTON busbar geometry, semiconductor technology or actual loop inductance. A named device alone does not describe the full interconnection.
02 / Voltage relation
A rate of current change creates an inductive voltage.
For a constant linear lumped inductance under the passive reference convention, vL = L di/dt.
For a constant linear lumped inductance under the passive reference convention, vL = L di/dt. Current enters the terminal labelled positive voltage. The sign therefore changes with the current slope, and interpreting it as an added terminal stress requires the surrounding circuit orientation.
A large current and a rapid current change are different quantities. A steady high current gives zero ideal inductive voltage even though it can create resistive heating. A lower current changing quickly can produce a substantial inductive term in the stated model.
03 / Arithmetic example
The derivative needs consistent units.
In a hypothetical linear model, assume L = 1 µH and current changes by 5 A over 1 µs at a constant slope.
In a hypothetical linear model, assume L = 1 µH and current changes by 5 A over 1 µs at a constant slope. The slope is 5 million A/s, giving a 5 V inductive voltage magnitude. These values are invented and describe no actual converter loop.
The 5 V term is not automatically the complete measured device peak or an acceptable voltage margin. Additional circuit voltages, parasitic capacitance, loss and transition behaviour influence that peak. The calculation isolates one contribution under explicit assumptions.
04 / Stored energy
A transient redistributes energy already in the field.
An ideal constant inductor stores magnetic energy ½Li².
An ideal constant inductor stores magnetic energy ½Li². When its current changes, energy must exchange with other paths or storage elements. This helps explain why a local voltage excursion can accompany a switching transition without representing newly created energy.
If capacitance participates, the network can exchange energy between magnetic and electric storage. Loss damps that exchange. An initial overshoot and later ringing are related possible waveform features, but they need not have the same amplitude, duration or physical contribution.
05 / Different metrics
Peak voltage and average dissipated power answer different questions.
A voltage peak records a local instantaneous quantity.
A voltage peak records a local instantaneous quantity. Event energy integrates an applicable power waveform, and average loss depends on repetition and operating conditions. A short high peak can therefore be important even where an average-loss figure appears modest.
The loss-mechanism guide and SOA guide retain these separate meanings. A waveform feature is not a whole-system efficiency result or an independently acceptable component stress.
06 / Model scope
The physical principle does not select a remedy.
A displayed transient also reflects the defined observation boundary and bandwidth.
A displayed transient also reflects the defined observation boundary and bandwidth. Its appearance alone cannot uniquely identify which inductive or capacitive path contributed. The Miller guide explains another simultaneous device-level interaction.
No MARSTON parasitic inductance, overshoot margin, snubber arrangement or switching waveform is specified. This page provides no layout modification, device replacement, gate adjustment, protection setting or live measurement procedure; it explains a restricted current-to-voltage relationship.
Related technical guides: Miller Capacitance, Gate Charge & Switching-Node Coupling · Diode Reverse Recovery: Stored Charge & Commutation · LCL Filter States, Free Oscillation & Boundary Conditions.
Review snubber transient energy and damping as an additional switching-transient topic, without assigning a circuit to a specific converter.
General background: Texas Instruments: current-loop parasitic inductance and dynamic voltage. These references explain industry concepts; they are separate from the applicable MARSTON product and project documents.
Key concepts and conditions.
Read each quantity with its reference, units and stated conditions. A general relation does not establish a model-specific value.
| Concept / quantity | Meaning / relation | Conditions / limits |
|---|---|---|
| Parasitic inductance | Model of a physical current path | Geometry and return path matter. |
| Current slew | di/dt in A/s | Not the same as current magnitude. |
| Inductive voltage | L di/dt in V | Constant linear model and passive reference convention. |
| Stored energy | ½Li² in J | Not identical to dissipated transition energy. |
| Waveform metric | Peak, event energy or average power | Time and circuit boundaries stay separate. |
Engineering questions.
Can an idealised wire still need inductance in a fast-transient model?
Yes. Physical current-path geometry can create an inductive contribution.
Does a large constant current create ideal L di/dt voltage?
No. Its derivative is zero, although other losses can remain.
Is 5 V the example’s guaranteed device overshoot?
No. It is only an invented inductive contribution.
Are overshoot and ringing the same measure?
No. They describe related but distinct waveform features.
Does low average loss prove a peak is acceptable?
No. Local stress and applicable conditions are separate.
Does this provide a snubber or gate-drive recipe?
No. No hardware remedy or operating adjustment is prescribed.
Technical enquiry
Discuss the technical question.
An electrical interconnection can contain inductance even when no separate inductor is drawn in the schematic. Share the model and applicable operating or design information for an engineering discussion.