How to choose multi-phase Buck coupling inductor? Load transient verification of 72–320nH, DCR and high temperature saturation current｜VOOHU

How to choose a multi-phase Buck coupling inductor? The key is not just to divide the output current by the number of phases and then find a product with a larger current. When the load of server motherboards, switches, graphics cards and high-current processor power supplies changes rapidly, common problems include output voltage sag or excessive overshoot, uneven current distribution among phases, early bending of the inductor waveform after high temperatures, and even protection action earlier than room temperature. To deal with these phenomena well, we need to look at the phase coupling method, inductance, DCR, high temperature saturation current, temperature rise capability and the load transient strategy of the controller at the same time.

VOOHU combined inductorCovering common choices such as 72nH, 150nH, 260nH and 320nH, it can be used in high-current scenarios such as multi-phase DC/DC, CPU/GPU power supply, server and switch power supply.WHPBU-100705N-R072L0、WHPBU-100710N-R15L0、WHPBU-131308N-R26L0andWHPBU-131308N-R32L0The four products have different emphases in sensing, DCR and high-temperature current capabilities, and are suitable for establishing sample comparisons under the same controller and PCB conditions. When selecting, first clarify whether you want to reduce steady-state ripple, improve load steps, or reduce copper loss and temperature rise, and then decide which set of parameters is closer to the target.

1. How to choose multi-phase Buck coupling inductor: first distinguish between ripple and transient targets

Phase-to-phase coupling changes the current change path, not just the number of devices

In discrete inductors, each phase mainly limits current changes by its own inductance; after adopting a coupling structure, changes in current in one phase will affect the other phase through the magnetic circuit. Reasonable coupling direction can suppress each phase ripple during steady-state interleaved operation, while retaining the current ramp-up speed required when the load changes. If the coupling direction, pin connections, or controller phase sequence do not match, the desired ripple cancellation may turn into additional current deviation. Therefore, when determining the solution, the controller phase sequence, the starting and ending ends of each winding, and the PCB network should be checked together.

If the inductance is too small or too large, it will leave characteristics on the waveform.

When the inductance is small, the current slope is faster, and the output current can be supplemented faster when the load suddenly increases, but the steady-state ripple, peak current and AC loss are usually more worthy of attention; when the inductance is large, the steady-state ripple is easier to suppress, but the current build-up speed will be slower, and a deeper voltage sag may occur during load steps. 72–320nH is not a simple ranking of advantages and disadvantages from small to large, but provides different adjustment space for controller frequency, input and output voltage difference, phase number and target transient.

The difference between multi-phase coupled inductors and discrete inductors depends on the actual waveform of the same board

When comparing two options, it is not advisable to just look at the individual inductance values on the data sheet. Each phase current ripple, phase relationship, output sag and overshoot, recovery time and device temperature should be recorded under the same input voltage, output voltage, switching frequency, load step and output capacitance conditions. Only when the test conditions are consistent can we judge whether the improvement comes from the coupling structure, changes in sensing, or the difference between controller compensation and PCB paths.

2. How to choose coupling inductor DCR and saturation current: average, peak and high temperature

DCR affects copper loss and also amplifies the current difference between phases.

The inductor copper loss can be estimated first using Pcu≈Irms²×DCR. After the current enters tens of amps, even if it is only a few tenths of a milliohm, the loss and temperature rise will become obvious. Multi-phase design also needs to pay attention to whether the DCR of each phase is close to the board-level path: differences in copper foil length, number of vias, pad narrowing and sampling point location will all be superimposed on the device DCR. If a certain phase is overheated for a long time, the average current of the phase and the voltage drop across the inductor should be measured at the same time to avoid just replacing the device and ignoring the bias current caused by wiring and sampling.

25℃ saturation current cannot replace high temperature peak value judgment

The saturation current is mainly used to determine the peak boundary during startup, load step and current limiting processes; the temperature rise current is closer to the thermal capacity during continuous operation, and the two cannot replace each other.WHPBU-100705N-R072L0The typical saturation current is 80A@25℃, 70A@100℃, andWHPBU-131308N-R32L0It is 50A@25℃, 35A@100℃. As the ambient temperature increases, the available peak headroom shrinks, so the 100°C data, device self-heating, and actual peak current should be evaluated together.

DCR and temperature rise current should also be compared in the same sensing range.

WHPBU-100710N-R15L0is 150nH, DCR is 0.18mΩ±10%, and the typical temperature rise current is 70A, which is suitable for priority evaluation of low copper loss and continuous high current;WHPBU-131308N-R26L0is 260nH, DCR is 0.32mΩ±9.4%, and the typical temperature rise current is 45A, which can establish an intermediate range between ripple, transient and thermal margin. For products with different structures, the current carrying capacity cannot be inferred solely based on the magnitude of the inductance. The DCR, 25°C and 100°C saturation current, and temperature rise current in the selection table need to be compared together.

3. How to choose multi-phase Buck load transient inductor: look at slope and current sharing

When the load suddenly increases, first check the sag depth, and then check whether the current of each phase rises synchronously.

When the load jumps from light load to heavy load quickly, the output capacitor first bears the current gap, and then the inductor current of each phase increases. If the sag is too deep and the waveforms of each phase remain uniform, it is usually necessary to continue to optimize the inductance, switching frequency, output capacitance and control loop bandwidth; if only one phase climbs slowly or the current is limited in advance, the phase sequence, sampling, drive and loop impedance of the phase should be checked. Putting the output voltage and all phase currents on the same time axis makes it easier to find the problem than just looking at the steady-state ripple.

Waveform bending and phase deviation are important signals in high temperature testing

When the peak current approaches the magnetic boundary, the decrease in inductance causes the current slope to accelerate, and the waveform may bend significantly. If this change only occurs at high temperatures or after heavy load for a long time, the device temperature, input voltage, output current and controller status should be recorded simultaneously. The turning points of each phase are obviously different, which may be related to thermal distribution, DCR, magnetic consistency or board-level impedance; first fix the conditions, and then change the load, air volume or sample one by one, the conclusion will be clearer.

Load dump also requires checking for overshoot and recovery.

When the load suddenly decreases, the energy in the inductor needs to be handled by both the output capacitor and the controller. The inductance, coupling method, phase shutdown sequence and compensation parameters all affect overshoot. The test should cover the process from light load to heavy load, heavy load to light load, and phase increase and decrease, and confirm that each phase can exit smoothly without abnormal reverse current, repeated starts and stops, or long-term ringing.

4. VOOHU 72–320nH combined inductor selection reference

The following four modelsVOOHU combined inductorForms a sample gradient from fast current response to stronger ripple suppression. You can first screen according to the sensing range recommended by the controller, then use DCR, high temperature saturation current and temperature rise current to narrow the range, and finally determine a more suitable product through load step and thermal tests on the same board.

If the design focus is on fast load response, you can first compare the 72nH and 150nH products, focusing on observing the peak value, ripple and temperature rise; if you pay more attention to the steady-state ripple, you can compare the 260nH and 320nH products, and pay attention to the difference in the 100°C saturation current from 45A to 35A. The 150nH product has outstanding DCR and temperature rise current performance, but the final selection still depends on the actual phase current, switching frequency, heat dissipation conditions and controller dynamics.

5. In the sample stage, ripple, load step and thermal stability are put into the same round of verification.

First fix the controller parameters and PCB conditions

Comparison of samples from the same round should use the same PCB, input and output conditions, switching frequency, output capacitance, compensation parameters and heat dissipation methods. After replacing the product, first confirm the installation direction and winding connection, and then gradually increase the load from light load to rated load. If the inductor, compensation, and output capacitance are changed at the same time, it is difficult to determine where the waveform improvement comes from.

Steady-state test simultaneously records ripple, current sharing and temperature rise

Record the average current of each phase, peak-to-peak ripple, output ripple and efficiency at multiple load points, and wait for the temperature to stabilize before reading the device surface temperature. When the temperature rise is high, you should not only pursue lower DCR, but also check the switching frequency, core losses, copper area, air ducts and thermal effects of adjacent MOSFETs. The multi-phase current difference should be judged based on the allowable range of the controller, and it should be confirmed that the deviation will not continue to expand with temperature.

Load steps cover maximum current changes and phase switching

The amplitude, rising edge, and repetition frequency of the load step should approximate the real application. Record output voltage sag, overshoot, recovery time, peak value of each phase and protection action, while covering phase increase, phase decrease and startup process. If a product has good transient performance at room temperature, but the waveform bends or the protection is advanced after high temperature, the high temperature saturation margin needs to be re-evaluated.

After thermal stabilization, retest efficiency and EMI.

Changes in inductance will change the current ripple, switch node waveform, and high-frequency current in the loop area, and efficiency and EMI may change accordingly. After the candidate product passes the transient and temperature rise tests, the efficiency, output ripple, conduction and radiation performance should also be retested at the upper and lower input limits and main load points. The products selected in this way not only have suitable parameters, but are also closer to the actual long-term working state of the whole machine.

6. Frequently Asked Questions (FAQ)

How to choose a multi-phase Buck coupling inductor, which parameter to look at first?

Start with the recommended sensing range and number of phases of the controller, and then check the continuous current, worst peak value, DCR, 100°C saturation current and temperature rise current of each phase. Then use load step, current sharing and thermal stability to test the convergence selection. Don't just look at the room temperature saturation current.

Can multi-phase coupled inductors directly replace discrete inductors?

It cannot be directly replaced by sensor and DCR alone. Also confirm the number of windings, pin orientation, coupling method, controller phase sequence, pad and PCB network, and re-verify the control loop, load transients, temperature rise and EMI.

Why is the temperature rise current of 72nH products lower than that of 150nH products?

The temperature rise current is determined by the winding structure, DCR, material, packaging and specified test conditions, and will not change unidirectionally with the inductance.WHPBU-100705N-R072L0withWHPBU-100710N-R15L0Belonging to different structures, complete parameters and actual thermal test results should be compared respectively when selecting models.

How to verify the temperature rise of combined inductor more reliably?

Continue to run under target PCB, air duct and enclosure conditions until the temperature stabilizes, while recording each phase RMS current, device surface temperature, ambient temperature and output indicators. Also cover input upper and lower limits, continuous overloading and phase switching to avoid using short-term readings from an open bench to replace the overall machine results.

7. Conclusion: Use the same set of waveforms to close the loop of selection judgment

The selection of multi-phase Buck coupling inductors requires a balance between ripple, load transients, copper losses and high temperature peaks. VOOHU WHPBU combination inductors are available in a range of 72–320nH, enabling clear sample gradients to be established around different controllers and current levels. By comparing each phase current, output voltage, device temperature and efficiency under the same set of working conditions, we can more quickly determine which product is more suitable for the current power supply, and reserve sufficient margin for subsequent complete machine verification.