Transient loads provide a workout for the power system

Testing a power system's transient response using the fast dynamic load described in this design idea can reveal many key operating characteristics. The voltage deviation resulting from a fast current step can provide insight into the regulator's phase margin. Furthermore, for power supplies located some distance from the point of load, transient testing can help determine the effective series interconnect inductance, shunt capacitance, and ESR. While the phase margin of commercial power supplies is typically verified by the supplier, adding remote sensing often destabilizes the supply. Interconnect inductance and load capacitance can introduce additional phase shift in the regulator's control loop feedback, impacting stability.

Many engineers have seen a low-frequency sine wave on the output of a regulator.

Performing transient testing on an assembled system provides a quick check of the system's dynamic regulation stability and accuracy. Most commercial dynamic electronic loads have rather slow current slew rates, which often limits their usefulness for testing faster regulator control loops, which typically reach steady state in 50µs or less after large load transients. Many high-power systems require current slew rates of 10A/µs or higher.

Figure 1 is an adaptation of the application note with some notable improvements. Maximum power has been increased to 150W, designed for 3.3V, 5V, and 12V regulator outputs. R1-R3 form a resistive load switched by a single N-channel low-side MOSFET. The size and population of the load resistors allow for a wide range of possible load combinations.

Figure 1 Schematic diagram of transient load tester

At the core of the circuit, MOSFET driver U1 with Schmitt trigger inputs drives the MOSFET and forms a relaxation oscillator with Q2, R8, R9, and C3. For the component values ​​shown, the duty cycle is approximately 5%, resulting in a cycle time of 20ms. Having a relatively low duty cycle allows for a moderate cooling solution.

R6 and R7, combined with the MOSFET's input capacitance, CISS, independently adjust the rise and fall times. With the values ​​shown, the rise and fall times are approximately 1µs. At this slew rate, the peak MOSFET gate current is approximately +110/-75mA, well below U1's maximum current rating of 1.4A. C2 can be added to further slow the edge rate. Due to the 1µs rise/fall times and the added dissipative damping of the relatively large gate resistor, MOSFET gate switching resonance is less noticeable. When the MOSFET turns off, R4 and C1 help dampen line resonance. The value of R4 is determined by the effective line inductance and input capacitance. A value of 0.5Ω has been shown to be effective for typical wiring scenarios.

One of the more convenient features of this implementation is the two-wire connection to the DUT. For 3.3V and 5V systems, a 12V boost converter is included to power the MOSFET driver and gate. No other connections or power supplies are required. The boost converter's output can provide approximately 350mA from a 3.3V input, which may limit the amount of current available to charge the MOSFET gate. Low-ESR aluminum capacitor C5 provides some initial gate-charging current for faster current edge rates. For 12V operation, a direct-connect version can be assembled by replacing C7 with a 0Ω resistor, disabling the boost converter. Some voltage drop will occur across L1 and D2, but this will not affect normal operation of the circuit.

The entire circuit fits comfortably on a 3" x 5" two-layer PCB, including the heat sink and small 12V fan. Operation is very simple, requiring only two wires to connect.

The tester leads must be short and low inductance to prevent ringing due to lead reactance. DUT connections should be made near the point of load or remote sense location. The tester common and voltage probe return leads should be connected in one location. This location should also be chosen to provide a low-impedance path back to the power supply.

Pressing momentary pushbutton PB1 activates the astable circuit, and the dynamic load begins switching. A static PSU load can be provided externally if desired. R5 and J2 provide a convenient, high-bandwidth means for measuring pulsed current. A straight 50Ω coaxial cable can be connected directly to the oscilloscope input for monitoring 1mV/A current. Voltage measurements should also be made near the point of load or a remote sense point and AC-coupled to a second oscilloscope input. Voltage probing must be done with caution. Probe inductance from distant ground/return leads can cause misleading measurements. A small series resistor (a few ohms) can be added to the probe tip to suppress high-frequency ringing from the probe's ESL.