Low-power electronic load for fast load transient testing

qwqwqw2088

Low-power electronic load for fast load transient testing [Copy link]

In the DCDC power supply test, the load transient test is a very important part. The load transient test can be used to quickly evaluate the stability and speed of the tested power supply. When selecting the DCDC converter chip, the load transient test performance is also an important reference for evaluating the dynamic performance of the chip. The figure below is a typical waveform of a DCDC converter load transient test. CH3 is the AC component of the output voltage, and CH4 is the load current. Note that the load current rising slope is not the same as the falling slope. A slower rising slope corresponds to a smaller voltage drop (Undershoot), while a steeper falling slope corresponds to a larger voltage overshoot (Overshoot).

Figure 1 Typical dynamic waveform of load

Load transient is usually tested using an electronic load (E-Load). As mentioned above, the load slew rate will have a key impact on the test results. However, due to the limitations of the internal circuit of the device, the di/dt that can be achieved by conventional electronic loads is not very high. In addition, due to factors such as the design of different manufacturers, the slew rates that can be achieved by electronic loads of different models are also different. As shown in Figure 2(a)(b) below, the two figures are models A and B, respectively. When the actual current rising slope is compared at the same setting of 2.5A/us, it can be seen that the actual current slew rate is much smaller than the set value, and the slew rates of different models are also different. This may lead to an ideal power supply transient test result or an objective performance evaluation between different chips. Therefore, it is of great engineering significance to design a simple and practical electronic load with a load slew rate that can meet the experimental requirements.

Figure 2 (a) Model A Figure 2 (b) Model B

To achieve a higher load jump rate, the conventional design idea is to use MOSFET to disconnect the load resistor. Although this method is simple to implement, it has an obvious disadvantage in practical application: since the switching process of MOSFET is generally in the hundreds of nanoseconds, the load current jump rate is mainly limited by the ESL (equivalent series inductance) of the selected load resistor. Generally, sliding rheostats are winding resistors, and their ESL is often large, so it is difficult to achieve a high jump rate. If an independent non-inductive power resistor is selected, assuming that the test needs to cover the load jump of 1.8V/3.3V/5V/12V at 0.1A/0.5A/1A/2A/3A, it is necessary to prepare up to 20 resistors with different resistance values. If the voltage/current combination is more complex, more resistors with different resistance values will be required, and the corresponding resistor must be replaced when the test voltage or load current changes, which is very troublesome.

In view of the shortcomings of the above traditional methods, this paper designs a low-power practical electronic load based on MOSFET. As shown in the figure below, the design mainly includes four parts: MOSFET, driver stage, power rail and pulse generator. Its basic working principle is: MOSFET is not in a conventional switching state, but makes it work in a constant current area. The pulse generator drives MOSFET through DRV8836 to generate a GS voltage with a certain amplitude and pulse width, thereby realizing the jump of drain current (load current). The amplitude of the load current can be controlled by adjusting the LDO output voltage, and the rise/fall slope of the load current can be controlled by adjusting the resistance value of the drive resistor.

Figure 3 System block diagram

There are a few points worth noting in the design:

Since MOSFET is in the constant current region, the drain current is controlled by the GS voltage. If the slope is adjusted by the traditional diode plus drive resistor, when the GS voltage and the drive voltage are less than the forward voltage drop of the diode, the diode will be equivalent to a high resistance, which will increase the time constant of the drive circuit and deteriorate the dynamics. Therefore, two half-bridges of DRV8836 are used here to achieve independent control of charge and discharge; the
actual load dynamic test requires a certain current A to jump to another current B, which can be decomposed into DC current (current A) and AC current (current B). This design only needs to consider the AC current (jump part), and the DC current only needs to connect an adjustable power resistor in parallel at both ends of the MOSFET;
to reduce the heating of the MOSFET, a lower pulse frequency (such as 10Hz) can be set, and a lower duty cycle can be matched accordingly;
to facilitate offline operation, the pulse generator part uses the LMC555 timer to build a pulse generation circuit. The following circuit realizes a pulse generator with constant frequency and adjustable duty cycle. The addition of two diodes separates the charge and discharge circuits. The charge and discharge time can be adjusted by adjusting R2, thereby achieving adjustable duty cycle. The charge and discharge time and pulse frequency are calculated as follows:

When actual conditions permit, a signal generator can also be used directly to generate a pulse signal.

To verify the feasibility of the design, a prototype is built based on the above design block diagram as shown below:

Figure 4 Prototype

The prototype uses an aluminum shell power resistor as a DC load, and adjusts the driving resistance through a multi-turn adjustable potentiometer to achieve different load jump slopes. A 5V/1A~6A load transient experiment is performed on the DCDC module under test. The experimental results are shown in the figure below. Channel 3 is the output voltage AC signal, and channel 4 is the load current:

Figure 5 (a) 125mA/us Figure 5 (b) 250mA/us

Figure 5(c) 2.5A/us Figure 5(d) 2.5A/us Zoom In

From the experimental results, we can see that the prototype can achieve load jump at a given (average) slope, and the slope and load size can be continuously adjusted by adjusting the drive resistance and LDO output voltage respectively. It is worth noting that the voltage drop (Undershoot) of the tested DCDC module at 250mA/us is 268mV, and at 2.5A/us it reaches 432mV, which shows that the load jump rate (Slew Rate) has a very obvious impact on the load transient test results.

Starting from the needs of actual load transient experiments, this paper analyzes the limitations of existing electronic loads, designs a simple and practical low-power electronic load based on the actual application needs, gives a specific system block diagram and design points, and builds a prototype for experimental verification. The experimental results show that the prototype can achieve load jumps at a given slope, which can better meet the needs of low-power load transient testing.

Author: Captain Luo