Discussion on Current Detection Circuit in Switching Power Supply

The circuit topology of the power switch circuit is divided into current mode control and voltage mode control. Current mode control has the advantages of fast dynamic response, simplified compensation circuit, large gain bandwidth, small output inductance, and easy current sharing, so it has been widely used. In the current mode control circuit, the current value needs to be measured accurately and efficiently, so the implementation of the current detection circuit becomes an important issue.

This article introduces the implementation method of the current detection circuit, and discusses the problems of current transformer saturation and secondary current droop that are often encountered in current detection. Finally, the current detection method in the boost circuit is analyzed using experimental results.

2 Implementation of current detection circuit

In the control circuit of the current loop, the current amplifier usually selects a larger gain. The advantage is that a smaller resistor can be selected to obtain a sufficient detection voltage, and the smaller the detection resistor, the smaller the loss.

There are two main methods for implementing current detection circuits: resistive sensing and current sense transformer detection.

There are two types of resistance detection, as shown in Figure 1 and Figure 2.

Figure 1 Resistance detection (grounding)

Figure 2 Resistance detection (not grounded)

When using Figure 1 to directly detect the current of the switch tube, a small RC filter circuit must be connected in parallel next to the detection resistor RS, as shown in Figure 3. Because when the switch tube is turned off, the collector capacitor discharges and generates a transient current spike on the current detection resistor. The pulse width and amplitude of this spike are often enough to lock the current amplifier, thereby causing the PWM circuit to fail.

Figure 3 Resistance detection circuit with filtering

However, in actual circuit design, especially when designing high-power, high-current circuits, the resistance detection method is not ideal, because the detection resistor has a large loss, reaching several watts or even more than ten watts; and it is difficult to find a resistor as small as a few hundred milliohms or tens of milliohms.

In fact, current transformer detection is practical in high-power circuits, as shown in Figure 4. Current transformer detection has a wide bandwidth while maintaining a good waveform. The current transformer also provides electrical isolation, and the detection current is small and the loss is small. The detection resistor can be a slightly larger value, such as a resistor of 10 to 20 ohms. The current transformer couples the entire transient current, including the DC component, to the detection resistor on the secondary side for measurement, but it also requires that the magnetic core can be reset normally every time the current pulse passes zero. Especially in average current mode control, current transformer detection is more applicable because the pulse current detected in average current mode control returns to zero in each switching cycle.

Figure 4 Boost circuit input inductor current value detection

In order to completely magnetically reset the current transformer, it is necessary to provide the core with equal and opposite volt-second products. In most control circuit topologies, the duty cycle is close to 100% when the current passes through zero, so the magnetic reset time when the current passes through zero accounts for only a small proportion of the switching cycle. To reset the core in a very short time, it is often necessary to add a large reverse bias to the current transformer, so when designing the current transformer circuit, a high-voltage diode should be used to couple between the secondary side of the current transformer and the detection resistor.

3 Methods to prevent current detection circuit saturation

If the core of the current transformer cannot be reset, it will lead to core saturation. Current transformer saturation is a very serious problem. First, the current value cannot be measured correctly, so effective current control cannot be performed; second, the current error amplifier always "thinks" that the current value is less than the set value, which will cause the current error amplifier to over-compensate and cause current waveform distortion.

Current transformer detection is most suitable for symmetrical circuits, such as push-pull circuits and full-bridge circuits. For single-ended circuits, especially boost circuits, there will be some problems that we must pay attention to. For boost circuits, the inductor current is the input current. Then, in the continuous current working mode, whether charging or discharging, the inductor current is always greater than zero, that is, a charging and discharging waveform is superimposed on the DC value. Therefore, the current transformer cannot be used to directly measure the input current of the boost circuit, because the inductor current cannot return to zero and the DC value is "lost"; and the current transformer cannot be saturated due to magnetic reset, thereby losing the overcurrent protection function, and the output produces overvoltage, etc. The same problem exists in the step-down circuit. The current transformer cannot be used to directly measure the output current.

The solution to this problem is to use two current transformers to measure the switch current and the diode current respectively. As shown in Figure 4, the actual inductor current is the synthesis of these two currents, so that each current transformer has enough time to reset. However, it should be noted that the turns ratio of the two current transformers should be the same to keep the current on the detection resistor RS symmetrical.

The power factor correction circuit generally adopts a boost circuit and uses a dual transformer for detection, but the current transformer is also particularly easy to saturate when the line current passes zero. Because the duty cycle at this time is about 100%, it is easy to cause the magnetic core to not have enough time to reset. For this reason, some measures can be taken in the external circuit to prevent the current transformer from saturating. For example, the current amplifier output clamp is used to limit its output voltage and further limit the duty cycle to less than 100%. The circuit is shown in Figure 5. The process of setting the clamp voltage is very simple. At the beginning, the current amplifier clamps at a relatively low value (about 4V), and the system starts to work, but the zero-crossing error is large; once the system is working normally, the clamp voltage will increase, the current transformer is close to saturation, and the clamp voltage rises to 6.5V at most (low voltage and large load) and the THD of the current is within an acceptable range (<10%) to limit the maximum duty cycle. The set clamp voltage cannot be too low, otherwise the current zero-crossing distortion will be large.

Figure 5 Current amplifier clamp circuit

If better characteristics are required or operation over a wide range is required, the circuit of Figure 6 can be used. This circuit will reverse-regulate the clamp voltage according to the line voltage.

Figure 6 Current amplifier clamp circuit with input voltage compensation

Each current pulse resets the core to overcome the core saturation method. In addition to improving the external circuit, the current detection circuit can also be improved. Generally, the current detection circuit is self-reset, that is, the energy stored in the core and the open circuit impedance of the current transformer are used to generate enough volt-second product in a short time to reset. However, when the duty cycle is greater than 50%, especially close to 100%, there may not be enough time to reset the core. At this time, in addition to the current amplifier output clamp, a forced reset circuit can also be used.

There are many circuits for forcing the core to reset, such as using an additional coil or a center-tapped coil, but the simplest method is to use the circuits shown in Figures 7 and 8 to force the core to reset. When a pulse current comes, there is no difference between the operation of the forced reset circuit and the self-reset circuit. When resetting, the current from VCC through Rr is added to the core reset current, the parasitic capacitor is quickly charged, the secondary voltage is reversed, the volt-second product increases, and the core reset speed is accelerated. If you need to get a negative detection voltage but don't want to use a negative voltage to force reset, use the circuit shown in Figure 8.

Figure 7: Forced reset circuit for detecting positive voltage

Figure 8: Negative voltage detection forced reset circuit

Another factor to consider for the core reset of the current detection circuit is the leakage inductance and distributed capacitance of the secondary coil. In order to reduce losses, a current transformer with a larger turn ratio is generally selected, but the larger the turn ratio, the larger the leakage inductance and distributed capacitance of the secondary coil. The leakage inductance affects the time for the current to rise and fall, and the distributed capacitance affects the bandwidth of the current transformer. And when the core is reset, the secondary inductance and distributed capacitance resonate. If the distributed capacitance is large, the resonant frequency is low and the period is long. Then, when the duty cycle is large and the core reset time is short, the secondary coil does not have enough time to release energy to reset the core. Therefore, you should try not to choose a current transformer with a large turn ratio.

4 Droop effect of current transformer

The pulse current on the secondary side of the current transformer must be equal to the current on the detection resistor by subtracting a magnetizing current generated by the pulse voltage on the secondary side of the current transformer winding, which increases linearly with time from zero. The magnitude of the magnetizing current is:

(1)

Where: US - secondary voltage; LS - secondary inductance; n - Ns/Np; Δt - current pulse width

At the beginning, the secondary current is n times the primary current, but as time goes by, the magnetizing current increases and the secondary current drops sharply. This is the droop effect of the current transformer. Therefore, in order to obtain a larger secondary detection voltage, it should not be achieved entirely by increasing the value of the detection resistor Rs, but also by reducing the secondary droop effect to increase the secondary pulse current. At the same time, a large value of Rs will make it difficult to reset the magnetic core.

As shown in formula (1), the larger the secondary inductance value, the smaller the droop effect; the smaller the turns ratio, the smaller the droop effect. However, it is best not to reduce the turns ratio by reducing the number of turns on the secondary side, because this will reduce the inductance of the secondary side. The number of turns on the primary side should be increased to reduce the turns ratio if space permits.

5 Experimental Results

In the power factor correction circuit, the detection circuit shown in Figure 4 is used, and the measures for preventing core saturation and reducing droop effect as described above are adopted. When the transformation ratio of the current transformer is 1:50, the secondary inductance is 30mH, the secondary voltage is 2V, and the current wave pulse width is 5μs, it is obtained that: relative to the detection current of more than ten amperes, the current drop effect is not obvious.

6 Conclusion

Current detection plays an important role in current control. Current detection is divided into resistance detection and current transformer detection. In order to reduce losses, current transformer detection is often used. In the design of the current transformer detection circuit, the influence of the circuit topology on the detection effect should be fully considered, and the saturation problem of the current transformer and the droop effect of the secondary current should be comprehensively considered to select the appropriate core reset circuit, turns ratio and detection resistor.