Circuit
Oscilloscope
Oscilloscope screen
This circuit provides latching and high-speed overcurrent protection for power-factor corrected (PFC) boost converters operating at dc voltages to 380 V. Overcurrent damage of PFC front-end boost converters can occur due to any kind of fault, such as a short circuit. But converter damage due to instantaneously high current levels, even for downstream units with inaccessible loads, can be avoided with a circuit that provides fast latching and handles voltage rails to 380 V.
In extreme cases, boost converters have been known to explode due to short-term overcurrent conditions. The downstream PFC pre-regulator poses a number of handling challenges because of the large size of the PFC circuit’s reservoir capacitor, the inherently poor current-limiting protection built into the boost converter, and the large short-term current levels possible with a 380-V rail.
The use of an additional protection circuit, placed between the output of the PFC circuitry and the input of a downstream converter (VBUS) (Fig. 1) rated to 380 V, can help prevent boost converter damage or even explosions due to mishandling or errors in manufacturing, debugging, and testing. In contrast to a simple load switch, this circuit must provide high common-mode-voltage current sensing, fast turnoff, high-power capability, and the ability to sense both ac and dc components because of the switched-mode load.
The protective circuitry consists of three sections: a load switch with on/off circuitry, a current-sensing circuit, and a comparator. The regulator network that includes transistor Q6 and diode D6 on the right-hand side of the circuit schematic form a high-side-referenced series regulator to maintain the VCC-VEE bias to amplifier U1 at about 60 V
Transistor Q7 converts the current sensed by resistor R12 into a ground-referenced voltage for the LTC6101HV current-sense amplifier. (More information on the LTC6101HV is available in a downloadable data sheet from Linear Technology at www.linear.com.) Diode D5 protects U1’s input from damage if R12 fails high. Components L1, C3, and R16 form a filter that removes the ac components superimposed on the sensing current while also minimizing any current spikes during startup.
Amplifier U2 compares the post filtered signal voltage, VCS, with a reference for any overcurrent events. During normal operation, outputs /Q and Q are high and low, respectively. Upon detecting an overcurrent event, outputs /Q and Q change states, light-emitting-diode LED1 is illuminated, and capacitor C6 is charged to latch the /Q and Q states. Once U2 is latched, it must be reset by manually pressing switch S1.
The load switch formed by transistors Q1 and Q2 is controlled by Q5, but its transition time is too slow to provide practical overcurrent protection for a 380-V rail. However, the network that includes devices Q3 and Q4 provides the switching speed required to guard against overcurrent damage. Upon the occurrence of an overcurrent event, Q3 will quickly short-circuit load switch Q1’s and Q2’s gate-source voltage (VGS) at high speed before Q5 can react.
To conserve power, the drive to transistor Q4 is ac coupled through capacitor C1, only until Q5 and its associated circuitry take over the power turn-off. After pressing switch S1, the charge stored in C1 is removed through R7, D3, and Q. The component values for the filtering network must be carefully chosen. Insufficient filtering can cause false triggering, while too much filtering can slow the switching response time and result in inadequate protection against overcurrent events.
An electronic load was used to test and tune the circuit (at 110 V due to the limited capabilities of the load) in the development stage (Fig. 2). When the load was switched to a short-circuit mode, the protective circuitry reacted almost instantly. As VCS reached 2 V, threshold voltage VCEQ4 switched on almost immediately while voltage VCEQ5 required only about 40 µs to switch state.
The circuit was also tested by simulating a fatal error within the downstream converter, an event that causes a large, short-term current flow that can eventually cause the converter to explode. By adding the protection circuitry, the converter is effectively safeguarded from damage even under such drastic conditions.
When the “fatal” event occurs, the input current IIn climbs rapidly while voltage VCS tries to catch up. After about a 10-µs delay, comparator Q (Fig. 3) is tripped and current IIn decays to zero. Once safe conditions are restored and switch S1 is thrown, the unit under test resumes normal operation.