Sub-Bandgap Reference Controls Sub-1V, 1.7A LDO Regulator

With the trend toward decreasing voltages, power management must keep pace to supply the necessary voltage and power levels. Achieving a precise sub-1V output with current devices is not feasible because most have reference voltage levels of either 1.25V or 2.5V. A sub-1V LDO (low dropout) regulator has become a vital component.

Figure 1: A precision sub-bandgap reference U1 in the feedback loop of pass transistor Q1 provides the necessary voltage margin to realize a 0.9V, 1.7A output LDO regulator.

For an accurate output voltage, the circuit's feedback network in Figure 1 uses a precision voltage reference. A voltage reference should not be used if its output voltage is within a 20% margin of the reference voltage. Because the LDO's output voltage is below 1.0V, the traditional TL431 and TLV431 programmable precision references are unacceptable. Instead, the ON Semiconductor NCP100 is selected for U1 because it supplies a 0.7V reference. With a low reference level, 0.9V is attainable with a greater than 20% margin.

In selecting the pass transistor, the choice between an NPN or a PNP must be made. Because this design is powered from a single supply, the NPN's dropout voltage minimum is the greater of either the saturation voltage from collector-to-emitter or the base-to-emitter ON voltage. For the PNP, the minimum voltage drop from input to output is the saturation voltage of emitter-to-collector.

To maximize the output current and minimize the voltage drop, this design uses the ON Semiconductor MBT35200MT1 PNP for Q1. This PNP has a 2.0A collector current, maximum emitter-to-collector saturation voltage of 0.31V, typical DC current gain of 200, and maximum emitter-to-base voltage of 0.875V.

Figure 1 illustrates the circuit schematic. Inverting the signal from the precision reference is necessary because of the PNP pass element. A small-signal NPN (Q2) inverts the signal. There is a voltage-differential issue with driving the base of an NPN, 0.6V to 0.7V turn on, from the NCP100's cathode, 0.9V minimum. To accommodate the differential, a diode (D1) shifts the voltage level.

A resistor (R5) from gate to ground performs two functions: pulls the gate to ground for turn off and provides a bias current through the diode to set a minimum voltage drop. If the minimum voltage drop is not set properly, the NPN will have a small base current that will be amplified by the NPN and the PNP, causing a voltage runaway at the output during light or no load. Because of the possibility of double amplification, the NPN’s base to emitter voltage needs to be very low.

In addition to Q2, to invert the control signal from the NCP100, R6 pulls up the gate to the input voltage to turn off the MBT35200MT1. R4 is an overcurrent protection resistor. R4 is determined by subtracting the minimum input voltage from the maximum V_BE of Q1 and the maximum V_SAT of Q2, then dividing by the base current of Q1.

A 1-µF capacitor (C3) is necessary for the NCP100's normal operation. It stabilizes the operation of the precision reference and has a negligible effect on the response time of the system. R1 and R2 comprise the resistor divider feedback network. C4 is necessary for the system’s fast transient response.

R3 provides the DC bias for the NCP100. The value of R3 is limited by the response of the system at low line and low load. If the value of R3 is too large, oscillations occur on the output. If R3 is too small, the output voltage will run away at high line and low load.

Figure 2: A 0.5A to 1.5A load transient applied to the circuit in Figure 1 results in small output voltage variation.