Home | Articles | Forum | Glossary | Books


It has been shown that to obtain the most efficient performance from bipolar power switching transistors, the base drive current must be correctly profiled to suit the characteristics of the transistor and the collector-current loading conditions. If the base drive current remains constant, problems can arise in applications where the collector current (load) is not constant.

When the drive current has been chosen for optimum performance at full load, if it then remains the same for light loading conditions, the excessive drive will give long storage times, which can lead to a loss of control in the following way. Under light loading (when narrow pulses are most required), the long storage time will give an excessively wide pulse.

The control circuit now reverts to a "squegging" control mode. (This is the cause of the well-known "frying-pan noise," a nondamaging instability common to many switchmode supplies at light loads.) Hence, to prevent overdrive and squegging when the load (collector current) is variable, it is better to make the amplitude of the base drive current proportional to the collector current. Many proportional drive circuits have been developed to meet this requirement.

A typical example follows.


FIG. 1 shows a typical proportional drive circuit applied to a single-ended forward converter. In this arrangement, a proportion of the collector current is current-transformer coupled by T1 into the base-emitter junction of the main switching transistor Q1, providing positive proportional feedback. The drive ratio Ib /Ic is defined by the turns ratio of the drive transformer P1/S1 to suit the gain characteristics of the transistor (typically a ratio of between 1/10 and 1/5 will be used).

Because the drive power during most of the "on" period is provided from the collector circuit, by the coupling from P1 to S1, the drive requirements from Q2 and the auxiliary drive circuit are quite small.


During the previous "off" period of Q1, energy has been stored in T1, since Q2, R1, and P2 have been conducting during this period. When Q2 turns off, the drive trans former T1 provides the initial turn-on of Q1 by transformer flyback action. Once Q1 is conducting, regenerative feedforward from P1 provides and maintains the drive to Q1.

Hence, Q2 is turned off for the conducting ("on") period of Q1, and on for the "off" period of Q1.

FIG. 1 Single-ended forward converter with single-ended proportional base drive circuit.


When Q2 is turned on again, at the end of a conducting period of Q1, the voltage on all windings is taken to near zero by the clamping action of Q2 and D1 across the clamp winding S2. The previous proportional drive current from P1 is now transformed into the loop S2, D1, and Q2, together with any reverse recovery current from the base-emitter junction of Q1 via S1 (less the current transformed from P2 as a result of conduction in R1). Hence the base drive is removed, and Q1 turns off.

As positive feedback from P1 to S1 is provided in this drive circuit, some care must be taken to prevent high-frequency parasitic oscillation of Q1 during its intended "off " state.

This is achieved by making the "off " state of Q1 the low-impedance "on" state of Q2, and by making the leakage inductance between S1 and S2 small. Consequently, any tendency for feedback from P1 to S1 will be clamped by the drive transistor Q2, D1, and S2, which will not allow the start of any winding to go positive.

To prevent Q2 from turning off when it should be on during the power-down phase (leading to loss of control during input power-down), the auxiliary supply to the drive circuit must be maintained during the system power-down phase. (Large capacitors may be required on the auxiliary supply lines.)


For the first part of the "on" period of the driver transistor Q2, D1 and S2 will be con ducting. However, when Q1 has turned off and the recovery current in the base-emitter junction of Q1 has fallen to zero, S2 and hence D1 will become reversed-biased as a result of the voltage applied to winding P2 via R1. The start of all windings will now go negative, and current will build up in winding P2, resetting the core back toward negative saturation.

At saturation, the current in P2 and Q2 is limited only by resistor R1, the voltage on all windings is zero, and the circuit has been reset, ready for the next "on" cycle.

The need for minimum leakage inductance between S1 and S2 tends to be incompatible with the need for primary-to-secondary isolation and creepage distance. Hence if T1 is used to provide such primary-to-secondary circuit isolation in direct-off-line applications, the transformer may need to be considerably larger than the power needs alone would dictate.


Where the range of input voltage and load are very wide, the circuit shown in Fig. 1 will have some limitations, as follows.

When the input voltage is low, the duty cycle will be large, and Q1 may be "on" for periods considerably exceeding 50% of the total period. Further, if the minimum load is small, L1 will be large to maintain continuous conduction in the output filter. Under these conditions, the collector current is small, but the "on" period is long.

During the long "on" period, a magnetizing current builds up in the drive transformer T1 as a result of the constant base drive voltage Vbe of Q1 that appears across winding S1. Since the drive transformer is a current transformer during this period, the magnetizing current is subtracted from the output current. Hence, the intended proportional drive ratio is not maintained throughout the long "on" period (the drive falls toward the end of the period). To minimize this effect, a large inductance is required in the drive transformer T1.

However, at the end of the "on" period, Q2 must reset the drive transformer core during the short "off " period that now remains. To allow a quick reset, the volts per turn on P2 must be large. This requires either a small number of turns on P2 (with a large reset current) or a large auxiliary voltage. In either case, the power loss on R1 will be relatively large.

Hence, a compromise must be made in inductance turns and auxiliary voltage that is difficult to optimize for wide-range control at high frequencies. This conflict can be solved by the circuit shown in Fig. 2.

In the circuit shown in Fig. 2, capacitor C1 charges rapidly when Q2 is off via R1 and Q3. Q3 will be turned on hard by the base drive loop P2, D2, R2 (the starts of all windings being positive when Q2 is "off " and Q1 "on").

FIG. 2 Single-ended forward converter with push-pull proportional base drive circuits.


When Q2 is turned on, the voltage across P2 is reversed, and the transferred current from S1 and P1 flows in the low-impedance loop provided by C1, P2, and Q2. The voltage on all windings is reversed rapidly, turning off Q1. At the same time, Q3 is turned off, so that as the core is reset and C1 discharges, only a small current is taken from the supply via R1, which is now much higher resistance than the similar resistor shown in Fig. 1.

If Q2 is "on" for a long period and C1 is fully discharged, a flywheel action will be provided by D1, preventing reversal of voltage on P2 by more than a diode drop. The turns ratio is such that Q1 will not be turned on under these conditions. Finally the core will return to a reset point defined by the current in R1.


When Q2 is turned off, the starts of all windings will go positive by flyback action, and Q1 will be turned on. Regenerative drive from P1 and S1 maintains the drive, holding Q1 and Q3 on and rapidly recharging C1. This action is maintained until Q2 is turned on again to complete the cycle. The advantage of this arrangement is that the core can be reset rapidly by using a high auxiliary supply voltage without excessive dissipation in R1 and Q2.

Hence, in this circuit the conflict between transformer inductance and reset requirements is much reduced; however, the inductance will be made only just large enough to limit the magnetizing current to acceptable limits. Sufficient drive must be available to ensure correct switching action under all conditions. If the magnetizing current component in the drive transformer is allowed to exceed the collector current, then positive feedback action will be lost.


If Q1 is a high-voltage transistor, it is probable that some shaping of the base drive current will be required for reliable and efficient operation.

FIG. 3 shows a suitable modification to the drive circuit in Fig. 2 for high voltage transistors; base drive shaping has been provided by R4, D3, C2, R3, and Lb.

FIG. 3 Push-pull-type proportional drive circuit with special drive current shaping for high-voltage transistors.


1. What are the major advantages of proportional drive?

2. Why does the drive transformer in a proportional drive circuit tend to be larger than the power requirements alone would indicate?

3. The maximum duty ratio for a transformer-coupled proportional drive circuit tends to be limited to less than 80%. Why is this?

4. What controls the minimum and maximum inductance of the proportional drive transformer?

Also see: Our other Switching Power Supply Guide

Top of Page

PREV.   NEXT   Guide Index HOME