Applications Engineering Staff
0-02 Fair Lawn Avenue - Fair Lawn, NJ 07410
The addition of a base-to-emitter inductor in the output stage of a PNP-NPN high-current driver imposes current fall time by a factor of three, compared with the conventional circuit without the inductor.
You can improve current fall time of single-diffused power transistors three-fold by using a speed-up inductor in place of conventional circuits. For example, under conditions where conventional turn-off circuit design results in collector current fall times of the order of 6usec, the inductive turn-off method produces fall times of about 2usec.
Single-diffused power transistors are well known for their great ruggedness, low VBE, and low VCE saturation voltage. These characteristics make them particularly useful in high-power applications. Unfortunately, they are often thought of as slow transistors and for that reason have been excluded from switching circuits in the 10 to 20kHz range. It is usual practice to provide current fall times of 5 to 10usec for most single-diffused transistors.
An inductor shunting the emitter and base of a single-diffused output transistor provides a considerable improvement in current fall time. The inductive turn-off method was investigated with the 2N5928 transistor, rated at 120V VCEO and an hFE greater than 10 at 100A.
In the basic circuit (Figure 1) with the inductor, the collector of a PNP driver connects to the base of an NPN output transistor. When the driver turns on, inductor current is zero and builds up linearly to the desired level during the on-time. At the instant of turn-on, all the driver current is supplied to the base of the output transistor and, subsequently, the inductor shunts some of the driver current. At the instant of turn-off, the supply of current from the driver is interrupted. The inductor current must then be supplied by reverse base current flowing from the transistor. Here, the inductor behaves as a high-impedance current source.
If enough energy (1/2LI^2) has been stored in the inductor for proper turn-off, the voltage across it rises rapidly to the reverse breakdown voltage of the emitter-base diode and remains at this level during the collector fall time. Although these conditions appear to violate the emitter maximum reverse voltage rating, the stored inductor energy is too small to cause damage or degradation, so the emitter-base diode is safe. Current fall time increases somewhat if the inductor's stored energy reduces to a level where the emitter-base diode does not break down. However, in these cases, the inductive turn-off is still very effective.
Figure 2 shows the effect of inductor stored energy on fall time. Optimum stored energy as a function of collector is shown in Figure 3. In many practical cases, the optimum inductor stored energy is reached when the inductor current is 5 to 10 percent of the collector current. Thus, excellent fall times can be achieved by operating with forced hFE between 10 and 20 in the off direction. Typical fall times under conditions of optimum off drive at collector currents from 20 to 80A are shown in Figure 4.
The data in Figures 2, 3 and 4 were taken with a conventional 30uh air core RF choke. Correct coil design parameters depend upon specific circuit conditions. Obviously, you must consider the on-time, peak current to which the inductor is to be charged and the voltage across the inductor during the on-time. Depending on the level of collector current, the voltage developed between emitter and base (VBE) will be between 0.9 and 1.2V. The inductor series resistance must be sufficiently low so the desired peak current flows with that voltage across it. Typical parameters are of the order of 10-100uh. Excessive shunt capacitance will increase the inductor voltage rise time. Therefore, good RF practice should be used in winding the coil.
In those cases where the emitter-base diode voltage is too low or the on-time is too short to provide sufficient stored energy in the inductance, a small resistor may be connected in series with the base to increase the voltage across the inductor by a few hundred millivolts. Alternatively, the emitter side of the inductor may be returned to ground or to a tap in the emitter load for emitter output circuits.
Care must be taken in the design of preceding stages so that the driver is not inadvertently turned on during the fall time when a high voltage is developed across the inductor. A PNP driver or pre-driver works very well. Any other means of floating the drive to the output transistor is usually satisfactory.
In many applications, rise and fall times significantly affect circuit performance. The collector rise time of the 2N5928 is related to gain-bandwidth product and is essentially independent of collector current amplitude in the 20 to 80A range. With a forced hFE of 5, the 2N5928 rise time is typically 2usec; rise time is typically 4usec with a forced hFE of 10.
Driver stages designed for a stage current gain of 10 in saturation very often provide much higher stage current gains during the turn-on transient when the driver is out of saturation. Therefore, the instantaneous available current drive during turn-on is usually greater than the average current available during the total on period. As a result, actual circuit rise time is often much better than simplified circuit calculations suggest.
The ruggedness of single-diffused power transistors has for many years made them the preferred choice for reliable operation. Experience has shown that while normal circuit conditions often place minimal demands on power output devices, abnormal drive or load conditions frequently increase the demands to levels that can easily destroy output transistors with marginal second breakdown capability. The low VBE and collector saturation voltages of the 2N5928-type transistors permits efficient operation at currents as high as 100A, and the inductive turn-off method now increases efficient switching frequency operation to 20kHz.
Back To Products Page