Push-Pull Converter

The push-pull converter is an isolated DC-DC converter that uses two switching transistors and a center-tapped transformer to achieve bidirectional flux excitation of the transformer core. Unlike the single-switch forward converter which uses only half the B-H curve, the push-pull converter drives the core in both directions, achieving full utilization of the transformer core and nearly doubling the power transfer capability for a given transformer size.

Push-pull converters are used in medium power applications from 100 W to several kilowatts, commonly found in DC-DC converters for automotive systems, telecom base stations, military power supplies, and industrial equipment where both isolation and high power density are required.

Figure 1: Push-pull converter showing center-tapped transformer, alternating switches Q1 and Q2, and output LC filter

Core Concept Explanation

The push-pull converter uses two transistors Q1 and Q2 that conduct alternately, connected to opposite ends of a center-tapped primary winding. When Q1 conducts, current flows through the upper half of the primary winding and the transformer core flux increases in one direction. When Q2 conducts, current flows through the lower half in the opposite direction, reversing the core flux. This alternating flux excitation uses both positive and negative portions of the B-H curve.

The center-tapped secondary winding produces a full-wave rectified output using diodes D1 and D2. When Q1 is ON, D1 conducts supplying energy to the load through output inductor Lf. When Q2 is ON, D2 conducts. During the dead time when both switches are OFF, both diodes conduct the freewheeling inductor current. The output ripple frequency is twice the switching frequency because two pulses are delivered per switching cycle.

The key circuit characteristic is that each switching transistor is subject to a voltage stress of 2 x Vin because when Q1 is ON, the full primary voltage Vin appears across Q1, and simultaneously the reflected voltage from Q2's side adds another Vin, placing 2Vin across Q2 (which is OFF). This double voltage stress requires higher voltage rated devices compared to bridge topologies.

Transformer Core Utilization

Since the push-pull converter drives flux in both directions (bipolar excitation), the transformer core operates on the full B-H loop, from +Bmax to -Bmax. This means the core can handle twice the flux change per cycle compared to unipolar excitation (as in the forward converter). For the same core material and size, a push-pull transformer can transfer more power per cycle than a forward converter transformer.

However, flux imbalance is a critical concern in push-pull converters. If Q1 and Q2 have slightly different saturation voltages or switching times, the volt-second product applied in each half cycle becomes unequal. This causes the average flux to drift in one direction over multiple cycles, potentially saturating the core on one side. Flux balancing techniques include current mode control, capacitive series coupling, or active flux monitoring.

Mathematical Expression

The output voltage of the push-pull converter is derived from volt-second balance on the output inductor. Since two current pulses are delivered per switching period (one from Q1 ON, one from Q2 ON), the effective duty cycle for output voltage calculation is doubled:

Vout = 2 x Vin x (Ns/Np) x D, where D is the duty cycle of each individual switch (each switch has D less than 0.5). The factor 2 arises because each half cycle contributes one pulse, and both pulses charge the output capacitor.

The constraint D less than 0.5 per switch is essential to prevent both switches from being ON simultaneously (shoot-through), which would short circuit the center-tapped primary and cause catastrophic current. A dead time is always inserted between Q1 turn-off and Q2 turn-on.

Practical Understanding

Because the output inductor sees current pulses at twice the switching frequency, the ripple current in Lf and the required filter capacitance are smaller than equivalent single-switch topologies. This allows smaller, cheaper output filter components, which is a significant advantage in cost-sensitive designs.

Current mode control is almost universally used in push-pull converters because it naturally provides cycle-by-cycle current limiting and inherently helps balance the flux by equalizing peak currents in each half cycle. Voltage mode control alone cannot prevent flux imbalance without additional compensation.

The push-pull topology is not suitable for high input voltage applications because the switch voltage stress of 2Vin would require very high voltage MOSFETs with poor RDS(on) performance. For high input voltages above 200 V, half-bridge and full-bridge converters are preferred because switch stress equals only Vin.

Example

Given:
Push-pull converter
Vin = 24 V, Ns/Np = 1/2, D = 0.4 per switch
Switching frequency fs = 100 kHz

Why this formula applies:
Push-pull output: Vout = 2 x Vin x (Ns/Np) x D
Two pulses per period, output ripple at 2 x fs

Formula:
Vout = 2 x Vin x (Ns/Np) x D
Switch voltage stress = 2 x Vin
Output ripple frequency = 2 x fs

Substitution:
Vout = 2 x 24 x (1/2) x 0.4
Vout = 2 x 24 x 0.5 x 0.4

Calculation:
Vout = 2 x 24 x 0.2 = 9.6 V
Switch stress = 2 x 24 = 48 V
Output ripple frequency = 2 x 100k = 200 kHz

Final Answer:
Vout = 9.6 V, Switch stress = 48 V, Ripple at 200 kHz

Exam Tip: Push-pull output voltage formula has a factor of 2: Vout = 2 x Vin x (Ns/Np) x D. Common GATE trap is forgetting this factor 2 and using Vout = Vin x (Ns/Np) x D as in the single-switch forward converter. Also remember: each switch D must be less than 0.5, and output ripple frequency = 2 x switching frequency due to two pulses per cycle.