Self-oscillating flyback converter - a type of a flyback converter (switch-mode power supply) which does not require a separate oscillator, but the switching of primary current is achieved by a built-in self-oscillatory action.¹⁾

The self-oscillation is achieved partially by a positive feedback from a tertiary winding of the coupled inductor (also referred to as a flyback transformer).

Self-oscillating flyback converters can be divided into three types:

1. Fixed on time, variable off time, which results with very low switching frequency at light load.
2. Fixed off time, variable on time, with low switching frequency with maximum load.
3. Variable on and off time, with variable switching frequency, but less than the two previous types.

The operating principle and diagrams given below are based on the description given in: Keith Billings, Switch Mode Power Supply Handbook²⁾

A self-oscillating flyback converter can be constructed with relatively few parts, without the need for dedicated integrated circuits, so that they can be an inexpensive alternative to other solutions.³⁾

Because of the self-oscillating mode the energy transfer is always complete (discontinuous mode). Also, current-mode control can be implemented, giving stable single-pole loop response. With good filter design and magnetic shielding of the transformer such converters can be used in applications demanding “quiet” converters.⁴⁾

An example of implementation is shown in Fig. 1. The windings L1, L2 and L3 are magnetically coupled (wound on the same magnetic core). After switching on, the capacitor C1 is charged and current will flow through R1 to turn on the transistor Q1. This action increases current in L1 and induces voltage in the feedback winding L3, which gives positive feedback and will accelerate turning on of Q1.

The base current of Q1 initially flows through C2 and then through D1 when the drive voltage is established.

After Q1 turns on, the primary current begins ramping up, as dictated by the value of primary voltage and inductance of primary winding L1 (as in normal flyback converter). This increasing current develops voltage across R4. When voltage on R4 reaches around 0.6 V it begins turning on Q2, which will divert majority of current away from the base of Q1, which as a consequence will begin to turn off. Reduction of primary current will cause voltage increase across L1 which will be passed by R5 to accelerate turning on Q2 and turning off Q1. Additionally, due to the flyback action negative voltage will be developed across L3 also helping to turn off Q1.

When all energy stored in the coupled inductor is transferred the voltages return to the initial state and the cycle repeats. With each new cycle the energy is transferred to the output whose voltage increases. When the output voltage exceeds the required level then the error amplifier A1 holds Q2 in “on” state, which stops Q1 from turning on and thus further energy transfer is blocked until the output voltage reduces sufficiently.

Fig. 1. Non-isolated self-oscillating flyback converter with primary current-mode control⁵⁾

^{S. Zurek, E-Magnetica.pl, CC-BY-4.0}

The circuit shown in Fig. 1 has inherent energy limiting capability, because the maximum current in the primary winding is limited to such a value, which develops around 0.6 V on R4. This will initiate the turn off action even without the feedback from the error amplifier on the secondary side. The same applies if a condition of magnetic saturation arises in the magnetic core, because it will result with sharp increase of the primary current.

The circuit can be also implemented as fully isolated - by using opto-coupler to pass the information from secondary to the primary side.⁶⁾

The circuit shown in Fig. 1 should run with at least 10% load, because otherwise the switching frequency might become very high. Dummy load by means of extra resistors might be required. Alternatively, at very light loads a rectangular signal modulation can be applied to the base of Q2, so that switching would be inhibited for some periods of time, thus resembling somewhat squegging operation.

• The component count is very low and thus the converter can be made inexpensively.
• Complete energy transfer (discontinuous mode).
• Self-limiting operation for over-current and magnetic saturation.

• The coupled inductor requires three windings, which increases its cost.
• The converter should be run with at least 10% of load.

• list of SMPS topologies