AMAZON multi-meters discounts AMAZON oscilloscope discounts
|
|
1. INTRODUCTION
Figure 1.1 shows the basic circuit of a triple-output flyback power supply.
The flyback unit combines the actions of an isolating transformer, an output inductor, and a flywheel diode with a single transformer. As a result of this magnetic integration, the circuit provides extremely cost-effective and efficient stabilized DC outputs.
The technique is particularly useful for multiple-output applications, where several semi-stabilized outputs are required from a single supply. The major disadvantage is that high ripple currents flow in transformer and output components, reducing their efficiency.
As a result of this limitation, the flyback converter is usually restricted to power levels below 150 W. The designer should note that it would be normal practice to include load line shaping components, (snubbers), to Q1 to maintain safe area operation.
2. EXPECTED PERFORMANCE
In the example shown in Fig. 1.1, the main output is closed-loop-controlled and is thus fully regulated. The auxiliary outputs are only semiregulated and may be expected to pro vide line and load regulation of the order of ±6%. Where better regulation is required, additional secondary regulators will be needed.
In flyback supplies, secondary regulators are often linear dissipative types, although switching regulators may be used for higher efficiency. For low-current outputs, the standard three-terminal IC regulators are particularly useful. The dissipation in the linear regulators is minimized as a result of the preregulation provided by the closed-loop control of the main output. In some applications, the closed-loop control regulation may be shared between two or more outputs.
Since the most cost-effective flyback converters will not have additional secondary regulators, overspecifying the requirements is a mistake. The essential attractions of this type of converter-simplicity and low cost-will be lost if additional circuitry is required to meet very critical specifications. For such applications, the designer might do well to consider one of the more sophisticated multiple-output topologies with their inherently higher performance.
FIG. 1.1 Power rectifier and converter section of a typical triple-output,
direct-off-line flyback (buck-boost) switchmode power supply.
2.1 Output Ripple and Noise
Where very low levels of output ripple are required, the addition of a small LC noise filter near the output terminals will often eliminate the need for expensive low-ESR capacitors in the main secondary reservoir positions.
For example, a typical 5-V 10-A supply may use the highest-quality low-ESR capacitors in positions C1, C2, and C3 of the single-stage filter shown in Fig. 1.1, but this will rarely give a ripple figure of less than 100 mV. However, it is relatively easy to keep ripple figures below 30 mV when low-cost standard electrolytic capacitors are used in positions C1, C2, and C3 by adding a high-frequency LC output filter. This approach can be very efficient and cost-effective. (See Part 1, Section. 20.) It should be understood that in a flyback converter the filter inductor can be quite small, since it is not required for energy storage (as it would be in a forward converter).
2.2 Synchronization
In fixed-frequency flyback units, some means of synchronizing the switching frequency to an external clock is often provided. This synchronization can lead to fewer interference problems in some applications.
3. OPERATING MODES
Two modes of operation are clearly identifiable in the flyback converter:
1. "Complete energy transfer" (discontinuous mode), in which all the energy that was stored in the transformer during an energy storage period ("on" period) is transferred to the output during the flyback period ("off" period).
2. "Incomplete energy transfer" (continuous mode), in which a part of the energy stored in the transformer at the end of an "on" period remains in the transformer at the beginning of the next "on" period.
3.1 Transfer Function
The small-signal transfer functions for these two operating modes are quite different, and they are dealt with separately in this section. In practice, when a wide range of input volt ages, output voltages, and load currents is required, the flyback converter will be required to operate (and be stable) in both complete and incomplete energy transfer modes, since both modes will be encountered at some point in the operating range.
As a result of the change in transfer function at the point where there is a move from one mode to the other, together with the merging into one component of the transformer, output inductor, and flywheel diode actions, flyback converters can be among the most difficult to design.
3.2 Current-Mode Control
The introduction of current-mode control to the pulse-width modulation action makes stabilizing the control-loop much easier, particularly for the complete energy transfer mode.
Hence current-mode control is recommended for flyback systems. However, current-mode control does not eliminate the stability problems inherent in the incomplete energy transfer mode, because of the "right-half-plane zero" in the transfer function. This will require the gain of the control loop to roll off at a low frequency, degrading the transient response.
4. OPERATING PRINCIPLES
Consider Fig. 1.1. In this circuit, the high-voltage-rectified 300-V DC line is switched across the primary winding P1 of a transformer, using a single switching device Q1. The control circuit has a fixed frequency, and the duty ratio of Q1 is adjusted to maintain the output voltage constant on the main output line. It will be shown that the unit may operate in a complete or incomplete energy transfer mode, depending on the duty ratio and load.
5 ENERGY STORAGE PHASE
The energy storage phase is best understood by considering the action of the basic single output flyback converter shown in Fig. 1.2.
FIG. 1.2 Simplified power section of a flyback (buck-boost) converter.
When transistor Q1 is turned on, the start of all windings on the transformer will go positive. The output rectifier diode D1 will be reverse-biased and will not conduct; there fore current will not flow in the secondary while Q1 is conducting.
During this energy storage phase only the primary winding is active, and the transformer may be treated as a simple series inductor; hence the circuit can be further simplified to that shown in Fig. 1.3a.
From Fig. 1.3a it is clear that when Q1 turns on, the primary current Ip will increase at a rate specified by
di_ p /dt=Vcc/ Lp
Where VCc =supply voltage
Lp = primary inductance
This equation shows that there will be a linear increase of primary current during the time Q1 is conducting, (ton). During this period the flux density in the core will increase from the residual value Br to its peak working value Bw. The corresponding current waveform and flux density changes are shown in Fig. 1.3b.
FIG. 1.3 (a) Equivalent primary circuit during the energy storage phase.
(b) Primary current waveform and magnetization during the energy storage phase.
6 ENERGY TRANSFER MODES (FLYBACK PHASE)
When Q1 turns off, the primary current must drop to zero. The transformer ampere-turns cannot change without a corresponding change in the flux density delta B. As the change in the flux density is now negative-going, the voltages will reverse on all windings (flyback action). The secondary rectifier diode D1 will conduct, and the magnetizing current will now transfer to the secondary. It will continue to flow from start to finish in the secondary winding. Hence, the secondary (flyback) current flows in the same direction in the windings as the original primary current, but has a magnitude defined by the turns ratio.
(The ampere-turns product remains constant.) Under steady-state conditions, the secondary induced emf (flyback voltage) must have a value in excess of the voltage on C1 (the output voltage) before diode D1 can conduct. At this time the flyback current will flow in the secondary winding starting at a maximum value Is, where Is = nIp. (n is the transformer turns ratio and Ip is the peak primary current at the instant of turn-off of Q1.) The flyback current will fall toward zero during the flyback period. Since during the flyback period Q1 is "off " and the primary is no longer conducting, the primary winding can now be neglected, and the circuit simplifies to that shown in Fig. 1.4a. The flyback secondary current waveform is shown in Fig. 1.4b.
FIG. 1.4 (a) Equivalent secondary circuit during the energy transfer phase (flyback period). (b) Secondary current waveform and magnetization during the flyback period.
For complete energy transfer conditions, the flyback period is always less than the "off " period, and the flux density in the core will fall from its peak value Bw to its residual value Br during the flyback period. The secondary current will also decay at a rate specified by the secondary voltage and secondary inductance; hence (neglecting the diode drop)
Di/dt=Vs/L s
Where:
Vs=supply voltage
Ls=primary inductance
7. FACTORS DEFINING OPERATING MODES
7.1 Complete Energy Transfer
If the flyback current reaches zero before the next "on" period of Q1, as shown in Fig. 1.5a, the system is operating in a complete energy transfer mode. That is, all the energy that was stored in the transformer primary inductance during the "on" period will have been transferred to the output circuit during the flyback period, before the next storage period starts.
If the flyback current does not reach zero before the next "on" period (Fig. 1.5b), then the system is operating in the incomplete energy transfer mode.
FIG. 1.5 (a) Primary current waveform Ip and secondary current waveforms
Is (discontinuous-mode operation). (b) Primary and secondary waveforms at the
transition from the discontinuous mode to the continuous or incomplete energy
transfer mode.
7.2 Incomplete Energy Transfer
If, in the circuit example shown in Fig. 1.2, the "on" period is increased and the "off " period correspondingly decreased, more energy is stored in the transformer during the "on" period.
For steady-state operation, this extra energy must be extracted during the "off " period. If the input and output voltages are to be maintained constant, it will be shown that the load current must be increased to remove the extra energy.
The slope of the input and output current characteristics cannot change, because the primary and secondary voltages and inductances are constant. Further, the equality of the forward and reverse volt-seconds applied to the transformer must be maintained for steady state conditions. Hence, for the increased "on" period, a new working condition will be established, as shown in Fig. 1.5b.
If the load is increased beyond this condition, the current will not be zero at the beginning of an "on" period, and an equal value will remain at the end of the "off " period (with due allowance for the turns ratio). This is known as continuous-mode operation or incomplete energy transfer, since a portion of the energy remains in the magnetic field at the end of a flyback period. Since the area under the secondary-current waveform is now greater by the DC component, the load current must be greater to maintain steady-state conditions.
Note: The behavior of the overall system should not be confused by the term "incomplete energy transfer," since, under steady-state conditions, all the energy input to the transformer during the "on" period will be transferred to the output during the flyback period.
In this example, a transition from complete to incomplete energy transfer was caused by increasing the "on" period. However, the following equation shows that the mode of operation is in fact controlled by four factors: input and output voltage, the mark space ratio, and the turns ratio of the transformer.
As previously mentioned, under steady-state conditions, the change in flux density during the "on" period must equal the return change in flux density during the flyback period. Hence:
delta B=Vt/N=Vs toff/Ns
It will be seen from this equation that the primary volt-seconds per turn must be equal to the secondary volt-seconds per turn if a stable working point for the flux density is to be established.
In the forward direction, the "on" period can be adjusted by the control circuit to define the peak primary current. However, during the flyback period, the output voltage and secondary turns are constant, and the active flyback period must self-adjust until a new stable working point for the transformer flux density is established. It can continue to do this until the flyback period extends to meet the beginning of the next "on" period (Fig. 1.5b).
At the critical point where the flyback current has just reached zero before the next "on" period, any further increase in duty ratio or load will result in the unit moving from the complete to the incomplete energy transfer mode. At this point, no further increase in pulse width is required to transfer more current, and the output impedance becomes very low. Hence, the transfer function of the converter changes to a low-impedance two-pole system.
8. TRANSFER FUNCTION ANOMALY
The flyback converter operating open loop in the complete energy transfer mode (discontinuous mode) has a simple single-pole transfer function and a high output impedance at the transformer secondary. (To transfer more power requires an increase in pulse width.)
When this system reverts to the incomplete energy transfer mode (continuous-mode operation), the transfer function is changed to a two-pole system with a low output impedance (the pulse width is only slightly increased when more power is demanded). Further, there is a "right-half-plane zero" in the transfer function, which will intro duce an extra 180° of phase shift at high frequency; this can cause instability. The loop stability must be checked for both modes of operation if it is possible for both modes to occur in normal use. To determine the need for this, consider light loading, normal loading, and short-circuit conditions. In many cases, although complete energy transfer may have been the design intention, incomplete transfer may occur under overload or short-circuit conditions at low input voltages, leading to instability.
9. TRANSFORMER THROUGHPUT CAPABILITY
It is sometimes assumed that a transformer operating in the complete energy transfer mode has greater transmissible power than the same transformer operating in incomplete transfer mode. (It sounds as if it should.) However, this is true only if the core gap remains unchanged.
Figure 1.6a and b shows how, by using a larger air gap, the same transformer may be made to transfer more power in the incomplete transfer mode than it did previously in the complete transfer mode (even with a smaller flux excursion). In applications in which the transformer is "core loss limited" (usually above 60 kHz for typical ferrite transformers), considerably more power may be transmitted in the incomplete energy transfer mode, because the reduced flux excursion results in lower core losses and reduced ripple currents in both primary and secondary.
Figure 1.6a shows the B/H curve for a core with a small air gap and a large flux density change. Figure 2.1.6b shows the B/H curve for the same core with a larger air gap and a smaller flux density change.
In general, the power available for transfer is given by:
P=fVc integral H dB
Where f = frequency
Ve=effective volume of core amd air gap
This power is proportional to the shaded area to the left of the B/H curve in Fig. 1.6; it is clearly larger for the example in Fig. 1.6b (the incomplete energy transfer case). Much of the extra energy is stored in the air gap; consequently, the size of the air gap will have a considerable effect upon the transmissible power. Because of the very high reluctance of the air gap, it is quite usual to have more energy stored in the gap than in the transformer core itself.
At the end of the "on" period, energy of ½ Lp I2^2 (I2 is the current corresponding to magnetization H2 in the figure) will have been stored in the transformer magnetic field. This energy, less the energy remaining in the core is ½ Lp I1^2 transferred to the output circuit each cycle.
In conclusion, the designer must choose the mode of operation depending on the performance required and the power to be transferred, be aware of the need to check the mode of operation under all possible loading conditions, and be prepared to design the control loop to deal with all realistic conditions.
FIG. 1.6 (a) Magnetization loop and energy transferred in a flyback converter
transformer when the core air gap is small (high-permeability magnetic path).
(b) Magnetization loop and energy transferred in a flyback converter transformer
when a large air gap is used in the core.
10. SPECIFICATION NOTES
The designer should be alert to the tendency for specifications to escalate. When a flyback converter is to be considered and potential requirements are large, costs are often particularly sensitive. The designer should establish with the customer the real limitations of the application. It may well be that a typical performance of 6% regulation on the auxiliary outputs of a multiple-output unit would be acceptable. This allows a semiregulated flyback system to be used. To guarantee a result of 5% (hardly better), a secondary regulator would be required, with consequent loss of efficiency and increased cost.
Very often specifications call for fixed frequency, or even a synchronized operating condition. This synchronization is often specified when a power supply is to be used for video display terminals or computer applications. Very often, in specifying such requirements, the user is making an assumption that the switching noise or magnetic field generated by the power supply will in some way interfere with the system performance. However, in a well-designed, well-filtered, and well-screened modern switching supply, the noise level is unlikely to be sufficiently high to cause interference. Moreover, in many cases, synchronization makes the noise even more noticeable.
In any event, synchronization is a poor substitute for eliminating the noise problem altogether.
If the specification calls for a fixed frequency or synchronization, the designer would do well to check this requirement with the user. Have available for demonstration a well screened variable-frequency unit. This should have a copper screen on the transformer and a second-stage output LC filter. If possible, try the sample in the actual application. The author has found that the user is often well satisfied with the result, and of course the cost of the supply in that case would be much lower.
In some applications in which a number of switching supplies are to be operated from the same input supply (more usual with DC-to-DC converters), the input filter requirements can be reduced by using synchronized and phase-shifted clock systems. This approach also eliminates low-frequency intermodulation components, and in this application the extra cost of a synchronized unit may well be justified.
Having fully researched the application, the designer is in a position to confidently select the most effective approach to meet the final specification requirements.
11. SPECIFICATION EXAMPLE FOR A 110-W DIRECT-OFF-LINE FLYBACK POWER SUPPLY
For the following example, a fixed-frequency single-ended bipolar flyback unit with three outputs and a power of 110 W is to be considered. It will be shown later that the same design approach is applicable to variable-frequency self-oscillating units.
Although most classical design approaches assume that the mode of operation will be either entirely complete energy transfer (discontinuous mode) or entirely incomplete energy transfer (continuous mode), in practice a system is unlikely to remain in either of these two modes for the complete range of operation. Consequently, in the simplified design approach used here, it will be assumed that both modes of operation will exist at some point within the working range. This approach also tends to yield higher efficiency, as the peak primary and secondary currents are reduced.
11.1 Specification
Output power:110 W
Input voltage range: 90-137/180-250 (user selectable)
Operating frequency: 30 kHz
Output voltages: 5 V, 10 A 12 V, 3 A -12 V, 2 A
Line and load regulation: 1% for main 5-V output 6% typical for a 40% load change (from 60% of nominal)
Output current range: 20% to full load
Output ripple and noise: 1% maximum
Output voltage centering: ± 1% on 5-V lines ± 3% on 12-V lines
Overload protection: By primary power limit and shutdown requiring power on/off reset cycle
Overvoltage protection: 5-V line only by converter shutdown, i.e., crowbar not required
11.2 Power Circuit
The above specification requirements can be met using a single-ended flyback system with out secondary regulators (see Fig. 1.1). To meet the need for dual input voltage by a link change, voltage doubling techniques can be employed for the input line rectifiers when they are set for 110-V operation. Consequently, the rectified DC line will be approximately 300 V for either 110-V or 220-V nominal inputs.
A voltage analogue of the primary current for primary power limiting is available across the emitter resistor R1. This waveform may also be used for control purposes when current mode control is to be used. A separate overvoltage protection circuit monitors the 5-V output and shuts the converter down in the event of a failure in the main control loop.
To meet the requirements for low output ripple, a two-stage LC filter will be fitted in this example. This type of filter will allow standard medium-grade electrolytic output capacitors to be used, giving a lower component cost. (A suitable filter is shown in Part 1.) The control circuit is assumed to be closed to the 5-V output to give the best regulation on this line. The details of the drive circuitry have been omitted; suitable systems will be found in Part 1, Sections. 15 and 16.
11.3 Transformer Design
The design of the transformer for this power supply is shown in Part 2, Section. 2.
12. QUIZ
From what family of converters is the flyback converter derived?
2. During which phase of operation is the energy transferred to the secondary in a flyback converter?
3. Describe the major advantages of the flyback technique.
4. Describe the major disadvantages of the flyback technique.
5. Why is the transformer utilization factor of a flyback converter often much lower than that of a push-pull system?
6. Under what operating conditions will the flyback converter give a core utilization factor similar to that of a forward converter?
7. Why is an output inductor not required in the flyback system?
8. Describe the two major modes of operation in the flyback converter.
9. What are the major differences in the transfer functions between continuous-and discontinuous-mode operation?
10. Why would an air gap normally be required in the core of a flyback transformer when ferrite core material is used?
11. Why is primary power limiting alone usually inadequate for full short-circuit protection of a flyback converter?
Also see: Our other Switching Power Supply Guide