Functions / Requirements of Direct-Off-Line SMPS -- CENTERING OF AUXILIARY OUTPUT VOLTAGES ON MULTIPLE-OUTPUT CONVERTERS

Home | Articles | Forum | Glossary | Books

AMAZON multi-meters discounts AMAZON oscilloscope discounts



1 INTRODUCTION

When more than one winding is used on a converter transformer to provide auxiliary outputs, a problem can sometimes arise in obtaining the correct output voltages.

Because the transformer turns can only be adjusted in increments of one turn (or in some cases a half turn) it may not be possible to get exact voltages on all outputs.

When output auxiliary regulators (often three-terminal series regulators) are to be used, the secondary output voltage error is generally not a problem. However, in many cases additional regulation is not provided, and it is desirable to "center" the output voltage (set it to an absolute value).

The following method describes a way of achieving this voltage adjustment in a loss free manner, using small saturable reactors.

2 EXAMPLE

Consider the triple-output forward-converter secondary circuit shown in FIG. 1. Assume that the 5-V output is a closed-loop regulated output, fully stabilized and adjusted.

There are two auxiliary 12-V outputs, positive and negative, which are now semiregulated as a result of the closed-loop control on the 5-V line. Assume that the regulation performance required from the 12-V outputs is such that additional series regulators would not normally be required (say, ±6%).

Further assume that to obtain 12 V out, the transformer in this example requires 11.5 turns and the half turn is not possible for flux balancing reasons. If 12 turns are used on the transformer, the output voltage on the 12-V lines will be high by approximately 0.7 V. (Remember, this output is obtained with a predefined pulse width which was set by the main control loop for the 5-V output.) Assume also that under these conditions, the pulse width is 15 Ms on and 18 Ms off, giving a total period of 33 Ms.

It is not possible to reduce the overall pulse width to obtain the correct output on the 12-V lines, as this will also reduce the 5-V output. If, on the other hand, the pulse width to the 12-V outputs could be reduced without changing the pulse width to the 5-V line, then it would be possible to produce the required output voltage on all lines. It is possible to achieve this with a saturable reactor.


FIG. 1 Saturating-core "centering inductors" applied to a Multiple-output push-pull converter.

3 SATURABLE REACTOR VOLTAGE ADJUSTMENT

Consider the effect of placing a saturable reactor toroid on the output lines from the transformer to the 12-V rectifiers D1 and D2.

These reactors L1 and L2 are selected and designed so that they take a time-delay period td to saturate, specified by…

The extra time delay td is introduced on the leading edge of the output power pulse by the saturable reactor, and the 12.7-V output would be adjusted back to 12 V.

It remains only to design the reactors to obtain the above conditions.

4 REACTOR DESIGN

Step 1, Selection of Material

From FIG. 1, it is clear that the cores will be set to saturation during the forward conduction of the output diodes D1 and D2, and to provide the same delay on the leading edge for the next "on" period, the cores must reset during the "off " period. When D1 and D2 are not conducting the "flywheel diodes" D3 and D4 are normally conducting. If a square loop material with a low remanence is chosen, the cores will often self-reset, the recovered charge of D1 and D2 being sufficient to provide the reset action. However, reset resistors R1 and R2 may be required in some applications.

A number of small square-loop ferrite toroids meet these requirements, and the TDK H5B2 material in a toroidal form is chosen for this example.

Step 2, Obtaining the Correct Delay Time

Prior to saturation, the wound toroid will conduct only magnetization current and, there fore, will be considered in its "off " state.

The time taken for the core to saturate when the "on" period starts (diodes forward biased) will depend on the applied voltage, the number of turns, the required flux density excursion, and the area of the core, as defined by the following equation: where td _ required time delay, Ms

Np _ turns

$B _ change in flux density from Br

to Bsat , T

Br _ flux remanence at H _ 0

Bs _ flux density at saturation, T

Ae _ effective area of core, mm2

Vs _ secondary voltage, V

In this example, the secondary voltage Vs applied to the core at the start of the "on" period may be calculated from the duty ratio and the output voltage as follows:

There are now two variables available for final voltage adjustments: turns and core area.

Assume, for convenience, that a single turn is to be used; that is, the output wire from the transformer is simply passed through the toroid. There is now only one variable, the core area, and the required core cross-sectional area may be calculated as follows:

This is a relatively large core, and for economy in low-current applications more turns may be used. For example, 5 turns and a core of 1b5 of the previous area will give the same delay.

The area would now be Ae _ 11.4 mm^2, and a TDK T7-14-3.5 or similar toroid would be suitable.

It may be necessary to fit resistors (R1, R2) across the rectifier diodes D1 and D2 to allow full restoration of the cores during the "off " period, as the leakage current and recovered charge from D1 and D2 may not be sufficient to guarantee full recovery of the cores during the nonconducting (reverse-voltage) period.

Note: This method of voltage adjustment will hold only for loads exceeding the magnetizing current of the saturable reactor; hence the voltage tends to rise at light loads. Where control is required to a very low current, it is better to use a small, high-permeability core with more turns, because the inductance increases as N2…while the delay is proportional to N (giving lower magnetization current and control at lower currents).

A further advantage of the saturable reactor used in this way is that it reduces the rectifier diode reverse recovery current, an important advantage in high-frequency forward and continuous-mode flyback converters.

5 QUIZ

1. What is meant by the term "centering" as applied to multiple-output converters?

2. Why is centering sometimes required in multiple-output applications?

3. Describe a method of non-dissipative voltage centering commonly used in duty ratio controlled converters.

4. Explain how saturable reactors L1 and L2 in FIG. 1 reduce the output voltages of the 12-V outputs.

5. Assume that the single-ended forward converter shown in FIG. 1 gives the required 5-V output when the duty ratio is 40% at a frequency of 25 kHz. The 5V secondary has 3 turns, the 12-V secondaries have 9 turns each, and the rectifier drop is 0.7 V. If L1 and L2 have 3 turns on a T8-16-4 H5B2 toroid core, calculate the output voltage with and without L1 and L2). Is there a better turns selection for 12 V?

Also see: Our other Switching Power Supply Guide

Top of Page

PREV. | Next | Guide Index | HOME