Mains voltage normalizer. Construction principles.

A.P. Krivetsky
The article describes a general approach to building an inexpensive voltage normalizer
an alternating current electrical network based on an autotransformer with 4 taps,
switched by triac keys. A method of decreasing the switching step using phase voltage boost is considered,
without changing the structure of the device.

Mains voltage, especially in weak suburban networks, is subject to significant fluctuations far beyond the standard tolerance of ± 10%. Fortunately, the dielectric strength of many household appliances allows them to withstand even voltages outside the 20 percent tolerance. Not all devices have power supplies that allow a wide range of supply voltages, so their normal operation is not guaranteed. In addition, overvoltage or undervoltage has a detrimental effect on their durability.

This problem is not new and exists in different countries. Not for nothing, for example, in Germany, for provincial areas, electronic and electromechanical devices are produced that maintain the voltage in the network of a residential building within acceptable limits.

The most compact devices are based on autotransformers with stepwise switching of windings. Periodicals [1-4] described many variants of such devices for self-production. In all these devices, an attempt is clearly visible to increase the number of steps to reduce the switching step (the author has seen up to 12). This leads to an unnecessary complication of the described devices. In addition, they irrationally use an autotransformer, which, as a rule, has a fixed connection to the network, where the voltage varies over a wide range, and a switchable load connection, where the voltage changes within narrow limits [4]. But, most importantly, in many such electronic devices [4 ... 6], the triac switches are switched incorrectly, which causes fair complaints from users, and sometimes the inoperability of stabilizers in real conditions.

In the article, according to the author, a more rational approach to the design of such devices is proposed.

In this case, it is more expedient to set the task of not stabilizing, but normalizing the mains voltage, i.e. keeping it within the standard 10% tolerance. This will simplify and reduce the cost of the device.

In fig. 1 shows a block diagram of the mains voltage normalizer. It consists of an autotransformer T, triac switches VS1 - VS5, VSд, a current sensor, a voltage sensor and a control system.

The autotransformer is connected to the network using triac keys. The power of the autotransformer at a maximum transformation ratio of 1.2 will be 20% of the maximum load power without taking into account the efficiency, since an autotransformer, unlike a transformer, transforms only a fraction of the power. The VS5 switch completely turns off the autotransformer when the transformation ratio is equal to 1. This increases the efficiency, since any transformer has, albeit a small, no-load current, at which hundreds of kilowatts of electricity are additionally consumed per year. The current of this key will also be about 20% from the current of the keys VS1 - VS4, again without taking into account the efficiency. An additional VSd key (shown in color) allows you to completely turn off the normalizer and the load in emergency situations, leaving monitoring the state of the network for automatic turning on. The use of an additional key allows you to reduce the voltage requirements for the remaining keys, because they turn out to be connected in series with VSd.

Структурная схема нормализатора сети

    The proposed transfer characteristic of the normalizer is shown in Fig. 2. The load is permanently connected to one of the taps, the network is connected through one of the switches (VS1 - VS4), depending on the input voltage and the selected range (d1 - d4, respectively). This makes it possible to reduce the redundancy of the autotransformer, because the entire winding is rated for maximum voltage. At reverse connection, for example [4,5], when the network is permanently connected to one of the taps, and the load is switched, the winding before the network tap must be designed for the maximum possible voltage. And so that the autotransformer could also be a step-up one, it would be necessary to add more windings after this one.

This figure shows the characteristics of the normalizer in 4 switchable ranges d1 - d4. For each range, the corresponding graph shows the transformation ratio of the autotransformer k and the switching points from one range to another (U2 - U7). When the mains voltage goes beyond the normalization range, the normalizer disconnects the load and its power section. As can be seen from the graph, the characteristic has significant switching hysteresis, which are achieved by overlapping ranges. This avoids frequent switching with small voltage fluctuations at the edge of any range. For the same purpose, the lower and upper turn-off voltages U1, U8 and turn-on voltages U1 *, U8 * noticeably differ at the boundaries of the normalization range. Additionally, measures can be taken based on time delays of reclosing and band switching. The characteristic is constructed in such a way that a slight violation by the output voltage of the standard tolerance field is allowed at the boundaries of the normalization range. This is done in order to expand the working range.

Передаточная характеристика нормализатора напряжения сети

     With a sinusoidal voltage, the triac switch can be opened at any time, and it will close only when the current through it becomes zero (if strictly - less than the holding current). With a reactive nature of the load, the phases of the current and voltage do not coincide. This is shown in Fig. 3, (IWITH - current with a capacitive nature of the load, IL - with inductive).Сдвиг фазы тока в зависимости от характера нагрузки    It can be seen from the diagrams that when the voltage crosses zero, the current in the general case may not be equal to zero. The phase difference between current and voltage will depend on the magnitude of the reactivity. Therefore, an attempt to switch the keys on the transition of voltage through zero [4, 5, 6] can lead to a short circuit of the autotransformer sections. For this reason, keys with a built-in zero detector cannot be used in this device. The simplest case for the proposed circuit occurs in the absence of an external load, then the key load is the inductance of the autotransformer winding. To avoid an accident in the absence of a current sensor, time delays are introduced. A sufficient (for any load) delay in turning on the next switch will lead to discontinuities in the sinusoid or even to skip a period.
    The control system during switching must take into account not only the voltage, but also the current. The current sensor will allow you to accurately track the moment of the current crossing through zero and carry out continuous switching of ranges for any type of load. The voltage sensor will not only measure the value of the input voltage, but also track the voltage transitions through zero, which will allow you to turn on the device correctly, without a sharp surge, when there is no current (all keys are turned off).

The response of the normalizer to voltage changes should be fast enough to quickly determine the nature of the response. Therefore, it is desirable to measure the voltage in each period. Moreover, the detector should reflect the energy of each period, and not the amplitude. In the household network, the sinusoid is often distorted.

As can be seen from the diagram in Fig. 1, voltage control is carried out only at the input. With the correct design of the autotransformer, the input and output voltages are rigidly connected by the transformation ratio, therefore, knowing the input, you can accurately predict the output voltage.

With the selected step of voltage change (~ 22V), range switching will be noticeable visually by lighting devices. This effect is partially mitigated by the fact that switching will occur infrequently due to the large hysteresis and switching delays within the normalization range. As practice shows, it turns out quite acceptable.

Another advantage of the proposed scheme for constructing the normalizer is that if the lower wire in the scheme is used as a zero, then the integrity of the neutral is not violated.

When the mains voltage goes beyond the normalization range, all switches must be turned off and the load is de-energized. For the reliability of the performance of protective functions, the normalizer must withstand in this state a prolonged exposure to line voltage of 380V. For this, it is not at all necessary to use all switches with a high voltage class. In the closed state, series-connected keys will withstand a higher voltage. It is useful to note that keys of a lower voltage class have a lower cost.

Схема подключения вольтодобавки

This circuit has one more non-obvious possibility, which will allow to reduce the switching step or introduce smooth regulation to build a stabilizer with output voltage feedback without increasing the number of autotransformer taps [7]. This possibility is incorporated in the properties of the triac key. The triac can be closed without waiting for the current to drop to zero, but by briefly applying a reverse polarity voltage to the open triac. This, however, will also lead to an artificial change in the direction of the current and its transition through zero.

Figure 4 shows a simplified diagram showing the distribution of polarities in different parts of it with a positive half-wave. Let's say the VS2 key is open and current flows through it through the winding w2. In this case, a voltage is induced in the winding w1 with the polarity shown in the figure. If at this time the VS1 key is opened, then the winding voltage w1 will be applied to the VS2 key in the same polarity as it has been on it so far. As a result, the winding w1 will close through both open keys. Now let's say that the VS1 key is open at first, while the polarities of the voltages induced in the windings are distributed in the same way. If you then open the VS2 key, the winding voltage w1 will be applied to the VS1 key in reverse polarity. This closes the VS1 key while leaving VS2 open. The autotransformer will go into boost mode. The same will happen with a negative half-wave. With a transformation ratio of 1.1, the output voltage will have the form shown in Fig. 5

Форма выходного напряжения при фазовом подключении вольтодобавки

   If, with a sinusoidal voltage waveform, smoothly change the phase of connecting the discrete voltage boost, then the effective value of the output voltage will also change smoothly. In this example, when the voltage boost connection phase changes from 0 to 180 degrees, the effective voltage at the output will change to 10%. This method of regulation is called phase boost.

Figure 6 shows the dependence of the change in the normalized effective voltage value on the voltage boost connection phase. It can be seen from the figure, in order to obtain a uniform step of change in the output voltage in 2.5% with a width of the 10% range, a voltage boost must be connected in the phases 66 °, 90 ° and 114 °.

The disadvantage of this method is the distortion of the output voltage waveform and the nonlinearity of the control characteristic. However, these disadvantages are not significant.

Зависимость нормализованного действующего напряжения от фазы включения вольтодобавки

     Ferroresonant stabilizers give much higher distortion. And the converters for uninterruptible power supply, which have become widespread in recent years, generally form a meander. Some of them introduce zero pauses between pulses, which allow to bring the level of harmonics up to 35%. In the circuit under consideration, when the waveform has the form shown in Fig. 5 (turning on the 10% voltage boost at 90 °), the amplitude of the first harmonic is 94% of the total amplitude, the third is 3.15%, the 5th and 7th is 1.1%. , 9th and 11th - 0.63%, the rest are less.

   The nonlinearity of the control characteristic can be taken into account in software if the control system is built on a microcontroller.

   Thus, the proposed structure in the presence of a programmable control system makes it possible to build network voltage normalizers based on a simple 4-stage autotransformer. At the same time, the total cost of the structure, which is mainly determined by the power unit, will be noticeably less than traditional multistage ones, due to a decrease in the number of powerful keys and the volume of the autotransformer. The emergence of the possibility of multi-stage or even smooth regulation will allow the stabilization of the output voltage.


  1. Yashchenko O. AC voltage stabilizer. - Radio No. 1, 1981, p. 10-12.
  2. Kagan A. Electronic-relay voltage stabilizer. - Radio No. 8, 1991, pp. 34-36.
  3. Koltsov V. Stabilizer from a laboratory autotransformer. To help the radio amateur: Collection. Issue 64. - M .: DOSAAF, 1979, p. 52-59.
  4. Koryakov S. Mains voltage stabilizer with microcontroller control. - Radio No. 8, 2002, pp. 26-29.
  5. Godin A. AC voltage stabilizer. - Radio No. 8, 2005, p. 33-36.
  6. Ozolin M. Improved control unit of the AC voltage stabilizer. - Radio No. 7, 2006, p. 34-35.
  7. Veresov G.P. Power supply for household electronic equipment. - M .: Radio and communication, 1983, p. 66-67.

Total comments: 0

leave a comment

Your email will not be published.