Several "fatal weaknesses" of high-frequency UPS are worth discussing (Part 2)

Figure 9 Equivalent circuit of neutral-to-ground voltage

It is almost unimaginable that there is a 22V voltage drop on the neutral line. If there is such a large voltage drop on the neutral line, there must be a problem. Because on the neutral line bus within the range of the UPS cabinet, under normal circumstances, there will never be a voltage drop of more than 3V, and it is generally less than 1V. There is another situation: due to the poor characteristics of the low-pass filter at the output end of the UPS, some high-order harmonics flow into the load. In fact, this is okay. The built-in power input end of the load machine is connected to a filter. First, the high-order harmonics are intercepted. The second level is the rectifier filter to intercept, and the third level is the DC converter. These three gates can shut out or eliminate any high-order harmonics or even interference. Precisely because the internal power supply of the load machine has such a powerful function, it is really groundless to label the neutral-ground voltage as an "interference load".

That is to say, there is no path that can add the neutral voltage and interference to the load. Moreover, the neutral voltage is not a source of interference. Of course, spatial interference is another matter and does not belong to the scope of discussion here.

(IV) After the mains power is cut off, the system efficiency of high-frequency UPS decreases by 2% when the battery is discharged

Some parts are very specific, and it seems that field measurements have been done. Unfortunately, he took some high-frequency UPS as all, and this conclusion still has loopholes. The following are several situations.

1. Single-phase low-power UPS

FIG10 shows a schematic circuit diagram of a general low-power high-frequency UPS. One of the characteristics of a high-frequency UPS is that the output isolation transformer is eliminated. Therefore, the transformer that accounts for the majority of the weight of the machine can be eliminated because a half-bridge inverter is used. However, the operation of the half-bridge inverter requires two DC power supplies, and the two DC power supplies of a low-power high-frequency UPS, especially the use of two sets of batteries, are too cumbersome. Therefore, the Boost circuit technology is used. As shown in the figure, the energy storage inductor L, the electronic switch S, the isolation diode VD2, and the virtual power capacitors C1 and C2 constitute a boost electronic transformer. When the mains is used for power supply, the rectifier ZL1 and the charger charge the battery pack GB, and the rectifier ZL2 supplies power to the main circuit. Since the 220V AC can only provide a DC voltage of about 300V, the half-bridge inverter requires two DC voltages of at least 310V. Therefore, the Boost circuit creates two virtual DC power supplies of about 400V connected in series on capacitors C1 and C2.

Figure 10 Principle circuit diagram of general single-phase low-power high-frequency UPS

When the mains power is cut off, the battery pack GB discharges. Generally, when the capacity is below 10kVA or below 30kVA, the voltage of the battery pack GB is relatively low, such as 3 12V, 4 12V... or even 10 12V. In short, the voltage is far from the working level of the half-bridge inverter. Therefore, it must still be increased to two 400V by the Boost circuit. That is to say, although the mains power supply stops, the work here is not like the power frequency UPS which only works by the inverter, and the Boost circuit must continue to work. In this way, the high-frequency machine has one more working link than the power frequency machine, so it consumes more energy than the power frequency inverter, even if the efficiency is reduced by 2%.

However, some questioners are distracted by comparing the electronic circuit part and are happy to find the "weakness" (the so-called fatal weakness) of the high-frequency UPS, but they forget that the output isolation transformer of the industrial frequency UPS is also working, as shown in Figure 11 (a). The power consumed by this transformer is far from being dismissed by 2%. The author once measured the output transformers of 4 imported 100kVA UPS at full load and found that the surface temperature of the 100kVA transformer core reached 90C, which is definitely not a phenomenon caused by 2kW power. (I hope this is not a common phenomenon). In short, actual measurements have found that the system efficiency of small-power high-frequency UPS is still higher.

Figure 11 Comparison of UPS output circuits between industrial frequency and high frequency machines

2. Medium and high power situation

In the case of medium and high power, the low 2% is not as low as the questioner said. Generally, in medium and high power high frequency structure UPS, the virtual power supply is far from meeting the requirements of large current output. At this time, the capacitor can only be used to supplement the problem of excessive internal resistance of the battery and the inability to provide the front current when the load changes suddenly. The subsequent large current still needs to be provided by a large-capacity battery pack, as shown in Figure 12. Whether it is a high-frequency UPS with two DC power supplies as shown in Figure 12 (a) or a high-frequency UPS with only one DC power supply as shown in Figure 12 (b), almost all of them use at least 32 12V batteries in series or batteries with similar voltages in series. The rated voltage of these battery packs is much higher than the peak voltage of 310V of AC 220V. Therefore, after the city power is cut off, the charging link also stops working, and only the capacity of the battery itself is used to maintain the set backup time until the battery voltage drops to the inverter shutdown voltage level. At this time, the shutdown voltage level is generally 320332V, which is exactly the same as the work of the power frequency UPS inverter, so this 2% does not exist. What really exists is the output transformer of the power frequency UPS. This transformer occupies nearly two-thirds of the space of the power frequency UPS and more than 2% of the power consumption. If you must say "fatal", you should look for it in the power frequency UPS. In fact, some people are making a fuss. Although the power frequency UPS has high power consumption, it has been working well for so many years, and no one says it is a fatal problem. Why is it that today, UPS that is more energy-efficient than the power frequency model is called "fatal"? Even in public, it is not prudent to openly shout that the high-frequency UPS has so many "fatal weaknesses". I don't know why there is such a big prejudice against products that adapt to the current national policy of energy conservation and emission reduction, and meet the requirements of data centers such as small size, light weight, new technology and low price.

(c) Schematic diagram of the output of a full-bridge inverter of a UPS with a DC power supply in a power frequency structure

Figure 12 Schematic diagram of inverter output of high-frequency UPS and industrial frequency UPS

(V) The external transformer of high-frequency UPS will damage the load

1. Why is an external isolation transformer required?

The elimination of the output isolation transformer is a major feature of the high-frequency UPS, and also a major advantage, because it reduces the system power consumption, volume, weight and price. But some people insist on adding the removed transformer. Of course, some users here also have such requirements, but most of the users' requirements are misled by some manufacturers. It is said that it is to reduce the neutral-ground voltage. Despite this, some questioners are still not at ease, saying that "the neutral-ground voltage is still high and continues to endanger the safe operation of electrical equipment." Even if we add an additional transformer to the high-frequency UPS according to the meaning of a certain place, as shown in Figure 13 (a), let's see how this conclusion works. You can compare the two circuits in Figure 13 (a) and (b). Now the outputs of the two inverters are connected to the transformer. It can be seen that the working mode of the two inverters is pulse width modulation, and the modulation frequency is similar, which can also be said to be the same. Therefore, there is no difference in the work of the inverter power tube; in order to send a sine wave voltage to the load, a low-pass filter must be added to filter out the high-frequency components during modulation and only allow the 50Hz sine wave to pass. It can be seen from the figure that both of them have this filtering link, but the harmonic filter of the high-frequency UPS is before the transformer, while the harmonic filter of the power frequency UPS is after the transformer, which means that the working links of the two are not only the same, but also the same. The difference is the position of the filtering link and the transformer. In this way, it can be seen that in the high-frequency UPS, the high-order harmonics are filtered out before the transformer and return to the negative end of the DC BUS through the zero line, that is, the high-order harmonics of the high-frequency UPS do not enter the primary winding of the transformer at all. The high-order harmonics of the power frequency UPS are filtered out after the transformer, in other words, they are filtered out near the load end. This raises a problem: According to a certain gentleman: the zero-ground voltage formed by the high-order harmonics close to the load cannot be added to the load, and does not affect the work of the load; on the contrary, the zero-ground voltage formed by the high-order harmonics far from the load will definitely be added to the load, continuing to endanger the safe operation of the load. The same circuit principle gives two different results. I don't know whether this result was obtained through analysis or measurement. It doesn't seem to make sense theoretically.

Figure 13 Harmonic path diagram when both types of UPS have transformers

Some places say that adding a transformer to a high-frequency UPS will bring hidden dangers that may burn out the equipment. It is also said that once a high-frequency UPS "has an output power outage or flash failure for some reason", the external isolation transformer will produce a "flyback transient spike voltage", which is enough to burn out IT equipment. When the input power is suddenly restored, it will cause the parallel system to be "severely overloaded", and so on. What is puzzling is that the same power supply link and the same function, but the industrial frequency model is replaced by a high-frequency model, with only one word difference, the results of the two are different. Does it mean that the industrial frequency UPS will not have an output power outage or flash failure? Even if it does, its transformer will not produce a "flyback transient spike voltage"? When the input power is suddenly restored, the industrial frequency UPS will not cause the parallel system to be "severely overloaded! Is the destructive power of the external isolation transformer inherent in the high-frequency UPS? Then again, the external transformer of this high-frequency UPS was added somewhere (the supplier never intended to do so), and so many "potential" "hidden dangers" were analyzed after it was added. That is, he was right to add the transformer, and the problem was that you added it incorrectly, and he was right to go around in circles. There is no need for an external transformer for high-frequency UPS. First of all, as mentioned above, the neutral voltage is not an interference source, and there is no channel to transmit the neutral voltage. What affects the electrical equipment is the common touch interference, the common How does normal mode interference enter the electrical equipment? Figure 14 shows the schematic diagram of normal mode interference and common mode interference. If the interference voltage is to work, there must be energy. The energy here is the power multiplied by the current and the voltage, that is, the interference source and the interfered object (electrical equipment) must form a current loop. As can be seen from Figure 14, the normal mode interference current is formed by the voltage between the live wire and the neutral wire, and can form a current loop with the power supply and the load. The common mode voltage (here is the neutral-to-ground voltage) is the voltage between the neutral wire and the ground wire, and it cannot form a closed loop of current with the electrical equipment. Neither the voltage nor the current has a channel to reach the electrical equipment, so how can we talk about interference? How can we talk about "endangering the safe operation of these electrical equipment"!

Figure 14 Schematic diagram of normal mode interference and common mode interference

What is puzzling is that the same transformer has so many "hidden dangers" when connected to the output of a high-frequency UPS inverter, but has a better ability to resist "impact" loads when connected to the output of a power-frequency UPS inverter. In fact, this is the characteristic of a reactor or choke. Putting aside the conceptual misunderstanding, even if this transformer is regarded as inductive, it is this inductive property that, in a certain sense, will cause "flyback transient spike voltage" that damages electrical equipment when used at the output of a high-frequency UPS inverter, while it has a better ability to resist "impact" loads when used at the output of a power-frequency UPS inverter. Not only that, it also becomes a "50Hz filter" connected across the UPS and the rectifier-filter type nonlinear load, which will greatly increase the UPS's ability to bear impact currents with a high peak ratio." It seems that this transformer is extremely intelligent! However, the author has encountered an example of a transformer and battery burning when connected to the output, and it was the power-frequency machine that burned. As shown in the following example.

For example, a manufacturing plant in Beijing used a 600kVA UPS power supply scheme as shown in Figure 15. Five 150kVA UPS were used in parallel for 4+1 redundancy, and the output end was a parallel connection of the secondary windings of five UPS output transformers. There was also a 300kVA transformer in the load, which can be said to be a layer of defense. However, when the 300kVA load transformer switch S was closed in battery mode, the load transformer had an instantaneous short circuit, which caused the UPS to partially burn and the battery pack to catch fire, burning more than 70 100AH ​​batteries in one fell swoop. The five transformers did not play any so-called "buffer" or "filter" role.

It is worth mentioning that some people say that transformers can resist interference, which is another basic concept problem. What devices can resist interference? People with basic circuit knowledge know that only nonlinear devices or inertial devices can resist interference. Transformers are nonlinear core devices that work in the linear region. Because of this, they make the transmission waveform undistorted. The key to winding a transformer is to strive to make the leakage inductance as small as possible, and zero leakage inductance is the best. A good transformer is almost a fully linear device. The characteristic of a linear circuit is to transmit waveforms without distortion - the output will copy the input waveform. This can be detected with a dual-trace oscilloscope. You can see it at a glance without any debate. Transformers with large leakage inductance are low-quality transformers because of their inductance, or even unqualified products, because they reduce the dynamic performance of the power supply output voltage. It is inappropriate for someone to take the phenomenon caused by the negative performance of unqualified products as a serious matter.

Figure 15 Schematic diagram of a semiconductor factory with 4+1 redundant parallel connection output connected to a transformer

Of course, in order to demodulate the sine wave from PWM, the output transformer of the special industrial frequency UPS intentionally leaves a little leakage inductance when winding the output transformer, in order to use this leakage inductance and the capacitor behind the transformer to form an LC filter. However, this leakage inductance is very small and does not affect the output dynamic performance of the UPS.

Figure 16 Schematic diagram of two types of UPS output and load connection

The transformer of the high-frequency UPS model mentioned above is useless, and its purpose is to promote the so-called high performance of the output transformer of the power frequency UPS model. Some people keep saying that the output transformer of this UPS is used to resist interference. What interference is resisted? Is it the interference from the front of the UPS output transformer or the interference from the load end? What is the purpose of resisting the so-called interference? Is it to protect the load behind or to protect the UPS inverter? You should know that the output voltage of the UPS inverter is a very good sine wave without interference; then there is only "resistance" from the interference of the load. But the so-called interference from the load end is caused by the normal operation of the load. Because the previous load equipment is mostly a rectifier filter load with a low input power factor, it has caused a certain degree of damage to the output voltage sine wave of the UPS, which is generally called "interference", and this so-called "interference" is the "result" of the voltage damage after the load works. The result of this damaged voltage is the largest at the load end. The farther the distance from the UPS output end to the load, the thinner the wire, and the more contacts passed, the greater the distortion; on the contrary, this distortion is the smallest at the UPS output end. This is not the result of the transformer being able to resist interference, but its original appearance. As shown in the upper and lower figures (a) and (b) of Figure 16, if two UPSs of the same power carry the same load, their UPS output ends are both good sine waves, but they become distorted waveforms at the load end, as shown in the two figures (a) and (b) of Figure 16. This is because the load's rectifier and filter circuit does not demand a sine wave current from the load, but a pulse current that is several times the average or effective value. This current must form a voltage drop with the distributed impedance of the transmission line. Since the pulse current is only formed near the peak of the sine voltage wave, this voltage drop is only formed near the peak. The peak value of the voltage wave reaching the load must be subtracted from the UO peak value by the voltage drop along the way, so the clipped distortion is formed. The waveform of the voltage UO at the output end of the UPS cabinet depends on the size of the UPS internal resistance, so the large distortion at the load end and the small distortion at the UPS end have nothing to do with the transformer, and it is not interference, let alone the result of the transformer's anti-interference. Moreover, whether it is a power frequency UPS or a high frequency UPS, the results in this regard are the same. As for the "burrs" on the cable between the UPS output and the load, they are caused by the nonlinearity of the load that destroys the voltage waveform and transmission, and are not so-called interference.

Figure 17 Schematic diagram of the relationship between the UPS output voltage reaching the load and the distance to the load

Since the interference amplitude at the UPS output port is very small, there is no need to resist it. The purpose of anti-interference is nothing more than to protect something. There are only two targets dealing with this output transformer here: the inverter in front and the electrical equipment in the back. As we have known before, this so-called interference is the result left after the load works normally, which belongs to the normal working range, so there is no need for protection; there are capacitors in front of the inverter in front, and the output voltage sine wave here is very good, there is no so-called "interference", and there is no need for the transformer to be aimless. Therefore, the transformer anti-interference advocated here is a "false shot" and "aimless shot". But if you don't know this principle, you will be shocked by this "false shot"!

In short, in the market that belittles high-frequency UPS, some propagandists use so-called "analysis" methods or the performance of substandard products to create some so-called "potential" and "hidden dangers" suspense to scare those who don't know the truth; they put the "advantages" of the same thing on the face of the power-frequency UPS, and put the so-called disadvantages on the head of the high-frequency UPS. They want to extend the market life of the power-frequency UPS for some time. Although it is not good for merchants to do this, it is understandable for livelihood. But as an academic discussion, it is unfair. Especially when they act as experts without understanding the performance of the machine, they create suspense without reason. Of course, there are many theoretical level and basic concept problems, but it is wrong to mislead users anyway. It should not go against the current national energy conservation and emission reduction policies. ■