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

Preface

At present, we have entered the era where high-frequency UPS gradually replaces industrial frequency UPS. Of course, the replacement process is not smooth. People have used industrial frequency UPS for decades and are familiar with this form of power supply. They cannot adapt to the sudden change of models all at once, so they are more comfortable to listen to the praises of industrial frequency UPS and are also easy to accept some criticisms of high-frequency UPS. They just hit it off. But they don't know that it has damaged the interests of users to a certain extent and is also contrary to the current national policy. We often hear such a saying: high-frequency UPS is a good thing, but because our system is very important and requires very high reliability of power supply, it is still reliable to use industrial frequency UPS. The implication is that high-frequency UPS is unreliable. Don't you know that reliability is designed, that is, the reliability of a machine depends on which level of reliability standard is adopted. To give a simple example, a 120120 axial flow fan commonly used in a UPS has a price of more than ten yuan and a price of more than one hundred yuan, with a price difference of nearly 10 times. Which one has higher reliability? It goes without saying that, of course, the reliability of more than one hundred yuan is higher. Another example is the 9315 series UPS of a certain brand, which is known as the "bid winner", meaning that it has the highest price every time it bids, but it also has the highest reliability in operation and is called the "iron machine" - it does not fail; while the PB4000 series of the same brand and the same power is much cheaper, but it also has more failures. Of course, users' concerns about high-frequency UPS are not groundless, and the basis is the misleading propaganda from certain aspects. Some even elevate these propaganda materials to "the fatal weakness of high-frequency structure UPS". Although only a few people raised the issue, the impact is quite large. It is posted on the Internet, as if many people wrote this article, and it has indeed affected many users, and even some technicians have been infected. In order to clarify these issues and give people a scientific view of the product, the following will discuss these aspects.

1. IGBT rectifier reliability is relatively low

There are two reasons for this view:

1. IGBT devices have a lower overload capacity than SCR devices

To prove this point, some people give an example of the overload capacity of two devices: SCR can be overloaded to 10 times the rated current for 20ms, while IGBT can only last for 20ms when overloaded to 10 times the rated current , which means that the overload capacity is 1000 times worse. Is it fair to say that the reliability of IGBT devices is not as good as SCR based on this point? This requires tracing back to why their overload capacity is different. Does it mean that the overload capacity of IGBT can only be 10 times 20ms ? Of course not. The device designer selects it based on its necessity. SCR is not a fully controlled device, that is, it can only be controlled to turn on in an AC circuit but not to turn off. Once the thyristor is turned on, it will not automatically turn off until the voltage or current passes through zero, as shown in the figure below Figure 1 (a). The working principle of this device determines that its overload capacity must not only be strong, but also must be able to withstand overload for a long time. For example, in Figure 1 (a), the SCR is triggered and turned on at time t2. If the corresponding time here is t2 = 1ms, and at this time the output end just happens to have an overcurrent or even more than 10 times, since there is no cut-off mechanism here, it must be able to withstand this overcurrent without damage within about 10ms before t3 (half cycle of 50Hz). Otherwise, if the overload tolerance time of this device is short, such as 1ms, the probability of damage to the device is too high and it cannot be used. But IGBT is different, because it can not only be turned on at any time but also turned off at any time. As shown in Figure 1 (b), it is turned on at t1 and turned off at t2. At present, the operating frequency of IGBT can reach up to 150kHz, that is, a turn-on and turn-off cycle is about 7ms , so 20ms is long enough for the time from the discovery of overload to the turn-off of IGBT. That is to say, the overload time of IGBT does not need to be so long. Even if the manufacturer extends its overload time by 1000 times, what is the use! Is it necessary to give an intercity train that can reach Tianjin Station from Beijing South Station in 30 minutes a 10-hour operating time margin?

Figure 1 Comparison of SCR and IGBT operation in rectifiers

At present, the modulation frequency of high-power UPS is mostly below 15 kHz. For example, 10kHz means 100 pulses per half cycle, and the width of each pulse is 0 ms < T < 100 ms . When overcurrent or short circuit occurs, IGBT can be turned off at any time. Since it can be turned off at any time, why should the overload time be so long? For example, there are two trains running between Beijing and Tianjin, one is a steam locomotive and the other is an electric locomotive. For safety reasons, it is stipulated that the steam locomotive should be overhauled every 4 hours, while the electric locomotive should be overhauled every 2 hours. Can it be said that the reliability of the steam locomotive is twice that of the electric locomotive? From the perspective of time, it seems so, but within 2 hours, the electric locomotive has run 4 round trips, while the steam locomotive has only run one single trip within 2 hours! Which one is more reliable? By the same token, isn’t it a bit far-fetched to compare the overload capacity of two devices with different shutdown mechanisms and performances?

2. It is said that because high-frequency UPS has not yet found a material with large magnetic flux, the temperature of its "boost inductor" is too high, which reduces reliability. It is even asserted that because of this (referring to the lack of a material with large magnetic flux), the UPS industry has not been able to manufacture a high-power high-frequency UPS with high enough reliability.

His original intention was to talk about the quality of the "boost inductor", but the material indicators proposed to improve the reliability of the inductor were those of the transformer. This basic concept problem confuses people: is it the inductor or the transformer? Because the main parameters of the materials selected for the two are completely different, the transformer requires a material with a large magnetic flux, which can be seen from the transformer winding calculation formula:

N---Number of turns of transformer winding

U---The voltage applied to the winding

f--- frequency of voltage

B---Magnetic induction intensity (corresponding to magnetic flux)

SC---Transformer core cross-sectional area

At present, there are many materials with large magnetic flux, such as the iron-cobalt-vanadium core that has been used by people for a long time, which has a large magnetic flux. The current cold-rolled steel strips and soft magnetic materials have very high magnetic flux. It can be seen from formula (1) that the larger the magnetic flux, the fewer winding turns are required, and the more copper is saved. However, the high-frequency machine structure UPS does not have a power transformer, so the requirement for materials with large magnetic flux is pointless. It seems that this does refer to inductors L1, L2 and L3, as shown in Figure 2. However, the calculation formula for inductance is different from that of the transformer, as shown in formula (2)

Figure 2: An electrical schematic diagram of a full IGBT UPS

L---Inductance

SC---Cross-sectional area of ​​inductor core

N---Number of turns of inductor winding

lC---Core magnetic path length

r---Relative magnetic permeability of core material

From this formula, we can see that there is no parameter called magnetic flux B. What is related to the inductor core is the relative magnetic permeability m _r . The larger the relative magnetic permeability, the larger the inductance. Currently, there are many materials with large relative magnetic permeability, but the most commonly used one is ferrite, commonly known as iron oxide.

Another basic concept is the problem of high inductor temperature. Anyone who has done circuit design knows that the temperature of the inductor can be controlled in the design and test, and it is also easy to solve this problem. Generally speaking, as long as the wire diameter of the winding is larger and the iron core is larger, it will be fine. This is a basic common sense for people who often work on circuits and is self-evident. How can it affect the production of high-power UPS? Besides, many manufacturers have made 500kVA high-frequency UPS, and some even made 1200kVA high-frequency UPS. Isn't it considered high power! Some manufacturers cannot make high-power high-frequency UPS with high reliability for a while, not because "materials with large magnetic flux have not been found yet", there are many technical problems here. And it is too arbitrary to say that one or two manufacturers cannot do this for the time being and say that it is the entire "UPS industry". If you can't do it yourself, you have to work hard, or acquire a company with this ability, so that the latecomers can catch up. What's the point of standing there complaining and making groundless accusations?

It seems inappropriate to call these seemingly plausible problems due to unclear concepts as "fatal weaknesses" and pin them on high-frequency UPS. It is mainly due to misunderstandings that the above two "arguments" are not appropriate.

(II) Some people believe that high-frequency UPS has the risk of "zero bias failure"

This problem is another so-called "fatal weakness". It means that high-frequency UPS will produce a phenomenon that "does not occur on other UPS models". This view is that when the upstream AC power supply (such as "input 1" to the backup generator "input 2") is switched through the ATS, the UPS output will form an output voltage flash of more than 8ms, as shown in Figure 3 (b). It is said that this can cause a data center computer room to be paralyzed for tens of minutes to several hours. The reason is that the dual power supply is

(a) Schematic diagram of the relative position of ATS switch and UPS

(b) Schematic diagram of an output voltage drop gap when the input power is switched by ATS

(c) Schematic diagram of the main circuit of half-bridge inverter and DC power supply

Figure 3 Principle circuit diagram of high frequency machine structure UPS inverter (I)

The midpoint potential of the source is ± 400V. Once a transient, unipolar DC bias voltage appears on the input power N line during UPS operation, it will be input to the inverter input terminal, causing "instantaneous DC overvoltage" and "instantaneous DC undervoltage" of the inverter, which will cause this "transient DC bias" fault.

In an AC circuit, there will be a "unipolar DC bias voltage". The so-called unipolarity, as the name implies, is either positive or negative. What is this DC bias voltage? How is it formed? The questioner did not explain it clearly. What this means is that when the upstream ATS is switched, the input rectifier boost link is instantly powered off, so the current on this section of the neutral line N is also interrupted. As shown in Figure 3 (a), the neutral line (dashed line N) from the mutual input cabinet to the UPS will induce a counter-electromotive force e in this section of the line , that is:

Where ——Null line self-inductance, Henry (H)

——Current change when power is off, ampere (A)

——Time change during power outage, seconds (s)

Whether this back EMF can pose a threat or even be "fatal" is useless if it is only qualitatively stated, and it is easy to mislead readers. It is necessary to know the magnitude of the back EMF to be convincing. In order to have a quantitative concept, assuming that the length of the neutral line from the UPS to the mutual input cabinet is l=30m=3000cm, and the diameter is d=0.6cm, then the distributed inductance Lo on this neutral line is:

The cross- sectional area of ​​the cable with a diameter of d=0.6cm=6mm ^is S= pr2 =28 ( mm2 ) . According to 10A / ( mm2 ) , ^{we take 300A here, and assume that the fastest} action time of ATS is 0.1s=100ms. Then the back electromotive force e value can be calculated by formula (3):

That is, the back electromotive force excited on the neutral line when the ATS switches is 0.15V. Of course, this calculation may not be very accurate, but it is not much different in terms of order of magnitude. Even if it is 10 times larger, it is only 1.5V, so some clues can be seen here. This kind of analysis somewhere is unfounded, using imaginary "hidden dangers" to scare people. In other words, the unipolar voltage excited on the neutral line when the upstream ATS switches is very small, which can neither cause output flashing, nor cause inverter overvoltage or undervoltage, nor cause power outages in the data center room for several hours. Besides, the neutral-ground voltage cannot be added to these places at all. And the output voltage flashing is not caused by this reason. This issue will be discussed later.

It is asserted somewhere that this unipolar neutral line voltage "will not appear in other UPS models". Does it mean that there is no neutral line in power frequency UPS? When the ATS is switched, the current on the neutral line from the mutual input cabinet to the UPS cabinet will also undergo a sudden change from full load (hypothesis) to zero, and the same back electromotive force will be generated on the neutral line because its neutral line is not a superconductor. How can it be said that "it will not appear in other UPS models"!

Here is another question about the basic knowledge of circuits, especially the working principle of UPS. How is the unipolar voltage on the neutral line (i.e., the DC bias of the N line) formed? Is the flash of output voltage caused by the so-called neutral line voltage? How does it cause the inverter to be over-voltage or under-voltage? Are these problems only found in high-frequency UPS, etc. In order to understand them clearly, let's discuss these issues one by one.

1. What does the neutral line voltage refer to? It is well known that we can only talk about current on a wire, not voltage, because voltage is the potential difference. Here, the concept of the N line voltage is proposed. Let's understand it as the neutral-ground voltage. Is it the voltage of point A to GE in Figure 3 (c) or the voltage of point B to GE? Because the voltages of these two points to the ground are different when there is a load. The voltage of point A to GE is the highest. This is the voltage drop on the entire neutral line in the UPS. In order to meet the wishes of a certain place, this highest value is temporarily taken. Will this cause the inverter to be "overvoltage" or "undervoltage"? What value can make the inverter overvoltage? Generally speaking, it must exceed the rated voltage by at least 10%. A certain place gives a rated working voltage of 400V. Even if 10% is considered overvoltage, the voltage on the neutral line must be at least 40V! The question is, is it possible for the neutral line to have such a high unipolar voltage? Generally speaking, the neutral line in most UPS will not exceed 2m, and the cross-sectional area is not small. Under any normal circumstances, let alone 40V, there will not even be 4V. Even if there is 4V, it doesn't mean that 404V is overvoltage and can damage the power tube. So it seems that the so-called unipolar voltage causing overvoltage does not actually exist! It is only a "potential" "danger". Besides, this neutral voltage cannot be added to the tube.

2. If the unipolar neutral voltage does not pose a hidden danger, how does the 8ms flash of the output voltage come about? Can it really cause a long power outage in the data center?

This is also basic knowledge that everyone who works with power supplies should have. As we all know, the internal resistance of the battery is relatively large. For example, when the upstream ATS is switched, there will be a sudden change in the internal load of the power supply. In addition, the dynamic performance of the battery is not very good, so it cannot respond to this sudden change in current quickly. Generally, when the UPS is working normally, the input rectifier supplies power to the inverter, and the battery pack is not only unloaded but also in a floating charge state. If the input end suddenly loses power, the battery pack must take over all the loads in time, but the strong current mutation is something that ordinary batteries cannot respond to. This will inevitably lead to an instantaneous current shortage state, which is the so-called instantaneous "flash-off" of the output voltage. In order to make up for this shortcoming, the designer has incorporated a capacitor of sufficient capacity after the battery pack or rectifier. Since the dynamic performance of the capacitor is much better than that of the battery, the transient front current is first compensated by the capacitor, and then the battery will continue the power current for a long time in the future. However, if the capacity of the capacitor connected in parallel with the battery is insufficient or the quality is poor, and it cannot meet the requirements of the sudden change in the front current, the output voltage will have a so-called "flash-off" gap. The smaller the capacitance of the capacitor, the deeper and wider the gap in the output voltage. So this output voltage gap has nothing to do with the so-called unipolar N-line voltage.

Moreover, this output voltage gap problem may exist on any UPS, and it is only found in unqualified products. Whether it is a high-frequency UPS or an industrial-frequency UPS, as long as it is a qualified product (not a product that cuts corners), this output flash phenomenon will not occur. For some reason, someone has forced this phenomenon, which can happen to anyone, on the neutral line voltage of a high-frequency UPS, which is another misunderstanding of the working principle of UPS.

3. Can an 8ms output voltage interruption really cause a data center to stop working?

Generally, qualified and functioning UPS will not have this phenomenon. Even if this 8ms flash failure hidden danger really occurs, is there any fatal danger? According to the actual test of IBM and HP on their PCs, after the mains power is cut off, the built-in power supply can still ensure that the machine works at full load for 50ms. This is mainly the additional effect obtained by the capacity of the filter capacitor determined by the circuit's requirements for the pulsation and stability of its internal DC power supply. In large-capacity machines, the capacitance is also increased proportionally. Therefore, the same effect should also be achieved. At least in many computer rooms, there are examples of 20ms power failure without any impact. At present, almost all electronic devices have built-in switching power supplies, and their task is to convert the input AC voltage into different types of DC voltage used by this device. The power supply circuit is shown in the left figure of Figure 4. C in the figure is the energy storage device. If this energy storage device does not have the ability to support the 8ms backup work of this device, it is probably not a qualified product. If you take unqualified products as an example, the result will not explain anything.

Figure 4 IT equipment and internal power supply main circuit

4. Are there other reasons that may cause unipolar voltage on the neutral line when the ATS switches?

The above "hidden dangers" are all from the assumption of the questioner about the unipolar neutral line voltage, which is probably another basic concept problem. First of all, ATS switching is a normal action, and ATS switching is nothing more than an instantaneous power outage. As we all know, for a qualified UPS, when the input end is instantly powered off due to ATS switching, the capacitor and battery will promptly supply sufficient power so that the load machine will not feel anything. In other words, there is no change in the voltage and current at the output end of the UPS. Then the current on the neutral line from the load to the battery pack will not change, and of course the voltage drop on this neutral line will not change. The neutral line between the neutral line and the input power supply has no power supply, no current, and no voltage, and even if there is a reverse electromotive force, it is very small, which has been mentioned above. In this way, the original neutral line voltage during operation will always be constant, and the so-called "unipolar" neutral line voltage hazard will not occur.

When the ATS switching process is completed and the UPS is connected to the input power supply again, the input rectifier is turned on to charge the capacitor and battery behind it, and also to power the inverter. At this time, since the load has not changed, the neutral line voltage to the right of point B in Figure 3 (b) remains unchanged, and the neutral line voltage to the left of point B is of course not zero. The voltage at point B is raised, and the neutral-ground voltage at this point is neither unipolar nor can it be added to the battery voltage, and at most it is less than 1V, and it has no effect. Somewhere it is said that the ATS switching process can lead to many serious consequences, but I don't know what machine it refers to. Even if there is such an example, I'm afraid the problem is not in the ATS switching, and other reasons must be found. Not to mention the so-called "hidden dangers".

All the above statements come from the middle neutral line tap between the two sets of DC power supplies. In fact, that is the original old circuit. It is always inconvenient to use two sets of batteries, so later a half-bridge inverter circuit that still uses one set of batteries was developed, as shown in Figure 5. In this circuit structure, another bridge arm is added to the inverter, as shown in the link in the dotted box in the figure, temporarily called the fourth bridge arm, composed of VT7 and VT8. In this way, the three-phase bridge arm can form a phase voltage output with a neutral line with the fourth bridge arm. In order to illustrate the problem, UC is taken as an example in the figure.

The path for UC to output positive half-wave voltage is: GB+ ® VT1 ® R _upper end ® VT8 ® GB-. The path for UC to output positive and negative wave voltage is: GB+ ® VT7 ® R lower end ® VT4 ® GB-. The same is true for the other two phases UA and UB . Since the three-phase is also designed to work according to the phase difference of 120 _° , the value between the line voltage and the phase voltage is also:

The relationship is 120 in sequence in phase. In this way, high-frequency UPS and industrial-frequency UPS use the same set of batteries. This reduces the number of shells in the equipment of the external battery pack by half. For example, 322=64 50AH/12V were used before, and now 32 100AH/12V can be used.

Figure 5 Main circuit schematic diagram of new high-frequency structure UPS

(III) "The neutral-to-ground voltage of high-frequency UPS is too high"

1. The mechanism of high “neutral-ground voltage”

Somewhere it is said that "high neutral-ground voltage" is also a "fatal weakness", and this view is also worth discussing. It is said that the high-frequency PWM interference voltage from the IGBT pulse width modulation rectifier and inverter is directly fed back to the neutral line of the UPS input power supply system and the output power supply system through the neutral line in the form of a "neutral-ground voltage" with a high amplitude value, thereby endangering the safe operation of electrical equipment. "It should be noted here that the IGBT inverters of the industrial frequency model and the high-frequency model UPS are the same devices, the same frequency, and the same working principle, so the "interference" should also be the same. This is not the case with the rectifier. The interference of the thyristor rectifier is much greater than that of the IGBT rectifier. Even 12-pulse rectification plus 11th harmonic filter (which increases considerable weight, volume and cost) generally cannot fully achieve the IGBT indicators. According to the statement here, the two interferences of the high-frequency machine can be directly added to the neutral line of the UPS input power supply system and the output power supply system, thereby endangering the safe operation of electrical equipment; the industrial frequency model with greater interference The two items of UPS cannot be added to these places? It is really incredible. As for how the neutral-ground voltage can be added to the electrical equipment, there will be a special discussion later. It is true that the neutral-ground voltage of the high-frequency UPS is sometimes a little "higher" than that of the industrial frequency UPS because there is no secondary grounding link of the output isolation transformer. This is because the circuit structure has an additional voltage drop of a tube in the single power supply structure, as shown in Figure 6 (a). The figure shows the high-order harmonic filter current path. Since the inverter works in pulse width modulation (PWM), that is, the sinusoidal wave voltage is "high frequency" modulated into a square wave form of varying widths for output, but because the load end requires a sinusoidal wave voltage, the PWM wave must be demodulated by a filter before reaching the load to filter out the high-frequency components in the PWM square wave and retain only the sinusoidal wave components. Therefore, this part of the high-order harmonic components will be sent back to the negative end of the power supply through the filter. Here we only take UC as an example to look at the high-order harmonic current path:

GB+ ® VT1 ® Low pass filter LC ® To neutral ® VT8 ® GB-

From here, we can see that since the neutral line passes through an IGBT tube at the VT7 or VT8 position, there is an additional tube voltage drop link on the neutral line, which increases the neutral-to-ground voltage. In the case of a dual DC power UPS, there is no VT7 or VT8 on the neutral line.

Figure 6. Current paths of high-order harmonic filters in two UPS structures

This link. But generally there is a distance between the battery and the machine, which lengthens the length of the neutral line and increases the voltage drop on the neutral line. Despite this, modern technology will make the neutral-to-ground voltage of the two high-frequency machine structure UPS less than 1V.

For the power frequency structure UPS, the output transformer makes it possible to reduce the neutral line voltage drop. As shown in Figure 6 (b), the high-order harmonic filter current path of the power frequency structure UPS is much shorter, because the return path of the high-order harmonic current is near and inside the transformer.

As for the conclusion that "IT equipment can only operate safely when the neutral-ground voltage is less than 1.5V", it is questionable because China Telecom has already broken through this forbidden zone. Actual tests have shown that the neutral-ground voltage has even reached 21V, and no abnormality has been found in more than 100 digital machines.

It should be noted that the factors that cause the voltage drop on the neutral line are not only high-order harmonics, but also three-phase load imbalance and neutral line resistance. Generally speaking, the neutral line current of a three-phase output power supply is mostly smaller than the phase line current of a single-phase output power supply. This is because the three-phase currents on the neutral line during three-phase output are the result of vector sums, and they have a mutual cancellation effect. Figure 7 shows the vector relationship of several of these situations. Figure 7 (a) shows the situation where the three-phase currents are equal, that is, I _A = I _B = I _C. In this case, it can be seen that the vector sum of any two-phase currents is equal to the third current value with opposite signs. Here, it is the vector sum of I _A and I _B , I _AB = - I _C , and the vector sum of the two is zero. At this time, the voltage drop on the neutral line depends only on the harmonic current and the neutral line resistance. This is also the case where the neutral-to-ground voltage is the smallest. Figure 7 (a) shows the situation where the current of phase A is small and the currents of phases B and C are equal and greater than the current of phase A, that is , I _B = I _C > I _A. It can be seen that at this time, the absolute value of the vector sum of _IA and IB is êIABê = ê -ICê _. The two cannot cancel each other out, _{so partial load current appears on the} neutral _{line. At this time ,} the current on the neutral line becomes the sum of partial load current and harmonic current, which increases the voltage drop on the neutral line. Figure 7 (c) shows the situation where the current of phase C is zero and the currents of phases B and A are equal, that is, IA _{= IB} , IC = 0 . From the vector sum in the figure, it can be seen that _{the vector sum of IA} and IB is _êIABê = êIAê ₌ êIBê . In _other words , in this case, the current on the neutral line is _{equal to} the _{current value of one phase. It can also be concluded that when only one phase voltage has a load, the current on the neutral line is also the current value of one phase.} And if the effect _{of harmonic} current is not considered, the maximum current on the neutral line does not exceed the current value of one phase. Of course, if the third harmonic and the higher harmonics of the third harmonic multiple are superimposed, the voltage drop on the neutral line will increase, and of course the neutral-to-ground voltage will also increase.

Figure 7 Vector relationship of three-phase current in several cases

Therefore, in order to prove his point, the questioner also gave the figure that the neutral-ground voltage of the power frequency UPS is 0.8V, while the neutral-ground voltage of the high frequency UPS is higher than 1.5V. In fact, this figure is meaningless and cannot explain anything, because the neutral-ground voltage can be easily reduced to 1V or even below 0.8V without a transformer. In the case of different combinations of the above-mentioned load current and harmonic current, the neutral-ground voltage is also different, and some are as high as 10V or more. Regardless of the power frequency UPS or the high frequency UPS, the neutral-ground voltage may be higher or lower than 1.5V.

2. Impact of neutral-ground voltage

Is the high neutral-ground voltage a "fatal weakness"? Ordinary users regard the neutral-ground voltage as a scourge, and they are terrified when it is mentioned. The person who raised the question added fuel to the fire and raised it to a "fatal" level. The author has already described the impact of the neutral-ground voltage in detail in many articles and books, so it is not a problem to repeat it here.

There are three factors that must be present to form interference: the interference source, the path for transmitting the interference, and the device being interfered with. These three are indispensable, so the discussion will start with these three.

(1) Is the neutral-to-ground voltage a source of interference?

If it is proved that the neutral-ground voltage is indeed the source of interference, the conclusion that the neutral-ground voltage interferes with the load and is even a "fatal" weakness may be established, and the high neutral-ground voltage of the high-frequency UPS is also to blame. In order to explain the neutral-ground voltage, we must first clarify what the neutral-ground voltage is. Figure 8 shows the location of the neutral-ground voltage. As can be seen from the figure, the neutral-ground voltage refers to the voltage between the lower end of the load and the ground. In the ideal wiring method, there is no current on the neutral line. It is just a reference point, so the entire neutral line is a zero potential. Generally, the neutral line and the ground line are connected to one point and grounded at the source end of the AC mains (such as a substation), as shown in Figure 8. In this way, it can be seen that the so-called neutral-ground voltage is the neutral line voltage formed by the neutral line current and the neutral line resistance. Figure 8 takes the A-phase power supply UA as an example. Obviously, if the load switch S is disconnected at this time, there is no load current, that is, Ia=0, then there is no current on the neutral line, and of course there is no voltage drop on the neutral line, and the neutral-ground voltage is also zero. When

Figure 8 Location and formation of neutral-ground voltage

After the switch S is closed, the load current Ia starts from UA and goes along the arrow direction through the switch Sloadneutral line resistanceback to the midpoint of the star transformer. It is worth noting that the load current Ia first flows through the load, and then enters the neutral line and returns to the midpoint after it comes out of the load. In other words, the load current Ia does work on the load first and then passes through the neutral line, that is, the voltage drop on the neutral line is the mark left on the neutral line by the return current after the work is done. Could it be that this mark will go back and reverse the result of the work done! For example, if you drive a stepper motor, the switch S is closed, the motor moves, and then a section of zero-to-ground voltage appears on the neutral line. Could this section of zero-to-ground voltage go back and prevent the motor from moving or make it move abnormally? Here is a basic concept: In fact, the zero-to-ground voltage appears and disappears at the same time as the load action, and there is no problem of affecting the subsequent action. ■