JPS5819168A - Controlling method for pulse width of inverter - Google Patents

Abstract

PURPOSE:To alter the output voltage of an inverter completely proportionally to a frequency by inverting the logic level of a control output signal at the time point of function value N-F(theta+DELTAtheta) and again inverting it at the time point of 2F(theta+DELTAtheta) from the previous point, thereby generating a control pulse. CONSTITUTION:When a function value F(theta+DELTAtheta) is outputted from an ROM5, a presettable timer 6 counts clock pulses E up to the prescribed value N, and the output signal of the timer 6 is inverted after the time proportional to N-F (theta+DELTAtheta) is elapsed from the time point when a modulation start signal B is outputted. On the other hand, a timer 7 counts from 0 upto N, and since the clock pulse H is switched from E to I from the previous inverting time, the pulse interval becomes twice value, a timer output signal K is again inverted after the time proportional to 2F(theta+DELTAtheta) is elapsed, thereby generating a pulse width modulation signal L. In this manner, an ideal V/F control can be performed.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a control method using pulse width modulation (hereinafter referred to as ?1M) K in a voltage source inverter control device.

PWM control uses square wave pulses obtained by rotating RK.
(By changing the pulse amplitude and thereby controlling the magnitude of the inverter output voltage, the output voltage is improved by eliminating lower harmonics.)

This is a control method that can improve the output waveform.

Conventionally, AWM control has generally used a method of comparing a constant frequency triangular wave with a reference signal of the inverter output voltage, and determining the width of the output pulse based on the magnitude relationship between these two signals. However, there is a drawback that the output voltage value changes from moment to moment due to the beat between the frequency of the triangular wave and the output reference. The reason why this phenomenon occurs is that pulse width modulation controls the average output value between modulation periods by using the duty cycle of the output pulse width for a certain period (modulation period). This occurs because in conventional hojo, this modulation period is given in terms of time.

If this conventional method k is used as an induction motor control KM, then
From the triangular wave frequency and output standard, which is 5 or k.

The output voltage value changes from moment to moment, which has the disadvantage of causing vibration, noise, current heat, etc.

On the other hand, recently read-only memory (below)

It's called ROM. ) has been announced to perform P W M @[email protected] advantage of using @1tOM is that complex functions can be easily obtained, so the circuit configuration can be simplified. Also, digitalization makes adjustments easier.

By using this FIOM, pulses are output at a predetermined phase of the output fundamental wave, so that beats do not occur as compared with the triangular wave having a constant line wave number. However, it is also true that many of the conventional methods using ROM have noticeable drawbacks that occur with digitalization.In other words, there are various methods using conventional POM. , it can be roughly divided into the following two methods. In the method of No. 18, the rise and fall of the pulse are set in advance according to the output frequency and output voltage.
r, which is controlled by electrical angle. The first method has a modulation signal with a frequency proportional to the output frequency, and the FIOM
A reference signal K is stored, and the rising and falling edges of the pulse are determined by comparing one of the signals. Among these, the first method is a counter that counts electrical angles, and the accuracy of the output waveform in electrical angles is determined by the number of bits K. Therefore, if the output frequency is low, it counts l bits of electrical angles. The time taken to hit becomes longer and the hourly accuracy decreases, and if the output voltage increases.

Even though the pulses are the same, they have different values as electrical angles, and because this information must be stored, a FIOM with a large storage capacity is required.In the case of this method, the output voltage When the reference is small, the output voltage reference itself may contain low-order harmonics.

It is an object of the present invention to provide a PWM control method that does not suffer from the drawbacks associated with digitization that have been problematic in the past.

The present invention will be described in detail below based on an illustrative embodiment.

<Principle Configuration> FIG. 1 shows the principle configuration of a PWM fIIll#li device for executing the PWM control method according to the present invention.

Figure 1 shows the output waveforms of each part.
The output frequency is set by the output setting tF/. That is, the output frequency setting value is given by the voltage signal A and is input to the /F converter.

V7. The converter tacho uses a pulse train with a frequency proportional to the input voltage signal (hereinafter referred to as saga start signal B).
) (I8 - see M in Figure (a)), this change start signal B is given to the n-ary counter.

The n-kin counter-tei is weird! Each time the start signal 1 is input, the count value C is incremented by one.

Here, n is a number determined by the number of pulses desired to be output. The count value C is! IOM! t as an addressing signal.

ROM jK is the sine wave function value sin for each constant electrical angle Δθ
The data D of θ is written in advance (th - fusion (b)'
), and the FIOMj sequentially outputs the data D at the address specified by the count value 0 given by the n-ary counter and sends it to the presettable timer 4. therefore,
This data D is a presettable tie! It is given as a preset value of 4.

The presettable timer 6 clears the output to 10# every time the variable 1mk-start signal 1 is applied, and loads the preset value (data D). Then, the clock pulse l1t ($J diagram (C
)), and when the count value ν reaches a predetermined number MK (Fig. 1 (phrase)), an output signal G'/' is generated and the counting operation is stopped (Fig. 1 (11)).

The modulation start signal B is input to tie-f7.

When this modulation start signal B or 4 is seen, the output signal is stopped at O# and a counting operation is started. It's fc and the count is %
Start from 0#. Clock pulse used for counting operation ■Clock pulse from Tsukuji Harita and Nerator J! And this clock pulse l is 7 times tK.
A pulse train whose frequency has been divided to a higher frequency is provided by switching from a switch IK. This switch t is controlled by the output signal GK of the brisetter pull timer, and when the output signal G is 10#, it supplies the clock pulse 1 to the timer, and when the output signal G reaches 17#, it switches to the pulse train generator and supplies the clock pulse in parallel. The count value J in the timer is a predetermined number Iff
When the timer output signal reaches 4t% (Figure 1 (f)),
Bind l# and stop counting operation (IJ-Figure (g))
.

The presettable timer output dragon horn G and timer output signal
A fiX-OR gate, )towc, is input. This 1riX-OR gate IO takes the exclusive OR of the output signals G and K and outputs the result as a PWM signal (FIG. 1)).

Here, the PWM signal obtained by one or more configurations and PW
The relationship with the output voltage of the M inverter will be described. It
, it is assumed that the output data D from ROMk is N sin θ, and the predetermined ff is
Count clock pulse curses up to 1N.

Let C be the pulse period of clock pulse 1.

The presettable counter output signal G becomes 91 at point II, when a time period of Tag (/-5 in#) has elapsed from the time when the change start signal B was applied. On the other hand, the timer 7 counts from O to Violet, and the clock pulse ■ after the presettable counter output signal G reaches 11# is switched from 1 to XK, so the pulse interval is one times the value, that is, -
Therefore, the timer output signal X becomes 1/# when the time equal to teoM(/-5in#)+JteroMginθ has elapsed since the modulation start signal B was given. In the end, the output of 'EX-Owl'-to is %11
Since the time until it becomes 1 is delayed by teaM (/-5in#), PWM from the odd 1m1ga start signal.
The time to the center of output signal O is.

Tc N (/-5in it) +knee mTc
1! , and remains constant even if the pulse radiation changes.

Furthermore, the interval between the centers of the PWM output signal B is such that the f-key start signal B is given every fixed period Tm, and.

Since the center of the PWM output signal B generated in each modulation cycle lags the modulation start signal B by a fixed time TeoM, it is equidistant from Tem. Namely.

The variable wI8 frequency is 4. and changes in proportion to the output frequency. On the other hand, the pulse width of PWM output signal B is RO
Therefore, the time is proportional to the content N5in# of the output data p of the MB, and is independent of the output frequency.

If the modulation cycle Tm is multiplied by M, the modulation period Tm becomes /4 ec, whereas KOTam-111# is negative, so the average output value during one period Tm is multiplied by M.

As a result, when the 71M inverter is operated according to the 1'WM control method described above, when one frequency is changed, the output voltage also changes automatically and proportionally. As a result, the magnitude of the output voltage can be changed completely linearly without causing excess or deficiency of voltage as in the conventional method.

The operation of the present invention will be further explained in detail using Figure 183.

Figure 183 represents the circuit operation of Figure 1 using an equivalent triangular wave Vt (see Figure aIIJ).

The height of the triangular wave Wt is the same as MWC. has a slope of 1 hour of friends singing! It changes from the maximum value to the minimum value in am, and goes from the small value to the maximum value in the Q interval TeoN. However, in the present invention, this triangular wave is the modulation start signal! This differs from the conventional method in that only one cycle occurs after the IK is triggered. l87rI! From J(a) to 5, the cycle of the triangular wave Vt matches the interval at which modulation start signal 1 is seen, and light 1! The 0-step wave Ws that continuously forms the triangular wave Wt is a function (data D) k generated by ROMB.
Priest.

Contents of the data file.

Nl1n# is generated every predetermined electrical angle j# (-tm). Therefore, in FIG. 3(a), the triangular wave Vt and the staircase wave W1 intersect after the lR1m1 start signal B is given.

N ・1n# ■ −N + No, no.・The time determined by the spirit 1. Here, if we calculate the t1st snow for 1 hour.

tl -To w (/-5in#)↑Snow-Tc M
(/+ein#) and (tl-t,)s-JTa)isln#. I set the output signal to '/# for this (tm-tm) period.
The AWM output signal shown in JJ (b) is obtained. The time relationship from the modulation start signal B to the rising edge and falling edge of the pulse is the same as that obtained in the circuit shown in Figure IJ1.

The J m(0), @) shows the case where the frequency is lower than the case shown in (IL), (fi). It can be seen that the symmetry of the output waveform is maintained even if the 0 frequency changes. .

This is an effect caused by the function (output data D) output by ROMk keeping a constant value during the modulation period tm, and if it is a continuous mourning sine wave, it will become asymmetrical. (Application to an actual 71M inverter) Figure 4 shows an example of a configuration when the control method of the present invention is actually applied to a voltage source inverter.
The same portion #IC as A/WJ is given the same reference numeral and the explanation thereof will be omitted.

In Figure 7, //a and //b are it in Figure 7.
Corresponding to the circuit busyness shown by, ti* is! 'The contents of the function data D stored in IOMj are -in (# +Jotsu, ii
b is a block diagram showing the case of co-#.

The modulation start signal 1 from the V,6 converter is sent to the counter/
, 7, and the counter /J counts the modulation start signal 1 and outputs a one-shot hull every time it counts the V variation.
Output to p o 1. This one-shot pulse Po is input to a six-success counter/l.

The logic circuit 1S inputs the output count value M1 of the hexadecimal ring counter #1, the 1'WM output signal La of the circuit/l turtle, and the PWM output signal L1 of the circuit //11, and performs a logical synthesis of these to obtain the result. Each transistor I U p of the Steinverter (Figure 111) that runs each PWM output signal number
, ”I P m W F and tr,,v,.

W, via a base drive circuit (not shown).

(Example) In the above explanation, the operating frequency was fixed, but 4
When changing the ratio, it is easier to change the voltage by using two converters for the clock generator J shown in Figure 1 and changing the voltage applied to 74 converters by 9 frequencies.

In addition to the above-mentioned PWM method, there is a commonly used method in which the duty cycle of the PWM output signal becomes SQ when the reference is zero. Even in this case, the present invention can be applied. In that case, as shown in *aa, the modulation start signal B to
When the time (/-5in9) has elapsed, the output logic level is changed from 10' to 11# and the first modulation start signal T
It is only necessary to invert the output logic level from 'l# to 'O# when aM (/+5in-) time has elapsed. In this example, the output logic level changes with one frequency change.
From the first signal B! Since only the ON time passes and the time after death changes, a constant PWM output signal can be obtained.

As described above, the present invention provides the following effects.

When an induction motor is driven by the FWM control method of the present invention, the output voltage changes completely in proportion to one frequency.
sm'/pHJl[l can be performed.

■ The electric motor vibrates because there is no excess or deficiency of voltage.

Does not cause current beat etc.

(2) Since the pulse width varies with time, it is possible to start from zero frequency, and it is possible to suppress the large current generated at start-up, unlike conventional systems that can only output up to KIH level.

(2) I'1OMK The data written is one for every seven pulses, and the output frequency changes to make the same data, and the voltage changes automatically, so the storage capacity of 1tOM can be small.

[Brief explanation of the drawing]

The 11th item is a block diagram showing the configuration of a control device for embodying the I'WM control method according to the present invention, shown in figure s1)
- 佽) are operation waveform diagrams of each part to explain the operation, 71C (&) - (A) are waveform diagrams to explain the action and effect of the present invention, and Fig. 4I shows the present invention as a voltage source inverter. Figure 3 is a circuit diagram of a voltage-source inverter, and Figures 11% and 6 (&) to (a) show other embodiments. It is a waveform diagram for explanation. c. Output frequency setter, c.) converter. J... Clock generator, Da... N-ary kakunta. S...ROM, Todoroki...presettable tie! ,the law of nature·
...Tie i, t...switch circuit! ···flip flop. 10-RX -OR game), //・iii road *//
a*//'b...Circuit.

Claims (1)

[Claims] 1. Using a modulation start signal with a frequency proportional to the output frequency command value of the inverter and a time reference signal with a predetermined frequency as input signals, the electrical angle - Noll#! Electrical angle (#+)#) which advances the value of 1ν(-) by a predetermined electrical angle #) #C key function value 1(
#10ΔI) A device that generates K flight. and a device for measuring time by counting the time reference signal on the pulse width side antic. Some value summer and said function value? Based on (#10)-)K, from the time when the modulation start signal is input -1 (e1)#)K
Invert the logic level of the control output signal after a proportional amount of time has elapsed. When a time proportional to 11 (#+j#)K has elapsed from the timing when this logic level is inverted as a reference, the control output logic level is again inverted and a control pulse is generated from k. Inverter pulse width control method with good image J, IP! ! In the method of writing the full claim. The value of the function ν (#) is changed to the function value 1 (#10) Pulse width control method. , y, @Claim 1) In the method described in jl. I ζ to change the value of the a number 1(-) to the function value ν(-+no)K is a tie-sender every time the ml'' key start signal is given, t or every time the modulation start signal is given. A method for controlling the pulse width of an inverter, characterized in that: L41 In the method #C described in Clause N of Claims N, Clause J! or Clause 3, at the point in time when the logic is turned and the bell button is reversed. The I tyda with reference to is the pulse-1 of the invert I, which is analogous to the point at which the logic level is inverted.
114 blessings. ! In the method according to claim 1, 2, or 3, the tie 5 which is based on the time when the logic level is inverted indicates that the first modulation start signal is input. Characteristic method for controlling pulse width m of an inverter. A,! Scope of request! ! 411g1I is an inverter that provides a 9-hour reference signal at a constant frequency within a predetermined frequency range of the 1-hour reference signal in the method described in IJ/, -, J, Reference, or 1. pulse width control method. 74I claim 1! In the method described in item gJ&/, item 1, item 3, item q, item S, or item 6. A method for controlling the pulse width of an inverter, the method comprising changing the frequency of a time reference signal as a function of an output frequency command value.