JPH07184372A - Resonance type power supply apparatus - Google Patents

Abstract

PURPOSE:To improve conversion efficiency and response characeristic and lower electromagnetic radiation by energizing a transformer in both directions at a constant frequency by self-oscillation, switching a main switch with zero- cross voltage and continuously generating a resonance current depending on a load by providing an intermediate capacitor in the output side. CONSTITUTION:FETs, Tr2a, Tr2b form a half-bridge together with a primary coil Lp, a resonance capacitor Cp, +Vp of DC power supply, GND1 and -Vp to energize a transformer 12 in both directions. Timers 11a, 11b are provided with a delay ON timer for delaying a feedback voltage from feeback wirings Lfa, Lfb and an OFF timer for short-circuiting the feedback voltage after the predetermined time. The OFF timer determines an oscillation frequency, the delay ON timer generates a dead time for FETs Tr2a, Tr2b and causes these transistors to execute zero cross switching. An intermediate capacitor Cm is provided in the output side to make an oscillating circuit composed of a primary winding Lp and a resosnce capacitor Cp self-oscillate to send a resonance current corresponding to a load to the secondary side.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a switching power supply, and more particularly to a structure of a switching power supply which continuously generates a resonance current having a magnitude corresponding to a load by a resonance system.

[0002]

2. Description of the Related Art Switching power supplies have been extensively researched, developed, and applied as a method for reducing the size and eliminating heat generation because of their good conversion efficiency with the development of electronic equipment around 1970. However, since the current and voltage are cut off at high speed, there are many harmonic components and electromagnetic radiation becomes a problem. From the beginning, a resonance type switching power supply with few high frequency components was proposed.

For example, US Pat. No. 3,529,228 (Sept., 197)
0) and U.S. Pat. No. 4,415,959 (Nov., 1983). However, since the resonance current ratio is large with respect to the output power and the component is burdened, it has not yet been applied to general use only in a specific field.

Since these two US patents are the same in principle, US Pat. No. 4,415,959 is taken as an example of the prior art and will be described with reference to FIGS. 1 and 2. The figure is slightly modified for convenience of explanation.

FIG. 1A is a basic circuit and shows an equivalent circuit of the resonance circuit of FIG. 1B. 1a is the input power supply on the primary side and represents the voltage Vp,
1b is a secondary side conversion power source of the transformer 2 and represents the voltage Vs. A switch 3 excites the transformer 2 with a constant width. Dr is a rectifying diode and Cr is a resonance capacitor. Lc is a choke
Coil, Cs is smoothing capacitor, Df is diode and cho
The energy stored in the coil is recirculated. Vout represents an output voltage and is an error signal from the reference value of Vout, which is not shown, but the output is stabilized by ON-OFF frequency modulation of the switch 3.

In the equivalent circuit of FIG. 1B, Lr represents the leakage inductance between the primary side and the secondary side of the transformer 2, and resonates with the resonance capacitor Cr. When the switch 3 is turned on and off as shown in Fig. 2 (a), the resonance current i1 = Vs / Z as shown in Fig. 2 (b).
・ SIN (ωt) flows. Where Z is the characteristic impedance Z
= √ (Lr / Cr), ω is the resonance angular frequency ω = 1 / √ (LrCr), and t is time. This resonance angular frequency ω is the resonance of the switch 3
Set to finish within the ON width.

The voltage Vc of the resonance capacitor Cr is Vc = Vs (1-CO
It changes with S (ωt)) and becomes maximum Vc = 2Vs at the end of resonance. Thereafter, the choke coil Lc and the smoothing capacitor Cs are charged while descending substantially linearly as shown in FIG. 2 (c).

The resonance type switching power supply according to the prior art of FIG. 1 changes the resonance generation period according to the load and discharges a constant voltage output. That is, the energy of one cycle is constant and the output is controlled by frequency modulation. However,
When no load or light load, the modulation frequency reaches audible sound,
The transformer 3 or the choke coil Lc has a defect that causes acoustic noise due to magnetic stress. To solve this, a dummy load is provided to make the modulation frequency higher than the audible sound. This results in wasteful power consumption and thus a reduction in input / output conversion efficiency.

At the time of heavy load, since the resonance current can be generated only discretely according to the basic principle shown in FIG. 2, the average current to the peak current ratio becomes 2.5 times or more as compared with the normal switching power supply. From this, so-called zero-cross switching has been realized, but there is a defect that the loss due to the ON resistance of the switch element rather increases and the conversion efficiency deteriorates. Furthermore, if the input power supply voltage Vp is a commercial power supply, the fluctuation rate is ± 1.
It is necessary to assume 5%, and the peak current will increase,
It has the drawback that it is expensive due to the need to select components with high permissible currents. Further, in U.S. Pat. No. 3,529,228, the operation of the switch is a separately excited method, so that there is a defect that the circuit becomes complicated and the cost becomes high.

Further, in both of the above-mentioned US patents, the switch realizes a zero current cross with respect to the load current, but the voltage resonance with the parasitic capacitance of the switch cannot be executed by the frequency modulation control and the power loss due to the forced switching. There is also a defect that the electromagnetic radiation due to the spike current increases.

From the above, the resonance type switching power supply according to the prior art is applied only in a special case in which the load and the input power supply voltage fluctuation are substantially constant, and the original characteristics are not produced.

[0012]

SUMMARY OF THE INVENTION The present invention has been made in view of the above problems, and an object of the present invention is to excite a transformer in both directions by using a self-excited type at a substantially constant frequency. By the main switch means pair continuously generating the zero cross voltage switching and the resonance current having a magnitude corresponding to the load, a high efficiency conversion characteristic and an output high speed response characteristic are obtained regardless of the input fluctuation, and It is an object of the present invention to provide a resonance type power supply device that reduces the electromagnetic radiation by reducing the burden on the components. Still another object is to provide a resonance type power supply device which facilitates a multi-power supply configuration by constantly exciting a transformer.

[0013]

According to the present invention, a transformer having a predetermined primary winding, a feedback winding, and a secondary winding insulates the primary side from the secondary side and oscillates at a substantially constant frequency on the primary side. In a resonance type power supply device in which a stabilized output voltage is obtained on the secondary side, a half-input power supply for exciting a transformer in both directions and a primary winding are provided. The main switch means pair, which is configured in a bridge configuration or a push-pull configuration with a single input power source and a primary winding, and the induced voltage in the feedback winding are applied so as to form a positive feedback loop, and the main switch means pair is alternated. Delay ON timer to turn on and O to turn off after a predetermined time
A pair of timers consisting of FF timers that set an almost constant oscillation frequency, and a basic configuration on the more primary side, characterized by the main switch means pair performing zero voltage / current cross switching, and a half bridge configuration. In the primary winding and in the push-pull configuration, it is connected in series with the secondary winding, and the resonance capacitor that resonates with the leak inductance of the transformer is inserted,
By providing an intermediate capacitor connected between the output terminals of the rectifying means connected to the secondary winding in a predetermined manner to charge up to the open-circuit induced voltage of the secondary winding, the leakage inductance equivalently connected in series can be obtained. It is characterized in that the applied voltage applied to both ends of the resonance capacitor is the difference voltage between the open induction voltage of the secondary winding and the charging voltage of the intermediate capacitor.

Since this differential voltage corresponds to the size of the load, the resonance current also corresponds to it and is continuously generated, so that the average and peak ratios are small, and the burden on the components is reduced and the cost is reduced. Will have characteristics. Stabilization of positive and negative buck-boost by arrangement and combination of PWM switch means with input of intermediate capacitor end, pulse width controlled, choke coil, diode, and smoothing capacitor as output end The output means for obtaining the output, the PWM control means for stabilizing the output voltage of the output means by giving a pulse width signal corresponding to the error from the reference value of the output means to the PWM switch means, and the secondary winding are different. A clock generator generated from another secondary winding, auxiliary means including an auxiliary power supply, and a means for obtaining a more stabilized output voltage, and by making the value of the intermediate capacitor larger than the resonance capacitor and smaller than the smoothing capacitor, It has a feature that a predetermined differential voltage is generated and an appropriate resonance current is immediately generated to obtain an output fast response characteristic. In addition, the fact that the zero voltage / current cross-switching is possible also significantly reduces the electromagnetic radiation. Further, since the transformer is always excited, it has a feature that it is easy to increase the number of power sources on the secondary side.

[0015]

FIG. 3 is a diagram showing a concrete circuit configuration of an embodiment of the present invention. The input circuit 10 is a circuit that double-rectifies an AC (commercial power supply) input to obtain DC + Vp and -Vp power supplies.
D1 is the intermediate potential and is also the ground of the primary side. Note that Vp is 110 to 160 volts, where Vp is a fluctuation rate of ± 15% in Japan.

FIG. 3 shows a self-oscillation type oscillation circuit having a substantially constant frequency formed on the primary side of the transformer, and the resonance capacitor connected in series with the primary winding and the leakage inductance of the transformer are caused to resonate to respond to a load. It is intended to improve the conversion efficiency and reduce the electromagnetic radiation by sending the current / voltage of the sine arc waveform to the secondary side and executing the zero voltage / current switching.

Reference numeral 12 is a transformer, which is a primary winding Lp and a feedback winding Lf.
a and Lfb, secondary windings Ls1, Lsa and Lsb. The winding name shall also indicate the self-inductance. The black dots on the winding indicate the polarity.

Tr2a and Tr2b are FETs (field effect transistors, but bipolar transistors may be used)
, Capacitor Cp, + Vp, GND connected in series with primary winding Lp
1 and -Vp to form a half bridge and transformer 12
Is excited in both directions. Dba and Dbb are FET body diodes. Cga and Cgb are the gate-source parasitic capacitance of the FET, which is several thousand pF in the FET of several ampere class.

Although not shown, the same degree of parasitic capacitance exists between the source and drain and between the drain and gate.

Dza and Dzb are zener diodes, and resistance Rb
A bias voltage is formed by 1a and Rb1b, and a bias voltage is applied to the gates of the FET Tr2a and Tr2b via the resistors Rb2a and Rb2b to ensure self-startability when the power is turned on. Select each resistance value that does not affect the oscillation frequency after self-starting and that power consumption can be ignored.

11a and 11b are timer pairs, and Tr2a and Tr2b are alternately arranged.
It is turned on and off, and includes a delay ON timer that delays the feedback voltage from the feedback windings Lfa and Lfb, and an OFF timer that short-circuits the feedback voltage after a predetermined time. The OFF timer determines the oscillation frequency.

The delay ON timer is a FET ON gate threshold voltage and C
Use g. Cga or Cgb is charged from zero to the feedback voltage with the feedback voltage via R1a or R1b. Diodes D1a and D
1b prevents shorting the bias voltage in the feedback winding Lf. Here, if the feedback voltage is Vf, the ON gate threshold voltage is Vh, and the delay amount τon of the delay ON timer is τon = CgR1 · ln (Vf /
(Vf-Vh)). Since Cg actually varies, an external capacitance is added to improve accuracy. This τon is provided to create a so-called dead time so that Tr2a and Tr2b do not penetrate, and the leakage inductance of the winding Lp and the FE
This is because so-called zero voltage cross switching is performed in which the FET is turned on when the source-drain voltage of the FET becomes zero due to voltage resonance with the drain-source capacitance Cds of T.

This zero-voltage cross switching prevents the consumption due to the drain-source parasitic capacitance Cds which extends over several watts (although it is proportional to the square of the oscillation frequency and the input voltage Vp), so that the conversion efficiency is improved. Also, the switching element
Thermal switching becomes easy because there is only switching loss due to ON resistance or ON saturation voltage.

The OFF timer will be explained by deleting the subscripts of a and b. The transistor Tr1 is turned on via the diode D3 by the time constant of the resistor R2 and the capacitor C1 to turn on the FET gate.
The FET is turned off by short-circuiting the source and source. A feedback voltage of −Vf is generated in the feedback winding Lf corresponding to the turned off FET, and a feedback voltage of + Vf is generated on the opposite side, and the next oscillation half cycle starts.

The diode D2 controls the charging voltage of the capacitor C1.
It is intended to quickly shift from Vf to −Vf. Dio
In order to set the ON time of the transistor Tr2 to a predetermined value so that the transistor D2 is not turned off immediately, the storage effect is promoted and at the same time, the reverse breakdown voltage between the base and the emitter is secured.

The time τoff of the OFF timer is the sum of the forward voltage of the diode D3 and the base-emitter saturation voltage of Tr1 (2 when the voltage of the capacitor C1 goes from −Vf to Vf).
Since it is trapped by Vd), τoff ＝ C1R2 ・ ln (2V
f / (Vf-2Vd)). Vf requires 6V or more for the FET to be fully turned ON, and 2Vd≈2 × .5 = 1V, so Vf> 2
It becomes Vd, which does not depend on the feedback voltage of τoff ≈ 0.693C1R2. Therefore, there is a feature that a stable oscillation frequency Fo = 1 / (2τoff) can be obtained regardless of the fluctuation of the input voltage.

FIG. 4 shows the state of each of these parts. (a) Tr2a and Tr2b alternate with dead time O
Indicates the timing for N. Accordingly, (g) shows the potential difference between the connection point P between Tr2a and Tr2b and GND1. In the transient state, the resonance arc is drawn and the zero switching operation is performed. (b) shows the charging voltage of C1 on the side where the FET pair is ON. (c) shows the ON timing of one of the Tr1 pairs. (d) shows one feedback voltage of the feedback winding Lf. (e) is the exciting current that flows in the winding Lp, and the first half of the half cycle is the body diode.
It flows to one side of Db and the stored electromagnetic energy is returned to the input power source, and in the latter half it is stored as electromagnetic energy from the FET that is turned on,
There is no subtraction energy loss. As will be described later, (f) is the primary-side converted current of the load current that is close to a sine arc. A sum current of the primary-side converted current and the exciting current flows in the primary winding.

The present invention has a feature that the FET pair Tr2 can perform a zero voltage / current switching operation to reduce switching loss. However, with respect to the exciting current, the zero-voltage switching is the only switching from the transistor during the peak to the body diode. Excitation current ie = -ie0 + Vp / Lp ・
Since it can be expressed as t or ie = ie0−Vp / Lp · t, Lp with a relatively large inductance is set and the relative value is lowered with respect to the load current. However, consider the relationship between the copper loss due to the increased number of turns and the resonance frequency described later.

The above is the embodiment of the primary side of the present invention which self-excites at a substantially constant oscillation frequency in the half bridge structure, but an embodiment for making the delay ON timer more accurate will be described with reference to FIG.

FIG. 5 shows one of the pair circuits, and the same numbers and reference numerals as in FIG. 3 have the same meanings below.

In FIG. 5, when Vf is generated in the feedback winding Lf similarly to the operation of the OFF timer described above, the transistor Tr3 is turned on at τon = .693C2R3, the current flows through the resistor Rb4, and the voltage drops to P.
Type FET Tr4 also conducts. Resistance R5 smaller than resistance R1
The gate capacitance is rapidly charged via the
Let N. Diodes D4 and D5 are reverse breakdown voltage compensation. Tr1
The OFF timer, which has a component of, turns off Tr3 and Tr4 with diode D6 after τoff = .693C1R2 and turns off T3 with diode D7.
Turn off r2. The resistor Rb3 has a small value so that the transistor Tr3 does not cause a delay due to the storage effect.

FIG. 6 shows operation waveforms. (a) shows the induced voltage of the feedback winding Lf, (b) shows the charging voltage of the capacitor C2, and (c) shows the ON timing of the delay ON timer.

Next, a resonance circuit composed of a transformer leakage inductance and a resonance capacitor will be described with reference to FIG. Figure 7
(a) is an equivalent circuit including the primary side and the secondary side.
Shows a secondary side conversion equivalent circuit. Primary winding Lp and secondary winding Lsa or Lsb (self-inductance is Ls)
It is well known that the equivalent circuit of the transformer 12 can be written as 20 if the coupling coefficient k with and the turn ratio is n: 1. Reference numeral 21 is an ideal transformer that does not require an exciting current and has a transformation ratio nk: 1 / k.
Ignoring the capacitor Cp, the open circuit voltage on the secondary side is Vs = kV
p / n. This voltage is the charging voltage Vm of the intermediate capacitor Cm
Will be the maximum of. From these, the equivalent circuit of FIG. 7 (b) is obtained by viewing the equivalent circuit of FIG. 7 from the secondary side. Resonance capacitor Cr = C
p / (nk) ^ 2, and the leak inductance which is the resonance inductance is Lr = (1-k ^ 2) Ls = (1-k ^ 2) Lp / (nk) ^ 2.

Here, the equivalent load resistance in terms of voltage Vm is Re,
Assuming that Vs−Vm has occurred, the resonance current ir and the charging voltage Vc of the resonance capacitor Cr are calculated as ir≈2 (Vs−Vm) / Zj ・ exp (-αt) ・ SIN (ωt) Vc≈ (Vs −Vm) (1 + 2exp (-αt) ・ COS (ωt)), and Vm due to the load is proportional to the difference voltage Vd = Vs−Vm and the resonance current and resonance voltage are continuously generated, so the peak Since the average current ratio (1.4 to 1.7: 1) is smaller than that of the conventional method, the components are not burdened and the cost is low.

Where ω = ωr√ ((Re / 2Z) ^ 2-1), ωr
= 1 / √ (CrLr), Zj = ωLr, α = Lr / 2Re.

Now, returning to FIG. 3, the output circuit 14 and the PWM
The control circuit 16 will be described. The figure shows an example of stabilizing the output Vout by step-down switching. The PWM switch 15 is controlled by the PWM control circuit 16 with a pulse width corresponding to the error from the reference value of the output Vout. Lc is cho
C coil and Cs are smoothing capacitors. Diode Df
Creates a loop that returns the energy stored in the choke coil Lc to the smoothing capacitor Cs. This output stabilization by switching is publicly known, and the present invention is not limited to the application of this output circuit, and may be a positive / negative step-up / down stabilized output circuit by changing the arrangement / combination of the above components.

The auxiliary circuit 13 includes a clock generator that generates a clock CK every half cycle of the oscillation frequency and an auxiliary power supply of the power supply voltage Vb. This auxiliary circuit supplies power to the PWM control circuit and clocks for operation timing.

FIG. 8 shows the operating state of each of the above parts. (a) is the clock CK of the auxiliary circuit 13, which is generated every half cycle of the oscillation frequency of the system. (b) to (g) show the state under light load, and (h) to (l) show the state under heavy load. (b)
And (h) are the difference voltage Vd = Vs-Vm, and (c) and (i) are the resonance capacitors.
The charging / discharging voltage of Cr, (d) and (j) are the rectified resonance current,
(e) and (k) are the switching time width of the PWM switch 15, and (g) is
(l) shows the current iLc flowing through the choke coil Lc. Of course, the average of ir and iLc is the same, and this is the output current Iout.

As described above, since Tr2a, Tr2b, Dra, and Drb switch at zero voltage / current, there is no wasted power consumption, and the conversion efficiency is improved. In the conventional technology, 80% is fine, whereas around 90% is obtained. The current also has a characteristic that electromagnetic radiation can be considerably reduced because harmonics can be reduced because the current is close to a sine arc although it is not accurate because it is multiplied by a damping term, and an unnecessary spike due to zero switching does not occur. However, since the PWM switch 15 of the output circuit 14 and the current switch of the diode Df have large values, a countermeasure against electromagnetic radiation is necessary.

Incidentally, in a general switching power supply of the type in which an output error signal is fed back to the primary side, since the transformer is not periodically excited, basically only a single power supply can be constructed. On the other hand, in the present invention, since the transformer is always excited at a substantially constant frequency, an induced voltage is generated in each secondary winding if the number of secondary windings is increased by the above-described method. It has the feature that a plurality of power supplies with different capacities can be easily constructed.

Although the PWM operation timing is shown in FIG. 8 in synchronization with the clock of the oscillation frequency of the system, it may be essentially different. However, if the frequencies are different, the difference frequency beat may be audible.

Next, in the above description of the present invention, a half-bridge structure was used to excite the transformer, but a case of implementing the push-pull structure in FIG. 9 will be described.

In this case, the input circuit 20 is only Vp. The pair of timers 21a and 21b are the same as those described with reference to FIGS. 3 and 5, and form a positive feedback loop. The FET Tr2a and Tr2b are the same as above, and the drain-source voltage is almost zero in the ON state and 2Vp in the OFF state. The configuration for obtaining a bias voltage for ensuring self-starting property is different from that in FIG. Bias voltage is made by Zener diode Dz and resistor Rb5, and Rb6a and Rb6a
b6b and give to each gate. The transformer 22 is different from the transformer 12 in that the primary winding is Lpa and Lpb and the secondary winding is single winding Ls.

With this, Tr2a and Tr2b are alternately turned on with a dead time as in FIG. 3, and the transformer 22 is excited in both directions at a substantially constant frequency. Since the resonance capacitor Cr cannot be inserted on the primary side in principle, connect it to the single wire winding Ls,
The equivalent circuit is the same as in FIG. Since the resonance capacitor Cr needs to flow current in both directions, it is bridge-rectified by the bridge rectifier 22 and sent to the intermediate capacitor Cm. Although not shown below,
By performing the same operation as in FIG. 3, a good switching power supply with high conversion efficiency characteristics and reduced electromagnetic radiation can be obtained.

Incidentally, although delayed from the description, the method of exciting the transformers in both rows has many defects in the prior art, because the circuit is often complicated and the circuit is complicated. Further, in the self-exciting method like the present invention, there are such as a Royer circuit and a Jensen circuit using a core having a square characteristic, but the oscillation frequency changes due to the fluctuation of the input power supply voltage, and the characteristics of the present invention There is a defect that cannot be obtained.

[0046]

According to the configuration of the present invention described above, the main switch pair is alternately switched with a dead time, and a self-excited type is provided. Since the resonance current is obtained, the input / output conversion efficiency can be improved and the output high-speed response characteristic can be obtained at a low cost, and the contribution of being able to provide the switching power supply with reduced electromagnetic radiation by the zero-cross switching operation is extremely large.

[Brief description of drawings]

FIG. 1 is a diagram showing an example of a simplified resonance type power supply according to a conventional technique.

FIG. 2 is a diagram showing operation waveforms in FIG.

FIG. 3 is a diagram showing a specific circuit configuration according to an embodiment of the present invention.

FIG. 4 is a diagram mainly showing operation waveforms and timing relationships on the primary side of FIG. 3;

FIG. 5 is a diagram showing another configuration of the timer according to the embodiment of the present invention.

FIG. 6 is a diagram showing operation waveforms in FIG.

FIG. 7 is a diagram showing an equivalent circuit around a resonance circuit according to an embodiment of the present invention.

FIG. 8 is a diagram mainly showing operation waveforms and timing relationships on the secondary side of FIG. 3;

FIG. 9 is a diagram showing another embodiment in which the method of exciting the transformer of the present invention is changed.

[Explanation of symbols]

 2, 12, 22 ... Transformer 10, 20 ... Input circuit 11a, 11b, 21a, 21b ... Timer 13 ... Auxiliary circuit forming auxiliary power and clock 14 ... Output circuit 16 ... PWM control circuit GND1 , GND2 ... Grounding of primary side and secondary side Tr2a, Tr2b ... Field effect transistor Lr ... Equivalent resonance inductance Cr, Cm ... Equivalent resonance capacitor and intermediate capacitor

Claims (1)

[Claims] 1. A transformer having a predetermined primary winding, a feedback winding, and a secondary winding insulates the primary side and the secondary side to form a loop that oscillates at a substantially constant frequency on the primary side, In a resonance type power supply device that obtains a stabilized output voltage on the secondary side, a) a half-bridge configuration with a positive and negative input power supply pair that excites the transformer in both directions and the primary winding, or Main switch means pair formed by push-pull construction with a single input power source and the primary winding b) The induced voltage of the feedback winding is applied so as to form a positive feedback loop, and the main switch means pair is alternately arranged. A timer pair consisting of a delay ON timer that turns ON and an OFF timer that turns OFF after a predetermined time. A pair of timers that set a substantially constant oscillation frequency. C) A primary winding in the half bridge configuration, and a secondary winding in the push-pull configuration. Series-connected, leak-inducer of the transformer Resonance capacitor that resonates with the closet d) An intermediate capacitor that is connected between the output terminals of the rectifying means that is connected to the secondary winding in a prescribed manner and charges to the open induced voltage of the secondary winding. Output means for obtaining a stabilized output of positive and negative buck-boost by arranging and combining the PWM switch means for controlling the pulse width, the choke coil, the diode, and the smoothing capacitor that becomes the output terminal f ) PWM control means for applying a pulse width signal corresponding to the error from the reference value of the output means to the PWM switch means to stabilize the output voltage of the output means g) Another two different from the secondary winding The intermediate capacitor is composed of a clock generator generated from the next winding and auxiliary means including an auxiliary power supply, and the capacitance of the intermediate capacitor is larger than the resonance capacitor and smaller than the smoothing capacitor. A resonance type power supply device for generating a continuous resonance current, wherein an applied voltage applied between a leakage inductance of the transformer that resonates in series with a capacitor has a magnitude corresponding to a load of the output means.