JPS59132218A - Switched capacitor - Google Patents

Abstract

PURPOSE:To improve remarkably the S/N of a power supply by increasing the capacitance of only required part among capacitors constituting a switched capacitor filter. CONSTITUTION:Since a noise signal N<5> of an operational amplifier OPA5 is increased further because of a positive feedback loop in which the signal is fed back via a feedback capacitor C30 of an operational amplifier OPA3, the S/N of a power supply is deteriorated. It is required to suppress a noise signal N3 generated in the operational amplifier OPA3 to improve the deterioration. Then, an integration capacitor CH3 connected between an input and an output of the operational amplifier OPA3 is increased. Further, a noise signal inputted to the operational amplifier OPA3 is suppressed at the same time. Thus, it is required to suppress a noise signal N1 generated in the operational amplifier OP1 and then to increase an integrating capacitor CH1.

Description

[Detailed Description of the Invention] (Technical Field of the Invention) This invention relates to a switched carrier filter (hereinafter referred to as
``s, c, and F'', and particularly relates to zero point forming types S, C, and F having zero points in the transfer function.

(Background of the invention) S, C, and F are composed of a combination of an operational amplifier, a switch, and a capacitor, and the integrator that forms the basic circuit has the configuration shown in Figures 1 to 3. It is used.

1 is the operation multiplier 14] unit, 2 , , ? , 4 , 5 ,
6 is a switch, CHi O” C1130 is an integral capacitor,
Cs1o-c83o are sampling capacitors, Vin”
out are the input and output signals, respectively.The integrator shown in Figure 1 is a parallel type negative phase integrator, the one in Figure 2 is a serial type positive phase integrator, and the one in Figure 3 is a serial type negative phase integrator. Each phase integrator is shown. 'S, C, and F are constructed by combining these integrators. As a basic property of S, C, and F, the input/output transfer function is expressed by the ratio of the sampling capacitance to the integral capacitance, so it is more important to design that ratio accurately than the value of each capacitance alone. Become.

Therefore, in view of the fact that conventional S, C, and F are realized as integrated circuits in most cases, a method is adopted in which the Zanzo ring capacitance and the integral capacitance are configured as integral multiples of the basic capacitance. ing.

The size of the basic intermediate capacitance itself does not affect the transfer function between input and output, so in the conventional S, C, and F, all integrators were designed to have the same basic unit capacitance (~.However, the power supply noise When considered based on the idea of designing the removal ratio (hereinafter referred to as p, s, R, R) to be high, it is not necessarily best to arrange all the basic trophic capacities to have the same size.

The power supply noise signal appearing at the outputs of S, C, and F via each operational amplifier is directly related to each sampling capacitance value and each integral capacitance value. Therefore, when designing S, C, and F in consideration of P, S, R, and R, in addition to the input/output transfer function, the transfer function from the power source to the output is also an important factor.

That is, depending on the value of the basic unit capacitance that constitutes the integrator, P,
The S, R, and R1% properties change significantly.

If only the improvement of P, s, R, R characteristics is considered, it is sufficient to increase the basic capacitance described above, but this inevitably increases the area occupied by the capacitance, which is disadvantageous especially when realized as an integrated circuit. becomes.

The inventors ('!, As a result of repeated various experiments in order to improve the P, S, R, R characteristics, all the capacitances constituting S, C, F were equal to P, S, R, It was discovered that this is not related to the R characteristic.

Based on this fact, in order to find out which capacitance is dominantly involved in the P, S, R, and R characteristics, noise sources are inserted into various parts of the circuit network, and the values of the capacitances are individually varied. Repeat the experiment even more.

As a result, we have identified the existence of a capacitance that greatly contributes to the improvement of P, S, R, , R characteristics. This invention was developed based on the knowledge of the inventors, and the conventional S
, F, and C, there is no known one that focuses on improving the P, S, R, and R characteristics.

(Object of the Invention) The object of the present invention is to improve the P, S, R, and R characteristics of S, C, and F while minimizing the increase in the area occupied by the capacitors that constitute S, C, and F. is to provide.

(Summary of the Invention) In this invention, the basic m-order capacitance of the capacitors constituting each integrator is appropriately changed depending on the integrator, and is composed of a combination of a plurality of operational amplifiers, switches, and capacitors. In the zero-point forming type switched canister filter, the band of the filter is varied by varying the ratio between the sampling capacitor and the integral capacitor connected to the operational amplifier, the sampling capacitor and the integral capacitor being is a set of a plurality of basic unit capacitors, and among the operational amplifiers involved in zero point formation, the operational amplifier whose output is directly connected to the input of another operational amplifier via a feedback capacitor. The basic level capacitance is relatively large compared to the basic unit capacitance of other operational amplifiers.

Hereinafter, this invention will be explained in detail based on examples.

(Embodiments of the Invention) First, prior to explaining the embodiments, the design procedure for S, C, and F will be explained. There are various methods for designing S, C, and F, but the explanation here will be based on the LC simulation method, which is designed to correspond to an LC filter.

FIG. 4 shows a circuit diagram of a fifth-order low-pass LC filter having two transmission zeros. The voltage at each node is v,
# v21 v3, and the current flowing into or out of each node is defined as shown by the arrow in Figure 2, and the relational expressions of these voltages and currents are obtained as Equations (1) to (7). holds true.

■ I r = (V inV 1 )−・・・・・・・・・・
·······················Mountain(
1) I2 = (VI V2 ) SC2+ (VI
Va ) −・−・・(2) SL2 Ia = (V2 Vs,) SC4+(V2 Va )
±・'・=−(3) SL.

14-V3・-・・・・・・・・・・・・・・・・
・・・・・・・・・・・・・・・・・・(4)v I-(r
+ −I 2 ) −・−==・=−=−=−・−−
---" (5) CI V2= (I, -I3)□ ・・・・・・・・・・・・
・・・・・・・・・・・・・・・・・・(6) C3 v3= (I3-I,)□ ・・・・・・・・・
(7) SC6 Here, S=, jω (ω: angular frequency).

Substituting equations (1) to (4) into equations (5) to (7) and arranging them respectively, we get ・・・・・・・・・・・・・・・(8) Here, R=G
It has something to do with swords.

Realizing these (8) to αO equations using an RC integrator 1. This is shown in Figure 5. The transmission zero point (attenuation pole) created by the combination of (C2-L2) and (C4, L4) shown in Figure 4 is the operational amplifier OPA 7 and OPA in Figure 5.
,? Also, a feedback capacitor C2 connecting between both operational amplifiers OPA J and OPA 3 and operational amplifiers OPA 3 and O
PA 5 and both operational amplifiers OPA 3. and a feedback capacitor C4 connecting between the OPAs 5 and 5. The feedback roof 0 due to these feedback capacitances C2 and C are both positive feedback. When the circuit shown in FIG. 5 is converted into S, C, and F, the circuit shown in FIG. 6 is realized. Among the capacitors shown in FIG. 6, CANCK is a sampling capacitor, CHI to CH5 are integral capacitors, and CI. ~C4G is a feedback capacitance.

The ratio of the sampling capacity and the integration capacity that make up the integrator is
Its value is determined according to the RC product of each integrator shown in FIG. 5, and it is determined by the element value (C+ ~
Cs * L2 + L4 + R).

S, C, and F shown in Fig. 6 designed in this way make the frequency characteristics (transfer function) of the LC filter shown in Fig. 4 the same, but generally when this is realized, each integrator is Since the dynamic ranges of the constituent operational amplifiers are not uniform, it is necessary to perform an operation called scaling to equalize the dynamic ranges.

FIG. 7 shows a circuit diagram after performing a scaling operation. Each operational amplifier OPA I~OPA
Let the output voltage levels of 5 be V1 to V, and when adjusting each output voltage level to 1/C1 to 1/α7 times, the capacitance connected to the output of the operational amplifier is C1.
Multiply by (i=1 to 5). For example, the output voltage level
Let's consider the change in capacity when doubling the capacity.

As mentioned above, the characteristics of S, C, and F are determined based on the ratio of the sampling capacitance and the integral capacitance, so the operational intensifier dJ
Considering the sampling capacitor connected to the input of PA5, the charge accumulated in the integral capacitor CH5 is the sum of the charges accumulated in the capacitors CI, CK, and CoO.

Therefore, the capacity ratio before scaling is CI:CK:C40:
If CH3-1=1:4:10, after the scaling operation, CI:CK:C40:CH5=1:0.
5:4:5, converted to integer ratio cr:CK:C40:CH
5=2:1:8:10. This means doubling the capacitances CI and C40 isometrically. Next, a method for designing S, C, and F in consideration of power supply noise signals will be explained.

Figure 8 shows the equivalent circuit of S, C, and F when it is assumed that a smear source noise signal appears at the output of each operational amplifier OPA1 to OPA5 making up the integrator due to a change in the power supply voltage. be.

Output of S, C, F■. The noise signals N1 to N5 of each operational amplifier appear superimposed on ut, and the magnitude of the noise signals N1 to N5 becomes smaller as the integral capacitance value thereof becomes larger. This is because the integral capacitor functions as a negative feedback in each operational amplifier. The noise signal N3 generated in the operational amplifier OPA3 is transmitted through the feedback capacitor C4.
Operation amplifier OP according to the capacitance ratio between 0 and integral capacitance CH5.
A5, the larger the capacitance value of the feedback capacitor C40 becomes due to the above-described scaling, the more the noise signal N5 contained in the operational amplifier OPA5 increases.

In addition, the noise signal N5 of the operational amplifier dJ OPA 5 forms a positive feedback loop that is fed back to the operational amplifier OPA 3 via the feedback capacitor C30, so the noise signal N5 becomes even larger, and P , S , R , which deteriorates the R characteristics. In order to improve this, it is necessary to suppress the noise signal N3 generated in the operational amplifier OPA3.

For this purpose, it is sufficient to increase the integral capacitance CH3 connected between the input and output of the operational amplifier OPA3.

It is also necessary to simultaneously suppress the noise signal input to the i/L operational amplifier OPA3.

For this purpose, it is sufficient to suppress the noise signal N1 generated in the operational amplifier OPA1, and therefore, it is necessary to increase the integral capacitance CHI based on the same concept as described above.

In the circuit shown in FIG. 8, five operational amplifiers are used to form the integrator, but the operational amplifiers OPA 1 and OPA 3 require an increase in their capacity.

Integral capacitance CH1, CH31CH connected to OPA 5
There are only 5. If this is expanded and applied more generally, among the operational amplifiers involved in the formation of the zero points of S, C, and F, the output of the operational amplifier will be connected to the input of the other operational amplifiers via the feedback capacitance. If the integral capacity of the operational amplifier directly connected to the amplifier is increased, the result will be 1C.

However, even if the integral capacity is increased, it cannot be increased by itself, and the ratio between the sampling capacity and the integral capacity must be maintained. For this purpose, the basic unit capacity constituting these may be increased.

FIG. 9 shows P between S, C, and F according to the conventional method and the present invention.
It is a characteristic diagram showing a comparison of S, R, and R. Curve A shows the case where the basic unit capacitances of the circuits shown in FIG. 7 are the same, and curve B shows the case where the basic unit capacitance of the necessary portion is set to double.

As is clear from the figure, improvements were seen in P, S, R, and R.
In particular, a significant improvement of about 7 to 8 dB was observed at frequencies around 3.5 KHz.

(Effects of the Invention) As described above in detail based on the embodiments, in this invention, the capacity of only the necessary portion of the capacity of S, C, and F is increased. ,R,R
All types (low-pass type, high-pass type) can dramatically improve the characteristics.

It can be used for S, C, and F of bandpass type).

[Brief explanation of the drawing]

Figures 1 to 3 show the basic circuit of an integrator used for S, C, and F. Figure 4 is a circuit diagram of a 5th-order low-pass LC filter. Figure 5 shows the circuit in Figure 4 as an RC integrator. Figure 6 is a circuit diagram of the circuit in Figure 5 realized as S, C, F. Figure 7 is a circuit diagram of S, C, F after scaling.
The figure is an equivalent circuit diagram of S, C, and F in consideration of the power supply noise signal.
, R, and R are compared. OPA 1 to OPA 5... operational amplifier, Cl0-C
40...Feedback capacitance, CA-CK/sampling capacitance, C
H2=CH, 5... Integral capacity. Patent applicant Oki Electric Industry Co., Ltd. Figure 1 CHl. Figure 3

Claims (1)

[Claims] By varying the ratio of the sampling capacitance and the integral capacitance connected to the plurality of arithmetic amplifiers [1j units, switches and capacitors] connected to the arithmetic amplifiers 11, the filter band can be adjusted. In the zero-point forming type switched capacitor filter that varies the value of The one operational amplifier that is directly connected to the input of another operational amplifier via is characterized in that the unit capacity is relatively larger than the unit capacity of the other operational amplifiers. Switched
Canozo Shita Filter.