JPH04340243A - Rc-error detection circuit and automatic time-constant adjustment circuit - Google Patents

Abstract

PURPOSE:To detect an RC error caused in an IC manufacturing process and to automatically adjust, by means of a detected value, variation in a time constant, which is caused by the RC error. CONSTITUTION:An RC-error detection circuit 10 and a time-constant varying circuit 50 are manufactured on the same semiconductor substrate. An electric current I which is generated from a current source 12 provided with a resistance component is charged in a capacitor CA; a changed voltage V1 is compared with a theoretical-value voltage VR. A correction voltage V as this comparison result expresses an RC error which is caused in an IC process. The correction voltage V is given to the time-constant varying circuit 50 via a control circuit 30, and its time constant is automatically adjusted.

Description

[Detailed description of the invention]

0001

[Industrial Application Field] The present invention relates to a circuit for detecting errors in resistance values of resistors and capacitances of capacitors (simply referred to as R and C) that occur in the manufacturing process of monolithic integrated circuits (hereinafter referred to as ICs), and The above RC
The present invention relates to a circuit that automatically adjusts using the detection output of an error detection circuit.

0002

2. Description of the Related Art In an IC, a resistor is formed by diffusion of impurities (diffused resistance), and a capacitor is formed by a technique such as forming a thin oxide film on a substrate and attaching a metal electrode thereon. Due to variations in diffusion, etc., variations in oxide film thickness, etc., variations in R and C occur from semiconductor wafer to semiconductor wafer.

[0003] Time constant circuit (with resistor and capacitor)
For example, in filter circuits and delay circuits, ±[(ΔR
)2 + (ΔC)2 ]1/2 error occurs.

[0004] In order to correct this error in the time constant, a variable time constant circuit is used which can control the time constant using an external voltage. As shown in Figure 1, this variable time constant circuit 5
0, the time constant was adjusted by adjusting the voltage applied from the outside using a variable resistor 59 such as a volume control at the time of factory shipment.

FIG. 2 shows an active filter as an example of a variable time constant circuit.

This active filter includes an operational amplifier 51 having transfer conductances gm1 and gm2, respectively.
, 52, capacitors (capacitors) C1, C2 for determining the time constant (cutoff frequency) along with transfer conductances gm1, gm2, variable gain amplifiers GCA1, GCA2 (hereinafter simply GC) for adjusting the time constant.
A1, GCA2), and buffer circuits 53, 5
4 and is manufactured using an IC process. The gain K of GCA1 and GCA2 is adjusted by a voltage applied from an external variable resistor 59.

This active filter has three input terminals (respectively connected to vA, vB and vC).
) and one output terminal (the output of which is denoted by vO ). The output vO is given by the following equation. However, K=1.

[0008]

[Equation 1] vO = (s2 C1 C2 vC +sC1 g
m2vB +gm1gm2vA )/
(s2 C1 C2 +sC1 gm2+gm1gm2
)...Formula 1

Inputs vA, VB and vC
By appropriately setting , it is possible to configure a low-pass filter, a high-pass filter, and a band-pass filter (hereinafter referred to as LPF, HPF, and BPF, respectively).

For example, an HPF is constructed by grounding the input terminals vA and vB. vC is the input. When vA = vB = 0, the following equation is obtained from equation 1.

[0011]

[Formula 2] vO /vC = (C1 C2 s2 )/(s2
C1 C2 +sC1 gm2+gm1gm2)
=ω0 2 /[s2 +(ω0
/Q)s+ω0 2 ]...Formula 2

0012

[Math. 3] ω0 = (gm1gm2/C1 C2)1/
2 ‥‥Formula 3

[0013]

[Equation 4] Q=(C2 gm1/C1 gm2)1/2
...Formula 4

[0014] Also, the LPF has inputs vB and vC
It is constructed by grounding the terminal of vA is the input. The following equation is obtained from equation 1 where vB = vC = 0.

[0015]

[Formula 5] vO /vA = (gm1gm2)/(s2 C1
C2 +sC1 gm2+gm1gm2)
=ω0 2 /[s2 +(ω0 /Q
)s+ω0 2 ]...Formula 5

[0016]

[Formula 6] ω0 = (gm1gm2/C1 C2)1/
2 ‥‥Formula 6

[0017]

[Equation 7] Q=(C2 gm1/C1 gm2)1/2
...Formula 7

Here, ω0 is the cutoff angular frequency, and Q is the sharpness.

A filter having arbitrary frequency characteristics can be obtained by setting the constants gm1, gm2, C1, and C2. As an example, a Butterworth filter has Q
It is obtained by determining each constant as =(1/2)1/2.

However, R (that is, gm1, gm2
), if there are process variations in C, it will not be possible to obtain the desired frequency characteristics.

Therefore, by adjusting the output currents of the operational amplifiers 51 and 52 using GCA1 and GCA2, respectively, the transfer conductances gm1 and gm2 are substantially changed to correct errors due to process variations.

[0022] The transfer conductance corrected by the gain K of GCA1 and GCA2 is g1 = Kg, respectively.
m1,g2=Kgm2.

Therefore, Equation 3 or Equation 6 is modified as follows.

[0024]

[Formula 8] ω0 = (g1 g2 /C1 C2)1/2
=(Kgm1Kgm2/C1 C2)1/2
=K(gm1gm2/C1 C2)1/2 ‥
‥Formula 8

[0025]

[Problems to be Solved by the Invention] As mentioned above, conventional variable time constant circuits require an adjustment process to change the externally set voltage to an appropriate value in order to obtain the desired time constant at the manufacturing or shipping stage. This led to an increase in costs. Further, due to RC variations that almost inevitably occur in the IC process, the relationship between the externally set voltage and the time constant is not uniform, making this IC difficult to use. This is particularly an important problem in ICs in which not only errors in the time constant are corrected using an externally set voltage, but also the time constant (frequency characteristics) itself is set to a desired value using an externally set voltage.

An object of the present invention is to provide a circuit for detecting RC errors occurring in an IC process.

Another object of the present invention is to provide a circuit that automatically adjusts a time constant error based on an RC error using a detection output of an RC error detection circuit in a variable time constant circuit.

[0028]

[Means for Solving the Problems] An RC error detection circuit according to the present invention includes a current source having a resistance component, a capacitor that is charged when a current output from the current source flows therein, and a It is equipped with a differential circuit that outputs a correction voltage representing the difference between the charging voltage of the capacitor and the theoretical value voltage corresponding to the charging voltage of the capacitor that would be obtained if there were no RC error. .

More specifically, the above RC error detection circuit is as follows:
A switching circuit that controls charging and discharging in the charging and discharging circuit consisting of the current source and the capacitor to be repeated at a predetermined cycle; It includes a sample-and-hold circuit for holding, a theoretical-value voltage generation circuit for generating the theoretical-value voltage, and the differential circuit that receives as input the voltage held by the sample-and-hold circuit and the theoretical-value voltage.

The time constant automatic adjustment circuit according to the present invention is applied to a time constant variable circuit whose time constant is variably controlled by an applied control voltage.

The variable time constant circuit and the RC error detection circuit are provided on one semiconductor substrate. The above RC error detection circuit consists of a current source with a resistance component, a capacitor that is charged by the inflow of the current output from this current source, and the charging voltage of the capacitor at a predetermined point after the start of charging, and when there is no RC error. A differential circuit is provided that outputs a correction voltage representing the difference between the theoretical value voltage corresponding to the charging voltage of the capacitor that would be obtained. The correction voltage is then applied to the variable time constant circuit as a control voltage via a control circuit.

Examples of the variable time constant circuit include a filter circuit and a delay circuit.

When the variable time constant circuit is a variable frequency active filter, it inputs the set voltage that defines the cutoff frequency of the filter and the correction voltage, and outputs a control voltage according to these two voltages. It is preferable to provide a control circuit for controlling the variable frequency active filter.

[0034]

[Operation] In the RC error detection circuit according to the present invention,
Since the current source includes a resistance component, the current generated from this current source is defined by the resistance value of the resistance component. The capacitor is charged by the output current of this current source flowing into the capacitor, so the charging voltage of the capacitor reflects the above error in the resistance value (R error) and the error in the capacitance value of the capacitor (C error). . therefore,
The difference between this charging voltage and the theoretical value voltage represents the RC error.

The time constant automatic adjustment circuit according to the present invention is applied to a time constant variable circuit whose time constant is variably controlled by an applied control voltage. This invention provides that the direction of variation in R of a plurality of resistors formed on one semiconductor wafer is in the same direction and the magnitude thereof is also approximately the same, and that a plurality of capacitors formed on one semiconductor wafer in the same manner It is assumed that the directions of variations in capacitance (capacitors) are in the same direction and the magnitudes are also approximately the same. Therefore, the RC error detection circuit is provided on the same semiconductor substrate as the variable time constant circuit.

RC output from the RC error detection circuit
By applying a correction voltage representing the error to the variable time constant circuit as a control voltage via the control circuit, the time constant error based on the RC error of the variable time constant circuit is compensated.

[0037]

According to the RC error detection circuit according to the present invention, a correction voltage representing the direction and magnitude of an RC error occurring in an IC process can be obtained. Therefore, this correction voltage can be effectively used to compensate for the time constant error generated based on the RC error in the variable time constant circuit.

According to the time constant automatic adjustment circuit according to the present invention, the time constant error of the time constant variable circuit based on the RC error caused by the IC process is automatically compensated, so that a time constant close to the design value or the theoretical value can be realized. can do. Since a process for adjusting the time constant of the variable time constant circuit is not necessary, costs can be reduced accordingly.

Further, in a circuit in which a desired time constant can be set by an externally applied voltage, the setting voltage can be set without considering a time constant error due to an RC error, making it easy to use.

Furthermore, since temperature changes in the time constant (errors due to temperature) due to temperature changes in R and C are compensated for at the same time as compensation for time constant errors based on R and C errors, time constant changes due to ambient temperature changes are compensated for. It is possible to realize a circuit with fewer

[0041]

Embodiment FIG. 3 shows an embodiment of the present invention.

A variable time constant circuit 50 and an RC error detection circuit 10 are provided in a monolithic IC circuit fabricated on one semiconductor substrate. The direction (increase or decrease in R) and the amount of variation in the plurality of resistors formed on one semiconductor wafer are almost the same. Similarly, the direction and amount of variation in capacitance of a plurality of capacitors formed on one semiconductor wafer are almost the same. The RC error detection circuit 10 includes a resistor and a capacitor, and its RC error is detected. The control circuit 30 corrects the RC error in the variable time constant circuit 50 using the detection output (corrected voltage output) representing the detected RC error.

The RC error detection circuit 10 includes a reset voltage power source 11, a constant current source 12, and a capacitor CA.
and a charging/discharging circuit composed of a switching circuit 13.

An example of the constant current source 12 is shown in FIG. This constant current source 12 is composed of transistors 21 and 22, resistors RR and RA that also serve as load resistance and bias resistance for these transistors 21 and 22, and a current mirror 23. Two resistors RR and transistor 2
1, the emitter potential of the transistor 22 becomes VCC/
2. VCC is a power supply voltage. Therefore, a current of VCC/2RA flows through the transistor 22, and this current is output from the current mirror 23. The output current I of the current source 12 is

[Equation 9] I=VCC/2
RA... is given by equation 9. It goes without saying that by exchanging the types of transistors (PNP and NPN) used in the constant current source 12, the direction of the output current can be reversed.

FIG. 5 shows the charging/discharging timing and charging/discharging voltage waveforms of the charging/discharging circuit.

By shifting the basic pulse P0 by one period and dividing the frequency by 2, two types of timing pulses P0 are obtained.
1 and P2 have been created. For example, a clamp pulse used in an image signal processing device is used as the basic pulse P0. In this case, 1 of the fundamental pulse P0
The period is equal to 1H (H is one horizontal scanning period). In this embodiment, one period of the basic pulse P0 is represented by t0. One timing pulse P1 is used as a charging/discharging control pulse for the charging/discharging circuit, and the other timing pulse P2 is used as a sampling pulse for charging/discharging voltage VX (terminal voltage of capacitor CA).

The switching circuit 13 is turned on by the timing pulse P1. As a result, the charge stored in the capacitor CA is rapidly discharged, and the terminal voltage VX of the capacitor CA becomes the reset voltage VB of the reset voltage power supply 11.

When the timing pulse P1 falls and the switching circuit 13 is turned off, the current I output from the current source 12 flows into the capacitor CA, and the capacitor CA is charged. This allows capacitor CA
The terminal voltage VX of increases.

When the switching circuit 13 is turned on by inputting the next timing pulse P1, the capacitor CA is discharged, and the terminal voltage VX of the capacitor CA is
returns to the reset voltage VB. The above operation is repeated at a cycle of 2t0.

The capacitor CA of this charge/discharge circuit is connected to the negative input terminal of the differential amplifier 15 via the switching circuit 17, and is connected to the hold capacitor 1.
Connected to 4. This capacitor 14 is externally connected outside the IC circuit. Differential amplifier 15
A theoretical value voltage VR (described in detail later) generated from a theoretical value voltage power supply 16 is applied to the positive input terminal of.

Switching circuit 17 is controlled by timing pulse P2. Timing pulse P2
When , the switching circuit 17 is turned on, and the terminal voltage VX (this is designated as VX0) of the capacitor CA when a time t0 has elapsed after the start of charging in the charging/discharging circuit is passed through the switching circuit 17. Provided and held by external hold capacitor 14.

This voltage VX0 sampled and held in the hold capacitor 14 is applied to the differential amplifier 1.
It is applied to the negative input terminal of 5. Therefore, the differential amplifier 15 outputs a correction voltage ΔV obtained by multiplying the difference between the voltage VX0 and the theoretical value voltage VR by its gain G.

[0053]

[Formula 10] ΔV=G (VR - VX0)...Formula 10

0
Now, the terminal voltage VX of the capacitor CA in the charging/discharging circuit is given by the following equation, where CA is the capacitance of the capacitor CA. t is the elapsed time from the start of charging.

[0055]

[Formula 11] VX = VB + It/CA...Formula 11

[
By substituting equation 9 into equation 11, the following equation is obtained.

[0057]

[Formula 12] VX = VB + VCCt/2RA CA
...Formula 12

Resistor RA and capacitor C
A includes the RC error of this IC circuit. The theoretical value (design value) of the resistance value of the resistor RA is R0, and the theoretical value (design value) of the capacitance of the capacitor CA is C0. I
If the errors in R and C occurring in the C manufacturing process are α and β, respectively, RA and CA are given by the following equations.

[0059]

[Formula 13] RA = αR0...Formula 13

[0060]

[Formula 14] CA = βC0...Formula 14

Substituting Equations 13 and 14 into Equation 12 yields the following.

[0062]

[Formula 15] VX = VB + VCCt/2αβR0 C0 ‥‥Formula 15

Voltage VX0 when t=t0 is given by the following equation.

[0064]

[Formula 16] VX0=VB +VCCt0 /2αβR0 C0 ‥
‥Formula 16

On the other hand, the theoretical voltage VR applied to the positive input terminal of the differential amplifier 15 is set to a value equal to the voltage VX0 that would be obtained if there were no RC error. That is,

[0066]

[Formula 17] VR = VB + VCCt0 /2R0 C
0...Formula 17

By substituting equations 16 and 17 into equation 10, the following equation is obtained.

[0068]

[Formula 18] ΔV=(GVCCt0 /2R0 C0) [1-
(1/αβ)] Formula 18

FIG. 6 shows a part of the charging waveform of the terminal voltage VX of the capacitor CA in the charging/discharging circuit. The waveform shown by a solid line is the case where there is no RC error, and the waveform shown by the broken line is the case where there is an RC error. It will be understood that the RC error generated in the manufacturing process of the IC circuit appears as the terminal voltage of the capacitor CA, and this appears in the correction voltage ΔV shown by Equation 18.

Let the time constant of the time constant variable circuit 50 be τm, and the resistance component and capacitance component that define this time constant τm be Rm and Cm, respectively.

[0071]

[Formula 19] τm = 1/Rm Cm...Formula 19

00
72] These Rm and Cm also include errors due to the IC manufacturing process, and these errors can be considered to be equal to the errors α and β of the resistors RA and CA included in the error detection circuit 10. Therefore, if the theoretical value of the time constant in the variable time constant circuit 50 is τ0, and the theoretical values of the resistance component and capacitance component are Rm0 and Cm0, respectively, then the following equation holds true.

[0073]

[Formula 20] τm = 1/Rm Cm = 1/αRm0βCm0
=(1/αβ)(1/Rm0Cm0) =(
1/αβ)τ0...Formula 20

On the other hand, the correction voltage ΔV is multiplied by the gain GC by the control circuit 30 and applied to the variable time constant circuit 50. It is assumed that the time constant τm of the variable time constant circuit 50 changes according to the voltage given from the control circuit 30 according to the following equation.

[0075]

[Formula 21] τm = τ0 (1+GC ΔV)...Formula 2
1

[0076] According to equations 20 and 21,

[Formula 22] ΔV=(1/GC) [(1/αβ)-1] Formula 22
is obtained.

By eliminating α and β from equations 18 and 22, the following equation is obtained.

[0078]

[Formula 23] G=-(1/GC)[1/(VCCt0/2R
0 C0 )]...Formula 23

It will be understood that by determining the gains G and GC to satisfy Equation 23, the time constant in the variable time constant circuit 50 is automatically adjusted to be equal to the theoretical value.

FIG. 7 shows an embodiment in which the RC error detection circuit and time constant automatic adjustment circuit according to the present invention are applied to a variable frequency active filter. In this figure, the same components as those shown in FIGS. 2 and 3 are denoted by the same reference numerals, and explanations thereof will be omitted. Also, the time chart shown in FIG. 5 applies as is.

The active filter can be configured as HPF, LPF, and BPF by appropriately setting the three input terminals (vA, vB, vC) as described above. In the variable frequency active filters of this embodiment, the cutoff frequencies of these filters can be arbitrarily set by a setting voltage VS applied from the outside.

That is, the control circuit 30 functions to control the cutoff frequency of the frequency variable active filter to match the set frequency, and to compensate for variations in the cutoff frequency due to RC errors caused by the IC process. .

A variable resistor 33 is provided outside the IC circuit, and the set voltage VS is set by this variable resistor 33. The set voltage VS is applied to the positive input terminal of the differential amplifier 31. A reference voltage ER is applied to the positive input terminal of the differential amplifier 31. Differential voltage VS -ER from differential amplifier 31
A voltage proportional to is output and applied to the negative input terminal of the differential amplifier 32.

A correction voltage ΔV output from the RC error detection circuit 10 is applied to the negative input terminal of the differential amplifier 32. The differential amplifier 32 multiplies the difference between the output voltage of the differential amplifier 31 and the correction voltage ΔV by a gain GC and outputs the multiplied difference, and provides the output voltage to GCA1 and GCA2. As a result, the cutoff frequency is automatically adjusted according to Equation 8.

In this way, the set voltage VS can be maintained regardless of the RC variations occurring in the IC process.
Since a filter with a cutoff frequency defined by is obtained, it is very easy to use.

[Brief explanation of drawings]

FIG. 1 is a circuit diagram showing a conventional time constant adjustment method in a variable time constant circuit.

FIG. 2 is a circuit diagram showing a conventional cutoff frequency adjustment method in an active filter.

FIG. 3 is a circuit diagram showing the configuration of an RC error detection circuit and a time constant automatic adjustment circuit according to the present invention.

FIG. 4 is a circuit diagram showing a specific configuration example of a current source.

FIG. 5 is a time chart showing charging/discharging timing and sampling timing in the RC error detection circuit.

FIG. 6 is a graph showing an example of a charging waveform.

FIG. 7 is a circuit diagram showing an embodiment in which the present invention is applied to a variable frequency active filter.

[Explanation of symbols]

10 RC error detection circuit 12 Constant current source 13, 17 Switching circuit 14 Hold capacitor 15 Differential amplifier 16 Theoretical value voltage generation circuit 30 Control circuit 50 Time constant variable circuit CA capacitor

Claims (4)

[Claims] [Claim 1] A current source with a resistance component, a capacitor that is charged by the inflow of current output from this current source, and the charging voltage of the capacitor at a predetermined point after the start of charging, assuming that there is no RC error. an RC error detection circuit, comprising: a differential circuit that outputs a correction voltage representing a difference between a theoretical value voltage corresponding to the charging voltage of the capacitor that would be obtained; 2. A switching circuit that controls charging and discharging in a charging and discharging circuit composed of the current source and a capacitor to be repeated at a predetermined cycle, and charging the capacitor after a predetermined period of time has elapsed from the start of charging in the charging and discharging circuit. Sample/Hold voltage
2. The differential circuit according to claim 1, further comprising a hold circuit, a theoretical voltage generation circuit that generates the theoretical voltage, and the differential circuit that receives the voltage held by the sample-and-hold circuit and the theoretical voltage as inputs. RC error detection circuit. [Claim 3] A time constant variable circuit whose time constant is variably controlled by an applied control voltage and an RC error detection circuit are provided on one semiconductor substrate, and the RC error detection circuit has a resistance component. A current source, a capacitor that is charged by the current output from this current source,
and a differential circuit that outputs a correction voltage representing the difference between the charging voltage of the capacitor at a predetermined time after the start of charging and the theoretical value voltage corresponding to the charging voltage of the capacitor that would be obtained when there is no RC error. , a time constant automatic adjustment circuit in which the correction voltage is applied to the time constant variable circuit as a control voltage via a control circuit. 4. The variable time constant circuit is a variable frequency active filter, and the control circuit receives as input a set voltage that defines a cutoff frequency of the filter and the correction voltage, and responds to these two types of voltages. 4. The time constant automatic adjustment circuit according to claim 3, wherein the time constant automatic adjustment circuit outputs the control voltage and applies it to the variable frequency active filter.