CN112332803B

CN112332803B - Active low-pass filter bandwidth calibration circuit

Info

Publication number: CN112332803B
Application number: CN202011190650.9A
Authority: CN
Inventors: 姚明; 王友华; 刘万福; 张凯; 张然; 李航标; 赵晓冬; 刘智卿
Original assignee: Southwest Electronic Technology Institute No 10 Institute of Cetc
Current assignee: Southwest Electronic Technology Institute No 10 Institute of Cetc
Priority date: 2020-10-30
Filing date: 2020-10-30
Publication date: 2023-02-03
Anticipated expiration: 2040-10-30
Also published as: CN112332803A

Abstract

The invention discloses a bandwidth calibration circuit of an active low-pass filter, aiming at adjusting the bandwidth in a wider range and keeping the bandwidth accurate and constant. The invention is realized by the following technical scheme: two sinusoidal signals with different amplitudes and frequencies are respectively generated at the input ends of the I/Q two paths of orthogonal channels, wherein the frequency of the I path of signals is the target bandwidth frequency to be calibrated. After the two paths of signals pass through the two paths of active filters, the peak value detection circuit detects amplitude information of the two paths of signals, the amplitude information is sent to the comparator to be compared, and a comparison result is sent to the finite-state machine to be processed. The bandwidth of the filter is detected by controlling and changing the capacitance value of an adjustable capacitance unit in the active filter through an algorithm in a finite-state machine. And detecting and iterating the bandwidth of the filter for multiple times according to the result output by the comparator until all the bits of the adjustable capacitor unit traverse, and stopping iterating. A more accurate filter bandwidth that is insensitive to circuit PVT variations may be obtained.

Description

Active low-pass filter bandwidth calibration circuit

Technical Field

The invention belongs to the technical field of filter calibration, and relates to a calibration circuit for an active low-pass filter bandwidth, which can be applied to a System On Chip (SOC) of a wireless radio frequency transceiver.

Background

Low-pass filtering (Low-pass filter) is a filtering method, in which the rule is that Low-frequency signals can normally pass through, and high-frequency signals exceeding a set threshold are blocked and attenuated. But the magnitude of the blocking and attenuation will vary depending on the frequency and filtering procedure (purpose). It is sometimes also called high-cut filter or top-cut filter. There are many types of low-pass filters. Among them, the most common are butterworth filters and chebyshev filters. The active low-pass filter circuit can eliminate high-frequency noise signals, retain direct current and low-frequency response signals, and has stronger high-frequency inhibition capability and higher response speed than an RC filter circuit. The active low pass filter circuit is usually composed of an integrated operational amplifier and passive element resistors and capacitors. It allows signals from zero to some cutoff frequency to pass unattenuated while suppressing signals at other frequencies. The active low-pass filter circuit can be used for filtering high-frequency interference signals and is widely applied to a digital-analog hybrid System On Chip (SOC) chip. In recent years, with the rapid advance of integrated circuit technology, circuit performance is continuously improved, and particularly with the continuous development of systems on a chip (SOC), a digital-analog hybrid circuit and a radio frequency integrated circuit can be integrated on one chip to realize complete transceiving and signal processing functions. At present, domestic and foreign researches on integrated active filters mainly focus on: wide band filter design, auto-tuned filter design, reconfigurable filters, current-mode filters, and filter designs incorporating several of the above features. In some applications requiring broadband, such as new IEEE communication standard, power consumption becomes a main consideration, a high gain-bandwidth product of an operational amplifier can suppress gain ripple in a passband, and a high gain-bandwidth product means high power consumption. The architecture of the transceiver is selected related to the communication standard and the performance requirement of the transceiver, on the premise of ensuring the performance, the transceiver is required to have the characteristics of low power consumption, high integration level, small size and low price, the radio frequency front end has a decisive effect on the performance of the transceiver, and the filter is generally positioned at the front stage of a post-stage variable gain amplifier of down-conversion and the front stage of a pre-distortion amplifier of an up-conversion mixer. The gain, bandwidth, noise and linearity of the filter directly affect the performance of the transceiver, and different transceivers often require different filter structures.

With the slow development of independent components to the high integration of systems, the integrated active filter is also developed from an independent chip to the high integration of complex systems. Thus, designing a high performance filter suitable for system integration becomes an urgent task. Currently, there are many implementation forms of integrated active filters, and active RC filters are more suitable for modern transceiver systems due to their high linearity and high dynamic range. In modern communication systems, the wanted signal is subject to interference by a large number of out-of-band signals, and therefore the receiver usually requires a channel selection filter to select the wanted signal. The circuit structure of the active RC filter with the characteristic of high dynamic range is selected, and the frog leaping type structure is utilized to synthesize the passive filter, so that the sensitivity of the filter characteristic to the feedback resistor and capacitor in the circuit is minimum. In an actual circuit, the phase and amplitude frequency characteristics are often greatly different from the ideal conditions under the influence of the integrated circuit device process, the ambient temperature and the device service life. With the development of CMOS Process and integrated circuit design, the size of the integrated circuit is continuously reduced, and the influence of PVT (Process and performance) fluctuation on the integrated circuit is more and more serious. After the integrated circuit enters deep submicron, the electric leakage of the transistor is increasingly serious, and the widely used portable electronic equipment is in a standby state for most of time. PVT fluctuations arise mainly from the manufacturing flow of the integrated circuit and the actual environment in which the circuit operates, and since PVT fluctuations affect the performance, stability and power consumption of the integrated circuit, they must be taken into account when designing the circuit. The bandwidth of the filter is mainly determined by its time constant RC. If the standard value of the filter design time constant is RC, due to uncertain factors such as process and temperature, the actual RC will deviate due to the influence of chip process deviation (Δ P), power supply voltage fluctuation (Δ V), temperature variation (Δ T), PVT for short, and aging of devices on the chip, which causes the bandwidth of the filter on the chip to deviate from the design requirement. And the offset bandwidth can be compensated and calibrated by adaptively adjusting the value of R or C through a proper algorithm. The bandwidth of the filter can be flexibly and accurately adjusted, and the application range of the transceiver is wider.

The invention discloses a bandwidth calibration circuit of an active low-pass filter, wherein the bandwidth of the filter can be calibrated according to the influence of PVT, and meanwhile, the bandwidth of the filter can be flexibly and accurately adjusted.

Disclosure of Invention

In order to solve the above problems, an object of the present invention is to provide an active low-pass filter bandwidth calibration circuit capable of adjusting a bandwidth in a wide range and maintaining a precise constant, in response to PVT influences.

The above object of the present invention can be achieved by the following technical solutions: an active low pass filter bandwidth calibration circuit comprising: I. the bandwidth-adjustable second-order active low-pass filter LPF arranged on the Q orthogonal channel, the two active low-pass filters LPF on the I, Q channel and the FSM connected in parallel and terminated with the two active low-pass filters LPF are characterized in that: two sinusoidal signals with different amplitudes and frequencies are respectively input to the input ends of two paths of orthogonal channels I, Q, the two paths of sinusoidal signals firstly respectively pass through bandwidth-adjustable second-order active low-pass filters LPF on I, Q two paths, the two paths of sinusoidal signals output by the LPFs are respectively sent to two peak value detection circuits on the output ends of I, Q two paths of LPFs, the amplitude information of the two paths of signals is detected and sent to a comparator which is commonly connected with the output ends of the two peak value detection circuits for comparison, and the comparator sends a comparison result to a Finite State Machine (FSM) connected with the output end of the comparator for processing; and the FSM controls and changes the capacitance value of the adjustable capacitor unit in the LPF through an algorithm circuit in the FSM according to the comparison result, detects and adjusts the bandwidth of the filter, approaches the amplitude value of the I path signal attenuated by the LPF to the amplitude value of the Q path signal, detects and continuously iterates the bandwidth of the LPF according to the result output by the comparator for many times until all bits of the adjustable capacitor unit are traversed, and stops iterating to obtain more accurate filter bandwidth insensitive to PVT change.

Compared with the prior art, the invention has the following beneficial effects:

by the calibration method, the bandwidth of the active low-pass filter can be adjusted in a wider range according to the filtering requirement of the input signal;

the invention can ensure that the design designated bandwidth of the active low-pass filter is not influenced by PVT and keeps accurate and constant.

Drawings

The invention is further described below with reference to the accompanying drawings.

Fig. 1 is a general circuit block diagram of the active low pass filter bandwidth calibration circuit of the present invention.

FIG. 2 is a graph comparing the spectrum of the output signal of the I/Q channel when the actual bandwidth of the LPF of FIG. 1 is equal to the calibration target bandwidth.

FIG. 3 is a graph of a comparison of the spectrum of the output signal of the I/Q channel when the actual bandwidth of the LPF of FIG. 1 is greater than the calibration target bandwidth.

FIG. 4 is a comparison graph of the spectrum of the output signal of the I/Q channel when the actual bandwidth of the LPF of FIG. 1 is less than the calibration target bandwidth.

Fig. 5 is a schematic diagram of the second-order active low-pass filter of the LPF of fig. 1.

Fig. 6 is a schematic diagram of the basic components of the tunable capacitor array of fig. 1.

Fig. 7 is a flow chart of the calibration algorithm of fig. 1.

Detailed Description

See fig. 1. In a preferred embodiment described below, an active low pass filter bandwidth calibration circuit comprises: I. the active low pass filter comprises a second-order active low pass filter LPF with adjustable bandwidth arranged on a Q orthogonal channel, two active low pass filters LPF on a I, Q channel and a finite state machine FSM connected in parallel and terminated with the two active low pass filters LPF. Two sinusoidal signals with different amplitudes and frequencies are respectively input to the input ends of two paths of orthogonal channels I, Q, the two paths of sinusoidal signals firstly respectively pass through bandwidth-adjustable second-order active low-pass filters LPF on I, Q two paths, the two paths of sinusoidal signals output by the LPFs are respectively sent to two peak value detection circuits on the output ends of I, Q two paths of LPFs, the amplitude information of the two paths of signals is detected and sent to a comparator which is commonly connected with the output ends of the two peak value detection circuits for comparison, and the comparator sends a comparison result to a Finite State Machine (FSM) connected with the output end of the comparator for processing; and the FSM controls and changes the capacitance value of the adjustable capacitor unit in the LPF through an algorithm circuit in the FSM according to the comparison result, detects and adjusts the bandwidth of the filter, approaches the amplitude value of the I path signal attenuated by the LPF to the amplitude value of the Q path signal, detects and continuously iterates the bandwidth of the LPF according to the result output by the comparator for many times until all bits of the adjustable capacitor unit are traversed, and stops iterating to obtain more accurate filter bandwidth insensitive to PVT change.

The frequency of the sinusoidal signal input from the input end of the I-path LPF is equal to the standard bandwidth BW = f to be calibrated _-3dB The same, the frequency of the sine signal input by the input end of the Q-path LPF is f _-3dB Of amplitude I of the sinusoidal signal

And (4) multiplying. The two inputs of the whole bandwidth calibration circuit I, Q are used as the calibrated sinusoidal signal utilization can be generated by the digital-to-analog converter (DAC) of the pre-stage circuit of the LPF. Frequency of the I-path sinusoidal signal and standard bandwidth BW = f to be calibrated _-3dB Same, amplitude VI = V ₁ (ii) a The frequency of the Q path sinusoidal signal is f _-3dB /32, amplitude

See fig. 2, 3 and 4. After the two paths of I/Q sinusoidal signals pass through the two paths of I/Q LPFs, if the actual bandwidth BW of the LPFs is ₀ Not affected by PVT, but still the standard bandwidth f to be calibrated _-3dB Then, since the LPF has a 3dB bandwidth, the gain will be reduced by 3dB, so that the I output signal amplitude VI _ out will be attenuated to

The Q path signal has a frequency far lower than f _-3dB The amplitude VQ _ out will not be attenuated after it passes through the LPF, and will still be

So that the I/Q signals are equal in amplitude after passing through the LPF, i.e., VI _ out = VQ _ out, as shown in fig. 2. However, if the LPF is affected by PVT, its actual bandwidth BW ₀ May be greater or less than f _-3dB Thus, is alignedIt is also possible that VI out is greater or less than

As shown in fig. 3 and 4, respectively.

Then, the amplitude information VI _ out and VQ _ out of the output signals of the two filters are detected by peak value detection circuits respectively and then are sent to a comparator for comparison. The comparator outputs 1 or 0 according to the magnitude conditions of VI _ out and VQ _ out, then the comparison result is sent to a finite-state machine for processing, and an algorithm in the finite-state machine controls and changes the capacitance value of an adjustable capacitor array unit in the active low-pass filter according to the comparison result, so that the bandwidth of the filter changes towards the direction of reducing the difference between VI _ out and VQ _ out. Thus, the bandwidth of the filter is continuously subjected to a plurality of detection iterations according to the result output by the comparator until the traversal is finished, so that the VI _ out can be close to the VQ _ out, even if the actual bandwidth BW of the filter ₀ Approaching the target bandwidth BW to be calibrated can obtain a more accurate bandwidth calibration result. The total number of iterations is equal to the total number of bits of the basic unit of the tunable capacitor array in the filter,

refer to fig. 5 and 6. As shown in fig. 5. The active low pass filter LPF may employ a second order active butterworth filter structure of an on-chip differential structure. The second-order active Butterworth filter is formed by cascading two stages of filter sections including transimpedance operational amplifiers OP1 and OP 2. The first stage filter section is symmetrically provided with two basic adjustable capacitor units C connected in parallel ₀ Form a total capacitance of 2C ₀ The capacitor array of (1); the second stage filter section is symmetrically provided with a single basic adjustable capacitor unit C ₀ Formed capacitor array and resistor R _F . The positive and negative output ends of the first-stage OP1 reverse proportion amplifying circuit pass through two crossed voltage-dividing resistors R _C The negative and positive input ends of the inverse proportional amplifying circuit of the second stage OP2 are connected in series, and the input end of the first stage OP1 passes through two resistors R respectively _A Connected in parallel to the output of the second stage OP 2. The 3dB bandwidth of the second order active butterworth filter can be expressed as:

it can be seen thatChanging the resistance or C ₀ The 3dB bandwidth of the filter can be adjusted. In the actual digital-analog mixed chip, the resistance is mostly a fixed value, and C ₀ Typically tunable, and its internal capacitor array is constructed as shown in fig. 6. The capacitance of the array is increased by a factor of M of 2 (M is from 0 to N), and is divided by a small fixed capacitor C _fix In addition, the other capacitor arrays are controlled by a switch control signal ADJ<N:0>And (5) controlling. If the switch control signal ADJ<N:0>The corresponding decimal integer is k (k is more than or equal to 0 and less than or equal to 2) ^N+1 -1) then C ₀ Can be represented as C _fix + k Δ C, so that a traversal of the capacitance values in the range of integer k times of the basic capacitance unit Δ C can be realized.

See fig. 7. In the primary calibration process, the calibration algorithm adopts a dichotomy control to change the adjustable capacitor unit according to the value output by the comparator, and the bandwidth of the filter is tested. Firstly, the capacitance value of the adjustable capacitance unit is set as the median value of the maximum capacitance variable range, and a switch control signal ADJ is initialized<N:0>=100 … 000, i.e. highest position 1, remaining bits remain 0. At this time, if the signal amplitude VI _ out output by the I channel LPF and the signal amplitude VQ _ out output by the Q channel LPF satisfy: VI _ out>VQ _ OUT, the comparator output result COMP _ OUT =1, indicating the current bandwidth BW of the filter ₀ Greater than the calibrated standard value f _-3dB Thus, the capacitance C needs to be increased ₀ To adjust BW down ₀ . Therefore, the next detection will be performed in the left half of the median of the maximum capacitance variable range, i.e., the high order region, which corresponds to a range larger than the current capacitance value. Thus, ADJ<N:0>The most significant bit of 1 is determined and the current register is saved. Similarly, if VI _ out<VQ _ OUT, COMP _ OUT =0, then ADJ<N:0>Will be set to 0 and the current register will be saved. The next highest detection is performed, still first ADJ<N:0>Setting the next highest position as 1, repeating the previous calibration process, judging whether the position is the lowest position, if so, ending the process, and sending a switch control signal ADJ<N:0>Is latched in a register that configures the filter bandwidth, otherwise the switch control signal ADJ is reset<N:0 > the current bit is 1, and the previous calibration procedure is repeated untilADJ<N:0>Is determined and then according to ADJ<N:0>Configuring the adjustable capacitor unit by the final code value to determine C ₀ To complete the calibration process. The values of VI _ out and VQ _ out are now approximately equal within the error tolerance. This error is determined by the decision error of the comparator itself and the total bit of the capacitor array. Finally, the actual bandwidth BW of the filter can be ensured ₀ And a calibration standard bandwidth f _-3dB And the influence of PVT change on the bandwidth of the filter is eliminated. Note that the active low pass filters of the I/Q channels are calibrated at the same time.

The foregoing is directed to the preferred embodiment of the present invention and it is noted that the above-mentioned embodiments illustrate rather than limit the invention, and that those skilled in the art will be able to design alternative embodiments without departing from the scope of the appended claims. It will be apparent to those skilled in the art that various modifications and improvements can be made without departing from the spirit and scope of the invention, and such modifications and improvements are also considered to be within the scope of the invention.

Claims

1. An active low pass filter bandwidth calibration circuit comprising: I. the two-order active low pass filter LPF with adjustable bandwidth arranged on the Q orthogonal channel, the two active low pass filters LPF on the I, Q channel and the finite state machine FSM connected in parallel and terminated are characterized in that: two sinusoidal signals with different amplitudes and frequencies are respectively input to the input ends of two paths of orthogonal channels I, Q, the two paths of sinusoidal signals firstly respectively pass through bandwidth-adjustable second-order active low-pass filters LPF on I, Q two paths, the two paths of sinusoidal signals output by the LPFs are respectively sent to two peak value detection circuits on the output ends of I, Q two paths of LPFs, the amplitude information of the two paths of signals is detected and sent to a comparator which is commonly connected with the output ends of the two peak value detection circuits for comparison, and the comparator sends a comparison result to a Finite State Machine (FSM) connected with the output end of the comparator for processing; and the FSM controls and changes the capacitance value of the adjustable capacitor unit in the LPF through an algorithm circuit in the FSM according to the comparison result, detects and adjusts the bandwidth of the filter, approaches the amplitude value of the I path signal attenuated by the LPF to the amplitude value of the Q path signal, detects and continuously iterates the bandwidth of the LPF according to the result output by the comparator for many times until all bits of the adjustable capacitor unit of the array are traversed, and stops iterating to obtain more accurate filter bandwidth insensitive to PVT change.

2. The active low pass filter bandwidth calibration circuit of claim 1, wherein: the peak detection circuits are connected to the output of the active low pass filter, the Finite State Machine (FSM) is connected to the output of the comparator, and the two peak detection circuits are connected between the comparators.

3. The active low pass filter bandwidth calibration circuit of claim 1, wherein: the two inputs of the whole bandwidth calibration circuit I, Q are used as the calibrated sinusoidal signal utilization generated by the digital-to-analog converter (DAC) of the pre-stage circuit of the LPF.

4. The active low pass filter bandwidth calibration circuit of claim 1, wherein: frequency of the I-path sinusoidal signal and standard bandwidth BW = f to be calibrated _-3dB Same, and amplitude VI = V ₁ (ii) a The frequency of the Q path sinusoidal signal is f _-3dB /32, amplitude

5. The active low pass filter bandwidth calibration circuit of claim 3, wherein: I. the Q signals are equal in amplitude after passing through the LPF, i.e. VI _ out = VQ _ out, and if the LPF is affected by PVT, its actual bandwidth BW is ₀ Greater or less than f _-3dB The corresponding VI _ out is also greater or less than

6. Active low-pass filtering according to claim 1A bandwidth calibration circuit, characterized by: the peak detection circuit detects amplitude information VI _ out and VQ _ out of output signals of the two filters, the amplitude information VI _ out and VQ _ out are sent to the comparator to be compared, the comparator outputs 1 or 0 according to the magnitude of the VI _ out and the VQ _ out, the comparison result is sent to the finite state machine FSM to be processed, an algorithm in the finite state machine FSM controls and changes the capacitance value of an adjustable capacitor array unit in the active low-pass filter according to the comparison result, so that the bandwidth of the filter changes towards the direction of reducing the difference between the VI _ out and the VQ _ out, the FSM continues to carry out detection iteration for multiple times according to the result output by the comparator on the bandwidth of the filter until the VI _ out approaches the VQ _ out, and the actual bandwidth BW of the filter approaches the VQ _ out ₀ And approaching the target bandwidth BW to be calibrated until the traversal is finished, and obtaining a more accurate bandwidth calibration result, wherein the total number of iterations is equal to the total bit number of basic units of the adjustable capacitor array in the filter.

7. The active low pass filter bandwidth calibration circuit of claim 1, wherein: the capacitance of the array is increased by a multiple of 2 to the power of M, and M takes a value from 0 to N, and is divided by a fixed capacitor C _fix In addition, the other capacitor arrays are controlled by a switch control signal ADJ<N:0>Control if the switch control signal ADJ<N:0>The corresponding decimal integer is k (k is more than or equal to 0 and less than or equal to 2) ^N+1 -1), then C ₀ Is represented as C _fix + k Δ C, thereby enabling the traversal of the capacitance values within an integer k times of the basic capacitance unit Δ C.

8. The active low pass filter bandwidth calibration circuit of claim 1, wherein: the active low pass filter LPF adopts a differential structure of a second-order active Butterworth filter which is formed by cascading two-stage filter sections comprising a transimpedance operational amplifier OP1 and an OP2 on a chip, wherein the first-stage filter section is symmetrically provided with two pairs of basic adjustable capacitor units C which are connected in parallel ₀ And a total capacitance of 2C ₀ The capacitor array of (1); the second stage filter section is symmetrically provided with a single basic adjustable capacitor unit C ₀ Formed capacitor array and resistor R _F 。

9. The active low pass filter bandwidth calibration circuit of claim 8, wherein: the positive and negative output ends of the first-stage OP1 reverse proportion amplification circuit pass through two crossed voltage-dividing resistors R _C The negative and positive input ends of the inverse proportional amplifying circuit of the second stage OP2 are connected in series, and the input end of the first stage OP1 passes through two resistors R respectively _A In parallel with the output of the second stage OP2, the 3dB bandwidth of the second order active butterworth filter is expressed as:

10. the active low pass filter bandwidth calibration circuit of claim 6, wherein: in the primary calibration process, the calibration algorithm adopts a dichotomy control to change the adjustable capacitor unit according to the value output by the comparator, the bandwidth of the filter is tested, firstly, the capacitance value of the adjustable capacitor unit is set as the median value of the maximum capacitance variable range, and a switch control signal ADJ is initialized<N:0>=100 … 000, i.e., the highest position 1, and the remaining bits are held at 0, if the signal amplitude VI _ out output by the I channel LPF and the signal amplitude VQ _ out output by the Q channel LPF satisfy: VI _ out>VQ _ OUT, the comparator outputs COMP _ OUT =1, and the next detection is performed in the left half of the median of the maximum capacitance variable range, i.e. the high-order region, which corresponds to a range larger than the current capacitance value, ADJ<N:0>Is determined to be 1, the current register is saved, and likewise, if VI _ out<VQ _ OUT, COMP _ OUT =0, then ADJ<N:0>The highest bit of the register is set to be 0, and the current register is saved; the next highest detection is performed, still first ADJ<N:0>Setting the next highest position as 1, repeating the previous calibration process, judging whether the position is the lowest position, if so, ending the process, and sending a switch control signal ADJ<N:0>Is latched in a register that configures the filter bandwidth, otherwise the switch control signal ADJ is reset<N:0>The previous calibration procedure is repeated until ADJ, with the current bit being 1<N:0>Is determined and then according to ADJ<N:0>Is adjustable in the final code value pairThe capacitor unit is configured to determine C ₀ The final value of (c), the calibration process is completed.