Terahertz band synchronous detection circuit

Abstract

To construct a wideband receiver using only one mixer, which can suppress fluctuations in detection output due to fluctuations in distance and jitter contained in an LO.SOLUTION: In a wideband receiver, a detection circuit is composed of: a mixer having an RF input, an LO input, and an IF output; a high-frequency oscillator that outputs a stable frequency; a multiplier that multiplies the output of the high-frequency oscillator to generate a local oscillation frequency; a phase shift circuit that can continuously change the amount of shift using a control voltage; and a feedback mechanism. The phase shift circuit is placed between the high-frequency oscillator and the multiplier, and the detection output is fed back to adjust the voltage to be applied to the phase shift circuit, thereby suppressing the effects of fluctuations in the distance between the transmitter and receiver and phase noise contained in the output of the high-frequency oscillator.SELECTED DRAWING: Figure 4

Description

[Technical Field]

[0001] The present invention relates to a method for configuring a detection circuit for terahertz band wireless communication, mainly at 100 GHz or higher. [Background technology]

[0002] In digital radio, a method such as heterodyne detection or homodyne detection is used to recover a baseband signal from a received signal. Heterodyne detection is a method in which the first mixer mixes the signal received with a local oscillator frequency that is different from the carrier wave, converts it to an intermediate frequency, and then demodulates the baseband signal with a second mixer. In recent years, a method called Low-IF, which uses a lower intermediate frequency, has also been attracting attention. Homodyne detection is a method of demodulating the baseband signal directly without passing through an intermediate frequency by mixing a local oscillator frequency that is the same frequency as the carrier wave of the received signal in a mixer. This method is also called direct conversion or Zero-IF. Summary of the Invention [Problem to be solved by the invention]

[0003] Homodyne detection mixes the received signal with a local oscillator signal (LO) in a mixer, but the detection output fluctuates if the antenna distance between the transmitter and receiver changes or if the LO signal contains jitter. The reason is that when the phase of the received signal and the LO are the same, the mixer output becomes sin(ωt) × sin(ωt) and the time average is maximized, but when the phase of the received signal is shifted by 90 degrees, the output becomes sin(ωt) × cos(ωt) and the time average becomes zero. This is because the phase shift fluctuates between 0 and 90 degrees (Figure 1). If the wavelength is λ and the change in distance between the transmitter and receiver is Δx, the phase of the signal received from the antenna will fluctuate by 2π Δx / λ. In other words, the detected output will alternate between maximum and minimum every time the distance between the transmitter and receiver fluctuates by half a wavelength. At the distance where the signal strength is at its minimum, the detected output will be almost zero, making it impossible to receive the signal at all (Figure 2). For this reason, homodyne detection typically uses quadrature detection, which uses a signal that is 90 degrees out of phase with the LO and performs detection using two mixers (Figure 3). However, in the terahertz band, mixers are very expensive and it is difficult to prepare two mixers with identical characteristics. Furthermore, high machining precision of approximately 10 micrometers is required to accurately match the distance between the splitter and the two mixers. It is also difficult to precisely shift the phase by 90 degrees for the two LOs.

[0004] The heterodyne detection method requires only one high-frequency mixer that operates in the terahertz band, but a bandpass filter must be used to avoid the problem of image interference (image reception). However, terahertz band filters are not easy to fabricate, and the receiving frequency is limited to the filter's transmission band of several GHz. It is not possible to construct a wideband receiver, such as one that can transmit from 220 GHz to 320 GHz.

[0005] An object of the present invention is to configure a wideband receiver that uses only one mixer and is capable of suppressing fluctuations in detection output due to fluctuations in distance and jitter contained in the LO. [Means for solving the problem]

[0006] Prepare a high-frequency oscillator, a phase shift circuit, a frequency multiplier, and a high-frequency mixer (Figure 4). A phase shift circuit is a circuit that can continuously change the phase of an input signal, and in this case we will assume that it can continuously change according to the applied analog voltage. The amount of shift that can be achieved is about 5 degrees. A frequency multiplier is a high-frequency component called an AMC (Amplifier Multiplexer Chain), which houses a multiplier and an amplifier in a single housing. For example, it can multiply a high frequency of 14 GHz to 20 GHz by N to output a high frequency of over 200 GHz. There are two types of high-frequency mixers: harmonic mixers and subharmonic mixers (SHM). A harmonic mixer multiplies the signal input from the LO port by the RF signal, while a subharmonic mixer multiplies the RF signal by twice the frequency of the signal input from the LO port. Either type of mixer can be used, but in the terahertz wave band, subharmonic mixers are easier to use and there are more types available. High-frequency oscillators require a signal source with stable frequency and extremely low phase noise, because the output of the high-frequency oscillator is multiplied by N in the AMC to generate the LO, and if the output of the high-frequency oscillator contains phase noise (jitter), the jitter will also be multiplied by N. The feature of this invention is that it utilizes the phenomenon that jitter in a high-frequency oscillator is magnified by multiplication, and inserts a phase shift circuit into the output of the high-frequency oscillator, which is an extremely stable frequency source, to compensate by significantly varying the phase of the LO. If the distance between the transmitter and receiver changes and the RF phase fluctuates, the LO phase can also be made to follow, preventing a decrease in the detection output. Since the transmitted and received data is digital, it contains a synchronization signal at regular intervals. The feedback control circuit sweeps the voltage applied to the phase shift circuit while receiving the synchronization signal, and determines the shift voltage that produces the maximum detection output. Once the shift voltage that produces the maximum detection output is determined, the voltage applied to the phase shift circuit is fixed at that voltage until the next synchronization signal period. By correcting the LO phase error for each synchronization signal period in this way, it is possible to suppress the effects of fluctuations in the distance between the transmitter and receiver, and phase noise contained in the high-frequency oscillator output. [Effects of the Invention]

[0007] In terahertz band detection circuits, it is possible to suppress fluctuations in receiver sensitivity due to fluctuations in the distance between the transmitter and receiver and jitter contained in the local oscillator. There is no need to prepare two mixers with identical characteristics, nor is a distributor required, and there is no requirement for precision in the wiring length from the antenna to the mixer or the distance from the AMC to the SHM. Because it is not a quadrature detection system, a 90-degree phase shifter is also not required. Because it is not a heterodyne configuration, there is no need for a bandpass filter for image rejection, making it possible to configure a wideband receiver with a simple configuration. For the above reasons, it is possible to significantly simplify and reduce the cost of a terahertz band receiving circuit. [Brief explanation of the drawings]

[0008] [Figure 1] Why is homodyne detection sensitive to the phase of the LO? [Figure 2] Why is homodyne detection sensitive to the distance between transmitter and receiver? [Figure 3] Conventional homodyne detection circuit configuration [Figure 4] Configuration of the detection circuit of the present invention [Figure 5] Configuration of the detection circuit of the present invention (representative diagram) [Figure 6] Baseband signal recovery viewed on the frequency axis [Figure 7] SHM structure [Figure 8] Details of APDP operation and detection in SHM [Figure 9] Fluctuations in detection output when the distance between the transmitter and receiver is changed [Figure 10] Configuration of a heterodyne detection circuit according to the prior art [Figure 11] How image interference occurs in heterodyne detection [Figure 12] Voltage-Controlled Continuous Phase Shift Circuit [Figure 13] Feedback Mechanism of the Present Invention [Figure 14] How to find the optimum phase shift voltage [Figure 15] Effect of phase correction according to the present invention DETAILED DESCRIPTION OF THE INVENTION

[0009] The structure of the present invention will be described with reference to Figure 5. The detection circuit of the present invention is made up of a master oscillator 1, a high frequency generator 2, a phase shift circuit 3, an AMC 4, an SHM 5, and a feedback mechanism 6. The master oscillator is a 10MHz accurate frequency source oscillator and is the reference clock for the high frequency oscillator. The high frequency generator multiplies the 10MHz reference input by 1600 to generate a 16GHz high frequency signal. The high-frequency signal is input to the phase shift circuit, where it is shifted by up to 5 degrees. The output of the phase shift circuit is connected to the input of the AMC, where it is multiplied by 8 to generate a 128GHz local oscillator (LO) frequency. The 128GHz LO is connected to the LO input of the SHM (Sub-Harmonic Mixer) through a waveguide. The SHM is a high frequency component with ports for RF input, LO input, and IF output. The result of multiplying the RF input by a frequency twice the LO input is output from the IF through a low-pass filter (LPF). The frequency of the RF input is RF and the LO input frequency is F LO Then, the output frequency of the IF is F RF -2×F LO This becomes:

[0010] For example, F RF is a 256GHz baseband signal, and F LO If the frequency is set to 128 GHz, the original baseband signal is restored from the IF output (Figure 6). The feedback mechanism consists of a high-speed ADC, a high-speed DAC, and an FPGA. In this example, the LO frequency is set to 256 GHz for ease of calculation, but in reality, it will operate in the same way at any frequency between 220 GHz and 320 GHz. In a typical homodyne detection circuit, a phase shift circuit is not inserted between the high-frequency oscillator and the frequency multiplier (AMC). In the present invention, a phase shift circuit is inserted here, but before explaining the reason, we will first briefly explain the operating principle of SHM.

[0011] (SHM operation explanation) The structure of the SHM is shown in Figure 7. The SHM consists of an antiparallel diode pair (APDP) 1, a low-pass filter 2, a stub 3, and other stubs and matching circuits.

[0012] The local oscillation frequency is input from the LO port. When the absolute value of this voltage is higher than the threshold voltage of the diode that makes up the APDP, the APDP turns ON, and when it is lower, it turns OFF. In this way, the APDP turns ON twice during one LO period. The period when APDP is ON is equivalent to the RF input being connected to GND, and the period when APDP is OFF, the RF input voltage is applied to the input side of the LPF. In SHM, when the phase difference between LO and RF is 0 degrees, the switch turns on between the maximum positive and negative values ​​as shown in Figure 8(a), and equal positive and negative voltages are applied to the input of the LPF. The high-frequency components are removed by the LPF, and zero is output from the IF output. When the phase difference between LO and RF is 45 degrees (90 degrees in terms of RF frequency), as shown in Figure 8(b), the switch turns on at the peak of the RF, so an asymmetric voltage is applied to the IF port, and the LPF output reaches its maximum value. This is how the SHM performs its detection operation, with the output being zero when the phase difference between RF and LO is 0 degrees and 90 degrees (0 degrees and 180 degrees when expressed in RF frequency), and maximum output when the phase difference is 45 degrees and 135 degrees (90 degrees and 270 degrees when measured in RF frequency). The RF is a signal input from the antenna, and its phase fluctuates depending on the distance between the transmitter and receiver. If the wavelength of the received signal is λ, then the phase changes by 180 degrees every λ / 2. At 256 GHz, the wavelength is 1.17 mm, so the phase rotates by 180 degrees every time the distance between the transmitting and receiving antennas (or more accurately, the distance between the transmitting and receiving mixers) fluctuates by 0.58 mm. The detected output is zero when the phase difference between the LO and RF is 0 degrees and 90 degrees, so the detected output alternates between zero and the maximum value, and zero and the negative maximum value, every 0.29 mm.

[0013] Figure 9 is a graph of the detected output using SHM. A 256 GHz transmitter and receiver were placed opposite each other on a movable stage, and the relationship between the distance between the transmitter and receiver and the detected output was plotted. Because the wavefront of the radio wave spreads out two-dimensionally, the received power should be inversely proportional to the square of the distance, and the voltage should be inversely proportional to the distance. Normally, there would not be fluctuations with repeated peaks and valleys for each wavelength, but for the reasons explained above, there are repeated maximum values ​​and zeros every λ / 2.

[0014] (Prior Art) The fluctuations in detection output described above have not been apparent in conventional microwave and millimeter-wave communications. This is because homodyne detection and heterodyne detection are used in microwave and millimeter-wave communications, but homodyne detection uses two mixers with LOs that are 90 degrees out of phase with each other, which avoids such fluctuations, and heterodyne detection does not cause fluctuations in reception strength due to distance. However, it is difficult to perform conventional homodyne detection using two mixers or heterodyne detection with terahertz waves. The reason for this is explained below.

[0015] (Difficulties in using quadrature detection circuits in the terahertz band) For detection circuits up to a few GHz, two mixers can be prepared and used for detection as shown in Figure 3. This method is called quadrature detection. The LOs given to the I-phase mixer and Q-phase mixer are out of phase with each other by 90 degrees, so when the I-phase mixer has a minimum output, the Q-phase mixer has a maximum output. Neither output will be zero. A quadrature detector is configured so that the received RF signal is split into two signals by a splitter and input to two mixers, but in the terahertz wave band, the wavelength is only about 1 mm, so the wiring length from the splitter to the two mixers requires extremely high precision. As mentioned above, the phase rotates 90 degrees at 0.29 mm, so the wiring needs to be about 0.01 mm. At this frequency, the wiring is a waveguide, and it might be more accurate to call it a pipe. It is not easy to adjust the length like with coaxial cable, so high precision is required in the mechanical processing and assembly when manufacturing the waveguide. Furthermore, signals that are 90 degrees out of phase with the LO are created using phase delay means such as hybrid circuits or rat-race circuits, but it is difficult to stably create signals with an accurate phase difference of 90 degrees and wire them to the mixer with an accuracy of 0.01 mm. Furthermore, SHMs that can be used in the terahertz band are very expensive, and using two mixers increases costs. Using two SHMs to implement a homodyne quadrature detection circuit is not easy from both a technical and cost perspective.

[0016] (Difficulties in using heterodyne detection circuits) Heterodyne detection is a method of converting to an intermediate frequency by creating a difference between the receiving frequency and the local oscillator frequency. This method is often used in the millimeter wave band and above because the amplitude of the detected output does not change even if the phase changes with distance. An example of the configuration of a heterodyne detection circuit is shown in Figure 10. If LO1 of 126 GHz is used to receive a 256 GHz RF signal, a baseband signal modulated at 4 GHz, which is the frequency difference between 2 x LO, is output from the IF port. The IF is usually amplified by an intermediate frequency amplifier and then fed to the second mixer. Since the IF is a waveform obtained by modulating the baseband signal at 4 GHz, to restore the baseband signal, 4 GHz is input to LO2 and it is detected by the second mixer. There are three problems with this method. If the frequency difference between RF and 2×LO1 is 4 GHz, (1) The IF bandwidth of the first mixer must be 4 GHz or higher. (2) The baseband signal bandwidth is within ±2 GHz. (3) Insert a bandpass filter that passes 256 GHz ± 2 GHz between the antenna and the SHM. The second problem imposes a restriction that the bandwidth of the baseband signal must be equal to or less than the intermediate frequency band.

[0017] Figure 11 shows how heterodyne detection works on the frequency axis. Let's assume that the received signal occupies a band of ±2 GHz centered around 256 GHz. The 256 GHz (±2 GHz) signal is downconverted to 4 GHz (±2 GHz) by SHM. However, the signal at 248 GHz (±2 GHz) is also downconverted to 4 GHz (±2 GHz), so in addition to the 256 GHz that you want to receive, you will also have sensitivity to 248 GHz. This phenomenon is called image interference, and 248 GHz is called the image frequency.

[0018] There are several ways to deal with image interference, but it is generally considered best to increase the selectivity of the high-frequency portion. For example, inserting a bandpass filter (BPF) between the antenna and mixer 1. The BPF has the characteristic of passing 256 GHz and blocking 248 GHz, but terahertz band filters are waveguides, and it is difficult to change the pass frequency. The receive frequency is almost fixed at the pass frequency of the filter and cannot be changed. (Note: At the time of filing this application, it was still difficult to utilize the 200-300 GHz band, and the situation was not such that there were so many wireless stations that interference would occur, so this filter was omitted in many terahertz wave experiments. However, in the future, when terahertz band wireless stations become widespread, a filter will likely become essential if the heterodyne method is used.)

[0019] (Features of the present invention) If the output of a high-frequency oscillator contains phase noise (jitter), the jitter of the LO generated by multiplication will be magnified by the multiplication. This invention takes advantage of the fact that the jitter of a high-frequency oscillator is magnified by multiplication, and inserts a phase shift circuit into the output of the high-frequency oscillator, which is an extremely stable frequency source, to control the phase of the LO mixed with the RF. When the RF phase fluctuates, the LO phase is also fluctuated accordingly, and controlled so that the IF output is maximized. Even if the phase shift amount is insufficient to achieve maximum output, it is possible to at least avoid zero output. When using an SHM mixer, controlling the shift amount so that the phase difference between RF and LO does not become 0 degrees can prevent the detection output from becoming zero. (When using a harmonic mixer instead of an SHM, control is performed so that the phase difference does not become 0. Since the operation is similar, the following explanation will only cover the case where an SHM is used.) The 16 GHz reference signal from the local oscillator is multiplied by 8 in the AMC and 2 in the SHM to become a 256 GHz reference signal, which is then multiplied by the RF signal to demodulate the baseband signal. At this point, if the original 16 GHz signal contained a phase error, the phase error will also be multiplied by 16 when it is multiplied by 16.

[0020] The configuration of the phase shift circuit used in this invention is shown in Figure 12. Hybrid circuit 1 generates 0-degree and 90-degree signals from a 16 GHz signal input. The 90-degree signal is pulled down to GND by HEMT 2, inserted in the capacitor-coupled section. This HEMT is used not for amplification but as a variable resistor controlled by the gate voltage. The resistance value decreases as the gate voltage approaches zero, thereby weakening the 90-degree signal. After the HEMT, the 0-degree and 90-degree signals are mixed using Wilkinson divider 3. In other words, by mixing the 0-degree signal and the moderately weakened 90-degree signal, a phase shift circuit is created in which the phase shift amount can be continuously controlled by the gate voltage. Although the phase shift amount that can be shifted by this circuit is approximately 5 degrees, it is amplified by 16 times by the multiplier circuit, so the LO phase can be shifted from 0 degrees to approximately 80 degrees. This circuit can be made using only a few resistors and capacitors and a HEMT, making it very low cost.

[0021] The detection output is fed back and controlled to determine the optimal phase shift voltage. Since the communication targeted by this invention is digital wireless, the transmitter transmits a synchronization signal at regular intervals. The receiver's FPGA and high-speed DAC change the phase shift amount in a stepped manner while receiving the synchronization signal. The operation of the feedback mechanism will be explained using Figures 13 and 14. The synchronization signal is transmitted for a period of 1 microsecond at a cycle of 1 millisecond. The synchronization signal is a 1 GHz sine wave. 1000 cycles of the sine wave are received in 1 microsecond. The FPGA of the receiving circuit uses a DAC to change the control voltage to the phase shift circuit in 32 steps, and measures the voltage amplitude of the detection output at each step with a high-speed ADC. After the voltage at each step has been output, the output voltage is fixed at the point where the amplitude is maximum.

[0022] Since the object of the present invention is to avoid the point where the detection output becomes zero, it is not necessary to control the amount of phase shift very finely. It is sufficient if it can be changed in 16 to 32 steps.

[0023] Figure 15 shows the difference in detection output when the LO phase is corrected using the present invention and when no phase correction is performed. The horizontal axis in Figure 15 is the distance between the transmitter and receiver antennas, and the vertical axis is the detection output. The dotted line shows the detection output when no correction is performed, which is zero every λ / 2 (≈0.59 mm). The thick solid line shows the case when phase correction is performed, and the phenomenon of the detection output becoming zero is suppressed.

[0024] When this circuit is operating normally and the phase shift is tracking, the LO is synchronized at the phase level with the 256 GHz RF. Therefore, if the output of the phase shift circuit is divided to create the ADC sampling clock, the baseband signal can be sampled at the peak positions of I and Q, i.e., direct oversampling is possible. At 256 GHz, λ / 4 is 0.29 mm, and if a synchronization signal is sent every millisecond, the receiver can be tracked even if it moves at a maximum speed of 29 centimeters per second.

Claims

1. a mixer having an RF input, an LO input and an IF output; A high-frequency oscillator that outputs a stable frequency; a multiplier that multiplies the output of the high frequency oscillator to generate a local oscillation frequency; A detection circuit comprising a phase shift circuit that can continuously change the amount of shift using a control voltage and a feedback mechanism, the phase shift circuit being disposed between the high frequency oscillator and the multiplier, and having the function of feeding back the detection output to adjust the voltage to be applied to the phase shift circuit.

2. 2. The detection circuit according to claim 1, wherein a sub-harmonic mixer is used instead of the mixer.

3. 3. The detection circuit according to claim 1, wherein the phase change caused by said phase shift circuit is discrete.

4. 3. The detection circuit according to claim 1 or 2, further comprising means for inputting a sweep waveform such as a ramp wave, sine wave, sawtooth wave or stepped waveform to a control input of said phase shift circuit at a timing when a periodically repeating characteristic pattern such as a synchronization signal is being transmitted or received, to determine a control voltage at which said detection output is maximized, and a feedback control mechanism for outputting the voltage at which the detection output determined by said means is maximized during a period when no synchronization signal is being received.

5. 3. The detection circuit according to claim 1 or 2, further comprising means for inputting a sweep waveform such as a ramp wave, sine wave, sawtooth wave or stepped waveform to the control input of said phase shift circuit at a timing when a periodically repeating characteristic pattern such as a synchronization signal is being transmitted or received, to determine a control voltage at which the detection output becomes zero, and a control mechanism for outputting a voltage that avoids the voltage at which the output determined by said means becomes zero during a period when no synchronization signal is being transmitted or received.

Images