A broadband four-core six-mode voltage-controlled oscillator
By designing a quad-core, six-mode voltage-controlled oscillator and utilizing a mode-switching capacitor network and a two-dimensional switching network, the phase noise and size issues of wideband voltage-controlled oscillators were solved, achieving a wide frequency tuning range and low phase noise.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2025-05-28
- Publication Date
- 2026-03-24
AI Technical Summary
Existing technologies make it difficult to design a voltage-controlled oscillator that can cover a wide frequency band and meet strict phase noise requirements. Traditional solutions suffer from low efficiency, large chip area, and deterioration of phase noise.
A quad-core, six-mode voltage-controlled oscillator is adopted. By setting up a mode-switching capacitor network and a two-dimensional switch network between the oscillator cores, the oscillator cores are coupled by electromagnetic coupling, and the size of the coupling capacitor is switched by controlling the voltage to achieve multi-mode operation.
It achieves a very wide frequency modulation range with low phase noise and a small size, making it suitable for integration at millimeter-wave frequencies and reducing the overall size and cost of the system.
Smart Images

Figure CN120639026B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to semiconductor devices, and more particularly to a broadband quad-core six-mode voltage-controlled oscillator. Background Technology
[0002] The development of radio design is shifting from single-mode systems to multi-mode systems, requiring systems capable of supporting multiple standards across multiple frequency bands. For example, today's 5G wireless communication demands wide bandwidth and high data rates, and needs to support multiple frequency bands such as 24.25-27.5GHz, 37-40.5GHz, 42.5-43.5GHz, and 50.4GHz-52.6GHz. A major challenge in designing these systems is whether it is possible to design a voltage-controlled oscillator (VCO) that can cover a wide bandwidth while meeting stringent phase noise requirements. Therefore, VCOs with a wide tunable range and low phase noise have become a current research hotspot.
[0003] One multi-band solution uses multiple LC oscillators at different frequencies, activating one VCO at a time. [1] This results in one or more inductors always being idle, leading to low efficiency and a significant increase in chip area; another technique is to control the inductor or capacitor value connected to the resonant cavity by controlling the MOS switch, thereby changing the resonant frequency. [2, 3] However, since a non-negligible on-resistance still exists when the switch is turned on, this impedance reduces the quality factor Q of the resonant cavity, thereby worsening the phase noise of the VCO. [4, 5] .
[0004] In addition, multimode VCO is also a technique for realizing broadband VCO. It uses a switching network to switch the oscillation frequency of the resonant cavity, thereby enabling the VCO to operate in different modes. [6] Furthermore, since no current flows through the switching network, it does not worsen phase noise, while achieving a wider bandwidth. This is a promising design approach for achieving broadband coverage and low-noise VCOs.
[0005] In summary, this invention proposes a quad-core, six-mode broadband voltage-controlled oscillator.
[0006] References
[0007] [1] A. Kral, F. Behbahani and A. A. Abidi, "RF-CMOS oscillators withswitched tuning," Proceedings of the IEEE 1998 Custom Integrated CircuitsConference (Cat. No.98CH36143), Santa Clara, CA, USA, 1998, pp. 555-558.
[0008] [2] Seong-Mo Yim and K. O. Kenneth, "Demonstration of a switchedresonator concept in a dual-band monolithic CMOS LC-tuned VCO," Proceedingsof the IEEE 2001 Custom Integrated Circuits Conference (Cat. No.01CH37169),San Diego, CA, USA, 2001, pp. 205-208.
[0009] [3] N. Da Dalt, E. Thaller, P. Gregorius and L. Gazsi, "A compacttriple-band low-jitter digital LC PLL with programmable coil in 130-nm CMOS,"in IEEE Journal of Solid-State Circuits, vol. 40, no. 7, pp. 1482-1490, July2005.
[0010] [4] D. Hauspie, E. -c. Park, J. Craninckx and B. Come, "Wideband VCOwith Simultaneous Switching of Frequency Band, Active Core and VaractorSize," 2006 Proceedings of the 32nd European Solid-State Circuits Conference,Montreaux, Switzerland, 2006, pp. 452-455.
[0011] [5] N. T. Tchamov, S. S. Broussev, I. S. Uzunov and K. K. Rantala, "Dual-Band LC VCO Architecture With a Fourth-Order Resonator," in IEEETransactions on Circuits and Systems II: Express Briefs, vol. 54, no. 3, pp.277-281, March 2007.
[0012] [6] G. Li and E. Afshari, "A Distributed Dual-Band LC OscillatorBased on Mode Switching," in IEEE Transactions on Microwave Theory andTechniques, vol. 59, no. 1, pp. 99-107, Jan. 2011.
[0013] [7] Jerng A, Sodini C G, The impact of device type and sizing onphase noise mechanisms [C]. Custom Integrated Circuits Conference, 2003.Proceedings of the IEEE 2003:547-550. Summary of the Invention
[0014] The purpose of this invention is to provide a broadband quad-core six-mode voltage-controlled oscillator to solve the problems existing in the prior art. It achieves a very wide frequency modulation range while maintaining low phase noise and has a small size, which has a significant advantage over traditional VCOs.
[0015] The broadband quad-core six-mode voltage-controlled oscillator described in this invention includes four oscillator cores, two mode-switching capacitor networks disposed between the oscillator cores, and a two-dimensional switching network that couples the oscillator cores together by electromagnetic coupling; the mode-switching capacitor network includes two symmetrically arranged coupling capacitors C. m Each of the coupling capacitors C m Each of them is connected to a mode-switching varactor C. mode_var The mode switching variable capacitance tube C mode_var Control voltage V ctrl control;
[0016] When the electrical coupling of the voltage-controlled oscillator is in odd mode, if the control voltage V ctrl =0, then the coupling capacitor C m And the maximum value mode switching varactor C mode_var Connected to the voltage-controlled oscillator circuit; if the control voltage V ctrl =1, then the coupling capacitor C m And the minimum value mode switching varactor C mode_var It is connected to the voltage-controlled oscillator circuit.
[0017] Preferably, the oscillator core consists of a cross-coupled transistor and a resonant network, wherein the cross-coupled transistor is configured with two NMOS transistors, and the resonant network consists of a parallel array of switched capacitors C. tank The system consists of an inductor L; the drains of the two NMOS transistors are respectively connected to the two ends of the inductor L, and the gate of one NMOS transistor is connected to the drain of the other NMOS transistor; the sources of the two NMOS transistors are grounded; and the switched capacitor array C is connected in parallel. tank One end connected to the inductor L is coupled to one of the capacitors C. m The two are connected, and the other end of the connection is coupled to another coupling capacitor C. m connect.
[0018] Preferably, the NMOS in the cross-coupled transistor is a deep N-well low threshold voltage transistor.
[0019] Preferably, the switched capacitor array C tank It includes an n-bit switched capacitor unit, each of which is configured with two inverters and two fixed capacitors nC. uAnd a switching transistor, the input of the inverter and the gate of the switching transistor are connected and input with a control level CAP_SW[n], the output of one inverter and one fixed capacitor nC u One end of the switch is connected to the drain of the switching transistor, and the output of the other inverter is connected to another fixed capacitor nC. u One end of the capacitor is connected to the source of the switching transistor, and one of the fixed capacitors nC in all the switched capacitor units is connected to the source of the switching transistor. u One end is connected as a switched capacitor array C tank One connection terminal, and another fixed capacitor nC in all the switched capacitor units. u One end is connected as a switched capacitor array C tank The other connection end.
[0020] Preferably, in the switched capacitor array C tank In this context, the size of the switching transistor is based on... Assign weights.
[0021] Preferably, the mode switching network includes two electrical coupling units E_SW and two magnetic coupling units M_SW; the four oscillator cores are core1, core2, core3, and core4, respectively. A mode switching capacitor network and an electrical coupling unit E_SW are provided between core1 and core2, and between core3 and core4. A magnetic coupling unit M_SW is provided between core1 and core3, and between core2 and core4.
[0022] Preferably, both the electrical coupling unit E_SW and the magnetic coupling unit M_SW are composed of a first NMOS switch, a second NMOS switch, a third NMOS switch, and a fourth NMOS switch. The gates of the first NMOS switch and the second NMOS switch are connected to form a coupled even-mode control pair, and the gates of the third NMOS switch and the fourth NMOS switch are connected to form a coupled odd-mode control pair. The drain of the first NMOS switch and the drain of the third NMOS switch are connected as a first signal terminal, the source of the first NMOS switch and the source of the fourth NMOS switch are connected as a third signal terminal, the drain of the second NMOS switch and the drain of the fourth NMOS switch are connected as a second signal terminal, and the source of the second NMOS switch and the source of the third NMOS switch are connected as a fourth signal terminal.
[0023] The first and second signal terminals are respectively input to the switched capacitor array C of an oscillator core corresponding to the coupling unit. tankThe two connection terminals, the third signal terminal and the fourth signal terminal, are respectively input to the switched capacitor array C of another oscillator core corresponding to the coupling unit. tank The two connecting ends.
[0024] Preferably, the first NMOS switch, the second NMOS switch, the third NMOS switch, and the fourth NMOS switch in the electrical coupling unit E_SW and the magnetic coupling unit M_SW are of the same size.
[0025] Preferably, the inductors L in core 1 and core 3 form a transformer in the VCO resonant cavity, and the inductors L in core 2 and core 4 form another transformer in the VCO resonant cavity.
[0026] Preferably, the transformer has a Y-axis mirror-symmetric structure.
[0027] The broadband quad-core six-mode voltage-controlled oscillator described in this invention has the following advantages:
[0028] 1. Achieving lower phase noise. This invention innovatively incorporates a mode-switching varactor on top of the electrical coupling, adding two more modes to the original quad-core, quad-mode architecture, thus achieving a wider frequency modulation range. Furthermore, the quad-core architecture employed results in lower phase noise.
[0029] 2. Achieve a wide frequency modulation range. Multimode VCOs can achieve a wider frequency coverage range by switching different operating modes of capacitors or transformers through a switching network.
[0030] 3. Achieve smaller size and higher integration. Multi-core, multi-mode VCOs have a compact structure and small size, are suitable for millimeter-wave frequencies, and can be integrated into smaller chips, achieving miniaturization and high integration. This helps reduce the overall system size and cost, while improving system reliability and stability. Attached Figure Description
[0031] Figure 1 This is a schematic diagram of the circuit structure of a broadband quad-core six-mode voltage-controlled oscillator described in this invention.
[0032] Figure 2 This is a schematic diagram of a two-dimensional switching network circuit for a broadband quad-core six-mode voltage-controlled oscillator described in this invention.
[0033] Figure 3 These are schematic diagrams of the core circuits of the oscillators with different structures according to the present invention.
[0034] Figure 4 This is a schematic diagram of the equivalent circuit of the core capacitor of the NMOS voltage-controlled oscillator of the present invention.
[0035] Figure 5This is a schematic diagram of the n-bit switched capacitor array circuit of the present invention.
[0036] Figure 6 This is a graph showing the effect of the switching transistor size weight on the switched capacitor array according to the present invention.
[0037] Figure 7 This is a schematic diagram of the mode switching circuit of a broadband quad-core six-mode voltage-controlled oscillator described in this invention.
[0038] Figure 8 The resonant cavity input impedance of the broadband quad-core six-mode voltage-controlled oscillator described in this invention exhibits conductivity as the switching transistor conducts. G ON change.
[0039] Figure 9 This is a parameter curve of the resonant cavity transformer of a broadband quad-core six-mode voltage-controlled oscillator described in this invention.
[0040] Figure 10 This is the layout of the resonant cavity transformer for a broadband quad-core six-mode voltage-controlled oscillator described in this invention.
[0041] Figure 11 This is the overall layout of a broadband quad-core six-mode voltage-controlled oscillator described in this invention.
[0042] Figure 12 This is a graph showing the electromagnetic post-simulation results of a broadband quad-core six-mode voltage-controlled oscillator described in this invention. Detailed Implementation
[0043] like Figures 1 to 2 The diagram shown is a circuit diagram of a broadband quad-core six-mode voltage-controlled oscillator according to the present invention. Specifically, the oscillation circuit architecture of the quad-core six-mode VCO (Voltage Control Oscillator) of the present invention is as follows: Figure 1 As shown, it includes four oscillator cores, namely core1 to core4, and two coupling capacitors C. m Mode switching variable capacitance tube C mode_var The mode-switching capacitor network is composed of [a specific type of capacitor]. Figure 1 In the diagram, M1 to M8 are NMOS transistors, which act as cross-coupled pairs. m I represents the transconductance value of the cross-coupled pair. ss C is the tail current; tank For a switched capacitor array; L is the resonant inductor, k is the coupling coefficient between the two inductors; C m This is the coupling capacitor, which acts as the mode switching capacitor for the oscillator; C mode_var For mode switching varactor; V ctrl This is the mode control voltage.
[0044] like Figure 2 The diagram shows a two-dimensional switching network corresponding to the mode switching switching network of a quad-core six-mode voltage-controlled oscillator. It uses electromagnetic coupling to couple the oscillator cores, and the mode switching is controlled by controlling the state of the two-dimensional switching network. Figure 2 V osc For the oscillation voltage, I osc For oscillating current; G M,e For the transconductance of a magnetically coupled even-mode switch, G M,o For the transconductance of a magnetically coupled odd-mode switch, G E,e For the transconductance of an electrically coupled dipole-mode switch, G E,o It is the transconductance of an electrically coupled odd-mode switch.
[0045] The biggest difference between the quad-core six-mode VCO of this invention and the quad-core four-mode VCO is that the coupling capacitor C of this invention... m Then, a control voltage is connected across it. V ctrl Control mode switching variable capacitor C mode_Var Mode switching variable capacitor C mode_Var and coupling capacitor C m Only when the electrical coupling is odd-mode, i.e. G E,e =0, G E,o = G ON When connected to the circuit; when electrically coupled, it is an even mode, that is G E,e = G ON , G E,o When the value is 0, no circuit is connected. Therefore, when the electrical couple is in odd mode, the control... V ctrl The high or low level can control the size of the capacitor connected to the resonant cavity, thus separating two modes in the odd mode of electrical coupling.
[0046]
[0047] Table 1 lists the electromagnetic coupling states from Mode 1 to Mode 6 and the oscillation frequencies of the oscillators. When the electromagnetic coupling is in odd mode... V ctrl =1 indicates V ctrl It is at a high level at this time. C mode_Var = C mode_VarMax ; V ctrl =0 indicates V ctrl It is at a low level at this time.C mode_Var = C mode_VarMin Mode 1 has the lowest frequency, and Mode 6 has the highest frequency.
[0048] The core CORE of the quad-core six-mode voltage-controlled oscillator of the present invention will be introduced below.
[0049] The core circuit of an oscillator typically consists of a cross-coupled transistor and a resonant network, where the cross-coupled transistor can be, for example, a cross-coupled transistor with ... Figure 3 The NMOS transistor shown in (a), or Figure 3 (b) PMOS transistor, or when Figure 3 (c) shows the complementary PMOS and NMOS transistors. The negative conductance of the cross-coupled pairs is marked in the figure.
[0050] In CMOS technology, the carrier mobility of NMOS transistors is 2 to 3 times that of PMOS transistors. To achieve the same negative resistance at the same power consumption, the transistor size in an NMOS structure is smaller than that in a PMOS structure, resulting in smaller parasitic capacitance and making it easier to obtain higher oscillation frequencies.
[0051] According to publicly available technology, such as papers [7] Because the overdrive voltage of the cross-coupled transistor in the oscillator is very large, the short-channel transistor will enter the velocity saturation region. NMOS transistors are more likely to enter the velocity saturation region than PMOS transistors. With the same equivalent negative conductance, PMOS transistors have lower leakage current thermal noise, resulting in better phase noise performance. Therefore, the PMOS structure is suitable for oscillators with lower frequencies and higher phase noise requirements.
[0052] The cross-complementary NMOS and PMOS structure can achieve a larger negative conductance, but the maximum output amplitude is only half of the supply voltage, and the DC level of the output node is unstable and susceptible to PVT variations. Therefore, this structure is only suitable for low-power oscillators.
[0053] In summary, NMOS transistors are more suitable for cross-coupled pairs.
[0054] To meet the requirements of high frequency and low phase noise, cross-coupling is particularly important in the design of NMOS transistors. In the millimeter-wave band, the parasitic capacitance of cross-coupling on NMOS transistors is even larger than that of varactor transistors. When the NMOS transistor is too small, the negative conductance it provides is too small, causing the oscillator core to fail to oscillate or the oscillation amplitude to be too small; when the NMOS transistor is too large, the maximum frequency decreases. Therefore, there is a trade-off between oscillation amplitude and maximum frequency in the design.
[0055] R gThis represents the gate resistance of each gate finger. R G Let N be the total gate resistance of the NMOS transistor, and N be the number of fingers. Therefore, we have... R G = R g / N ;at the same time W t It is the total width of the NMOS transistor. W It is the width of a finger, so there is W = W t / N .because R g ∝ W ,so R G ∝ W 2 It is evident that the smaller the finger width, the lower the gate loss and the better the phase noise of the oscillator core. However, while reducing the finger width, the number of fingers N increases, the parasitic capacitance of the NMOS transistor increases, and the maximum frequency and tuning range decrease. Therefore, there is a trade-off between phase noise and maximum frequency.
[0056] Figure 4 The figure shows the equivalent capacitance distribution of the core of an NMOS voltage-controlled oscillator. C gd When connected in reverse, according to the principle of Miller capacitance, the gate-drain parasitic capacitance contributes 2( C gd1 + C gd2 For each NMOS transistor, the total capacitance of the oscillation node is... C equ = C var + C gs +4 C gd + C db .
[0057] Ultimately, considering the trade-offs between the oscillator core's highest oscillation frequency and phase noise / oscillation amplitude, the NMOS in the voltage-controlled oscillator core of this invention employs a deep-N-well low-threshold voltage transistor. dnwlvt Because deep N-well transistors have better isolation and less noise introduced by the substrate; transistors with low threshold voltages can have higher negative conductance at the same gate-source voltage.
[0058] Next, regarding the switched capacitor array C in a quad-core six-mode VOC... tank Let me introduce it.
[0059] Figure 5 The circuit diagram shows an n-bit switched capacitor array, where (a) the switching transistor size ratio is 1:2:4...; (b) the switching transistor size ratio is... . Figure 5 (a) illustrates the structure of a commonly used n-bit switched capacitor array and the dimensions of each capacitor and switching transistor. In the figure, CAP_SW[n] represents the control level of the nth switched capacitor unit. When the control level is high, the source and drain of the switching transistor are pulled low by the inverter, and the switching transistor is turned on. At this time, the fixed capacitor nC... u A resonant network is connected in series; when the control level is low, the source and drain are pulled high by the inverter, the switching transistor is turned off, and the fixed capacitor nC... u No resonant network is connected. Furthermore, to ensure frequency coverage of the switched capacitor array, the size of the n switched capacitor units needs to be allocated with a weight of 2, such as... Figure 5 As shown in (a) and 5(b).
[0060] When the nth switching transistor is turned on, because it operates in the linear region, its on-resistance formula is:
[0061] (3.1)
[0062] Because the on-resistance and the fixed capacitance nC u When connected in series with a resonant network, the smaller the on-resistance, the larger the Q value of the switched capacitor array. Therefore, increasing the aspect ratio of the switching transistor can increase the Q value of the switched capacitor array, but the parasitic capacitance of the switching transistor will also increase, thereby reducing the frequency tuning range (TR) of the switched capacitor array. It is evident that the design of the switching transistor size is crucial.
[0063] Common n-bit switching transistor sizes are allocated with a weight of 2, known as constant. RC Design, such as Figure 5 As shown in (a). Because according to the Q-value formula, the nth switched capacitor unit conducts... and shut down The formula is as follows:
[0064] (3.2)
[0065] In Equation 3.2 R nCu Indicates the fixed capacitance nC u Its own series resistance, This represents the total parasitic capacitance when the nth switching capacitor is not connected. According to formulas 3.1 and 3.2, constant... RC In the design, when the switching capacitor unit is turned on, each bit... They are approximately the same. Furthermore, the overall quality factor is higher when the entire switched capacitor array is connected. for:
[0066] (3.3)
[0067] Only when each bit When they are nearly the same, Only then will it reach its maximum. At this point, because the switching transistors at higher positions are larger, when all switching transistors are off and the switched capacitor array is completely disconnected, the excessively large parasitic capacitance of the switching transistors will be connected to the resonant cavity, lowering the highest resonant frequency of the resonant cavity; when the switched capacitors are fully connected, the parasitic capacitance of the switching transistors is not connected to the resonant cavity, and the lowest resonant frequency does not change. In summary, in constant RC The design achieved a good Q value, but reduced the frequency tuning range to some extent.
[0068] Therefore, in this scheme, in order to obtain a wider frequency tuning range and a higher oscillation frequency, the size of the n=3 bit switch is... The assignment of weights is called Design. According to equations 3.2 and 3.3, The weighting of switching transistor size will decrease However, due to The decrease, and On the contrary, it will increase, and the frequency tuning range of the switched capacitor array will also increase. The formula for the frequency tuning range is as follows:
[0069] (3.4)
[0070] C max This represents the total capacitance when the switched capacitor array is fully connected. C min This represents the total capacitance when the switched capacitor array is completely disconnected.
[0071] Figure 6 The effect of switch size weights on switched capacitor arrays is shown, where (a) represents the effect of capacitance value and (b) represents the effect of Q value.
[0072] Because this design uses a 3-bit control switched capacitor array, Figure 6 The CAP_SW value on the horizontal axis ranges from 0 to 7. 0 represents no switching capacitor connected at all, and 7 represents the switching capacitor fully connected. Figure 6(a) It can be seen that the change in the size weight of the switching transistor only affects C min It has almost no impact. C max ,therefore The design features a wider tuning range; from Figure 6 (b) It can be seen that, The designed switched capacitor array has a worse Q value.
[0073]
[0074] Table 2 shows that, While switched capacitor arrays have a lower Q-value, they can achieve a wider frequency tuning range. Therefore, to obtain a large frequency tuning range, this design selects... Architecture.
[0075] The specific circuit diagram of the mode switching network for the quad-core six-mode voltage-controlled oscillator is as follows: Figure 7 As shown.
[0076] The mode switching network includes two electrically coupled units E_SW and two magnetically coupled units M_SW. A mode switching capacitor network and an electrically coupled unit E_SW are provided between core1 and core2, and between core3 and core4. A magnetically coupled unit M_SW is provided between core1 and core3, and between core2 and core4.
[0077] Specifically, both the electrical coupling unit E_SW and the magnetic coupling unit M_SW are composed of a first NMOS switch, a second NMOS switch, a third NMOS switch, and a fourth NMOS switch. The gates of the first and second NMOS switches are connected to form a coupled even-mode control pair, and the gates of the third and fourth NMOS switches are connected to form a coupled odd-mode control pair. The drains of the first and third NMOS switches are connected as the first signal terminal, the sources of the first and fourth NMOS switches are connected as the third signal terminal, the drains of the second and fourth NMOS switches are connected as the second signal terminal, and the sources of the second and third NMOS switches are connected as the fourth signal terminal. The first and second signal terminals are respectively input to the switched capacitor array C of an oscillator core corresponding to the coupling unit. tank The two connection terminals, the third signal terminal, and the fourth signal terminal are respectively input to the switched capacitor array C of another oscillator core corresponding to the coupling unit. tank The two connecting ends.
[0078] exist Figure 7In the middle, C M represent Figure 1 C in m and C mode_var The mode-switching capacitor network consists of M, which represents... Figure 1 The coupling between the two inductors, EN ei It is an NMOS switch that controls the electrically coupled dipole mode, EN oi It is an NMOS switch that controls the electrically coupled odd-mode; MN ei It is an NMOS switch that controls magnetically coupled even modes, MN oi These are NMOS switches that control the odd-mode magnetic coupling. Here, i is 1 or 2. For example, in an electrically coupled unit E_SW, the first, second, third, and fourth NMOS switches are EN, respectively. e1 EN e2 EN o1 EN o2 The conduction transconductances of the magnetic coupling unit M_SW and the electrical coupling unit E_SW correspond to... Figure 2 G in M and G E Because NMOS transistors have higher electron mobility than PMOS transistors of the same size, they have higher on-conductivity. Therefore, NMOS transistors are better suited for suppressing oscillations in other undesirable modes when used as mode-switching switches. It should be noted that... Figure 7 In the diagram, the circuit diagram on the left represents an electrically coupled unit E_SW. Similarly, the circuit diagram on the right only represents a magnetically coupled unit M_SW.
[0079] Figure 8 The input impedance of the resonant cavity is related to the conductance of the switching transistor. G ON The graph shows the changes, where (a) represents mode 1; (b) represents mode 2; (c) represents mode 3; (d) represents mode 4; (e) represents mode 5; and (f) represents mode 6.
[0080] Figure 8 Demonstrates the conductance of the mode switching switch. G ON The effect on the input impedance of the resonant cavity. When G ON When boosted, it enhances the suppression of resonant impedance in other non-oscillatory modes. When G ON When = 1.0 ms, the other undesirable modes are hardly suppressed, thus there is a risk of oscillation in the other modes; when G ONWhen the impedance is 5.7 ms or 7.8 ms, the other undesirable modes are well suppressed, and only the impedance peaks of the desired modes are seen on the input impedance diagram.
[0081] In this scheme, cross-coupling affects the NMOS transistor. =20ms, based on the oscillation condition of the oscillator under cross-coupling. Therefore, the resonant cavity impedance at the undesired oscillation mode frequency should be less than 50Ω. According to... Figure 8 It can be seen that when G ON This condition can be met when the current is 5.7 ms. To ensure circuit symmetry, the NMOS switches controlling the electrical and magnetic coupling are of the same size.
[0082] Finally, regarding the implementation scheme of the transformer.
[0083] To achieve a wide frequency coverage and low phase noise, an ideal transformer was simulated and its parameters adjusted. The transformer in the VCO resonant cavity needs to be a center-tapped differential inductor with a single inductance of 150pH and a Q value greater than 20 at 30GHz. The two inductors in the transformer are symmetrical, with a coupling coefficient of k = -0.38. The transformer simulation results obtained through EMX simulation are shown below. Figure 9 As shown, Figure 9 In the figure, (a) is the inductance curve; (b) is the Q value curve; and (c) is the coupling coefficient curve.
[0084] After EMX simulation modeling, the above indicators were basically achieved. In the range of 20GHz to 40GHz, the inductance value ranged from 146.3 to 154.6 pH, the Q value ranged from 22.6 to 30.2, and the coupling coefficient k varied from -0.392 to -0.378.
[0085] Since the two inductors are identical, the transformer is designed with a Y-axis mirror-symmetric structure, overlapping a portion of the area to achieve a coupling coefficient of -0.38. Due to the low inductance requirement, a single-turn coil is sufficient. The thickest, highest layer of metal is used to improve the Q value. Testing revealed that, except for the overlapping portion, using two layers of metal with identical traces connected vias significantly improves the Q value. Increasing the trace width within a certain range also improves the Q value. The layout implementation of this transformer is shown in [link to transformer layout]. Figure 10 .
[0086] The circuit described above was designed using SMIC 40nm process technology in Cadence Virtuoso. Figure 11For the overall VCO layout, the VCO core area measures 402 µm × 146 µm, the output buffer measures 160 µm × 200 µm, and the total area including DC and RF pads is 900 µm × 900 µm.
[0087] Figure 12 The achieved electromagnetic post-simulation performance is shown. The frequency variation range for each mode is as follows: Figure 12 As shown in (a), a continuous tuning range (TR) from 18.67 GHz to 50.28 GHz is illustrated, representing 91.8% of the tuning range. The frequency range for each mode is detailed in Table 3, showing a wide overlap between the various modes. Figure 12 (b) Phase noise (PN) at typical frequencies is shown in the six modes. At a 1 MHz frequency offset, the phase noise varies from -97.0 dBc / Hz to -111.2 dBc / Hz, with the worst phase noise at which the Q values of the inductor and varactor diodes degrade when the VCO frequency exceeds 50 GHz. At a supply voltage of 700 mV, the power consumption of the VCO core is less than 15 mW in all six modes. Figure 12 (c) and Figure 12 (d) The factor of quality (FoM) and combined factor of quality (FoMT) are shown for each of the six modes. At a frequency offset of 1 MHz, the maximum FoM is 185.0 dBc / Hz, and the corresponding maximum FoMT is 204.2 dBc / Hz. , .
[0088] It is worth noting that, despite the lack of optimized mode spacing and the underutilization of a significant portion of the low-frequency range, the presence of six modes achieves a wider frequency range than a traditional four-mode VCO. The buffered output maintains over -8 dBm of output power across the entire frequency bandwidth with a power consumption of 15 mW, ensuring that the VCO waveform is not overwhelmed by noise during testing.
[0089]
[0090] For those skilled in the art, various other corresponding changes and modifications can be made based on the technical solutions and concepts described above, and all such changes and modifications should fall within the protection scope of the claims of this invention.
Claims
1. A broadband quad-core six-mode voltage-controlled oscillator, comprising four oscillator cores, two mode-switching capacitor networks disposed between the oscillator cores, and a mode-switching switch network that couples the oscillator cores together by electromagnetic coupling; the mode-switching capacitor network includes two symmetrically arranged coupling capacitors C m Its characteristics are, Each of the coupling capacitors C m Each of them is connected to a mode-switching varactor C. mode_var The mode switching variable capacitance tube C mode_var Control voltage V ctrl control; When the electrical coupling of the voltage-controlled oscillator is in odd mode, if the control voltage V ctrl =0, then the coupling capacitor C m And the maximum value mode switching varactor C mode_var Connected to the voltage-controlled oscillator circuit; if the control voltage V ctrl =1, then the coupling capacitor C m And the minimum value mode switching varactor C mode_var Connect to the voltage-controlled oscillator circuit; The oscillator core consists of cross-coupled transistors and a resonant network. The cross-coupled transistors are configured with two NMOS transistors, and the resonant network consists of a parallel array of switched capacitors C. tank The system consists of an inductor L; the drains of the two NMOS transistors are respectively connected to the two ends of the inductor L, and the gate of one NMOS transistor is connected to the drain of the other NMOS transistor; the sources of the two NMOS transistors are grounded; and the switched capacitor array C is connected in parallel. tank One end connected to the inductor L is coupled to one of the capacitors C. m The two are connected, and the other end of the connection is coupled to another coupling capacitor C. m connect; The switched capacitor array C tank It includes an n-bit switched capacitor unit, each of which is configured with two inverters and two fixed capacitors nC. u And a switching transistor, the input of the inverter and the gate of the switching transistor are connected and input with a control level CAP_SW[n], the output of one inverter and one fixed capacitor nC u One end of the switch is connected to the drain of the switching transistor, and the output of the other inverter is connected to another fixed capacitor nC. u One end of the capacitor is connected to the source of the switching transistor, and one of the fixed capacitors nC in all the switched capacitor units is connected to the source of the switching transistor. u One end is connected as a switched capacitor array C tank One connection terminal, and another fixed capacitor nC in all the switched capacitor units. u One end is connected as a switched capacitor array C tank The other connection end; In the switched capacitor array C tank In this context, the size of the switching transistor is based on... Assign weights.
2. The broadband quad-core six-mode voltage-controlled oscillator according to claim 1, characterized in that, The NMOS in the cross-coupled transistor is a deep N-well low threshold voltage transistor.
3. The broadband quad-core six-mode voltage-controlled oscillator according to claim 1, characterized in that, The mode switching network includes two electrical coupling units E_SW and two magnetic coupling units M_SW; the four oscillator cores are core1, core2, core3, and core4, respectively. A mode switching capacitor network and an electrical coupling unit E_SW are provided between core1 and core2, and between core3 and core4. A magnetic coupling unit M_SW is provided between core1 and core3, and between core2 and core4.
4. The broadband quad-core six-mode voltage-controlled oscillator according to claim 3, characterized in that, The electrical coupling unit E_SW and the magnetic coupling unit M_SW are both composed of a first NMOS switch, a second NMOS switch, a third NMOS switch, and a fourth NMOS switch. The gates of the first NMOS switch and the second NMOS switch are connected to form a coupled even-mode control pair, and the gates of the third NMOS switch and the fourth NMOS switch are connected to form a coupled odd-mode control pair. The drain of the first NMOS switch and the drain of the third NMOS switch are connected as a first signal terminal, the source of the first NMOS switch and the source of the fourth NMOS switch are connected as a third signal terminal, the drain of the second NMOS switch and the drain of the fourth NMOS switch are connected as a second signal terminal, and the source of the second NMOS switch and the source of the third NMOS switch are connected as a fourth signal terminal. The first and second signal terminals are respectively input to the switched capacitor array C of an oscillator core corresponding to the coupling unit. tank The two connection terminals, the third signal terminal and the fourth signal terminal, are respectively input to the switched capacitor array C of another oscillator core corresponding to the coupling unit. tank The two connecting ends.
5. A broadband quad-core six-mode voltage-controlled oscillator according to claim 4, characterized in that, The first NMOS switch, second NMOS switch, third NMOS switch, and fourth NMOS switch in the electrical coupling unit E_SW and the magnetic coupling unit M_SW have the same size.
6. The broadband quad-core six-mode voltage-controlled oscillator according to claim 3, characterized in that, The inductors L in core 1 and core 3 form a transformer in the VCO resonant cavity, and the inductors L in core 2 and core 4 form another transformer in the VCO resonant cavity.
7. A broadband quad-core six-mode voltage-controlled oscillator according to claim 6, characterized in that, The transformer has a Y-axis mirror-symmetric structure.
Citation Information
Patent Citations
Four-mode oscillator based on electromagnetic hybrid coupling
CN111293981A
Control method and device of switched capacitor circuit, equipment and medium
CN118432540A