-
晶体谐振器广泛应用于振荡器设计[1]、时频控制[2]等领域。晶体谐振器也应用于传感器领域,如温度传感器[3]、石英晶体微天平[4-7]等。
-
图 1a为晶体谐振器BVD(Butterworth-Van Dyke)等效模型串联负载电容${C_L}$。其中${R_1}$为动态电阻、${L_1}$为动态电感、${C_1}$为动态电容、${C_0}$为静态电容。通常认为晶体谐振器BVD等效模型的谐振频率只与动态电感${L_1}$和动态电容${C_1}$有关,并且有:
$${f_r} = \frac{1}{{2{\rm{ \mathit{ π} }}}}\sqrt {\frac{1}{{{L_1}{C_1}}}} $$ (1) 式(1)中只包含${L_1}$和${C_1}$,因而,通过频率只能求解${L_1}$和${C_1}$。通过推导,本文得到BVD等效模型零相位频率的精确形式为:
$${f_ \pm } = \frac{1}{{2{\rm{ \mathit{ π} }}}}\sqrt {{\rho _1} \pm \sqrt {{\rho _1}^2 - \frac{1}{{{L_1}^2{C_1}^2}}\left( {1 + \frac{{{C_1}}}{{{C_0}}}} \right)} } $$ (2) 式中,
$${\rho _1} = \frac{1}{{{L_1}{C_1}}} + \frac{1}{{2{L_1}{C_0}}} - \frac{{{R_1}^2}}{{2{L_1}^2}}$$ (3) 式(2)中取正号即得到反谐振频率[13-14]${f_a}$,取负号即得到谐振频率${f_r}$。由式(2)可知,频率和4个等效参数都有关。因此,可以通过频率求解4个等效参数,而不仅仅是两个等效参数。由于有4个待解参数,除了${f_a}$、${f_r}$之外,还需要两个频率,以便于构建四元方程组。这里选择负载反谐振频率${f_A}$与负载谐振频率${f_R}$。
为了简化求解负载反谐振频率与负载谐振频率的过程,先做一个等效变换。图 1a模型可以等效变换为图 1b模型的形式,并且这种变换不受谐振器参数限制[15],因而低Q值谐振器也适用于此种变换。变换之后,${R_{1x}}$、${L_{1x}}$、${C_{1x}}$、${C_{0x}}$是${R_1}$、${L_1}$、${C_1}$、${C_0}$、${C_L}$的函数。将变换关系带入式(2),即得到负载零相位频率为:
$${f'_ \pm } = \frac{1}{{2{\rm{ \mathit{ π} }}}}\sqrt {{\rho _1} \pm \sqrt {{\rho _1}^2 - \frac{1}{{{L_1}^2{C_1}^2}}\left( {1 + \frac{{{C_1}}}{{{C_0}}}} \right)\left( {1 + \frac{{{C_1}}}{{{C_0} + {C_L}}}} \right)} } $$ (4) 式中,
$${\rho _1} = \frac{1}{{{L_1}{C_1}}} + \frac{1}{{2{L_1}({C_0} + {C_L})}} + \frac{1}{{2{L_1}{C_0}}} - \frac{{{R_1}^2}}{{2{L_1}^2}}$$ (5) 式(4)中取正号即得到负载反谐振频率${f_A}$,取负号即得到负载谐振频率${f_R}$。
-
由${f_a}$、${f_r}$、${f_A}$、${f_R}$即可构建四元方程组。观察式(2)可见,${f_a}$、${f_r}$存在内在联系,其平方和就是式(3)的形式,即:
$${f_a}^2 + {f_r}^2{\rm{ = }}\frac{1}{{2{{\rm{ \mathit{ π} }}^2}}}\left( {\frac{1}{{{L_1}{C_1}}} + \frac{1}{{2{L_1}{C_0}}} - \frac{{{R_1}^2}}{{2{L_1}^2}}} \right)$$ (6) 同理可得:
$${f_A}^2 + {f_R}^2 = \frac{1}{{2{{\rm{ \mathit{ π} }}^2}}}\left[ {\frac{1}{{{L_1}{C_1}}} + \frac{1}{{2{L_1}({C_0} + {C_L})}} + \frac{1}{{2{L_1}{C_0}}} - \frac{{{R_1}^2}}{{2{L_1}^2}}} \right]$$ (7) 将式(6)带入式(2)得到:
$${f_r}^2{f_a}^2 = \frac{1}{{4{{\rm{ \mathit{ π} }}^2}}}\frac{1}{{{L_1}^2{C_1}^2}}\left( {1 + \frac{{{C_1}}}{{{C_0}}}} \right)$$ (8) 同理可得:
$${f_R}^2{f_A}^2{\rm{ = }}\frac{1}{{4{{\rm{ \mathit{ π} }}^2}}}\frac{1}{{{L_1}^2{C_1}^2}}\left( {1 + \frac{{{C_1}}}{{{C_0}}}} \right)\left( {1 + \frac{{{C_1}}}{{{C_0} + {C_L}}}} \right)$$ (9) 式(6)~式(9)组成新的方程组,便可简化计算等效参数的过程。求解上述方程组,可依次得到:
$${C_0}{\rm{ = }}\frac{{{C_L}}}{{\frac{{{f_R}^2{f_A}^2 - {f_r}^2{f_a}^2}}{{4{{\rm{ \mathit{ π} }}^2}{{({f_A}^2{\rm{ + }}{f_R}^2 - {f_a}^2 - {f_r}^2)}^2}}} - \frac{1}{{\left( {\frac{{{f_R}^2{f_A}^2}}{{{f_r}^2{f_a}^2}} - 1} \right)}} - 1}}$$ (10) $${C_1}{\rm{ = }}\left( {\frac{{{f_R}^2{f_A}^2}}{{{f_r}^2{f_a}^2}} - 1} \right)({C_0} + {C_L})$$ (11) $${L_1}{\rm{ = }}\frac{1}{{4{{\rm{ \mathit{ π} }}^2}}}\frac{1}{{{f_r}{f_a}{C_1}}}\sqrt {1{\rm{ + }}\frac{{{C_1}}}{{{C_0}}}} $$ (12) $${R_1} = {L_1}\sqrt {\frac{2}{{{L_1}{C_1}}} + \frac{1}{{{L_1}{C_0}}} - 4{{\rm{ \mathit{ π} }}^2}({f_a}^2{\rm{ + }}{f_r}^2)} $$ (13) -
图 4是安捷伦网络分析仪E5062A实测的晶体谐振器相位-频率曲线,所测晶体为10 MHz、SC切石英晶体。供应商提供的参数为:${R_1}$=69.78 Ω,${L_1}$= 1 407.29 mH,${C_1}$=0.18 pF,${C_0}$=2.11 pF。其Q值较大,因而本文方法测量结果应该和标称结果很接近。图中,相位偏移为-45.7°,这是由于导线长度等问题产生的。
由图 4可见,相位-频率曲线在谐振点附近的斜率很大,在1 Hz以内相位可以变化几度。如,距离${f_a}$最近两点的相位分别为10.02°和-5.57°。它们和零相位仍有较大差距,所以本文采用线性插值的方法来间接测量零相位频率。此时测得:${f_r}$=9 999 717.0 Hz,${f_a}$=10 000 095.3 Hz。同理可得,${f_R}$=9 999 957.9 Hz,${f_A}$=10 000 088.1 Hz。同时,为了消弱接触电阻对测量的影响,本文串联了一个电阻${R_X}$=27 Ω。
实测过程中发现存在一个随机漂移频率${f_x}$使${f_R}$=${f_R}$+${f_x}$,${f_A}$=${f_A}$+${f_x}$。本文选择相频曲线在${f_r}$和${f_a}$处的导数方程来修正此随机频率。导数方程为:
$${\left. {\frac{{\partial ({\rm{Pashe}}(f)}}{{\partial f}}} \right|_{f = {f_r}}} = {\mathit{\Delta} _1}$$ (14) $${\left. {\frac{{\partial ({\rm{Pashe}}(f))}}{{\partial f}}} \right|_{f = {f_a}}} = {\mathit{\Delta} _2}$$ (15) 式中,
$${\rm{Pashe}}(f) = \frac{{180}}{{\rm{ \mathit{ π} }}}{\rm atan}\left( {\frac{{{\rm{imag(}}G(f){\rm{)}}}}{{{\rm{real(}}G(f){\rm{)}}}}} \right)$$ (16) $$G(f){\rm{ = }}\frac{1}{{Z(f) + {R_0} + {R_X}}}$$ (17) 式中,Z是谐振器的阻抗;${R_0}$是网络分析仪的内阻,这里${R_0}$=100 Ω。
由式(10)~式(15)组成了六元非线性方程组,即可求解等效参数以及随机频率${f_x}$和负载电容${C_L}$。本文采用二维搜索法求解此方程组。
其具体的过程为:
1) 和有限元解法类似,首先生成一个包含${f_x}$和${C_L}$的二维网格。${f_x}$初始值为${f_{x\_{\rm{start}}}}$,间距为${f_{{\rm{\_bu}}}}$。${C_L}$初始值为${C_{L\_{\rm{start}}}}$,间距为${C_{L\_{\rm{bu}}}}$。则网格上的点为(${f_{x\_{\rm{start}}}}+N{f_{{\rm{\_bu}}}}$, ${C_{L\_{\rm{start}}}}+M{C_{L\_{\rm{bu}}}}$)。
2) 将${f_x}$=${f_{x\_{\rm{start}}}}+N{f_{{\rm{\_bu}}}}$,${C_L}$=${C_{L\_{\rm{start}}}}+M{C_{L\_{\rm{bu}}}}$,以及测量的${f_a}$、${f_r}$、${f_A}$、${f_R}$,带入式(10)~式(13)计算等效参数。
3) 由等效参数计算相频曲线,由相频曲线得到谐振点的导数${\gamma _{{\rm{1NM}}}}$和${\gamma _{{\rm{2NM}}}}$。
4) 计算误差:${\beta _{{\rm{1NM}}}}$=${\gamma _{{\rm{1NM}}}}$–,${\beta _{{\rm{2NM}}}}$=${\gamma _{{\rm{2NM}}}}$–。则,网格上的每个点(${f_{x\_{\rm{start}}}}+N{f_{{\rm{\_bu}}}}$, ${C_{L\_{\rm{start}}}}+M{C_{L\_{\rm{bu}}}}$)都有对应点(${\beta _{{\rm{1NM}}}}$, ${\beta _{{\rm{2NM}}}}$)。
5) 寻找(${\beta _{{\rm{1NM}}}}$, ${\beta _{{\rm{2NM}}}}$)中的两组点(${\beta _{{\rm{1nm}}}}$, ${\beta _{{\rm{2nm}}}}$)和(${\beta _{{\rm{1}}xy}}$, ${\beta _{{\rm{2}}xy}}$)同时满足下列两个条件:${\beta _{{\rm{1nm}}}}$${\beta _{{\rm{1}}xy}}$ < 0,${\beta _{{\rm{2nm}}}}$${\beta _{{\rm{2}}xy}}$ < 0。其中,(${\beta _{{\rm{1nm}}}}$, ${\beta _{{\rm{2nm}}}}$)对应点(${f_{x\_{\rm{start}}}}+n~{f_{{\rm{\_bu}}}}$, ${C_{L\_{\rm{start}}}}+m~{C_{L\_{\rm{bu}}}}$),(${\beta _{{\rm{1}}xy}}$, ${\beta _{{\rm{2}}xy}}$)对应点(${f_{x\_{\rm{start}}}}$ $x~{f_{{\rm{\_bu}}}}$, ${C_{L\_{\rm{start}}}}+y{C_{L\_{\rm{bu}}}}$)。
6) 寻找点(${f_{x\_{\rm{start}}}}+n{f_{{\rm{\_bu}}}}$, ${C_{L\_{\rm{start}}}}+m{C_{L\_{\rm{bu}}}}$)与点(${f_{x\_{\rm{start}}}}+x{f_{{\rm{\_bu}}}}$, ${C_{L\_{\rm{start}}}}+y{C_{L\_{\rm{bu}}}}$)距离最近的两个点。
7) 判断这两个点的距离是否足够小,并满足精度要求。满足计算结束,不满足以这两个点为基础重新设置${f_x}$和${C_L}$的范围。缩小间距${f_{{\rm{\_bu}}}}$、${C_{L\_{\rm{bu}}}}$再次生成网格。重复流程2)~6),直到满足要求为止。
上述过程的流程图如图 5所示。
最终求解的参数为:${C_L}$=1.271 pF,${f_x}$=7.195 Hz,${R_1}$=72.826 Ω,${L_1}$=1 384.587 mH,${C_1}$=0.183 fF,${C_0}$=2.417 pF。通过等效参数,复现的相频曲线如图 6所示。
串联180 Ω电阻,即${R_X}$=180 Ω时所测量参数为:${C_L}$=1.259 pF,${f_x}$=5.466 Hz,${R_1}$=73.907 Ω,${L_1}$= 1 410.757 mH,${C_1}$=0.180 fF,${C_0}$=2.381 pF。
两组实验的测量值和标称值都很接近,${R_X}$= 180 Ω时更接近标称值。这可能是因为还有接触电阻等不确定电阻的存在,串联电阻越大,不确定电阻的影响越小。测量值与标称值之间略有差距,可能是由于测试环境(如测试温度、测试夹具的静电容等)与供应商测试环境不一致。
Measurement of Quartz Crystal Resonator Parameters Based on the Precise Derivation of Zero Phase Frequency
-
摘要: 晶体谐振器等效参数的测量方法很多,工程上通常利用谐振频率和负载谐振频率来求解等效参数。该文推导了谐振频率、负载谐振频率、反谐振频率和负载反谐振频率的精确形式,并以此为基础求解晶体谐振器的等效参数。ADS仿真实验表明,该方法在理论上正确。利用相位-频率曲线在谐振点与反谐振点的导数构建非线性方程组,解决实测实验中的频率随机游动问题。采用二维搜索法求解非线性方程组。实测结果表明,该方法测量的等效参数和供应商提供的等效参数基本一致。该方法没有采用近似计算,不仅适用于高Q值晶体谐振器,也适用于低Q值谐振器,因此,该方法也能应用于传感器领域,如温度传感器、石英晶体微天平等。Abstract: There are a lot of methods to measure the equivalent parameters of the crystal resonator. In engineering, the resonant frequency and the load resonant frequency are usually used to calculate the equivalent parameters. In this paper, the precise form of the resonant frequency, the load resonant frequency, the anti-resonant frequency and the load anti-resonant frequency are deduced, which serve as the basis to calculate the equivalent parameters of crystal resonators. Advanced design system (ADS) simulation show that this method is correct in theory. In this paper, the nonlinear equations are constructed by the derivative of the phase-frequency-curve at the point of resonance and anti-resonance to solve the problem of random walk of frequency in the experiment. Then, the two-dimensional search method is used to solve those nonlinear equations. The experimental results show that the equivalent parameters measured by this way are basically the same as those provided by manufacturer. This method does not adopt approximate calculation, namely, it is not only suitable for high Q crystal resonators, but also suited to low Q resonators. Therefore, it can also be applied to the field of sensors, such as temperature sensors, quartz crystal microbalances, and so forth.
-
[1] HUANG X H, LIU D, WANG Y, et al. 100 MHz low-phase-noise microprocessor temperature-compensated crystal oscillator[J]. IEEE Transactions on Circuits and Systems Ⅱ-Express Briefs, 2015, 62(7):636-640. doi: 10.1109/TCSII.2015.2415652 [2] ZHOU W, WANG B, GAO J N, et al. AMCXO and its test system[J]. IEEE Transactions on Ultrasonics Ferroelectrics and Frequency Control, 2004, 51(9):1050-1053. doi: 10.1109/TUFFC.2004.1334837 [3] LIU B Q, HAN T, ZHANG C R. Error correction method for passive and wireless resonant SAW temperature sensor[J]. IEEE Sensors Journal, 2015, 15(6):3608-3614. doi: 10.1109/JSEN.2015.2394776 [4] YAO Y, XUE Y J. Influence of the oxygen content on the humidity sensing properties of functionalized graphene films based on bulk acoustic wave humidity sensors[J]. Sensors and Actuators B-Chemical, 2016, 222:755-762. doi: 10.1016/j.snb.2015.08.121 [5] YAO Y, XUE Y J. Impedance analysis of quartz crystal microbalance humidity sensors based on nanodiamond/graphene oxide nanocomposite film[J]. Sensors and Actuators B-Chemical, 2015, 211:52-58. doi: 10.1016/j.snb.2014.12.134 [6] YAO Y, CHEN X D, MA W Y, et al. Quartz crystal microbalance humidity sensors based on nanodiamond sensing films[J]. IEEE Transactions on Nanotechnology, 2014, 13(2):386-393. doi: 10.1109/TNANO.2014.2305986 [7] YAO Y, CHEN X D, LI X Y, et al. Investigation of the stability of QCM humidity sensor using graphene oxide as sensing films[J]. Sensors and Actuators B-Chemical, 2014, 191:779-783. doi: 10.1016/j.snb.2013.10.076 [8] IEC. Measurement of quartz crystal unit parameters by zero phase technique in a pi-network. Part 2: Phase offset method for measurement of motional capacitance of quartz crystal units: IEC60444-2[S]. London: International Electrotechnical Commission, 1980. [9] 董政洁, 刘桂礼, 王艳林, 等.基于S参数传输法的石英晶体静电容测量[J].仪器仪表学报, 2011, 32(4):892-896. http://www.cqvip.com/QK/94550X/201104/37575933.html DONG Zheng-jie, LIU Gui-li, WANG Yan-lin, et al. Quartz crystal static capacitance measurement based on S parameter transmission method[J]. Chinese Journal of Scientific Instrument, 2011, 32(4):892-896. http://www.cqvip.com/QK/94550X/201104/37575933.html [10] 刘解华, 张其善, 杨军, 等.高频高精度石英晶体元件电参数测量技术研究[J].仪器仪表学报, 2006, 27(5):447-450. https://www.cnki.com.cn/huiyi-YQYB200911001066.html LIU Xie-hua, ZHANG Qi-shan, YANG Jun, et al. Research on measurement with high frequency and precision for electric parameters of Quartz crystal units[J]. Chinese Journal of Scientific Instrument, 2006, 27(5):447-450. https://www.cnki.com.cn/huiyi-YQYB200911001066.html [11] 吴兵. 石英晶体微天平电学参数获取及在物理吸附上的应用[D]. 合肥: 中国科技大学物理电子学, 2008. http://cdmd.cnki.com.cn/Article/CDMD-10358-2008091848.htm WU Bing. Acquirement of eleetrieal parameters of quartz crystal microbalance and its application in physisorption[D]. Hefei: School of physics, University of Science and Technology of China, 2008. http://cdmd.cnki.com.cn/Article/CDMD-10358-2008091848.htm [12] LIU D, HUANG X H, HU J G, et al. Measurement of quartz crystal unit parameters based on the precise derivation of zero phase frequency[J]. Electronics Letters, 2017, 53(3):142-144. doi: 10.1049/el.2016.4320 [13] BOTTOM V E. Introduce to crystal unite design[M]. New York:Van Nostrand Reinhold Company, 1987. [14] 中华人民共和国信息产业部. 石英晶体元件·电子元器件质量评定体系规范·第1部分: 总规范: GB/T 12273-1996[S]. 北京: 电子工业部标准化研究所, 1996. MIIT. Quartz crystal units-a specification in the quality assessment system for electronic components, part 1: Generic specification: GB/T 12273-1996[S]. Beijing: China Electronics Standardization Institute, 1996. [15] HUANG X H, LIU D, WANG Y, et al. Precise derivation for the equivalent circuit parameters of a crystal resonator with series capacitor[J]. IEEE Transactions on Ultrasonics Ferroelectrics and Frequency Control, 2012, 59(6):1316-1317. doi: 10.1109/TUFFC.2012.2323 [16] HUANG X H, CHEN P P, FU W, et al. Prediction, simulation, and verification of the phase noise in 80 MHz low-phase-noise crystal oscillators[J]. IEEE Transactions on Ultrasonics Ferroelectrics and Frequency Control, 2015, 62(9):1599-1604. doi: 10.1109/TUFFC.2015.007153