低功耗智能便携式温度测量仪的研究翻译.doc_第1页
低功耗智能便携式温度测量仪的研究翻译.doc_第2页
低功耗智能便携式温度测量仪的研究翻译.doc_第3页
低功耗智能便携式温度测量仪的研究翻译.doc_第4页
低功耗智能便携式温度测量仪的研究翻译.doc_第5页
已阅读5页,还剩11页未读 继续免费阅读

下载本文档

版权说明:本文档由用户提供并上传,收益归属内容提供方,若内容存在侵权,请进行举报或认领

文档简介

An ultra-low-power CMOS temperature sensor for RFID applicationsAbstract: An ultra-low-power CMOS temperature sensor with analog-to-digital readout circuitry for RFID applications was implemented in a 0.18m CMOS process. To achieve ultra-low power consumption, an error model is proposed and the corresponding novel temperature sensor front-end with a new double-measure method is presented .Analog-to-digital conversion is accomplished by a sigma-delta converter. The complete system consumes only 26A 1.8 V for continuous operation and achieves an accuracy of 0.65C from 20 to 120C after calibration at one temperature.Key words: CMOS temperature sensor; error analysis; digital output; low power; RFIDDOI: 10.1088/1674-4926/30/4/045003 EEACC: 1205; 1065H1. IntroductionRFID technology has attracted profound attention. Not content with object identification only, people are seeking for more functions provided by RFID systems. One of the most at tractive applications is to realize real-time temperature sensing by integrating a temperature sensor into the RFID tag. However, there exist some challenging features for the sensor. First, the temperature sensor must be on-chip and easy to calibrate to minimize the cost. Second, the power consumption of the sensor is stringently limited. Third, the temperature sensor should achieve high accuracy in the whole temperature sensing range. Several papers have reported temperature sensor designs. However, these can either not be realized in low-cost CMOS technology or they consume too much power for continuous operation. In this paper, a novel design of a CMOS temperature sensor, which offers low power, low cost and high accuracy, is proposed. An error model for this sensor is developed. Also, the architectural improvements of the temperature sensor, which includes a sensor front end and a sigma-delta ADC, is given.2. Error modelThe most important candidate for a CMOS temperature sensor is the bipolar transistor, which can be realized as a parasitic device in standard CMOS technology. Almost all CMOS temperature sensors follow the principle of the general architecture of the band gap reference as illustrated in Fig. 1.The difference of the two base-emitter voltages is direct proportional to the absolute temperature (PTAT), as given byTherefore, the temperature can be measured by comparing with the reference voltage, which is theoretically temperature independent. Nevertheless, several errors are inherent in such a circuit, and thus the accuracy greatly depends on the process parameters. A typical test result for the errors of CMOS temperature sensor is presented in Ref. and listed in Table 1.As shown, the errors of this band gap reference circuit are due to the offset of the opamp (Vos), the spread in the absolute value of the baseemitter voltage (Vbe), the spread in the absolute value of the resistors (Rsh), the mismatch of the bipolar transistor (Vos, Q), and the mismatch of the resistors (R), respectively.A temperature sensor error model was built according to the test results described in Table 1 with the error sources being expressed aswhere VT= kT/q. The difference between, representing the (Vbe) of the bipolar transistor Q1, and ,representing the (Vbe) of the bipolar transistor Q2, denotes the mismatch of the bipolar transistors (Vos,Q).(Rsh) : R = R(1 + R):(Rsh) of resistors R1 and R2, respectively, denotes the mismatch of the resistors (R).The simulation of the errors (Vos) and (Vos, Q) are illustrated in Figs. 2(a) and 2(b), respectively, when setting the ratio of I1 and I2 to be a typical value of eight. The results matched well with the data given in Table 1. Figure 2(c) shows (Vos) and (Vos, Q) as functions of temperature. As shown,the error varies 2.5 C in the range of 20 to 120 C. Based on Eqs. (2)(4), the reference voltage and the temperature dependent voltage illustrated in Fig. 1 are given bywhere A is a coefficient. It is shown that the reference voltage is affected by (Vos), (Vos, Q), (R), and (Vbe),while the temperature dependent voltage is only affected by (Vos), (Vos, Q), and (R).3. Architecture improvement As illustrated in Fig. 3, the temperature sensor front end generates the temperature dependent voltage , which isthe input signal of a second-order sigma-delta ADC. The reference voltage for ADC is generated from the RFID tag. After ADC digitalizes , the digital circuit calculates the difference of for two working modes of the front end circuit.3.1 CMOS temperature sensor front endtemperature sensor front-end, as illustrated in Fig. 4. The output voltage is also temperature dependent and given byThe Switch 2 controls the value of I1. I1 is four times larger than I2 when the Switch 2 is closed. While the Switch 2 is open, I1 is three times larger than I2. So the difference of between two working modes isThe result implies a good performance of this new double measure method. The only error of the voltage difference, which is proportional to the absolute temperature, comes from the mismatch of the resistances, namely (R), which is well controlled in CMOS technology.However, is digitalized by ADC by comparing it with . So the accuracy of also depends on the precision of the reference voltage, which is provided by a band gap circuit in the RFID tag. As seen from Eq. (4), it is difficult to keep constant over a large temperature range and usually requires complex calibrations such as second order curvature correction. Fortunately, the fact that temperature sensing and recording is discrete in time alleviates this demand. By comparing for two working modes with the same reference voltage, the difference voltage can be precisely measured when is equal in just two cases rather than requiring to be exactly the same over the whole measured temperature range. Thus, a calibration for is nolonger necessary. As is shown in Eq. (6), the proposed temperature sensor front end has the advantage of better immunity to the main errors of (Vos), (Vos, Q), (Vbe), and (Rsh). A sigma-delta ADC following the circuit is designed to digitize thefor two working modes.3.2 AD converterThe sigma-delta ADC has been proven to be very suitable in low frequency, high performance cases. In our application, the signal almost acts as a DC signal. A second-order sigma-delta ADC is chosen for its low complexity and moderate accuracy of 10 bits.Nevertheless, the offset in the sigma-delta modulator would influence the DC signal. As shown in Fig. 6, the output voltage would include the offset voltage of the opamp when sampling the input signal in a sample-and-hold (S-H) circuit.The output voltage is given bybeing closed or open, we can writeThe offset voltage of the opamp is cancelled automatically and does not induce additional errors when converting the analog DC signal to the digital domain.3.3 Architecture error simulationThe proposed architecture overcomes the error sources of (Vos), (Vos, Q), (Vbe), and Rsh) described in the error model. As illustrated in Fig.7, the temperature error in the most interested range of 20 to 50 C is negligible. The remaining error (R) caused a 0.65 C variation in the range of 20 to 120 C due to the mismatch of R1 and R2.4. Measurement resultsThe whole CMOS temperature sensor is implemented in SMIC 0.18m process. The micrograph of this proposed sensor is shown in Fig. 8. The complete circuit occupies 0.9 mm2 (excluding bond pads).The measured results for the proposed temperature sensor, which are illustrated in Fig. 9, matches well with the simulation results. After calibration at one temperature, the variation of the error is less than 0.65 C from 20 to 120 C. To meet the stringent limit of power consumption of the RFID tag, the sensor with digital output consumes only 26 A1.8V for continuous operation, of which 6 A is consumed by the sensor front-end and 20 A by the sigma-delta converter.Table 2 compares the power consumption, the inaccuracy and calibration condition with previous works. Many temperature sensors provide an additional low-power working mode in which the sensor is normally shut down and only powers up when necessary. As the time for one temperature measurement is usually only several ten microseconds, the average power consumption is far less than the peak one. For continuous operation the power consumption should be limited to 50 A for RFID applications. Our work also provides an additional low-power working mode controlled by Switch 1 in Fig. 4.The presented temperature sensor performs as well as sensors presented in previous works over a wide temperature range, but achieves much lower power consumption for continuous operation.5. ConclusionThe temperature sensor with digital output for RFID applications offers limited power consumption, low cost and high accuracy. A novel design of a CMOS temperature sensor is presented. Compared with traditional power-hungry error cancellation technologies such as Dynamic Element Matching(DEM) of bias current, the proposed sensor front-end with a second-order sigma-delta ADC overcomes the circuit errors by means of a new double-measure method and consumes only 26A 1.8 V for continuous operation. After calibration at one temperature, the sensor achieves an accuracy of 0.65 C in the measured temperature range from 20 to 120 C.AcknowledgmentsThe authors would like to thank Lee Yang of SMIC (Shanghai) for chip fabrication support.References1 Soto M A, Sahu P K, Rueck C, et al. Distributed temperature sensor system based on Raman scattering using correlationcodes. Electron Lett, 2007, 43(16): 8622 Bakker A, Huijsing J H. Micropower CMOS temperature sensor with digital output. IEEE J Solid-State Circuits, 1996,31(7): 9333 Tuthill M. A switched-current, switched-capacitor temperature sensor in 0.6um CMOS. IEEE J Solid-State Circuits, 1998,33(7): 11174 Pertijs M A P, Niederkorn A, Ma X, et al. A CMOS smart temperature sensor with a 3_ inaccuracy of 0:5 C from 50 Cto +120 C. IEEE J Solid-State Circuits, 2005, 40(2): 4545 Bakker A, Huijsing J H. A low-cost high-accuracy CMOSsmart temperature sensor. Proceedings of the 25th European ESSCIRC, 1999: 302045003-4毕业设计(论文)外文资料翻译系 别: 电子信息系 专 业: 自动化 班 级: 姓 名: 学 号: 外文出处: CNKI中国知网 附 件: 1. 原文; 2. 译文 年月 RFID应用的超低功耗CMOS温度传感器摘要:超低功耗的CMOS读出电路模拟到数字温度传感器的RFID应用阳离子在0.18微米CMOS工艺实现了实现超低功耗,这是一个错误模型提出并提出了相应的新型温度传感器的前端有一个新的双测量方法。模拟到数字的转换是通过一个A/D转换器。完整的系统只消耗26A,1.8 V连续运行,实现从-20到120 C,校准后精度0.65 C。关键词:CMOS温度传感器;误差分析;数字输出,低功耗,RFIDDOI: 10.1088/1674-4926/30/4/045003 专业代码:1205;1065H1.介绍RFID技术已经吸引了深刻的关注,不只关注对象标识的内容,人们正在寻求RFID系统提供更多的功能。最吸引的应用是实现实时的温度传感通过集成一个温度传感器的RFID标签。操作方法以往任何时候都存在一些具有挑战性的传感器功能。首先,温度传感器必须在芯片和易于校准最大限度地降低成本。第二,功耗被严格限制。第三,温度传感器应在整个温度感应范围达到高精确度。有几篇论文报道了温度传感器的设计。然而,这些可以不实现低成本的CMOS技术或他们消耗过多的能量不断运行。在本文中,设计新颖的CMOS温度传感器,它提供了低功耗,低成本和高准确性。这种传感器的误差模型开发是先进的。此外,温度的架构改进的传感器,其中包括一个传感器前端和一个给出的ADC。2.误差模型最重要的CMOS温度候选人是双极型晶体管,它可以作为一个实现面值在标准CMOS工艺设备。几乎所有的CMOS温度传感器遵循的一般原则因此,可以测量温度比较PTAT与参考电压,这是理论上的温度独立。然而,固有几个错误在电路上。两个基极发射极电压的区别是直接绝对温度成正比(PTAT)的,给予在这样的电路,从而准确性大大取决于工艺参数。一个典型的错误的测试结果CMOS温度传感器号。如上所示,这种带隙基准电路的误差由于偏移运算放大器,在传播绝对值电压,电阻的绝对值,双极晶体管不匹配,不匹配的电阻。根据温度传感器的误差模型测试结果的误差来源是其中VT = kT/q。表双极晶体管Q1,代表(Vbe)双极晶体管Q2。不匹配双极晶体管(Vos, Q)。同样,是不同的,(Rsh)是R1 ,R2阻值的误差,表示不匹配电阻(R)。(Vos),(Vos, Q)的仿真是错误的,当设置I1和I2为典型的电压,结果显示为表1,显示了(Vos),(Vos, Q)的功能电压。误差范围为正负2.5摄氏度,变化范围为负20到正120摄氏度。根据上式(2)到(4),参考电压和温度依赖电压图1。其中A是一个系数,参考电压受(Vos),(Vos, Q),(R)的影响。但温度电压只受(Vos),(Vos, Q),(R)的的影响。3.架构改进如图3所示,温度传感器前端产生的温度依赖电压,ADC参考压来自RFID标签。后ADC数字化,数字电路的计算两种工作模式的差异前前端电路。(1) CMOS温度传感器前端为了实现高精确度和低功耗,我们提出了一个新型温度传感器前端,如图所示4,输出电压也取决于温度和给予开关2控制I1,I1是I2的4倍是开关2关闭,当开关2打开,I1是I2的3倍,电压在不同模式下为结果意味着这个新的双性能好测量方法。唯一错误的电压不同,这是绝对温度成正比,来自(R) 。然而经过ADC离散化与比较,的精度也取决于RFID提供的参考电压。图4是图6的的采样和保持电路的二阶调制。保持温度的恒定很困难,通常需要复杂的校准,如二阶曲率校正。幸运的是,该温度下的遥感和录音时间是离散的事实减轻这方面的需求。通过比较两个工作模式与基准电压,电压差,测量时,是等于两个房案,而不是要求的是在整个测得的温度范围内完全一样。因此,为的校准是必要的。如式所示(6),建议温度下的传感器前端有更好的抗干扰优势。主要错误(Vos),(Vos, Q),(Rsh)的ADC电路设计,数字化两种工作模式。(2) AD转换器ADC已被证明是非常适合在低频率,高性能的情况下。在我们的应用阳离子,信号几乎作为一个直流信号。选择一个第二或ADC,其复杂度低和中等精度的10位。然而,在ADC调制器的偏移会影响直流信号。正如图6所示。输出电压时,将包括运放的失调电压采样输入信号在采样和保持电路。输出电压由下式给出当Vos为开关2时,开关开与关,如图4,我们可以写成运放失调电压被自动取消,并不会引起额外直流模拟信号转换到数字域的错误。(3) 误差仿真模块目的是在误差模块中克服误差来源(Vos),(Vos, Q),(Rsh),在仿真图7中,温度误差

温馨提示

  • 1. 本站所有资源如无特殊说明,都需要本地电脑安装OFFICE2007和PDF阅读器。图纸软件为CAD,CAXA,PROE,UG,SolidWorks等.压缩文件请下载最新的WinRAR软件解压。
  • 2. 本站的文档不包含任何第三方提供的附件图纸等,如果需要附件,请联系上传者。文件的所有权益归上传用户所有。
  • 3. 本站RAR压缩包中若带图纸,网页内容里面会有图纸预览,若没有图纸预览就没有图纸。
  • 4. 未经权益所有人同意不得将文件中的内容挪作商业或盈利用途。
  • 5. 人人文库网仅提供信息存储空间,仅对用户上传内容的表现方式做保护处理,对用户上传分享的文档内容本身不做任何修改或编辑,并不能对任何下载内容负责。
  • 6. 下载文件中如有侵权或不适当内容,请与我们联系,我们立即纠正。
  • 7. 本站不保证下载资源的准确性、安全性和完整性, 同时也不承担用户因使用这些下载资源对自己和他人造成任何形式的伤害或损失。

评论

0/150

提交评论