sensors Article
A Monolithic Multisensor Microchip with Complete On-Chip RF Front-End Massimo Merenda 1,2, * 1
2
*
ID
, Corrado Felini 1 and Francesco G. Della Corte 1,2
Department of Information Engineering, Infrastructure and Sustainable Energy (DIIES), “Mediterranea” University, Via Graziella Loc. Feo di Vito, 89124 Reggio Calabria, Italy;
[email protected] (C.F.);
[email protected] (F.G.D.C.) HWA srl-Spin Off dell’Università Mediterranea di Reggio Calabria, Via Reggio Campi II tr. 135, 89126 Reggio Calabria, Italy Correspondence:
[email protected]; Tel.: +39-0965-1693-441
Received: 9 November 2017; Accepted: 29 December 2017; Published: 2 January 2018
Abstract: In this paper, a new wireless sensor, designed for a 0.35 µm CMOS technology, is presented. The microchip was designed to be placed on an object for the continuous remote monitoring of its temperature and illumination state. The temperature sensor is based on the temperature dependence of the I-V characteristics of bipolar transistors available in CMOS technology, while the illumination sensor is an integrated p-n junction photodiode. An on-chip 2.5 GHz transmitter, coupled to a mm-sized dipole radiating element fabricated on the same microchip and made in the top metal layer of the same die, sends the collected data wirelessly to a radio receiver using an On-Off Keying (OOK) modulation pattern. Keywords: CMOS sensors; On-Chip-Antenna; LC oscillator; temperature sensor; light sensor; Internet of Things
1. Introduction Microchip sensors have the advantage that the transducers can be monolithically integrated with signal amplification and conditioning circuitry, with clear advantages in terms of reduced size, noise immunity and production costs. With the advent of the Internet-of-Things [1], these devices are expected to gain in popularity, as they can be considered the link between the world surrounding an object they are attached to and electronics. Of course, sensors that are readily deployable and can be easily accessed, possibly in a wireless manner, are highly desirable [2,3]. Nevertheless, the application of an external antenna to a microchip to allow wireless communication is a costly process and moreover leads to the de-miniaturization of the device [4]. Therefore, the concept of an on-chip-antenna (OCA), fabricated during the CMOS process, which eliminates the need for external transmission lines and sophisticated packaging, seems to be the simplest approach to warrant short range data transmission together with low costs and reliability. Usually, this task is obtained by integrating spiral inductors as the radiating elements [5–8]. In this paper, we present a full custom microchip integrating a dipole RF radiator fed by an LC oscillator and an RF power amplifier. The choice of a dipole antenna allowed the design of almost one-dimensional form factor microchips, while loop antennas require square devices. The device includes a temperature sensor, a light sensor and a general-purpose input and was designed for ambient and processes monitoring when temperature, illumination or other output signals need to be kept under control. Two examples of possible application scopes are photovoltaic cells in solar modules and light emitting diodes. By means of the integrated 2.5 GHz transmitter, the device wirelessly sends the collected data to a base station, placed up to a few meters away. Sensors 2018, 18, 110; doi:10.3390/s18010110
www.mdpi.com/journal/sensors
Sensors 2018, 18, 110
2 of 15
Sensors 2018, 18, 110
2 of 16
The microchip was fabricated by a silicon foundry (Austria Micro Systems), with a standard and low Micro with a standard and [9], lowwhich cost demonstrates 0.35 µm CMOS (C35B4M3) [9], which cost 0.35Systems), µm CMOS process (C35B4M3) thatprocess even low cost technologies are demonstrates that even low cost technologies are suitable for this kind of fully integrated sensors. suitable for this kind of fully integrated sensors. The paper paper is organized as follows. The follows. In InSection Section22we wepresent presentthe themicrochip microchiparchitecture architecturewith witha integrated ageneral generaldescription descriptionofofits itsinternal internalorganization. organization.In InSection Section 33 we we provide provide details about the integrated sensors, the the analogue analogue front-end, front-end, the the control control logic logic and and the the RF RF section. section. In Section 4 we we provide provide sensors, experimental results results on on the the microchip microchipoperation. operation. Finally, Finally,the theconclusions conclusionsare aredrawn drawnininSection Section5.5. experimental 2. 2. System Architecture Architecture The of the theproposed proposeddevice deviceisisshown shown Figure 1. The analogue section includes The architecture of inin Figure 1. The analogue section includes two two sensors, namely a temperature sensor and a light sensor. The former is designed to provide sensors, namely a temperature sensor and a light sensor. The former is designed to provide a a“proportional “proportionaltotoabsolute absolutetemperature” temperature”(PTAT) (PTAT) output output and guarantees an optimal linearity; the latter latter is is in in fact fact an an integrated integrated solar solar cell, cell, designed designed to to provide provide aa useful useful output output under under normal normal outdoor outdoor daylight daylight irradiation. irradiation. Sensor outputs are first amplified to bring their their respective respective signals signals to to useful useful levels levels and and then then digitalized digitalized by by means means of of aa single single analogue-to-digital analogue-to-digital converter converter (ADC). (ADC). The The ADC ADC is is time-shared time-shared between between either two or three sources, namely the PTAT, PTAT, the the light light sensor sensor and and in in case case aa generic generic input input applied applied. applied between between two two input input pads pads of the microchip, where an additional external sensor can be applied. The The serial serial conversion conversion timing timing isis provided providedby bythe thedigital digitalsection. section.
Figure 1. Wireless multisensor architecture. Figure 1. Wireless multisensor architecture.
The ADC serial outputs are afterwards loaded into a parallel-in/serial-output (PISO) 2 × 8 (or 3 Theshift ADC serial outputs afterwards loaded into parallel-in/serial-output (PISO) 2×8 × 8) bits register. Once theare measurements are done anda the PISO loaded, its content is serialized (or 3 × 8) bits shift register. Once the measurements are done and the PISO loaded, its content is for transmission at a bit-rate of 3 kbps. The microchip is also internally provided with an integrated serialized for transmission a bit-rate of 3 kbps. The also internally provided with RF transmitter, composed ofatan LC-oscillator, turned onmicrochip and off byisthe bitstream and an RF power an integrated RF transmitter, composed of an LC-oscillator, turned on and off by the bitstream and amplifier (PA), in turn energizing an integrated small dipole antenna. an RF power amplifier (PA), in turn energizing an integrated small dipole antenna. 3. Microchip Blocks Description 3. Microchip Blocks Description 3.1. Temperature TemperatureSensor Sensor 3.1. Low cost cost high-performance high-performance temperature temperature sensors sensors are are increasingly increasingly required. required. Besides, Besides, itit is is needed needed Low to reduce their power consumption. This concept is necessary to use the sensor in battery operated to reduce their power consumption. This concept is necessary to use the sensor in battery operated systems, to to decrease decrease the the errors errors produced produced by by self-heating self-heating and and to to be be able able to to add add the the sensor sensor to to aa chip chip systems, without causing a significant increase in the power consumption of the complete system [10]. without causing a significant increase in the power consumption of the complete system [10]. An integrated integrated CMOS CMOS temperature temperature sensor sensor perfectly perfectly fits fits low low cost, cost, high high performance performance and and low low An power consumption. consumption. power Several temperature sensors topologies have been proposed in CMOS technology. Most of the systems reported in literature, including the one presented hereafter, exploit a PTAT scheme, realized by means of coupled bipolar junction transistors (BJT) [11,12].
Boltzmann constant, q the electronic charge and the ideality factor, if two identical transistors ( = = ) are biased at collector currents of and respectively, the difference between the emitter-base voltages of the two BJTs has a linear dependence on T, according to the following equation. Sensors 2018, 18, 110
∆
=
−
=
ln
−
ln
=
ln
,
3(2) of 15
where = 1 is assumed. Several temperature sensors proposed of the Our sensor was designed totopologies provide a have high been sensitivity andina CMOS signal technology. level tuned Most in order to systems reported in literature, including the one presented hereafter, exploit a PTAT scheme, realized simplify the Analog-to-Digital conversion (ADC) in the temperature range from room temperature by means to 100 °C. of coupled bipolar junction transistors (BJT) [11,12]. In fact, assuming for PNP consists BJTs the collector current given by:a PTAT voltage generator and a The temperature sensor of two building blocks: differential amplifier, as shown in Figure 2. qV ( EB ) IC =with IS e ηkT , (1) To obtain a proportional output signal respect to absolute temperature, two diodeconnected BJTs, driven by different currents, were realized. The two PNP transistors Q1 and Q2, both with Is theinsaturation therealized emitter-base junction,on VEB emitter-base Boltzmann working the activecurrent region,ofare side-by-side thethe chip and have voltage, the samek the geometry and constant, q the electronic charge and η the ideality factor, if two identical transistors ( I = I = IS ) S1 S2 orientation and therefore can be assumed identical. The ratio between their collector currents is fixed are biased at collector currents of I and Io respectively, the difference between the emitter-base re fM1 and M2. Assuming in principle negligible channel length by the different W/L aspect ratios of voltages of the two BJTs has a linear onIoT,that according to following modulation effects [13], Q2 is drivendependence by a current is the , where n is equation. the channel width sizing of the MOS ratio WM1/WM2 while LM1 = LM2. In this work n=10I was chosen for an appropriate Ire f Io kT re f (2) ∆Vbetween − VEB2 =inVtthe ln measured − Vt ln temperature = ln range, and linearity of the transistor as trade-off sensitivity EB = VEB1 Is Is q Io output signal. The current is generated through an integrated Bandgap current reference [14] where η =reference 1 is assumed. with Our supply andwas temperature compensation that deliversand a constant of 11.5 µA, showing a sensor designed to provide a high sensitivity a signal current level tuned in order to simplify temperature dependency of about 800 in ppm/°C. As shown by the this temperature the Analog-to-Digital conversion (ADC) the temperature range fromsimulations, room temperature to 100 ◦ C. dependency has no practical impactofontwo thebuilding performance two sensors. The source designed The temperature sensor consists blocks:ofa the PTAT voltage generator and aisdifferential to supply as each diode-connected amplifier, shown in Figure 2. BJT with currents that are independent from .
Figure with differential differential amplifier. amplifier. Figure 2. 2. Architecture Architecture of of the the integrated integrated PTAT PTAT with
To obtain a proportional output signal with respect to absolute temperature, two diode-connected BJTs, driven by different currents, were realized. The two PNP transistors Q1 and Q2, both working in the active region, are realized side-by-side on the chip and have the same geometry and orientation and therefore can be assumed identical. The ratio between their collector currents is fixed by the different W/L aspect ratios of M1 and M2. Assuming in principle negligible channel length modulation effects [13], Q2 is driven by a current Io that is n × Ire f , where n is the channel width ratio WM1 /WM2 while LM1 = LM2 . In this work n=10 was chosen for an appropriate sizing of the MOS transistor as trade-off between sensitivity in the measured temperature range and linearity of the output signal. The reference current Ire f is generated through an integrated Bandgap current reference [14] with supply and temperature compensation that delivers a constant current of 11.5 µA, showing a temperature dependency of about 800 ppm/◦ C. As shown by the simulations, this temperature
Sensors 2018, 18, 110
4 of 15
dependency has no practical impact on the performance of the two sensors. The source is designed to Sensors 2018, 18,diode-connected 110 4 of 16 supply each BJT with currents that are independent from Vdd . In Figure 3 the simulated emitter-base voltages of transistors Q1 and Q2 at various temperatures In Figure 3 the simulated emitter-base of transistors Q1toand Q2 at various temperatures are shown. Their behaviour shows a negativevoltages derivative with respect temperature. From simulations, are shown. Their behaviour shows a negative derivative with respect to temperature. From it is possible to calculate the sensor output at each absolute temperature: simulations, it is possible to calculate the sensor output at each absolute temperature: VPTAT ) = V−+ −, V− , ( ()T=
(3) (3)
◦C In the the same same Figure, Figure,the thedifferential differentialoutput outputofofthe thePTAT PTAT temperature range from to 100 In in in thethe temperature range from 0 to0 100 °Cshown. is shown. is
Figure 3. Emitter-base Emitter-base voltage voltage of of transistors transistors Q1 Q1 and and Q2 Q2 at various temperatures temperatures (right scale) and differential output of the PTAT (left scale) scale) in in the the temperature temperature range from 0 to to 100 100 ◦°C. C. The lines are guides to the eye.
In the thetemperature temperature range of interest, the estimated consumption ofPNP the transistors two PNP In range of interest, the estimated power power consumption of the two transistors is 7.7 µW and 66.5 µW respectively for Q1 and Q2, obtained from simulations. This is 7.7 µW and 66.5 µW respectively for Q1 and Q2, obtained from simulations. This power power might might cause self-heating of the sensor and therefore the PTAT circuit is activated only for a few cause self-heating of the sensor and therefore the PTAT circuit is activated only for a few milliseconds milliseconds through the Bandgap circuit at regular intervals. through the Bandgap circuit at regular intervals. The amplifier amplifier stage stage was was designed designed through through extensive extensive parametric parametric simulations simulations aimed aimed at at obtaining obtaining The output signal output signal levels levels that that could could be be reasonably reasonably applied applied to to the the ADC, ADC, while while preserving preserving aa good goodlinearity. linearity. Here, M3 and M4 are used as active load, instead of using high value resistors, whose implementation Here, M3 and M4 are used as active load, instead of using high value resistors, whose implementation is wasteful in terms is wasteful in terms of of area area consumption. consumption. M5 M5 and and M6 M6 have have been been sized sized to to achieve achieve the the maximum maximum output output dynamics for the differential amplifier. Being the PTAT output differential, the topology requested dynamics for the differential amplifier. Being the PTAT output differential, the topology requested for for the amplifier is differential as well. Mhas tail has the role of a current limiter for the amplifier, in order the amplifier is differential as well. Mtail the role of a current limiter for the amplifier, in order to to reduce power consumption. reduce power consumption. 3.2. Light 3.2. Light Sensor Sensor The proposed proposed solution solution for design of the CMOS CMOS light light sensor sensor is is based based on on planar planar p-n p-n junction junction The for the the design of the formed by the P-type substrate and the N-type well. This structure could be figured out as a micro formed by the P-type substrate and the N-type well. This structure could be figured out as a micro solar solarproducing cell producing a current proportional to surface irradiation. The outermost metal layer (Metalcell a current proportional to surface irradiation. The outermost metal layer (Metal-1 for the 1 for the CMOS technology) usedfingers to build collect photocurrent at the The CMOS technology) was used was to build thatfingers collectthat photocurrent at the cathode. Thecathode. structure of structure of the light sensor is shown in Figure 4. the light sensor is shown in Figure 4.
Sensors 2018, 18, 110
5 of 15
Sensors 2018, 18, 110
5 of 16
Sensors 2018, 18, 110
5 of 16
Figure 2. Structure of the realized photodetector: (a) top view; (b) cross section (figure not in scale). Figure 4. Structure of the realized photodetector: (a) top view; (b) cross section (figure not in scale). Figure 2. Structure of the realized photodetector: (a) top view; (b) cross section (figure not in scale).
In order to convert the current signal into a voltage signal, it is necessary to use a transimpedance In order to convert current signal a voltage signal, isFigure necessary to use a transimpedance amplifier (TIA), designed according tointo theinto depicted in 5. to use In order to convert the the current signal aschematic voltage signal, it isit necessary a transimpedance amplifier (TIA), designed according to the schematic depicted in Figure 5. amplifier (TIA), designed according to the schematic depicted in Figure 5.
Figure 5. Integrated light sensor with transimpedance amplifier. Figure 5. Integrated lightlight sensor withwith transimpedance amplifier. Figure 5. Integrated sensor transimpedance amplifier.
The integrated solar cell area (300 × 200 µm2) was chosen to generate a photocurrent in a range that the linear of (300 the amplifier to chosen 22 µA).toBy assuming an indicative Thesuits integrated solarregion cell area × 200 µm2(up ) was generate a photocurrent in amaximum range 2 ) was chosen to generate a photocurrent in a range The integrated solar cell area (300 × 200 µm 2 10%, the areaof was in such way under 1000 W/man theindicative cell wouldmaximum still provide thatefficiency suits theoflinear region thechosen amplifier (upa to 22that µA). By assuming suits the linear region of the amplifier (up to µA).1000 Byan assuming anwould indicative maximum 2 the cell athat photocurrent within the chosen said limits . The TIAthat is 22 based on integrated MOSFET operational efficiency of 10%, the area was in such a way under W/m still provide 2 the cell would still provide efficiency of 10%, the area was chosen in such a way that under 1000 W/m amplifier. This operational amplifier of 100 dB with an output resistance a photocurrent within the said limits. performs The TIA an is open basedloop on gain an integrated MOSFET operational a photocurrent within the said limits. The TIA is based on an integrated MOSFET operational amplifier. of 0.3 ΩThis andoperational can operateamplifier in the temperature range 120dB°C. Theansimulated TIA output amplifier. performs an openfrom loop−40 gain°Cofto100 with output resistance This operational amplifier performs an open loop gain of 100 dB with an output resistance of 0.3 Ω is shown Figure 6. The feedback resistor Rfeed from value−40 is 100 kΩ. of 0.3 Ω and in can operate in the temperature range °C to 120 °C. The simulated TIA output and can inThe the temperature rangeRfrom −40 ◦ C to 120 ◦ C. The simulated TIA output is shown is shown inoperate Figure 6. feedback resistor feed value is 100 kΩ. in Figure 6. The feedback resistor Rfeed value is 100 kΩ.
Sensors 2018, 18, 110 Sensors 2018, 18, 110
6 of 15 6 of 16
Figure 3. Simulated characteristic of the transimpedance amplifier output vs photocurrent. Figure 6. Simulated characteristic of the transimpedance amplifier output vs photocurrent.
3.3. Clock 3.3. Clock Conventional current-starved single-ended ring oscillators are usually employed for realizing Conventional current-starved single-ended ring oscillators are usually employed for realizing wide tuning range, voltage controlled oscillators that act as local clock for the digital circuitry of wide tuning range, voltage controlled oscillators that act as local clock for the digital circuitry of ASICs ASICs by exploiting the dependence of the large signal oscillation frequency fOsc on the bias current by exploiting the dependence of the large signal oscillation frequency fOsc on the bias current Ib [15,16] Ib [15,16] that is determined by: that is determined by: 1 1 Ib fOsc , == ≈ ≈ , (4)(4) 2 2Nt D2 2NOscC where the oscillation oscillation amplitude amplitude where N N is is the the number number of of inverters invertersused, used,ttDD the the time time delay delay of of each each stage, stage,VVOsc Osc the and C the load capacitance. and C the load capacitance. Furthermore, Furthermore, in in aa ring ring oscillator oscillator topology topology the the small-signal small-signal amplitude amplitude of of the the oscillating oscillating signal signal can can be expressed as be expressed as ! √ | As | − 2 | As | 3 VOut (t) = vOut (0) exp| | − 2 ωs t cos ωs t , (5) | |√3 2 2 (5) ( )= (0) exp cos , 2 2 where vOut (0) is the initial condition at the ring oscillator output, As the gain of each amplifier stage where vOut(0) is the initial condition at the ring oscillator output, As the gain of each amplifier stage and ω s the −3 dB bandwidth of each stage [13]. The first term of (5) represents the time-varying and ωs the −3 dB bandwidth of each stage [13]. The first term of (5) represents the time-varying amplitude of the oscillations. Therefore, the settling time can be expressed to a first approximation as amplitude of the oscillations. Therefore, the settling time can be expressed to a first approximation as 2 2 0.9 0.9VOsc , t= = ln s ln , (6)(6) v(0) − 2) ω s Out (0) (| (||A−s |2)
ring oscillator oscillator (RO), at aa given given initial initial condition, condition, depends only on the The settling time of a ring amplifier stages stagescharacteristics characteristics and bandwidth). In particular, when the oscillator is amplifier (gain(gain and bandwidth). In particular, when the oscillator is implemented implemented by exploiting CMOS inverters, thegain small-signal of is a single stage is given by by exploiting CMOS inverters, the small-signal of a singlegain stage given by =− g +g A s = − gm n + gm p r o n k r o p
(7) (7)
where ron and rop are the output resistances of the MOS devices. However, an increase of the where ron and rop of arethe theinverting output resistances thenecessarily MOS devices. an of increase of the transconductance stages doesofnot implyHowever, a reduction the oscillator transconductance of the ainverting does not necessarily imply a reduction ofand the oscillator settling settling time because higher stages gain requires more current consumption thus larger size time because a higher gain requires more current consumption and thus larger size components with components with increased parasitics that tend to decrease ωs. To obtain a fast start-up oscillator, increased parasitics that tend to decrease ω s . and To obtain fast start-up forset. low power required for low power clocks, bias current henceadevices sizes oscillator, have to berequired properly clocks, currentthe and hence devices sizes have carrier to be properly set. frequency (Wp/Wn = 4, C1 = C2 = C3 In bias this work, oscillator provides a stable of 10 MHz = 790 pF). For achieving this purpose, the same bandgap current reference introduced in Section 3.1
Sensors 2018, 18, 110
7 of 15
of216 the oscillator provides a stable carrier of 10 MHz frequency (Wp /Wn = 4, C1 =7 C = 7 of 16 C3 = 790 pF). For achieving this purpose, the same bandgap current reference introduced in Section 3.1 has used,providing providinghigh highrobustness robustness against supply voltage temperature fluctuations. A has been been used, against supply voltage andand temperature fluctuations. A clock has been used, providing high robustness against supply voltage and temperature fluctuations. A clock signal of 5 kHz is thus obtained using a frequency divider (f in/fout = 2000). signal of 5 kHz is thus obtained using a frequency divider (fin /fout = 2000). clock signal of 5 kHz is thus obtained using a frequency divider (fin/fout = 2000). SensorsIn 2018, 110 this18,work, Sensors 2018, 18, 110
3.4. 3.4. Digital DigitalSection Section 3.4. Digital Section In InFigure Figure 77the theblocks blocks of ofthe thedigital digitalsection sectionare areshown. shown. An AnADC ADCwas wasused usedto toconvert convertthe theanalogue analogue In Figure 7 the blocks of the digital section are shown. An ADC was used to convert the analogue signals signals produced produced by by the the sensors; sensors; aa parallel-in-serial-output parallel-in-serial-output (PISO) (PISO) shift shift register register was was implemented implemented to to signals produced by the sensors; a parallel-in-serial-output (PISO) shift register was implemented to serialize the parallel output of the ADC; a Load signal is necessary to activate the loading by the PISO. serialize the parallel output of the ADC; a Load signal is necessary to activate the loading by the PISO. serialize the parallel output of the ADC; a Load signal is necessary to activate the loading by the PISO. In In this this work, work, an an integrated integrated 8-bit 8-bit successive successive approximation approximation ADC ADC was was used, used, available available in in the the standard standard In this work, an integrated 8-bit successive approximation ADC was used, available in the standard cell cell library library of of the the AMS AMS technology. technology. cell library of the AMS technology.
Figure Figure 7. 7. Block Blockdiagram diagramof ofthe theDigital Digital Section. Section. Figure 7. Block diagram of the Digital Section.
The PISO output normally consists of 44 ××8-bit used however when the The 8-bitwords. words.Five Fivewords wordsare are The PISO PISO output output normally normally consists consists of of 4 × 8-bit words. Five words are used used however however when when the the external sensor isis also selected. The first and the last words are used to identify Start and Stop of external sensor also selected. The first and the last words are used to identify Start and external sensor is also selected. The first and the last words are used to identify Start and Stop Stop of of modulation. They can additionally contain a microchip read-only identifier. The entire sequence of modulation. modulation. They They can can additionally additionally contain contain aa microchip microchip read-only read-only identifier. identifier. The The entire entire sequence sequence of of bits is depicted in Figure 8. Between the first and the last 8-bits, there are the digital values of the bits is depicted in Figure 8. Between the first and the last 8-bits, there are the digital values of the PTAT bits is depicted in Figure 8. Between the first and the last 8-bits, there are the digital values of the PTAT and light sensor (payload). and light (payload). PTAT andsensor light sensor (payload).
Figure 8. Serialized output of 32-bits PISO. Figure 8. Serialized output of 32-bits PISO. Figure 8. Serialized output of 32-bits PISO.
The Load command generator was designed to generate a high-level every 50 cycles of the 5 kHz The Load command generator was designed to generate a high-level every 50 cycles of the 5 kHz Clock.The Load command generator was designed to generate a high-level every 50 cycles of the Clock. 5 kHz Clock. 3.5. RF Section 3.5. RF Section 3.5. RF In Section this microchip sensor, a directly On-Off-Keying (OOK) modulated oscillator-based In this microchip sensor, a directly On-Off-Keying (OOK) modulated oscillator-based transmitter was used, sensor, as depicted in Figure 9. The transmitter topologyoscillator-based exploits a cross-coupled In this microchip a directly On-Off-Keying (OOK) modulated transmitter transmitter was used, as depicted in Figure 9. The transmitter topology exploits a cross-coupled complementary LC oscillator running a frequency of 2.5 GHz. Thea cross-coupled differential output is coupled was used, as depicted in Figure 9. The at transmitter topology exploits complementary complementary LC oscillator running at a frequency of 2.5 GHz. The differential output is coupled to a common source differential power amplifier and buffered on an integrated dipole antenna. A LC oscillator running at a frequency of 2.5 GHz. The differential output is coupled to a common source to a common source differential power amplifier and buffered on an integrated dipole antenna. A start-up circuit allows the duty cycling of transmitterdipole and the implementation of the allows OOK differential power amplifier and buffered onthe an integrated antenna. A start-up circuit start-up circuit allows the duty cycling of the transmitter and the implementation of the OOK modulation. the duty cycling of the transmitter and the implementation of the OOK modulation. modulation.
Sensors 2018, 18, 110 Sensors 2018, 18, 110
8 of 15 8 of 16
Figure 4. Circuit schematic of the implemented transmitter. Figure 9. Circuit schematic of the implemented transmitter.
LC cross-coupled oscillators are suitable circuits to be used in radio frequency systems. These LCofcross-coupled oscillators are suitable circuits be used in radio frequency systems. These kinds oscillators offer high frequency stability overtotemperature, voltage and process tolerances kinds of oscillators offer high frequency stability over temperature, voltage and process tolerances and and low phase noise [13]. low phase noise [13]. In this oscillator topology, the oscillation frequency is set by an inductors-capacitors (LC) In this oscillatortotopology, the oscillation frequency is set by an inductors-capacitors (LC) network network according the formula: according to the formula: = ,1 (8) f = √ √ , (8) 2π LC Despite the simplicity of the cross-coupled differential scheme, the complementary architecture Despite the simplicity the cross-coupleduseful differential scheme, thethe complementary architecture features a higher negativeoftransconductance for overcoming resonating circuit losses, features a higher negative transconductance useful for overcoming the resonating circuit losses, essentially due to the parasitic resistance of the LC tank and sustaining oscillations [17]. In particular essentially duerelationship to the parasitic resistance of the the following must be verified [18]:LC tank and sustaining oscillations [17]. In particular the following relationship must be verified [18]: ≥ , (9) 2 ≥ Rtankdifferential , (9) is the transconductance provided couple and is the parasitic where Gm by the resistance of the LC tank. whereThe Gmcomplementary is the transconductance provided by the differential Rtankfactors is the (parasitic topology allows oscillations to start alsocouple for lowand quality ) of resistance of the LC tank. the resonating circuits, which is typical for fully integrated oscillators. In fact, the inductor shows a complementary topologyinallows oscillations to start also for low quality poorThe quality factor , typically a range 4–20 for enhanced structures [19,20]. factors (Qtank ) of the resonating circuits, which is typical for fully integrated oscillators. In fact, thewe inductor shows asquare poor The resonating circuit employs L = 4 nH and C = 1 pF. In particular, used 4-turns quality factor Q , typically in a range 4–20 for enhanced structures [19,20]. L fabricated in the top aluminium metal layer and polysilicon capacitors. The spiral inductors The resonating circuit employs L = of 4 nH C GHz = 1 pF. In particular, we used 4-turns square spiral employed coils exhibit a quality factor 4.7and at 2.5 [9]. inductors fabricated in the top aluminium metal layer and polysilicon capacitors. The employed coils Each transistor has a W/L ratio set to achieve the needed negative transconductance for exhibit a quality factor of 4.7 at 2.5 GHz [9]. sustaining oscillations. Furthermore, the sizes of the cross-coupled pairs, together with the current has a W/L ratio achieve the needed for sustaining tail, Each weretransistor chosen to properly biasset theto differential power negative amplifier.transconductance Table 1 summarizes the key oscillations. Furthermore, the sizes of the cross-coupled pairs, together with the current tail, were dimensions of each transistor. chosen to properly bias the differential power amplifier. Table 1 summarizes the key dimensions of each transistor. Table 1. Transistor dimensions of differential cross coupled LC oscillator. Table 1. Transistor coupled Mtail LC oscillator. MP1dimensions MP2 of differential MN1 cross MN2
W 72 μm 0.35 μm LMP1 W
72 µm
72 μm MP2 μm 0.35 72 µm
72 μm MN1 0.35 μm 72 µm
72 μm 100 μm MN20.35 μm Mtail 0.35 μm 72 µm
100 µm
L 0.35 µm (PA) based 0.35 µm 0.35 0.35 µm 0.35 µm A linear power amplifier on a class B µm bridge configuration amplification stage was implemented to feed the small dipole antenna. The differential MOSFET pairs were accurately sized for maximizing the power delivered to the antenna alternatively through the Mpa3-Mpa2 and Mpa4Mpa1 paths. A W/L ratio of 65 µm/0.35 µm was set for all Mpa transistors. The estimated power gain
Sensors 2018, 18, 110
9 of 15
A linear power amplifier (PA) based on a class B bridge configuration amplification stage was implemented to feed the small dipole antenna. The differential MOSFET pairs were accurately sized for maximizing the power delivered to the antenna alternatively through the Mpa3-Mpa2 and Mpa4-Mpa1 paths. A W/L ratio of 65 µm/0.35 µm was set for all Mpa transistors. The estimated power gain is Sensors 2018, 18, 110 9 of 16 2.3 dB with a total harmonic distortion (THD) of 0.13%, estimated from simulations. This stage is turned Mtail3 only during distortion transmission. is 2.3ondBbywith a total harmonic (THD) of 0.13%, estimated from simulations. This stage is The on-chip dipole radiating element is made within the top aluminium metal layer, having turned on by Mtail3 only during transmission. a thickness 0.925 µm. This antenna requires a differential excitation, whichmetal suits the PAhaving scheme. Theof on-chip dipole radiating element is made within the top aluminium layer, a For a resonating antenna, the linear dimensions must be of thewhich ordersuits of one halfscheme. of the carrier thickness of 0.925 µm. This antenna requires a differential excitation, the PA wavelength. operations at 2.4dimensions GHz, a well-tuned antenna should For aTherefore, resonatingfor antenna, the linear must be dipole of the order of one halfhave of thedimensions carrier wavelength. Therefore, for operations at 2.4size GHz, a the well-tuned antenna have of the order of centimetres. However, the chip limits antennadipole dimension, so itshould is not possible dimensions of the of centimetres. However, chip size limits the antenna dimension, so ait small is even approaching theorder resonating conditions in our the case. The characteristic parameters of such not possible even approaching resonating conditions in our case. The parameters of of antenna are obviously very poor,the with an estimated gain of about −50 dBcharacteristic and a radiation efficiency small are obviously very poor, with estimated gain of about the such ordera of 1%antenna [21], enough however to allow shortanrange transmissions [6].−50 dB and a radiation efficiency of the order of 1% [21], enough however to allow short range transmissions [6].
4. Experimental Results
4. Experimental Results
The circuital sections were first separately characterized. In particular, tests were made on the The circuital sections were first separately characterized. In particular, tests were made on the analogue sensors (temperature and light), the digital conversion section with data serialization and analogue sensors (temperature and light), the digital conversion section with data serialization and finally on the RF transmitter. finally on the RF transmitter. A microphotograph of the die is shown in Figure 10a. The area of the realized chip is 1 mm × 4 mm. A microphotograph of the die is shown in Figure 10a. The area of the realized chip is 1 mm × 4 After fabrication, for a complete characterization, thethetest packagedinin 44-pin TQFP mm. After fabrication, for a complete characterization, testchips chips were were packaged 44-pin TQFP packages (Figure 10b) which give easy access to all input and output nodes of the circuit. Bare packages (Figure 10b) which give easy access to all input and output nodes of the circuit. Bare diesdies samples were also tested after toaalimited limitednumber number bondpads, samples were also tested afterultrasonic ultrasonicwire wire bonding bonding to ofof bondpads, e.g.,e.g., for for RF RF transmission tests. transmission tests.
(a)
(b) Figure 10. (a) of theofrealized chip. The of the microchip are highlighted; Figure 10. Microphotograph (a) Microphotograph the realized chip.main Thesections main sections of the microchip are (b) Microphotograph of an encapsulated microchip. microchip. highlighted; (b) Microphotograph of an encapsulated
Temperature Sensor 4.1. 4.1. Temperature Sensor To investigate performanceof ofthe the PTAT PTAT sensor, sensor, the in in a thermostatic To investigate thethe performance the whole wholechip chipwas wastested tested a thermostatic oven, using a standard PT100 as reference as shown in Figures 11 and 12. oven, using a standard PT100 as reference as shown in Figures 11 and 12.
Sensors 2018, 18, 110
10 of 15
Sensors 2018, 18, 110
10 of 16
Sensors 2018, 18, 110
10 of 16
Figure Experimentalsetup setup used used to PTAT performances. Figure 11.11. Experimental to characterize characterizethe the PTAT performances. Figure 11. Experimental setup used to characterize the PTAT performances.
12. Picture thesetup setup used thethe characterization of the temperature sensor. FigureFigure 12. Picture ofofthe usedforfor characterization of the temperature sensor.
Figurein12.Figure Picture the integrated setup usedPTAT for thepresents characterization of the temperature sensor. As shown 13,ofthe a good linearity (R2 = 0.99962) in the
2 = 0.99962) in the Astemperature shown inrange Figure 13, the integrated presents a good linearity (Ranalogue of interest (20 °C–90 °C) with PTAT a sensitivity of 19.1 mV/°C, measured at the ◦ ◦ ◦ 2 As range shown Figurethe13, the integrated presents a 19.1 good (R the = lighting 0.99962) the amplifier output. Inside climatic chamber it PTAT was possible to change inmV/ alinearity smallC, range temperature ofininterest (20 C–90 C) with a sensitivity of measured at theinanalogue conditions in combination with temperature. In particular, a modification of the LED array output, temperature range of interest (20 °C–90 °C) with a sensitivity of 19.1 mV/°C, measured at the analogue amplifier output. Inside the climatic chamber it was possible to change in a small range the lighting 2 approximately, was applied showing no effect at all on the temperature sensor from 10 output. to 50 W/m amplifier Inside the climatic chamber In it was possible atomodification change in a small range thearray lighting conditions in combination with temperature. particular, of the LED output, output. However, it can be expected that higher irradiations would induce heating of the device. conditions in combination with temperature. In particular, a modification of the LED array output, 2 from 10 to 50 W/m approximately, was applied showing no effect at all on the temperature sensor This specific implementation of the sensor does not include a calibration circuitry, which could from 10 to 2018, 50 W/m sensor Sensors 18, 1102 approximately, was applied showing no effect at all on the temperature 11 of 16 be however obtained in a final version by e.g., laser trimming acting on I ref or other standard output. However, it can be expected that higher irradiations would induce heating of the device.
output. However, it can be expected that higher irradiations would induce heating of the device. techniques. This specific implementation of the sensor does not include a calibration circuitry, which could be however obtained in a final version by e.g., laser trimming acting on Iref or other standard techniques.
Figure 13.Figure Experimental characteristic of theofPTAT sensor, obtained using analogue 13. Experimental characteristic the PTAT sensor, obtained using analogueoutput. output. The The simulated simulated characteristic also provided comparison. error between simulated and characteristic is also provided foriscomparison. Theforerror between The simulated and experimental characteristic experimental characteristic is also added. is also added. Table 2 summarizes the PTAT key parameters.
This specific implementation of the sensor does not include a calibration circuitry, which could be Table 2. Performance summary of proposed PTAT sensor. however obtained in a final version by e.g., laser trimming acting on Iref or other standard techniques. Power Drawn Current Area Temperature Sensitivity Table 2 summarizes the PTAT key parameters. Linearity Supply (V) 3.3
(μA) 150
(mm2) 0.022
Range (°C) 25–90
(mV/°C) 19.1
0.99962
4.2. Light Sensor To examine the performance of the light sensor, the whole chip was exposed to direct sunlight and the photocurrent was measured at several irradiations (IRR), continuously monitored with a
Sensors 2018, 18, 110
11 of 15
Table 2. Performance summary of proposed PTAT sensor. Power Supply (V)
Drawn Current (µA)
Area (mm2 )
Temperature Range (◦ C)
Sensitivity (mV/◦ C)
Linearity
3.3
150
0.022
25–90
19.1
0.99962
4.2. Light Sensor To examine the performance of the light sensor, the whole chip was exposed to direct sunlight and the photocurrent was measured at several irradiations (IRR), continuously monitored with a hand-held solar power meter. Photocurrent has been measured using an Agilent Technologies 4155C Semiconductor parameter analyser. As shown in Figure 14 the light sensor presents a good linearity (R2 = 0.99915) in the range of interest of a typical sunny day. The sensitivity is calculated as follows: SCel +Tia =
∆VTI A = 16.96 mV/W/m2 , ∆IRR
(10)
In this specific implementation of the sensor, an additional calibration feature was not considered, though it would be possible by e.g., laser trimming of Rfeed . Temperature variations can affect the light sensor output. This dependence is shown in Figure 15, which demonstrates that, due to the amplifier non-idealities, the light reading saturates at lower and lower photocurrent as the temperature increases. Sensors 2018, 2018, 18, 18,values 110 12 of of 16 16 Sensors 110 12
Figure 14. Characteristic of the irradiation sensor, obtained using analogue Figure 14. Characteristic Characteristic of the the irradiation sensor, sensor, obtained using analogue output. output. Figure 14. of irradiation obtained using analogue output.
Figure 15. Light sensor output at different temperatures. Figure 15. Light Light sensor outputvs. vs.photocurrent photocurrent at at different different temperatures. Figure 15. sensor output vs. photocurrent temperatures. 4.3. RF RF Transmitter Transmitter Performance Performance 4.3. The RF RF performance performance of of the the implemented implemented transmitter transmitter was was characterized characterized by by analysing analysing the the The transmitted signal signal received received by by aa 0.4–3 0.4–3 GHz GHz bandwidth, bandwidth, 55 dBi dBi logarithmic logarithmic periodic periodic dipole dipole antenna antenna transmitted (model Rhode&Schwarz Rhode&Schwarz HL040) HL040) connected connected in in turn turn to to aa Rhode&Schwarz Rhode&Schwarz FSV FSV Spectrum Spectrum Analyser. Analyser. (model The antenna was placed at a distance of 1 m from the microchip. The antenna was placed at a distance of 1 m from the microchip. The experimental experimental measurements measurements show show aa continuous continuous wave wave (CW) (CW) signal, signal, produced produced by by the the crosscrossThe coupled LC LC oscillator, oscillator, at at about about 2.54 2.54 GHz GHz as as shown shown in in Figure Figure 16. 16. coupled
Sensors 2018, 18, 110
12 of 15
4.3. RF Transmitter Performance The RF performance of the implemented transmitter was characterized by analysing the transmitted signal received by a 0.4–3 GHz bandwidth, 5 dBi logarithmic periodic dipole antenna (model Rhode&Schwarz HL040) connected in turn to a Rhode&Schwarz FSV Spectrum Analyser. The antenna was placed at a distance of 1 m from the microchip. The experimental measurements show a continuous wave (CW) signal, produced by the Sensors 2018, LC 18, 110 13 of 16 cross-coupled oscillator, at about 2.54 GHz as shown in Figure 16. Sensors 2018, 18, 110
13 of 16
Figure 16.16. Frequency graphofofthe thereceived received signal. Figure Frequencyspectrum spectrum graph signal.
The proposed oscillator showsa aphase phase noise of − −90 dBc/Hz at 1MHz offset fromfrom the centre The proposed LC LC oscillator shows noise of at 1MHz offset the centre Figure 16. Frequency spectrum graph of90 thedBc/Hz received signal. frequency, as shown in Figure 17. The phase noise is estimated as follows: frequency, as shown in Figure 17. The phase noise is estimated as follows:
The proposed LC oscillator shows a )phase(noise) of −90 dBc/Hz at 1MHz offset from the centre (∆ ) − − 10 = ( (11) , frequency, as shown in Figure 17. The phase noise is estimated as follows: RBW L(phase ∆ f )dBc = Ps (dBm Pc (dBm 10log ) −noise ) −level is the noise, is the power in10a band of 1, Hz at a certain offset (11) where (∆ ) )− ( ) − 10 (∆ ) = ( ,1Hz (11) frequency from the carrier central frequency f0, is the power level of the carrier at f0 and RBW is theL(bandwidth resolution ofnoise, the spectrum analyser. where ∆ f )(∆ phase Ps isisthe noise levelinina aband band 1 Hz a certain is the phase noise, the noise power power level of of 1 Hz at aat certain offsetoffset where dBc)is the
frequency fromfrom the carrier central frequency f 0 ,f0P, c is isthe andRBW RBWisis the frequency the carrier central frequency thepower powerlevel levelof ofthe the carrier carrier at f00and the bandwidth resolution the spectrum analyser. bandwidth resolution of the of spectrum analyser.
Figure 17. Phase noise of the received signal.
The LC transmitter shows a temperature sensitivity Δf/ΔT =signal. −0.25 MHz/°C Figure of the thereceived received signal. Figure17. 17.Phase Phase noise noise of 4.4. Whole System Operation
Thetransmitter LC transmitter shows a temperaturesensitivity sensitivity Δf/ΔT = −0.25 MHz/°C The LC shows a temperature ∆f /∆T = −0.25 MHz/◦ C.
The operation of the whole system was tested employing a commercial hand-held IC-R20 (Icom 4.4. WholeInc., System Operation America 12,421 Willows Road NE, Kirkland, WA, USA) receiving unit placed at a distance of 1 m, although the transmitted RF system signal was to a distance of 1.5 m (Table The The operation of the whole wasclearly tested detectable employingup a commercial hand-held IC-R203). (Icom assembled device was tested under the sun at a temperature of 24 °C and an irradiation level of 1125 America Inc., 12,421 Willows Road NE, Kirkland, WA, USA) receiving unit placed at a distance of 1 m, although the transmitted RF signal was clearly detectable up to a distance of 1.5 m (Table 3). The
Sensors 2018, 18, 110
13 of 15
4.4. Whole System Operation The operation of the whole system was tested employing a commercial hand-held IC-R20 (Icom America Inc., 12,421 Willows Road NE, Kirkland, WA, USA) receiving unit placed at a distance of 1 m, although the transmitted RF signal was clearly detectable up to a distance of 1.5 m (Table 3). The assembled device was tested under the sun at a temperature of 24 ◦ C and an irradiation level of 1125 W/m2 . In agreement with the previous characterization, the output voltage of the PTAT and light sensor were 0.763 V and 1.971 V respectively. Table 3. Comparison table with state-of-the-art sensors. Ref.
Area (mm × mm)
Antenna Type
Reading Distance
Temperature Range (◦ C)
This work 1×4 Dipole 1.5 m U. Calpa et al. [5] 1.5 × 1.5 Inductor 110 cm Sensors 2018,Kim 18, 110 Boram et al. [22] 7.5 × 7.5 Inductor 5 mm Arsalan et al. [8] 3.2 × 1.5 Spiral 3 cm 2. In W/mJun agreement the×previous characterization, Yin et al. [23] with 0.9 1.25 External 4the m output
Additional Sensors
25–90 light 15–35 no 14 of 16 24–40 no 27–35 no voltage −20–30 of the PTAT no and light
sensor were 0.763 V and 1.971 V respectively. Thetransmitted transmitteddata databitstream bitstreamisisshown shownininFigure Figure1818(top) (top)together togetherwith withthe therecorded recordedsignal signalatat The theoutput outputofofthe theradio radioreceiver receiver(bottom). (bottom).Figure Figure1818(bottom) (bottom)shows showssignal signalstability stabilityissues issuesthat thatcould could the be suppressed by properly designing the AC-coupling of the demodulated signal to a low-frequency be suppressed by properly designing the AC-coupling of the demodulated signal to a low-frequency (audio)amplifier amplifierinstead insteadofofthat thatintegrated integratedininthe theused usedcommercial commercialreceiver. receiver.The Thereceived receivedvalues valuesare are (audio) 2 ◦ 2 23.4C, °C,1118 1118W/m W/m .. 23.4
Figure Sampletransmitted transmittedcode code(top) (top)and anddemodulated demodulated signal (bottom) recorded the output Figure 18.18.Sample signal (bottom) asas recorded atat the output the receiver. The stream contains data produced PTAT and light sensors. ofof the receiver. The bitbit stream contains thethe data produced byby thethe PTAT and thethe light sensors.
5. Conclusions
Table 3. Comparison table with state-of-the-art sensors.
Area Antenna Reading Temperature Additional In this paper, Ref. a full custom microchip realized in 0.35 µm CMOS technology, designed for ambient Type a temperature Distance andRange (°C) Sensors × mm) It features (mm and processes monitoring, was proposed. an illumination sensor. This work 1×4 Dipole m 25–90 light The temperature sensor features a sensitivity of 19.1 mV/◦ C and1.5 a linearity 0.99962 in the temperature ◦ ◦ 1.5 ×illumination 1.5 Inductor 110 cm 15–35 rangeU.ofCalpa 20 Cettoal.90[5]C, while the sensor, designed to operate in the range fromno 0 to 2 , shows 2 with a linearity Boram et al.a[22] 7.516.96 × 7.5 mV/W/m Inductor 5 mm of 0.99915.24–40 no 1200 W/mKim sensitivity of Arsalan et al. 2.5 [8] GHz transmitter 3.2 × 1.5with an On-Chip-Antenna Spiral 3 cm An integrated (OCA)-made27–35 in the top metal no layer Yindie-transmits et al. [23] 0.9 × 1.25 Externalby means 4m −20–30 no of theJun same the collected data wirelessly of an on-off amplitude modulation 5. Conclusions In this paper, a full custom microchip realized in 0.35 µm CMOS technology, designed for ambient and processes monitoring, was proposed. It features a temperature and an illumination sensor. The temperature sensor features a sensitivity of 19.1 mV/°C and a linearity 0.99962 in the
Sensors 2018, 18, 110
14 of 15
scheme, at a bit rate of 3 kbit/s. The low bit rate was chosen in order to obtain that the decoded signal at the receiver output were in the audible band, as the receiver is a commercial low-cost radio receiving unit. However, simulations indicate that bit rates as high as 100 kbit/s could be easily obtained with the same technology. Acknowledgments: This work was partially funded by the Italian Government—Ministry of Economic Development, under the call “Progetti di Innovazione Industriale per l’Efficienza Energetica”, Project code C31J09000420005. Author Contributions: M.M. and C.F. conceived and designed the microsensor and performed the experimental tests; F.G.D.C. conceived the experiment and supervised the design. All authors contributed equally in writing the paper. Conflicts of Interest: The authors declare no conflict of interest.
References 1. 2. 3. 4. 5.
6. 7. 8.
9. 10. 11. 12. 13. 14. 15.
16. 17. 18.
Atzori, L.; Iera, A.; Morabito, G. The Internet of Things: A survey. Comput. Netw. 2010, 54, 2787–2805. [CrossRef] Eder, C.; Valente, V.; Donaldson, N.; Demosthenous, A. A CMOS smart temperature and humidity sensor with combined readout. Sensors 2014, 14, 17192–17211. [CrossRef] [PubMed] Rekhi, A.S.; Khuri-Yakub, B.T.; Arbabian, A. Wireless Power Transfer to Millimeter-Sized Nodes Using Airborne Ultrasound. IEEE Trans. Ultrason. Ferroelectr. Freq. Control 2017, 64, 1526–1541. [CrossRef] [PubMed] Zaid, J.; Abdulhadi, A.; Kesavan, A.; Belaizi, Y.; Denidni, T.A. Multiport circular polarized RFID-tag antenna for UHF sensor applications. Sensors 2017, 17, 1576. [CrossRef] [PubMed] Unigarro Calpa, E.; Achury Florian, A.; Rairez Rodriguez, F.; Sacristan, J.; Bohorquez, J.; Segura-Quijano, F.E. Towards Fully-integrated Low-Cost Inductive Powered CMOS Wireless Temperature Sensor. IEEE Trans. Ind. Electron. 2017, 64, 8718–8727. [CrossRef] Zito, F.; Aquilino, F.; Fragomeni, L.; Merenda, M.; Della Corte, F.G. CMOS wireless temperature sensor with integrated radiating element. Sens. Actuators A 2010, 158, 169–175. [CrossRef] Aquilino, F.; Della Corte, F.G.; Merenda, M.; Zito, F. Fully-integrated wireless temperature sensor with on-chip antenna. In Proceedings of the 2008 IEEE Sensors, Lecce, Italy, 26–29 October 2008; pp. 760–763. Arsalan, M.; Ouda, M.H.; Marnat, L.; Ahmad, T.J.; Shamim, A.; Salama, K.N. A 5.2 GHz, 0.5 mW RF powered wireless sensor with dual on-chip antennas for implantable intraocular pressure monitoring. In Proceedings of the IEEE MTT-S International Microwave Symposium Digest (MTT), Seattle, WA, USA, 2–7 June 2013. [CrossRef] Datasheet AMS Foundry. 2017. Available online: http://asic.ams.com/ (accessed on 31 December 2017). Bakker, A.; Huijsing, J.H. Micropower CMOS temperature sensor with digital output. IEEE J. Solid-State Circuits 1996, 31, 933–937. [CrossRef] Meijer, G.C.M.; Wang, G.; Fruett, F. Temperature Sensors and Voltage References Implemented in CMOS Technology. IEEE Sens. J. 2001, 1, 225–234. [CrossRef] Pertijs, M.A.P.; Meijer, G.C.M.; Huijsing, J.H. Precision Temperature Measurement Using CMOS Substrate PNP Transistors. IEEE Sens. J. 2004, 4, 294–300. [CrossRef] Razavi, B. Design of Analog CMOS Integrated Circuits; McGraw-Hill: New York, NY, USA, 2001; Volume 1, p. 684. Yoo, C.; Park, J. CMOS current reference with supply and temperature compensation. Electron. Lett. 2007, 43, 1422. [CrossRef] Zhao, X.; Chebli, R.; Sawan, M. A wide tuning range voltage-controlled ring oscillator dedicated to ultrasound transmitter. In Proceedings of the 16th International Conference on Microelectronics, Tunis, Tunisia, 6–8 December 2004; pp. 313–316. [CrossRef] Sun, L.; Kwasniewski, T.A. A 1.25-GHz 0.35-µm monolithic CMOS PLL based on a multiphase ring oscillator. IEEE J. Solid-State Circuits 2001, 36, 910–916. [CrossRef] Hajimiri, A.; Lee, T.H. Design issues in CMOS differential LC oscillators. IEEE J. Solid-State Circuits 1999, 34, 717–724. [CrossRef] Bunch, R.; Raman, S. A 0.35 pm CMOS 2.5 GHz Complementary -GM VCO Using PMOS Inversion Mode Varactors. In Proceedings of the IEEE Radio Frequency Integrated Circuits (RFIC) Symposium, Phoenix, AZ, USA, 20–22 May 2001; Digest of Papers; pp. 49–52.
Sensors 2018, 18, 110
19.
20. 21. 22.
23.
15 of 15
Haobijam, G.; Paily, R. Quality factor enhancement of CMOS inductor with pyramidal winding of metal turns. In Proceedings of the 2007 International Workshop on Physics of Semiconductor Devices, Mumbai, India, 16–20 December 2007; pp. 729–732. [CrossRef] Craninckx, J.; Steyaert, M.S. A 1.8 GHz low phase-noise CMOS VCO using optimized hollow spiral inductors. IEEE J. Solid-State Circuits 1997, 32, 736–744. [CrossRef] Aquilino, F.; Della Corte, F.G. On-Chip Integrated Antenna Structures for Biomedical Implantable Sensors. Procedia Chem. 2009, 1, 513–516. [CrossRef] Kim, B.; Nakazato, K. Dual data pulse width modulator for wireless Simultaneous Measurement of Redox Potential and Temperature using a Single RFID Chip. In Proceedings of the 19th IEEE International Conference on Electronics, Circuits and Systems, Seville, Spain, 9–12 December 2012; pp. 368–371. [CrossRef] Yin, J.; Yi, J.; Law, M.K.; Ling, Y.; Lee, M.C.; Ng, K.P.; Gao, B.; Luong, H.C.; Bermak, A.; Chan, M.; Ki, W.H.; Tsui, C.Y.; Yuen, M. A system-on-chip EPC Gen-2 passive UHF RFID tag with embedded temperature sensor. IEEE J. Solid-State Circuits 2010, 45, 2404–2420. [CrossRef] © 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
