1275782 九、發明說明： 【發明所屬之技術領域】 本案係為一種數位溫度感測系統，尤指一種應用 於中央處理器、晶片組、電池保護IC、環境監控系 統、機電系統及其他附有溫度感測電路之應用電路或 系統之數位溫度感測系統。 【先前技術】 溫度感測之技術眾多，相關前案如：00415623 「晶片型金屬電阻溫度感測器之端電極的構造」、 00461961 「壓阻式溫度感測器」及00387640 「導 線與溫度感測器的熱敏電阻之間的防潮連接結構」皆 係以熱敏電阻來做溫度感測，利用溫度改變使各壓阻 元件的電阻值變化被轉換為電壓信號輸出；00309661 「具有溫度感測器之可場效控制的功率半導體元件」 則是利用雙載子電晶體的基、射極兩端電壓對溫度而 改變的特性來偵測溫度；US2002147564 「Digital temperature sensor (DTS) system to monitor temperature in a memory subsystem」及 US200407Π83 「Integrated digital temperature sensor」乃係利用隨 溫度改變之電壓或電流來做溫度感測，而後以類比至 數位轉換器（analog_to-digital converter)將電壓或電 流訊號轉換成數位信號，大略之電路方塊圖如圖一所 示，主要構件為溫度感測器11、參考源電路12、及 5 1275782 類比至數位轉換器13。其中溫度感測器11通常是以 特殊製程之垂直式雙極接面電晶體或純CMOS製程 之寄生性基底或水平式雙極接面電晶體（parasitic substrate or lateral bipolar transistor)為主體的電路所 構成，其功能主要用來產生隨溫度改變之電壓或電流; 而參考源電路12之功能則為產生與溫度無關的能隙 參考電壓或電流源；至於類比至數位轉換器13則將溫 度感測器11與參考源電路12兩者之電壓差或電流差 轉換成數位信號。 在溫度感測器11方面，因為垂直式雙極接面電 晶體，需要特殊製程才有辦法量產，所以對於成本有 相當的影響；而寄生性基底雙極接面電晶體雖然是採 用CMOS製程，但其元件特性在製程上難以掌控， 往往預期與實際有不小的差距，對於日後生產有一定 的困難；若改以水平式雙極接面電晶體替代，則雖較 易掌控其特性，但其順向電流放大增益較小，所製作 之溫度感測電路通常會有較差之操作特性。另外，不 論採用何種雙極電晶體，其溫度至電壓或電流之轉換 特徵曲線皆存在有相當曲率的問題，因此為了減少誤 差，此數位式溫度感測器需要額外的電路來做校正而 造成晶片面積及功率消耗的增加。 至於參考源電路12的溫度特性也會隨著晶片溫 度的起伏而跟著改變，其準確度自然無法與將晶片外 參考源保持於恆溫之精密溫度測量系統相比，而且其 6 1275782 ，路依然需要雙極電晶體來實現，故存在與溫度感測 器11相同的問題。 至於類比至數位轉換器13則係數位式溫度感測 盗的數位轉換核d數位式溫度❹以之解析度取 決於類比至數位轉換器13的有效輸出位元數，以測 量範圍過1G(rC且溫度解析度優於G.rc之電路為 通常需要使用超過1Q位元以上之類比至數位轉 換盗13,往往會佔用更大的面積並消耗更多的功 =而成本、功率消耗以及準確度乃是單W溫度感 測系統規格中最為重要的三個項目。 因此，如何改善上述習知手段之缺失，係為發展 本案之主要目的。 【發明内容】 /本案係關於一種新型數位溫度感測系統，用以將 待測溫度轉換成對應之數位輸出，該系統包含一溫度 感測器，作為該數位溫度感測系統之前端電路，= 產生隨溫度改變之時間訊號；—時間至數位轉換電 ^電連接於該溫度感測器之輸出端，用以將該溫度 感測器所轉換之時間訊號轉換成對應之數位訊號。 本案所提出之新型數位溫度感測系統之電路方 塊圖如圖二所示，其中溫度感測器21係用來產生隨 溫度改變之時間訊號，其輸出端電連接於時間至 轉換電路22之輸入端，時間至數位轉換電路22可作 1275782 為该數位溫度感測系統之後端電路，用以將前 感測器所產生之時間訊號轉換成對應之數：： 出。 | 根據上述構想’第三圖⑷為本案所提出之全新 溫度感測器2】之方塊圖及其工作時序圖之第—實施 STA二3尸，延遲線211 ’用以將起始轉換訊號 START延遲-段時間而後電連接至互斥或開犯之 第-輸入端，·-導線212，將將起始轉換訊號$丁術 至互斥或間213之第二輪入端；互斥或開213 用來擁取第一延遲線211之延遲時間做為脈衝 (Pulse)型輸出訊號。後端之時間至數位轉換器U可 依需要取第一延遲線211及導線212之訊號Pa、？b 充當其輸入訊號，或直接採用互斥或開213之輪出脈 衝Pout充當其輸入訊號。 1該第-實施例電路之操作原理如下·· S始轉換 A就START分別電連接於第一延遲線211之輸入端 及導線212之一端’而該第一延遲線211之輸出端電 連接於互斥或閘213之第一輸入端，該導線212之另 一端電連接於該互斥或閘213之第二輸入端，而該互 斥或閉之輸出端即為溫度感測器21之脈衝型輸出訊 號輸出端。 ^根據上述構想，本第一實施例之第一延遲線211 係由複數個延遲元件串接而成，作為溫度感測之主 脰’導線212相較之下受溫度之影響十分輕微，而互 1275782 斥或閘213即用來取START訊號行經第一延遲線2ιι 及導線212之延遲時間差做為脈衝型輸出訊號。 根據上述構想，第一實施例之溫度感測器21可 產生與溫度呈線性相關之脈衝寬度訊號。此實施例係 利用溫度改變、延遲元件導通電流隨之改變、進而造 成延遲時間改變之電路原理來設計；欲啟動溫度感測 器21，只需在溫度感測器21之輸入端輸入一步階信 號START便可。由於第一延遲線2 i!會造成步階信 唬延遲一段時間之後再送進互斥或閘213之第一輸 入端，而導線212並不會造成明顯之步階信號時間延 遲’故互斥或閘213兩輸入端之步階信號會有明顯之 時間差’其差值恰好是第„延遲線211之延遲時間， 經由互斥或問之邏輯運作便可產生一對應之輸出脈 衝·ρ^υτ其時間寬度乃是第一延遲線211之延遲時 間，右延遲兀件具有正溫度係數，當溫度上升時其延 遲時間會隨之增長，則第—延遲線211所造成之步階 信號延遲時間亦會隨之上升，意即第一延遲線叫 及導線2U之輸出訊號Ρα、ρβ間之時間差距會隨之 f大：；樣地，經由互斥或㈣後所產生之脈衝 P_兔度亦隨溫度而上升，即溫度感測電路21 生時間寬度與溫度成正比例之時間訊號，其溫度特徵 曲線如第四圖之實線所示。反之，若延遲元件具有負 溫度係數，則溫度感測器21 、 .^ ? ^ 」屋生4間見度與溫度 成反比例之時間訊號，其溫度特徵曲線如第四圖之虛 1275782 線所示。 簡易之延遲元件可由反閘（NOT)為之，其傳輸延 遲時間可以表示成(L/W)Cl ~ 1，身又 1*月1275782 IX. Description of the invention: [Technical field of invention] This case is a digital temperature sensing system, especially one applied to a central processing unit, a chipset, a battery protection IC, an environmental monitoring system, an electromechanical system, and other temperature-attached A digital temperature sensing system for an application circuit or system of a sensing circuit. [Prior Art] There are many techniques for temperature sensing, such as: 00415623 "Structure of the terminal electrode of the chip type metal resistance temperature sensor", 00461961 "Presistance temperature sensor" and 00387640 "Wire and temperature sense" The moisture-proof connection structure between the thermistors of the detector is based on the thermistor for temperature sensing, and the temperature change is used to change the resistance value of each piezoresistive element to be converted into a voltage signal output; 00309661 "with temperature sensing The field-controllable power semiconductor component of the device uses the characteristics of the voltage across the base of the bipolar transistor and the temperature of the emitter to change the temperature; US2002147564 "Digital temperature sensor (DTS) system to monitor temperature "in a memory subsystem" and US200407Π83 "Integrated digital temperature sensor" is a temperature or current that changes with temperature for temperature sensing, and then converts a voltage or current signal into a digital analog to analog converter (analog_to-digital converter) Signal, roughly the circuit block diagram shown in Figure 1, the main component is the temperature sensor 11 Reference source circuit 12, and 51,275,782 analog to digital converter 13. The temperature sensor 11 is usually a circuit with a special process of a vertical bipolar junction transistor or a pure CMOS process parasitic substrate or a horizontal bipolar junction transistor (parasitic substrate or lateral bipolar transistor). The function is mainly used to generate a voltage or current that changes with temperature; and the function of the reference source circuit 12 is to generate a temperature-independent bandgap reference voltage or current source; as for the analog to digital converter 13, the temperature is sensed. The voltage difference or current difference between the comparator 11 and the reference source circuit 12 is converted into a digital signal. In the temperature sensor 11, because the vertical bipolar junction transistor requires a special process to mass production, it has a considerable impact on the cost; while the parasitic substrate bipolar junction transistor is in a CMOS process. However, its component characteristics are difficult to control in the process. It is often expected that there is a big gap between the actual and the actual production. If it is replaced by a horizontal bipolar junction transistor, it is easier to control its characteristics. However, the forward current amplification gain is small, and the temperature sensing circuit produced usually has poor operating characteristics. In addition, regardless of the bipolar transistor used, the temperature-to-voltage or current conversion characteristic curve has a considerable curvature problem. Therefore, in order to reduce the error, the digital temperature sensor requires an additional circuit for correction. Increase in wafer area and power consumption. As for the temperature characteristic of the reference source circuit 12, it also changes with the fluctuation of the temperature of the wafer, and the accuracy thereof naturally cannot be compared with the precision temperature measurement system that maintains the external reference source of the wafer at a constant temperature, and its 6 1275782 still needs The bipolar transistor is implemented, so there is the same problem as the temperature sensor 11. As for the analog-to-digital converter 13, the coefficient-type temperature sensing pirated digital conversion core d-digit temperature ❹ depends on the analog output to the number of valid output bits of the digital converter 13, to measure the range over 1G (rC And the temperature resolution is better than G.rc. The circuit usually needs to use more than 1Q bit to analog to digital conversion. It often takes up more space and consumes more work. Cost and power consumption and accuracy. It is the three most important items in the specification of single-W temperature sensing system. Therefore, how to improve the lack of the above-mentioned conventional means is the main purpose of developing this case. [Summary] This case is about a new type of digital temperature sensing. a system for converting a temperature to be measured into a corresponding digital output, the system comprising a temperature sensor as a front end circuit of the digital temperature sensing system, = generating a time signal that changes with temperature; - time to digital conversion ^ is electrically connected to the output end of the temperature sensor for converting the time signal converted by the temperature sensor into a corresponding digital signal. The circuit block diagram of the type digital temperature sensing system is shown in FIG. 2, wherein the temperature sensor 21 is used to generate a time signal that changes with temperature, and the output end thereof is electrically connected to the input of the time to the conversion circuit 22, time to The digital conversion circuit 22 can be used as a rear circuit of the digital temperature sensing system for converting the time signal generated by the front sensor into a corresponding number:: out. According to the above concept, the third figure (4) is the case. The block diagram of the proposed new temperature sensor 2 and the operation timing diagram thereof - the implementation of the STA 2 3 corpse, the delay line 211 ' is used to delay the initial conversion signal START - the time period and then electrically connected to the mutual exclusion or The first input terminal, the --wire 212, will switch the initial conversion signal to the second round of the mutual exclusion or interval 213; the exclusive or open 213 is used to capture the first delay line 211. The delay time is used as a pulse type output signal. The time from the back end to the digital converter U can take the signals Pa, ?b of the first delay line 211 and the wire 212 as their input signals, or directly adopt the mutual exclusion or Open 213 round pulse Pout to act as Input signal 1. The operation principle of the circuit of the first embodiment is as follows: · S start conversion A is electrically connected to the input end of the first delay line 211 and one end of the wire 212 respectively and the output end of the first delay line 211 Electrically connected to the first input end of the mutex or gate 213, the other end of the wire 212 is electrically connected to the second input end of the mutex or gate 213, and the mutually exclusive or closed output is a temperature sensor According to the above concept, the first delay line 211 of the first embodiment is formed by a series of delay elements connected in series as the main sense of temperature sensing. The temperature is very slight, and the mutual 1275782 repulsion or gate 213 is used to take the delay time difference of the START signal through the first delay line 2ιι and the wire 212 as a pulse type output signal. According to the above concept, the temperature sensor 21 of the first embodiment can generate a pulse width signal linearly related to temperature. This embodiment is designed by using a circuit principle in which the temperature change, the delay current of the delay element is changed, and the delay time is changed. To activate the temperature sensor 21, only one step signal is input at the input end of the temperature sensor 21. START is OK. Since the first delay line 2 i! causes the step signal delay to be delayed for a period of time and then sent to the first input of the mutex or gate 213, the wire 212 does not cause significant step signal time delay. The step signal of the two input terminals of the gate 213 has a significant time difference', and the difference is exactly the delay time of the „delay line 211, and a corresponding output pulse ρ^υτ can be generated by the logic operation of the mutual exclusion or the question. The time width is the delay time of the first delay line 211, and the right delay element has a positive temperature coefficient, and the delay time will increase as the temperature rises, and the delay time of the step signal caused by the first delay line 211 will also be With the rise, it means that the time difference between the first delay line and the output signal Ρα, ρβ of the wire 2U will be f:; the sample, the pulse generated by mutual exclusion or (4) P_ rabbit degree also The temperature rises, that is, the temperature sensing circuit 21 generates a time signal whose time width is proportional to the temperature, and the temperature characteristic curve is shown by the solid line of the fourth figure. Conversely, if the delay element has a negative temperature coefficient, the temperature sensor twenty one , .^ ^ ^ "The time difference between the 4 and the temperature of the house is inversely proportional to the temperature. The temperature characteristic curve is shown in the figure 1275782 of the fourth figure. The simple delay element can be reversed (NOT), and its transmission delay time can be expressed as (L/W)Cl ~ 1, body and 1* month
p~ mC0X(vdd-vt) t~^J 下數位邏輯電路之供應電壓VDD遠大於臨界電壓 VT，故傳輸延遲時間受遷移率μ之影響較大，然而遷 移率與溫度之關係為// = «|5^ = -1.2〜-2.()’意即溫度越 高、遷移率越小、反閘之延遲時間越長，所以由反閘 所組成之延遲線具有正溫度係數；至於具負溫度係數 之延遲元件之一實施例如第五圖所示，該延遲元件由 圖中虛線部份之反閘搭配虛線以外之溫度補償電路 所組成，由於臨界電壓&之公式為匕(r) = Fr(7；) + a(r-7；) 其中〜-3._1>7。尺，其大小會隨溫度之增加而減少，意 即以二極體方式連接（diode-connected)之電晶體 PI、N1之波極（drain)與源極（source)電壓差會隨溫度 之增加而減少，此時加諸電阻性負載Load兩端之電 壓會增加，使其導通電流變大，藉由PI、P2及N1、 N2此二組電流鏡將該電流導入反閘之P型及N型電 晶體，以抵補反閘隨溫度增加而變小之導通電流，若 抵補電流之效果過大（等同於過度補償），便會讓延遲 元件之導通電流會隨溫度之增加而變大，意即其延遲 時間會隨隨溫度之增加而變小，形成具有負溫度徐數 之延遲元件。 需要注意的是，若後端之時間至數位轉換器22 10 Ϊ275782 、 茜要兩組步階訊號做為輸入，並取二者之時間差值作 為時間轉換依據，則可將第三圖（a)之互斥或閘213 去除，直接輸出第一延遲線211及導線212之訊號 pA、PB便可，此乃溫度感測器21第一實施例之另一 變形。 本案溫度感測器21第二實施例之電路方塊圖及 其工作時序圖如第三圖（b)所示，第二實施例主要用 • 來修正第一實施例在量測溫度下限（或上限，端視溫 度係數之正負而定）所產生的脈衝時間寬度過寬之缺 失。其電路包含：一第一延遲線211，用以將起始轉換 汛號START延遲一段時間而後電連接至互斥或閘 213之第一輸入端；一第二延遲線2112，將將起始轉 換訊號START延遲另一段時間而後電連接至互斥或 閘213之第二輸入端，第二延遲線2112較佳之實施 例乃由具溫度補償延遲元件所組，透過如第五圖之溫 • 度補償電路來降低延遲元件之溫度敏感性（thermal sensitivity); —互斥或閘213，用來擷取第一延遲線 211與第二延遲線2112之延遲時間差值做為脈衝輸 出吼號Pout。後端之時間至數位轉換器可依需要 取第一延遲線211及第二延遲線2112之訊號^小充 當其輸入訊號，或直接採用互斥或閘213之輸出脈衝 充當其輸入訊號。 根據上述構想，第二實施例之溫度感測電路21 可產生時間寬度與溫度成比例之時間訊號，且在其所 11 1275782 需量測溫度規格下限所產生之脈衝寬度可較接近於 零。在第一實施例中，只有在接近絕對零度(〇°κ)時第 一延遲線211所造成之時間延遲才有可能接近於 零，亦即只有在接近絕對零度時，第一實施例之輸出 時間寬度（Pa、Ρβ之時間差或Ρ〇υτ)才會接近於零丨通常 我們所需量測溫度規格的下限會遠高於絕對零度’此 時輸出之時間寬度會存有嚴重之偏移（offset)的問 題。由於溫度感測電路21之後端電路以時間至數位 轉換器作為數位輸出轉換之用，寬度偏移將使時間至 數位轉換器的轉換時間變長，並且需要較多的輸出位 元數才得以存下最後的轉換結果。 與第一實施例相較’此第二實施例之目的主要在 於設計出合適的時間寬度扣抵電路，讓輸出時間訊號 在所需測量之溫度下限得到趨近於零之時間寬度。為 此，第二實施例較佳是以具溫度補償電路之第二延遲 線2112來取代第一實施例之導線212，由於經過溫 度補償電路之第二延遲線2112具有較低之溫敏係數 (thermal coefficient)，其延遲時間可取決於延遲線的 長度（等同於延遲線内之延遲元件個數），而且較不受 溫度改變之影響。第二實施例之運作原理與第一實施 例相仿，於起始時輸入一步階啟始信號START，該 步階信號經第一延遲線211延遲一段時間後送入互 斥或閘213之第一輸入端，若第一延遲線211由類似 反閘之類的正溫度係數延遲元件組成，則該延遲時間 12 1275782 將與待測溫度呈正比例關係；同時，該步階信號亦經 第二延遲線2112延遲另一段時間後送入互斥或閘 213之第一輸入端，且該另一延遲時間較不受溫度改 變之影響；最後，經由互斥或閘213之邏輯運作便可 產生一對應之輸出脈衝訊號Pm。當溫度上升時，具 正=度係數之第一延遲線211所造成之步階信號延 遲時間亦隨之上升，而具溫度補償電路之第二延遲線 2112所造成之步階信號延遲時間改變較小，故經由 互斥或間213後所產生之脈衝時間寬度亦會隨溫 上升。 興第 貝訑例相較，經由適當之設計，在所需量 測溫度規格之下限讓第一延遲線211與第二延遲線 2112所造成之步階㈣延料間彼此接近，此時輸 出時間寬度將趨近於零，可有效修正輸出時間寬产在 測量溫度下限偏移量過大之缺失，如第六圖⑷所 不二原本第一延遲線211在溫度下限具有相當大之偏 移量，經由第二延遲線2U2做適當的扣抵後，所產 生之輸出時間寬度在溫度下限只具有小量之偏移。 二需要注意的是，請看第六圖（b)，經由適當之設 1 口人將第—延遲、線2112所造成之延遲時間增加 (專同於增加延遲線内之延遲元件數目），經由互斥 3 兩延遲線之延遲時間差值，可得出一隨 ’升而見度減少之時間信號。此為具負溫度係數 之溫度感測電路21的又-實施例。同理，若 13 1275782 遲線211具負溫度係數，我們亦可經由調整第二延遲 線2112之延遲時間來產生具正溫度係數或具負溫度 係數之溫度感測電路21。 需要注意的是，若後端之時間至數位轉換器22 需要兩組步階訊號做為輸入，並取二者之時間差值作 為時間轉換依據，則可將第三圖（b)之互斥或閘213 去除，直接輸出第一延遲線211及第二延遲線2112 之訊號PA、PB便可，此乃溫度感測器21第二實施例 之另一變形。 第三圖（c)為本案溫度感測器21第三實施例之電 路功能方塊圖，其包含一與絕對溫度成比例 (proportional to absolute temperature 5 簡稱 PTAT)之 電流源214，其輸出端電連接於控制電路216之第一 輸入端，作為溫度感測之用；一定電流源（constant current source)215，其輸出端電連接於控制電路216 之第二輸入端，其電流值大略與溫度無關；控制電路 216之第一輸入端電連接於PTAT電流源214，第二 輸入端電連接於定電流源215，輸出端電連接於時間 擷取電路217，其功用為將PTAT電流源與定電流源 做電壓轉換並將兩者做比較及判斷；時間擷取電路 217係電連接於控制電路216之輸出端，作為時間擷 取之用。 其詳細操作原理如第七圖（a)所示，PTAT電流源 214與定電流源215分別輸入控制電路216，第一階 14 1275782 段:控制電路216先將PTAT電流源214做定時間充 電或放電（第六圖（a)以充電示意），由於pTAT電流源 214隨溫度產生一線性增加或下降的改變，所以定時 間充電或放電所造成之電壓值亦會隨溫度改變而有 所不同（如圖所示，T〇、T!代表不同溫度）(3第二階端： 控制電路216利用定電流源215作反向放電或充電的 動作（第六圖（a)以放電示意），由於定電流源不隨溫度 而改變，故放電之斜率相同；然而因為不同溫度τ〇、 乃所造成之電壓準位不同（如圖所示，在τ〇、丁1溫度 下刀別產生電壓V 〇、V i)，故經由定電流源2丨5放電 所所需之時間自然不同，分別標示為砣、^ ’接著時 間擷取電路217擷取定電流源215放電所需之時間作 為輸出時間訊號。由於PTAT電流源214隨溫度而線 性改變，定時間充電所造成之電壓準位亦線性改變， 故、、’工由疋笔源215放電後，由時間擷取電路217 所得之時間亦線性改變。故此溫度感測電路21之第 二貫施例隨溫度可產生一線性增加或下降的時間訊 號。 第二較佳實施例之另一操作原理可由第七圖（的來表 示，PTAT電流源214會隨溫度產生一線性增加或減少之電 流，控制電路216將此電流經過一阻抗元件（未於圖中標 示），產生一隨溫度線性增加或減少之電壓(如圖所示，在 ο T! ’皿度下为別產生電壓v 〇、v ο，定電流源215 之大小不受溫度改變之影響，故經由—電容性树(未於圖 15 1275782 中標示）做充電之斜率固定，其充電電壓值隨時間增加而綠 性改變。控制電路216持續比較PTAT電流源所轉換之電: 與定電流源215之充電電壓，直到二者相等為止。時間擷ς 電路217則用於擷取定電流源215所需之充電時間做為輪出 時間訊號。由於PTAT f流源214隨溫度而線性改變，^過 一阻抗元件所產生之電壓準位亦呈線性改變，故定電流= 215所需之充電時間亦隨溫度呈線性改變，所產生之時= 號寬度會隨溫度改變而呈線性增加或減少。 本案應用之時餘轉換器係—種將時間訊號之宽度 轉換成對驗位_讀鏡路。f知之實财式主要有計 數-法循％脈衝缩減法、雙斜率法、游標尺法、内插法等； '、目的&係將日㈣訊雜換為相對應之數位輪出訊號。 【實施方式】 、^為本案之新型數位溫度感測系统之電路 功能方塊圖，其包 w 溫度而線性變化之二：^1，用 之日守間。fl號Ρ_，以及時間至數位轉 t路22;時間至數位轉換電路22可作為此新变數 二度感測系统之後端電路，用以將前端之溫度感測 、 之日守間机號轉換成相對應之數位訊號。 ^第二圖（a)為本案溫度感測器21之第一實施例之 =路功能方塊圖及其卫作時序圖，其包含-第一延遲 線211，一導岣， 諍、、果212, 一互斥或閘213作為時間寬度 16 1275782 擷取電路之用。其巾該導線較佳者係為—金屬延遲 、、表&着/JHL度上升，該第一延遲線211將START作 號延遲之時間亦隨之而線性上升，然導線212 ^ START信號延遲之時間受溫度改變之影響並不顯 著。故兩線所造成STAR丁之延遲時間，經由互斥或 閘213作時間寬度擷取後之時間信號寬度亦隨溫度 上升而線性增加。即此第一實施例之溫度感測器經由 S?T信號觸發後，可得一隨溫度上升而線性增加 之日寸間Λ號。此時間訊號除脈衝寬度外，亦包括上述 二訊號之時間差寬度。 第二圖（b)為本案溫度感測器21第二實施例之電 路功能方塊圖及其工作時序圖，其包含一第一延遲線 2li ’ 一具溫度補償電路之第二延遲線2112，一互斥 或閘213作為時間寬度擷取電路之用。與第一實施例 相，’此第二實施例之目的主要在於設計出合適的脈 衝寬度扣抵電路’讓輪出脈衝(P_)在所需測量之溫 度下限（或上限）得到趨近於零之脈衝寬度。減少作為 後端電路之時間至數位轉換電路之轉換時間及所需 輸出位元數。隨着溫度上升，第-延遲線211將 START信號延遲之時間亦隨之而線性上升，然具溫 度補彳貝電路之第二^遲線2112亦將START信號延遲 #奴日=間’但溫度上升時其所造成之延遲時間並未顯 著改、交。故兩延遲線所造成START之延遲時間，經 由互斥或閘213後之時間信號寬度亦隨溫度上升而 17 1275782 線性增加。此第二實施例之溫度感測器21經由 START信號觸發後，可得一隨溫度上升而線性增力口 之時間信號。請看第六圖（a)，在適當的設計下，在 所需規格之溫度下限，其時間信號寬度可趨近於零。 此外，請看第六圖（b)，吾人亦可將具溫度補償 電路之第二延遲線2112作一適當修正，在所需規格 之溫度範圍内，讓它所造成之延遲時間大於第一延遲 線211，經由互斥或閘213做時間脈衝寬度的擷取， 與第六圖（a)不同的是，此時我們可得一隨溫度上升 而線性減少之時間寬度信號。在所需規格之溫度上 限，其時間信號寬度亦可趨近於零。 第三圖（c)為本案溫度感測器21之第三實施例之 電路功能方塊圖，其包含一 PTAT電流源214，電連 接於控制電路216之第一輸入端，作為溫度感測之 用，其電流隨溫度改變呈線性變化；一定電流源 215，電連接於控制電路216之第二輸入端，其電流 值與溫度不大相關；控制電路216之第一輸入端電連 接於PTAT電流源214之輸出，第二輸入端電連接於 定電流源215之輸出，而216之輸出端電連接於時間 擷取電路217，控制電路216將PTAT電流源214做 固定時間充電（或放電），而後透過定電流源215做反 向放電（或充電）所得到之時間經時間擷取電路217時 間擷後輸出，其輸出時間寬度可隨溫度改變呈線性變 化，相關訊號與時序如第七圖（a)所示。 18 1275782 另外又如第七圖（b)所示，控制電路216可將 PTAT電流源214透過電阻性元件（未於圖中標示）轉 換成與溫度成線性正比（或反比，端視其溫度係數之 正負而定）之電壓，再讓定電流源215對電容性元件 (未於圖中標不）充電’其電壓隨時間增加但約略與溫 度無關，而後將該二電壓做比較，以便讓時間擷取電 路217擷取定電流源215所需之充電時間作為輸出時 間訊號，其寬度將可隨溫度改變而呈線性增加或下 降。其中該控制電路216較佳係為一電壓比較器與一 電阻性元件、一電容性元件與充電控制開關所組成。 本案應用之時間至數位轉換器係一種將時間訊號之寬度 轉換成對應數位輸出之轉換電路。習知之實現方式主要有計 數器法、循環脈衝縮減法、雙斜率法、游標尺法、内插法等； 其目的皆係將時間訊號轉換為相對應之數位輸出訊號。 綜上所述，本案之數位溫度感測系統可應用於中 央處理器、晶片組、電池保護1C、環境監控系統、 機電系統及其他附有溫度感測電路之應用電子電路 中。本案之溫度感測器可產生對溫度信號呈線性相關 之時間信號，此新顆電路無雙極接面電晶體’故無須 額外校正電路（意味晶片面積及功率消耗上升）以及 曲率校正技術來降低誤差，故可大大改善設計的困難 度、節省成本及功率消耗，有效達成發展本案之目 的。特別是當應用於可攜式系統，可有效降低成本及 19 1275782 功率消耗。因此，本案實為一新穎、進步及實用之創 作，爰依法提出申請。 本案得由熟悉本技藝之人士任施匠思而為諸般 修飾，然皆不脫如附申請專利範圍所欲保護者。 【圖式簡單說明】 第一圖：習知數位式溫度感測系統器之電路功能方塊 圖。 第二圖：本案之數位溫度感測系統之電路功能方塊 圖。 第三圖（a):本案溫度感測器之第一實施例之電路功能 方塊圖。 第三圖（b):本案溫度感測器之第二實施例之電路功能 方塊圖。 第三圖（c):本案溫度感測器之第三實施例之電路功能 方塊圖。 第四圖：本案溫度感測器第一實施例之溫度特徵曲線示 意圖。 第五圖：具溫度補償之延遲元件架構示意圖。 第六圖（a):本案溫度感測器第二實施例之工作示意圖 A 〇 第六圖（b):本案溫度感測器之第二實施例之工作示意 圖B。 第七圖（a):本案溫度感測器第三實施例之工作示意圖 20 1275782 A 〇 第七圖（b):本案溫度感測器之第三實施例之工作示意 圖B 〇 【主要元件符號說明】 S TART:電路觸發啟動信號p~ mC0X(vdd-vt) t~^J The supply voltage VDD of the digital logic circuit is much larger than the threshold voltage VT, so the propagation delay time is greatly affected by the mobility μ, but the relationship between mobility and temperature is /// «|5^ = -1.2~-2.()' means that the higher the temperature, the smaller the mobility, and the longer the delay time of the reverse gate, so the delay line composed of the reverse gate has a positive temperature coefficient; One of the delay elements of the temperature coefficient is implemented as shown in the fifth figure. The delay element is composed of a temperature compensation circuit other than the broken line of the broken line in the figure and the dotted line, because the formula of the threshold voltage & 匕(r) = Fr(7;) + a(r-7;) where ~-3._1>7. The size of the ruler decreases with the increase of temperature, which means that the voltage difference between the diode and the source of the diode-connected transistor PI, N1 increases with temperature. When the voltage is reduced, the voltage across the resistive load is increased, and the conduction current is increased. The currents of the two groups of current mirrors, PI, P2, and N1, N2, are introduced into the reverse gate P-type and N. The type of transistor, in order to compensate for the on-state current that becomes smaller as the temperature increases with the increase of the temperature, if the effect of the offset current is too large (equivalent to over-compensation), the conduction current of the delay element will become larger as the temperature increases, that is, The delay time becomes smaller as the temperature increases, forming a delay element having a negative temperature. It should be noted that if the back-end time is up to the digital converter 22 10 Ϊ 275782 and two sets of step signals are used as inputs, and the time difference between the two is used as the time conversion basis, the third picture (a) The mutual exclusion or the gate 213 is removed, and the signals pA and PB of the first delay line 211 and the wire 212 are directly outputted, which is another modification of the first embodiment of the temperature sensor 21. The circuit block diagram of the second embodiment of the temperature sensor 21 of the present invention and its operation timing chart are as shown in the third figure (b). The second embodiment mainly uses the first embodiment to correct the lower limit (or upper limit) of the measurement temperature of the first embodiment. , depending on the positive and negative temperature coefficient, the resulting pulse time width is too wide. The circuit includes: a first delay line 211 for delaying the initial conversion signal START for a period of time and then electrically connecting to the first input of the mutex or gate 213; a second delay line 2112, which will initiate the conversion The signal START is delayed for a further period of time and then electrically connected to the second input of the mutex or gate 213. The preferred embodiment of the second delay line 2112 is comprised of a temperature compensated delay element, which is compensated by temperature compensation as shown in FIG. The circuit is configured to reduce the thermal sensitivity of the delay element; the repulsion or gate 213 is used to capture the delay time difference between the first delay line 211 and the second delay line 2112 as the pulse output nickname Pout. The back-end time-to-digital converter can take the signal of the first delay line 211 and the second delay line 2112 as needed to input the signal, or directly use the output pulse of the mutex or gate 213 as its input signal. According to the above concept, the temperature sensing circuit 21 of the second embodiment can generate a time signal whose time width is proportional to the temperature, and the pulse width generated by the lower limit of the temperature specification required by the 11 1275782 can be closer to zero. In the first embodiment, the time delay caused by the first delay line 211 is only close to zero when approaching absolute zero (〇°κ), that is, the output of the first embodiment only when approaching absolute zero. The time width (Pa, Ρβ time difference or Ρ〇υτ) will be close to zero. Usually, the lower limit of the temperature specification we need to measure will be much higher than the absolute zero. At this time, there will be a serious offset in the output time width ( Offset) problem. Since the rear end circuit of the temperature sensing circuit 21 uses a time-to-digital converter as a digital output conversion, the width offset will make the conversion time of the time-to-digital converter longer, and more output bit numbers are required to be stored. The last conversion result. Compared with the first embodiment, the purpose of this second embodiment is mainly to design a suitable time width decoupling circuit, so that the output time signal is brought to a time width close to zero at the lower limit of the temperature to be measured. To this end, the second embodiment preferably replaces the wire 212 of the first embodiment with a second delay line 2112 having a temperature compensation circuit, since the second delay line 2112 through the temperature compensation circuit has a lower temperature sensitivity coefficient ( Thermal coefficient), the delay time may depend on the length of the delay line (equivalent to the number of delay elements in the delay line) and is less affected by temperature changes. The operation principle of the second embodiment is similar to that of the first embodiment. At the beginning, a one-step start signal START is input, and the step signal is delayed by the first delay line 211 for a period of time and then sent to the first of the mutual exclusion or gate 213. At the input end, if the first delay line 211 is composed of a positive temperature coefficient delay element such as a reverse gate, the delay time 12 1275782 will be proportional to the temperature to be measured; meanwhile, the step signal is also passed through the second delay line. 2112 is delayed for another period of time and sent to the first input of the mutex or gate 213, and the other delay time is less affected by the temperature change; finally, a logical operation of the mutex or gate 213 can generate a corresponding The pulse signal Pm is output. When the temperature rises, the delay time of the step signal caused by the first delay line 211 having a positive coefficient is also increased, and the delay time of the step signal caused by the second delay line 2112 of the temperature compensation circuit is changed. Small, so the pulse time width generated by mutual exclusion or after 213 will also increase with temperature. In the case of the Xingdubei case, the appropriate delay is used to make the step (4) caused by the first delay line 211 and the second delay line 2112 close to each other at the lower limit of the required measured temperature specification. The width will be close to zero, which can effectively correct the lack of excessive output offset at the lower limit of the measured temperature. For example, the original first delay line 211 has a considerable offset at the lower temperature limit, as shown in the sixth figure (4). After a proper deduction via the second delay line 2U2, the resulting output time width has only a small amount of offset at the lower temperature limit. Second, please note that, in Figure 6 (b), increase the delay time caused by the first delay and line 2112 by appropriately setting the number of people (specifically increasing the number of delay elements in the delay line). The time difference between the two delay lines of the two delay lines can be used to obtain a time signal with a decrease in visibility. This is a further embodiment of a temperature sensing circuit 21 having a negative temperature coefficient. Similarly, if the 13 1275782 late line 211 has a negative temperature coefficient, we can also generate a temperature sensing circuit 21 having a positive temperature coefficient or a negative temperature coefficient by adjusting the delay time of the second delay line 2112. It should be noted that if the back-end time to the digital converter 22 requires two sets of step signals as input, and the time difference between the two is used as the time conversion basis, the third figure (b) can be mutually exclusive. Alternatively, the gate 213 is removed, and the signals PA and PB of the first delay line 211 and the second delay line 2112 are directly outputted, which is another modification of the second embodiment of the temperature sensor 21. The third figure (c) is a circuit function block diagram of the third embodiment of the temperature sensor 21 of the present invention, which includes a current source 214 proportional to absolute temperature (PTAT), and the output end thereof is electrically connected. The first input end of the control circuit 216 is used for temperature sensing; a constant current source 215 has an output terminal electrically connected to the second input end of the control circuit 216, and the current value is substantially independent of temperature; The first input end of the control circuit 216 is electrically connected to the PTAT current source 214, the second input end is electrically connected to the constant current source 215, and the output end is electrically connected to the time extraction circuit 217, and the function is to connect the PTAT current source and the constant current source. Voltage conversion is performed and the two are compared and judged; the time extraction circuit 217 is electrically connected to the output of the control circuit 216 for time extraction. The detailed operation principle is as shown in the seventh figure (a), the PTAT current source 214 and the constant current source 215 are respectively input to the control circuit 216, the first stage 14 1275782 segment: the control circuit 216 first charges the PTAT current source 214 for a fixed time or Discharge (figure (a) is indicated by charging), since the pTAT current source 214 changes linearly with temperature or decreases, the voltage value caused by charging or discharging at a fixed time will also vary with temperature ( As shown, T〇, T! represents different temperatures) (3 second-order end: control circuit 216 uses the constant current source 215 for reverse discharge or charging action (sixth figure (a) is indicated by discharge), due to The constant current source does not change with temperature, so the slope of the discharge is the same; however, the voltage level caused by the different temperatures τ〇 is different (as shown in the figure, the voltage is generated at the temperature of τ〇 and 丁1) , V i), so the time required for discharging through the constant current source 2丨5 is naturally different, and is respectively labeled as 砣, ^ ', and then the time extraction circuit 217 takes the time required for the constant current source 215 to discharge as the output time signal. Due to PTAT current source 21 4 linearly changes with temperature, the voltage level caused by charging at a fixed time also changes linearly. Therefore, the time obtained by the time extraction circuit 217 also changes linearly after the discharge of the pen source 215. Therefore, the temperature sensing is performed. The second embodiment of the circuit 21 produces a linearly increasing or decreasing time signal with temperature. Another operational principle of the second preferred embodiment can be represented by the seventh figure (the PTAT current source 214 produces a line with temperature). The current is increased or decreased, and the control circuit 216 passes this current through an impedance element (not shown) to produce a voltage that linearly increases or decreases with temperature (as shown, at ο T! ' Do not generate voltages v 〇, v ο, the size of the constant current source 215 is not affected by the temperature change, so the slope of the charging is fixed via the -capacitive tree (not shown in Figure 12 1275782), and the charging voltage value increases with time. The green circuit changes. The control circuit 216 continuously compares the power converted by the PTAT current source with the charging voltage of the constant current source 215 until the two are equal. The time 电路 circuit 217 is used to draw the constant current source 215. The required charging time is used as the turn-off time signal. Since the PTAT f current source 214 changes linearly with temperature, the voltage level generated by an impedance element also changes linearly, so the charging time required for constant current = 215 is also Linearly change with temperature, when generated = the width of the number will increase or decrease linearly with the change of temperature. In this case, the time converter is used to convert the width of the time signal into the pair check position. Knowing the real financial type mainly includes counting-method-based %-pulse reduction method, double-slope method, vernier method, interpolation method, etc.; ', purpose& is to change the day (four) signal to the corresponding digital rotation signal. Embodiments], ^ is the circuit function block diagram of the novel digital temperature sensing system of the present case, which has a linear change of w w temperature: ^1, used for the day. Fl number Ρ _, and time to digital to t way 22; time to digital conversion circuit 22 can be used as the rear end circuit of the new variable second degree sensing system for the temperature sensing of the front end, the day-to-day tracking number conversion Into the corresponding digital signal. ^图图 (a) is a block diagram of the first embodiment of the temperature sensor 21 of the present embodiment and its guard timing chart, which includes - a first delay line 211, a guide, a 诤, a fruit 212 , a mutex or gate 213 is used as a time width 16 1275782 to capture the circuit. Preferably, the wire of the towel is a metal delay, and the table & /JHL degree rises, and the time delay of the first delay line 211 to delay the START is also linearly increased, and the wire 212 ^ START signal is delayed. The time is not significantly affected by temperature changes. Therefore, the delay time of STAR caused by the two lines increases linearly with the temperature rise after the mutual exclusion or gate 213 is used for the time width. That is, after the temperature sensor of the first embodiment is triggered by the S?T signal, an inter-day nickname which linearly increases as the temperature rises can be obtained. In addition to the pulse width, the time signal also includes the time difference width of the above two signals. FIG. 2B is a circuit function block diagram of the second embodiment of the present temperature sensor 21 and an operation timing chart thereof, which includes a first delay line 2li 'a second delay line 2112 of a temperature compensation circuit, Mutual exclusion or gate 213 is used as a time width capture circuit. In contrast to the first embodiment, the purpose of this second embodiment is mainly to design a suitable pulse width buckle circuit to allow the round pulse (P_) to approach zero at the lower temperature limit (or upper limit) of the desired measurement. Pulse width. Reduce the conversion time and the number of output bits required as the time of the back-end circuit to the digital conversion circuit. As the temperature rises, the delay time of the START signal from the first delay line 211 also rises linearly, and the second delay line 2112 of the temperature compensation mussel circuit also delays the START signal #奴日=间' but the temperature The delay time caused by the rise did not change significantly. Therefore, the delay time of the START caused by the two delay lines increases linearly with the time signal width after the mutual exclusion or gate 213 increases with temperature. After the temperature sensor 21 of the second embodiment is triggered by the START signal, a time signal of the linear boosting port as the temperature rises is obtained. Looking at Figure 6 (a), with the proper design, the time signal width can approach zero at the lower temperature limit of the required specification. In addition, please refer to the sixth figure (b). We can also make a proper correction of the second delay line 2112 with the temperature compensation circuit, so that the delay time caused by it is greater than the first delay within the temperature range of the required specification. Line 211, the time pulse width is extracted via the mutex or gate 213. Unlike the sixth graph (a), at this time, we can obtain a time width signal that linearly decreases as the temperature rises. At the upper temperature limit of the required specification, the time signal width can also approach zero. FIG. 3(c) is a circuit functional block diagram of the third embodiment of the temperature sensor 21 of the present invention, which includes a PTAT current source 214 electrically connected to the first input end of the control circuit 216 for temperature sensing. The current varies linearly with temperature; a constant current source 215 is electrically connected to the second input of the control circuit 216, and the current value is not related to the temperature; the first input of the control circuit 216 is electrically connected to the PTAT current source. The output of 214 is electrically connected to the output of the constant current source 215, and the output of the 216 is electrically connected to the time extraction circuit 217. The control circuit 216 charges (or discharges) the PTAT current source 214 for a fixed time, and then The time obtained by the reverse current discharge (or charging) by the constant current source 215 is outputted by the time 电路 circuit 217, and the output time width can be linearly changed with the temperature change, and the correlation signal and timing are as shown in the seventh figure (a). ) shown. 18 1275782 In addition, as shown in the seventh diagram (b), the control circuit 216 can convert the PTAT current source 214 through a resistive element (not shown) to be linearly proportional to the temperature (or inversely proportional to its temperature coefficient). The voltage is positive and negative, and then the constant current source 215 charges the capacitive component (not shown in the figure). The voltage increases with time but is approximately independent of temperature, and then the two voltages are compared to allow time 撷The circuit 217 takes the charging time required for the constant current source 215 as an output time signal whose width will increase or decrease linearly with temperature. The control circuit 216 is preferably a voltage comparator and a resistive component, a capacitive component and a charge control switch. The time-to-digital converter used in this case is a conversion circuit that converts the width of the time signal into a corresponding digital output. The implementation methods of the conventional methods mainly include a counter method, a cyclic pulse reduction method, a double slope method, a vernier scale method, an interpolation method, etc.; the purpose is to convert a time signal into a corresponding digital output signal. In summary, the digital temperature sensing system of the present invention can be applied to a central processor, a chipset, a battery protection 1C, an environmental monitoring system, an electromechanical system, and other application electronic circuits with temperature sensing circuits. The temperature sensor of this case can generate a time signal linearly related to the temperature signal. This new circuit has no bipolar junction transistor 'so no additional correction circuit (meaning wafer area and power consumption rise) and curvature correction technology to reduce the error. Therefore, the difficulty of design, cost saving and power consumption can be greatly improved, and the purpose of developing the case can be effectively achieved. Especially when applied to portable systems, it can effectively reduce the cost and power consumption of 19 1275782. Therefore, this case is a novel, progressive and practical creation, and it is submitted in accordance with the law. This case has been modified by people who are familiar with the art, but it is not intended to be protected by the scope of the patent application. [Simple description of the diagram] The first figure: the circuit function block diagram of the conventional digital temperature sensing system. The second picture: the circuit function block diagram of the digital temperature sensing system in this case. Fig. 3(a) is a block diagram showing the circuit function of the first embodiment of the temperature sensor of the present invention. Fig. 3(b) is a block diagram showing the circuit function of the second embodiment of the temperature sensor of the present invention. Fig. 3(c) is a block diagram showing the circuit function of the third embodiment of the temperature sensor of the present invention. Fourth figure: The temperature characteristic curve of the first embodiment of the temperature sensor of the present invention is shown. Figure 5: Schematic diagram of the delay component architecture with temperature compensation. Fig. 6(a): Schematic diagram of the operation of the second embodiment of the temperature sensor of the present invention A 第六 Fig. 6(b): Schematic diagram of the operation of the second embodiment of the temperature sensor of the present invention. Figure 7 (a): Working diagram of the third embodiment of the temperature sensor of the present invention 20 1275782 A 〇 seventh diagram (b): operation diagram of the third embodiment of the temperature sensor of the present invention B 〇 [main component symbol description 】 S TART: circuit trigger start signal
Pcut:溫度感測器之脈衝式時間輸出訊號 PA:第一延遲線之輸出訊號 PB:第二延遲線之輸出訊號 11:溫度感測器 12:參考源電路 13:類比至數位轉換器 21:溫度感測器 22:時間至數位轉換電路 211:第一延遲線 212:導線 213:互斥或閘 2112:具溫度補償之第二延遲線 214:PTAT電流源 215:定電流源 216:控制電路 217:時間擷取電路 21Pcut: pulsed time output signal of temperature sensor PA: output signal of first delay line PB: output signal of second delay line 11: temperature sensor 12: reference source circuit 13: analog to digital converter 21: Temperature sensor 22: time to digital conversion circuit 211: first delay line 212: wire 213: mutual exclusion or gate 2112: temperature compensated second delay line 214: PTAT current source 215: constant current source 216: control circuit 217: Time capture circuit 21