TWI261414B - An improved analogue selector - Google Patents

An improved analogue selector Download PDF

Info

Publication number
TWI261414B
TWI261414B TW92126541A TW92126541A TWI261414B TW I261414 B TWI261414 B TW I261414B TW 92126541 A TW92126541 A TW 92126541A TW 92126541 A TW92126541 A TW 92126541A TW I261414 B TWI261414 B TW I261414B
Authority
TW
Taiwan
Prior art keywords
input
selector
switch
scope
gain
Prior art date
Application number
TW92126541A
Other languages
Chinese (zh)
Other versions
TW200419904A (en
Inventor
John Laurence Pennock
Edward Mauger Granville
Original Assignee
Wolfson Microelectronics Plc
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority claimed from GB0227553A external-priority patent/GB2395849B/en
Application filed by Wolfson Microelectronics Plc filed Critical Wolfson Microelectronics Plc
Publication of TW200419904A publication Critical patent/TW200419904A/en
Application granted granted Critical
Publication of TWI261414B publication Critical patent/TWI261414B/en

Links

Landscapes

  • Amplifiers (AREA)
  • Electronic Switches (AREA)
  • Control Of Amplification And Gain Control (AREA)

Abstract

The present invention relates to analogue selectors, also known as analogue multiplexers, for audio equipment input source selection. The present invention provides an analogue input selector for selecting one of a number of analogue source devices. The selector comprises a plurality of input connections for coupling to the input sources, each said input connection being connected to a gain controller having at least zero and one or more non-zero gains, said input connection not including a semiconductor switch. Preferably the gain controllers are implemented using MOSFETs which are held at a voltage which is independent of the signal voltage.

Description

1261414 玫、發明說明: 【發明所屬之技術々員城】 發明領域 本赉明疋有關於類比選擇器,亦被知為類比多工器, 5尤其但並不僅由於積體電路中實施,用於例如高傳真(hi-fi) 放大為之音訊設備。 發明背景 類比輸入選擇器具有許多種應用,包括例如使用於音 10訊娛樂設備中,例如:家庭用高傳真音響設備,車輛中娱 樂設備以及手提式設備,其在正常情況下提供音訊來源之 選擇,例如心播放器、磁帶卡、或無線電收音機。此等 設備典型地將音訊輸入信號連接至一或多個輸出:例如: 耳機、擴音機、(磁帶)錄音機以及數位器。此典型地使用輸 15入選擇器庫實施,其一個用於各輸出。 隨著設備微型化之增加,此等類比選擇器典型地使用 MOSFET切換技術以積體電路(IC)實施。使用例如為 之MOSFET是因為其相較於其他型式之電晶體切換技術為 低之漏電電流。第1圖顯示用於類比輸入選擇器之典型電路 20配置,其具有可設計之增益。因為由不同型式類比來源所 提供不同之信號位準,所以使用可設計或可變輪入择兴 例如,5己錄卡可以長:供數個毫Krms,而來自音欵—〜備— 準“線,,輸出提供典型的2Vms。對各輸入來源設有1之標 選擇器開關,且此cm〇s增益選擇器開關是根據 所選擇之 1261414 、、局“原所提供之信號位準而配置。因此,不論其輸入來 、可此颂比選擇器之輸出信號位準應該類似。 此種配置的問題是,在IC中使用cm〇s型式之開關會 5大致如同於第2圖中所說明之取決於信號位準之失真。這是 :〇 S開關之電喊取決於從及極至閘極與從源極至 2 a壓’料電壓在選擇11開關巾並非恒定,因為;:及 ”壓與源極電壓隨著信號上升與下降,而問極電壓是與 固定之電源電壓有關。 H丨…此問題之解決讀是增加選㈣開關之大小,以便減 ^它1對於在電路巾其練定㈣阻抗之輸人阻抗,並且 ^牛低跨及極/源極之電壓,並且因此降低此CMOS開關 ^阻任何調變之相對效應。然而,較大的開關須要更多的 -材料,其增加成本且亦增加非所欲之寄生電容,其例如 會使得電源抑制退化。 15 對此問題所建議之其他解決方案是在由M. Sato, K.1261414 玫,发明说明: [Technology Employee City to which the invention belongs] Field of the Invention The present invention relates to an analog selector, also known as an analog multiplexer, 5 especially but not only implemented in an integrated circuit, For example, high-fidelity (hi-fi) is amplified into an audio device. BACKGROUND OF THE INVENTION Analog input selectors have a variety of applications including, for example, use in audio 10 entertainment devices, such as home high-fidelity audio equipment, in-vehicle entertainment equipment, and hand-held equipment, which provide a choice of audio sources under normal conditions. , for example, a heart player, a tape card, or a radio. These devices typically connect an audio input signal to one or more outputs: for example: headphones, amplifiers, (tape) recorders, and digitizers. This is typically implemented using an input selector library, one for each output. As device miniaturization increases, such analog selectors are typically implemented in integrated circuits (ICs) using MOSFET switching techniques. The MOSFET is used, for example, because it has a low leakage current compared to other types of transistor switching techniques. Figure 1 shows a typical circuit 20 configuration for an analog input selector with a design gain. Because different signal levels are provided by different types of analog sources, the use of designable or variable wheeling options, for example, 5 card cards can be long: for a few milli-Krms, and from the sound-----" Line, the output provides a typical 2Vms. There is a 1 selector switch for each input source, and the cm〇s gain selector switch is configured according to the selected 1261414, the “original signal level” . Therefore, regardless of its input, the output signal level of the selector should be similar. The problem with this configuration is that the use of the cm〇s type of switch in the IC will be substantially as distorted by the signal level as illustrated in Figure 2. This is: 〇S switch's electric shouting depends on the voltage from the pole to the gate and from the source to 2 a. The material voltage is not constant in the selection of the 11 switch, because; and the voltage and source voltage rise with the signal. And drop, and ask the pole voltage is related to the fixed power supply voltage. H丨... The solution to this problem is to increase the size of the selected (four) switch, so as to reduce the input impedance of the impedance in the circuit board. And ^Niu low-span and pole/source voltage, and thus reduce the relative effect of any modulation of this CMOS switch. However, larger switches require more - material, which increases cost and increases unwanted The parasitic capacitance, which for example causes the power supply to degrade. 15 Other solutions suggested for this problem are at M. Sato, K.

Suzuki,Τ· Suzuki 以及Κ· Akutsu所著之“A Volume and Tone Control IC for h卜Fi Audio”,IEEE J Solid-State Circuits,Vol 16, No· 6, Dec 1981. p682-688中說明。此文件說明對閘 極電極使用非常高的薄片電阻材料。然後將此局部閘極電 2〇 t以“號頻率以電容方式輕合至通道,以維持大致怪定之 閘極通道電壓。但此須要不尋常之製造步驟,其增加成本 並且此種處理並不容易實施。其在當從一通道切換至另一 通道日寸亦導致延遲,並且因此並未採用此觀念。 另一個解決方案是在T L. Bro〇ks等人所著之“a 1261414Suzuki, Τ·Suzuki and Κ·Akutsu, "A Volume and Tone Control IC for h Bu Fi Audio", IEEE J Solid-State Circuits, Vol 16, No. 6, Dec 1981. p682-688. This document shows the use of very high sheet resistance materials for the gate electrodes. Then, the local gate is electrically connected to the channel at a frequency of "number" to maintain a substantially strange gate channel voltage. However, this requires an unusual manufacturing step, which increases the cost and the processing is not It is easy to implement. It also causes delays when switching from one channel to another, and therefore does not adopt this concept. Another solution is in T L. Bro〇ks et al. "a 1261414

Cascaded Sigma-Delta Pipeline A/D Converter with 1.25MHz Signal Bandwidth^ in IEEE Journal of Solid-StateCascaded Sigma-Delta Pipeline A/D Converter with 1.25MHz Signal Bandwidth^ in IEEE Journal of Solid-State

Circuits,Vol.32, Νο·12, Dec 1997 中說明,尤其請參考 1901-2 頁。其使用充電粟結構,而取決於先前所取樣之信號位準 5而在電源電壓上有效地增加閘極電壓,以嘗試維持源極/汲 極對閘極實質上恆定之電壓差異。然而,此結構導致須要 鬲速日t脈’其本身可以為添加至信號之失真與切換暫態之 來源。其電路之額外複雑性亦會增加晶片面積,並且因此 增加成本。此涉及以充電泵之方式將電壓提升高於正常之 1〇供應電壓,對於開關與驅動此開關之任何電路,會造成可 靠度之風險。 Γ ^^明内容3 發明概要 一般而言,本發明之目的為提供一種類比選擇器,盆 中藉由將祕或祕端子保持如定或與㈣無社電壓 例如接地«,而將使用作為電晶體開關之例_〇sfet 15 之電晶體之閘極對源極或閘極對汲極之電壓保持恆定。、言 意味著電壓不隨著信號電壓改變而改蠻。 ° 艾此精由以用於各 輸入連接之增魏取代連接⑽各“連接料益級之間 選擇開關而達成。此用於所選擇輪入通道之增益級以其構 成之m〇sfet開關之一導通(。η)而操作,並且二 其他與電源有關之電壓,以便提供實皙 5 、、f疋之源極與沒 極至閘極電壓。藉由去除浮動電壓選擇開關。而實質上去 除隨著源極至閘極以及汲極至閘極電壓之 、、 ^ 文交之開關電阻 20 1261414 所造成之失真。 為了說明的目的,使用汲極、源極以及閘極是為了簡 單與清楚之說明,但其用意為包括在非-FET開關中具有類 似功能之相對應結構。例如,此名詞汲極 '源極與閘極之 5 用意為包括:在雙載子接面電晶體中集極、射極以及基極 結構,以及在其他型式半導體裝置中具有相同功能之類似 結構。 在本發明之觀點中尤其提供如申請專利範圍第1項之 類比選擇器。 10 此種配置規定,此類比選擇器之輸入連接級(stage)並 不須要用於各輸入之選擇器開關,並且因此去除失真之重 要來源。此切換功能現在由輸入增益級或控制器實施。然 而典型地使用大的選擇器開關,以便減少取決於信號位準 之失真,並且藉由佈局預防措施而進一步增加其實體大 15 小,以允許它們承受直接施加之ESD與閂鎖應力,藉由本 配置將此等去除,亦可節省寶貴之1C材料。 此等增益級或控制器較佳包括一或多個半導體開關, 其各具有連接至選擇器接地之源極或汲極。 此配置提供給開關固定之源極至閘極電壓以及汲極至 20 閘極電壓,以作為相對於電源電壓而非信號電壓之實質接 地。由於可以使用小的開關結構,此種配置使用相當少之 的1C材料,因為此等開關實質上並未承載信號電流,且並 不直接曝露於外部ESD或閂鎖應力。 在一實施例中,此選擇器包括例如為運算放大器(op 1261414 amp)之放大器而具有輸入與輸出,且各增益級或控制器更 包括:連接介於放大器輸出與各輸入連接之間之阻抗鏈, 此等開關是連接介於此鏈與放大器輸入之間。以替代之方 式,各增盈級可以包括專用放大器,以致於將每個通道之 5整個運异放大級輸入級重覆,並且如所須要地切換入。 將一阻抗鏈典型地執行作為例如電阻器之一系列離散 阻抗I置,沿著此例如在離散阻抗之間連接之鏈,在不同 之點可以掏取不同組合之阻抗。 此允許使用例如一系列電阻器之簡單阻抗鏈選擇多增 10盈因數。此等開關是連接介於電阻器連接與放大器輸入之 間如果使得此等開關均為切斷(open)(零增益),則此將輸 入信號切斷。 15 在曰代的配置中,一或多個增益控制器可以包括:僅 /、们開關其具有連接介於此開關與各輸入之間之阻抗· 以,連接介於此開關放大器輸出之間之其他阻抗。此種配 置提供兩種設定:零(靖)與非,(導通);並朋此作 、、屯之輸人選擇☆、而沒有可調整之非·零增益。 … — 衣直,具當在苓增益時用於將> 曰皿級短路至參考電壓。此降低在輸人電 ㈣ 之可能性。 j U f相 &衣置車乂佳包括連接介於各阻抗鏈與參考带题 3之=。:玫大器較佳具有連接至參考電壓之第::入 :作:積體電路執行時,此短路裝外 的開關,其社地從料他恤路裝置„連接至= 20 1261414 電壓。 由於此等選擇器開關實質上並未通過信號電流,可以 實體上為小。對此短路開關電阻須求亦不大,因為它們並 不在直接的信號通路中。在此兩種情形中NMOS或PMOS開 5 關即足夠,其占用少量之1C材料。 本發明之第二觀點為提供一種多工可程式增益放大 器。其包括多個輸入連接用於連接至多個輸入來源。各輸 入連接是連接至可程式增益級,其具有至少為零以及一或 多個非零增益。各可程式增益級較佳具有一或多個半導體 10 開關,各具有連接至選擇器之接地的非-閘極輸入。各可程 式增益級較佳具有阻抗鏈,其連接介於放大器輸出與各輸 入連接之間。此等開關是連接介於鏈與放大器輸入之間。 本發明其他的實施與優點將由以下實施例之說明而為 明顯。 15 以下將參考所附圖式詳細說明本發明之實施例,其只 作為例子且並無意用於限制。 圖式簡單說明 第1圖顯示用於類比選擇器之已知結構; 第2a-2d圖顯示CMOS裝置之概要圖式,以及用於此裝 20 置各種操作情況之導電特徵圖形; 第3圖顯示根據本發明實施例用於類比選擇器之結構; 第4圖顯示根據另一實施例之結構; 第5圖顯示第4圖結構之加強; 第6圖顯示第5圖結構之另一加強;以及 10 1261414 "°貝示用於參考電壓源之緩衝電路。 【實施冬式】 較佳實施例之祥細說明 4蒼考第1圖,其顯示已知之類比選擇器或多工器結 )構。將數個選择器l〇i、l〇ii典型地併入音訊設備等中,以 便將來自例如CD播放器與微音器;若干類輸入來源1之信 號連接至數個輪出。各選擇器10包括多個輸入連接11用於 連接至輸入來源1。各輸入來源1亦連接至其他選擇器10, 以便允許在音訊設備之輸入來源與輸出之間之多路切換。 10 各連接10具有選擇或輸入開關12a-12c,其一側使用各 別輸入連接lla-llc而連接至各別輸入來源ia-ic。各輸入開 關12a-12c之另一側連接至電阻器鏈或串13,其包括數個串 聯之離散電阻器13a-13n。此鏈13之另一端連接至運算放大 器(op amp)15之輸出(Vo),其亦提供來自選擇器電路1〇之輸 15出。從電阻器鏈13分接出若干增益開關143-141^,其各來 自電阻器13a-13n中兩個電阻器之間之連接。此等開關14之 另一端連接至運算放大器15之反相輸入㈠。 藉由改變此等增益開關14之導通/切斷(on/off)狀態,此 由運算放大器15所提供之信號增益是藉由此電阻器/開關 20 網路所提供之負回饋配置而改變。因此,對於高位準輸入 來源1,可以降低此運算放大器之增益,以致於不論輸入來 源之位準如何,可以由選擇器10提供正常化之輸出信號位 準(Vo) 〇 典型地只將此等開關14a-—關閉(close)。例如, 1261414 如果開關14b為關閉,則此增益為:Circuits, Vol.32, Νο·12, Dec 1997, especially on page 1901-2. It uses a charged die structure to effectively increase the gate voltage across the supply voltage depending on the previously sampled signal level 5 in an attempt to maintain a source/thinance voltage differential that is substantially constant to the gate. However, this structure results in the need for an idle time t pulse 'which itself can be a source of distortion and switching transients added to the signal. The extra reproducibility of the circuit also increases the chip area and therefore increases the cost. This involves raising the voltage higher than the normal supply voltage by means of a charge pump, which creates a risk of reliability for any circuit that switches and drives the switch. SUMMARY OF THE INVENTION Summary of the Invention In general, it is an object of the present invention to provide an analog selector that will be used in the basin by holding the secret or secret terminal as determined or with (4) a non-social voltage such as grounding « In the case of a transistor switch, the gate of the transistor of 〇sfet 15 is kept constant for the voltage of the source or gate to the drain. , meaning that the voltage does not change with the signal voltage. ° Ai Jingjing is achieved by replacing the connection between each input connection (10) for each input connection. This is used to select the switch between the gain stages of the selected input channel. This is used for the m〇sfet switch of the gain stage of the selected wheel input channel. Operates with a turn-on (.n) and two other voltages associated with the power supply to provide the source and the gate-to-gate voltage of the real 5, f疋. By removing the floating voltage select switch, substantially removing Distortion caused by the source-to-gate and drain-to-gate voltages, and the switching resistance of the switch 12 1261414. For illustrative purposes, the drain, source, and gate are used for simplicity and clarity. Description, but intended to include a corresponding structure having a similar function in a non-FET switch. For example, the term 'source and gate 5' is intended to include: collecting in a double-carrier junction transistor , an emitter and a base structure, and a similar structure having the same function in other types of semiconductor devices. In particular, an analog selector as in the first aspect of the patent application is provided in the viewpoint of the present invention. The analog connection stage input stage does not require a selector switch for each input, and therefore removes an important source of distortion. This switching function is now implemented by the input gain stage or controller. However, large selections are typically used. Switch to reduce distortion depending on the signal level, and further increase the physical size by 15 by the layout precaution to allow them to withstand the directly applied ESD and latching stress, which is removed by this configuration. The valuable 1C material can be saved. These gain stages or controllers preferably include one or more semiconductor switches each having a source or drain connected to the selector ground. This configuration provides the source to the gate of the switch. The pole voltage and the drain voltage to 20 gate voltage are used as the substantial grounding relative to the supply voltage rather than the signal voltage. Since a small switching structure can be used, this configuration uses a relatively small amount of 1C material because these switches are essentially The signal current is not carried and is not directly exposed to external ESD or latch stress. In an embodiment, the selector includes, for example An input and an output for an amplifier of an operational amplifier (op 1261414 amp), and each gain stage or controller further includes: an impedance chain connected between the output of the amplifier and each input connection, the switches are connected between the chains Between the amplifier inputs and the amplifier inputs, each of the gain stages may include a dedicated amplifier such that the entire input amplifier stage of each of the channels is repeated and switched in as needed. Typically implemented as a series of discrete impedance I, such as a resistor, along which, for example, a chain connected between discrete impedances, different combinations of impedances can be drawn at different points. This allows for the use of, for example, a series of resistors. The impedance chain selection is increased by a factor of 10. These switches are connected between the resistor connection and the amplifier input. If these switches are both open (zero gain), the input signal is cut off. 15 In a deuterated configuration, one or more of the gain controllers may include: only /, the switch has an impedance connected between the switch and each input, and the connection is between the output of the switching amplifier Other impedance. This configuration provides two settings: zero (jing) and non- (conduct); and this is done, the choice of the input is ☆, and there is no adjustable non-zero gain. ... — Straight, used to short the > class to the reference voltage when gaining. This reduces the possibility of losing electricity (4). j U f phase & clothing 乂 包括 includes the connection between the impedance chain and the reference band 3 =. The rose device preferably has a connection to the reference voltage::in: for: when the integrated circuit is executed, the short circuit is installed outside the switch, and the social device is connected to the device „connected to = 20 1261414 voltage. These selector switches do not substantially pass the signal current and can be physically small. The short-circuit switch resistance is not required to be large because they are not in the direct signal path. In either case, the NMOS or PMOS is turned on. 5 is sufficient to occupy a small amount of 1C material. A second aspect of the present invention is to provide a multiplex programmable gain amplifier that includes a plurality of input connections for connection to a plurality of input sources. Each input connection is connected to a programmable A gain stage having at least zero and one or more non-zero gains. Each programmable gain stage preferably has one or more semiconductor 10 switches, each having a non-gate input connected to the ground of the selector. The program gain stage preferably has an impedance chain connected between the amplifier output and each input connection. These switches are connected between the chain and the amplifier input. Other implementations and advantages of the present invention The embodiments of the present invention will be described in detail below with reference to the accompanying drawings, which are illustrated by way of example only, and not by way of limitation. FIG. Known structure; Figures 2a-2d show schematic views of a CMOS device, and conductive features for various operations of the device; Figure 3 shows a structure for an analog selector in accordance with an embodiment of the present invention; Figure 4 shows the structure according to another embodiment; Figure 5 shows the enhancement of the structure of Figure 4; Figure 6 shows another enhancement of the structure of Figure 5; and 10 1261414 "° indicates the reference voltage source Buffer circuit. [Implementation of winter style] The detailed description of the preferred embodiment is shown in Fig. 1, which shows a known analog selector or multiplexer structure. Several selectors l〇i, l〇 Ii is typically incorporated into an audio device or the like to connect signals from, for example, a CD player to a microphone; several types of input sources 1 to a number of rounds. Each selector 10 includes a plurality of input connections 11 for connection to Input source 1. Each input source 1 Also connected to other selectors 10 to allow for multiplexing between the input source and output of the audio device. 10 Each connection 10 has a select or input switch 12a-12c with a separate input connection 11a-llc on one side. Connected to a respective input source ia-ic. The other side of each input switch 12a-12c is connected to a resistor chain or string 13 comprising a plurality of discrete resistors 13a-13n connected in series. The other end of the chain 13 is connected to The output of the operational amplifier (op amp) 15 (Vo), which also provides the output 15 from the selector circuit 1. A plurality of gain switches 143-141^ are tapped from the resistor chain 13, each from the resistor 13a- The connection between the two resistors in 13n. The other end of these switches 14 is coupled to the inverting input (I) of operational amplifier 15. By varying the on/off state of the gain switches 14, the signal gain provided by the operational amplifier 15 is varied by the negative feedback configuration provided by the resistor/switch 20 network. Therefore, for a high level input source 1, the gain of the operational amplifier can be reduced so that the normalized output signal level (Vo) can be provided by the selector 10 regardless of the level of the input source, typically only such Switch 14a - close. For example, 1261414 If switch 14b is off, then the gain is:

Vout/Vin = -(R13 c+R 13 d+... +R13 n)/(Rs+R 12 c+R 13 a+R 13 b) Vout=在節點(Vo)之輸出信號電壓振幅 Vm=由相同信號電壓源(lc)所產生之輸入信號電壓振幅 5 Rs=相同之信號源阻抗(lc與2c) R12c、R14b = M0SFET開關之導通電阻(12c與14b) R13a、R13b=各電阻器Ri3a、R13b之電阻 在理論上可以將此等開關之一個以上關閉, 而將笔阻器鍵13之數個區段短路,以便提供額外之增益設 10定。然而,在實際上藉由一次只選擇一個關閉而將操作保 持簡單。 第2a圖顯示CMOS傳送閘裝置,其包括PMOS部份與 NMOS部份,且典型地使用於以上所說明選擇器1〇之開關 實施(12、14)中。此裝置(12、14)包括:源極S與汲極D,其 15連接至選擇器10中之其他電路元件,以及PMOS與NMOS閘 極Gp與Gn,以控制此裝置_^PM〇s與NM0S部份之導電。 當PMOS閘極Gp為高電位(+ve供應,vdd)且NMOS閘極 Gn為低電位(-ve供應,接地)時,則此裝置之nm〇S與PMOS 部份均為切斷(off)。當此PMOS閘極是在低電位且NMOS閘 20 極是在高電位時,此兩個部份均導通。此外,當;:及極D與源 極S為接地時(或小於在接地電位以上之pmos臨界電壓 Vtp)’此PMOS部份被切斷(0的,以及當;;及極與源極是在vdd 時(或小於此低於+ve電力供應从仙之NMOS臨界電壓Vtn), 此NMOS部份被切斷。此等特性是顯示於第2b圖之導電圖 12 1261414 形中。可以看出此汲極-源極導電隨著汲極與源極電壓而改 變。換句話說,此裝置之導電特性雖著在汲極與源極之電 壓而改變。 請再參考第1圖,可以看出此選擇開閉裝置12a-12c之 5 及極電壓與源極電壓隨著輸入信號改變,其造成電路之增 益隨著輸入信號變高與變低而改變(如同在以上例中R12c 相對於Rs+R13a+R13b而改變),即,造成取決於信號之電 壓增益調變。此導致輸出信號之失真。 第2c圖顯示使用於可攜式音訊設備中CMOS傳輸閘之 10 特徵,其電源電壓(Vdd)典型地低於非可攜式設備,以便保 護電池使用壽命。可以看出此裝置對於輸入電壓之某些位 準而言實際上為切斷,造成更多的失真。 此問題可以藉由對矽製造過程添加一些步驟而緩和至 某種程度,其允許所選擇電晶體之臨界電壓從正常之 15 800mV降低至例如300mV。此包括此低臨界值電晶體之 CMOS開關將具有如同於第2d圖中所示之一些改良之特 徵。然而,當超過溫度與正常製造公差,此種電晶體在 Vin(在節點11a、llb...) = Vdd或接地時可以開始導通或至少 大量漏電。如果將大的音訊信號施加至未經選擇之輸入, 20 或如果意外或故意將輸入短路至接地或電源,或甚至當將 此設備對外部端子裝附或移除時之瞬間電壓,此可導致此 干擾信號重疊至所選擇之音訊信號上之可聽見之穿透。尤 其此效應對於純粹機械性開關不會發生,則此對許多使用 者是無法接受的。 13 1261414 第3圖顯示根據本發明第一實施例之類比選擇器2〇。此 選擇器20各經由輸入連接21a-21c連接至若干輸入來源 la-lc。各輸入連接2la-21c連接至各增益控制器22心22c。各 化ϋ控制22a-22c包括各阻抗鍵,其具有例如電阻器之若 5干離散阻抗23a_23n。增益開關243-241^從介於電阻器 23a-23n至運异放大器25之反相輸入㈠之間之阻抗鏈分接 出。此運异放大器25之輸出(Vo)連接至各增益控制器22(在 黾阻為23η),以幵》成用於輸入信號放大之負回饋迴路。藉 由控制此等增盈開關24,可將不同數目之電阻器23切換入 1〇回饋迴路與信號輸入通路中,以控制輸入信號增益。此增 益是如同以上說明相對於第1圖之電路而決定。 藉由將增益控制器22之所有增益開關24開啟(open),此 來自相對應輸入來源1之信號無法傳送至運算放大器25,並 且因此未被放大(零増益)。藉由僅使一個增益控制 器(22c) 15具有關閉(cl〇sed)之增益開關(較佳僅為24ca-24cn之一),而 僅將相對應之輸入來源1(:傳送至用於放大之運算放大器 25。因此,此增益開關24亦作為輸入選擇器運作。 此等多個增益控制器22各連接各別的輸入來源1,以實 施輪入選擇功能以及增益控制功能,並且在如此作之過程 中去除對於在第1圖配置中所使用選擇器或輸人開關(12)之 員要。這去除了取決於信號位準之增益,以及因此有關的 輪出信號失真。 因為各增益開關24是連接至高阻抗運算放大器之反相 輪入㈠,匕們並不傳送任何信號電流(除了用於驅動運算放 14 1261414 大⑽之小輪入電容㈠pf)所須之可忽略數量之外)。而且由於 =開關連接至接地,各CM〇·關之源極電壓與沒極電壓實 為艮定,且與供應電壓相關而並非與信號位準相關, 5因此其導通電阻與信號無關。因而可以使則、的CMOS開關 ^ Ό、屯阻态鏈23一起占用少量之1C材料。在事實上當 電晶體導通時,其源極與沒極電壓為已知且值定,因此可 以蘭〇邮細開關取代CMOS開關,而進一步節省所省 略之電晶體與其冗餘驅動電路以及内部連接之面積。而 且,如果此等開關均為NM〇s,則此輸入信號可以振盈超 · 1〇過正的供應電壓,而不會將PMOS(目前省略)之寄生汲極_ 基板二極體順向偏壓,而避免在開機或關機時電路可能鎖 定或損壞,且允許較大的尖峰輸入信號。 此在第1圖中之輸入開關通常須要通過輸入信號,其具 有從接近正供應電壓至接近負供應電壓之電壓,因此此開 - 15關必須為CMOS開關,即,並聯之nm〇^pm〇s。然而, 本實施例之增益選擇開關與短路開關具有固定之汲極與源 極電壓,因此對於各開關只須要NMOS或PMOS,而可將電 _ 晶體之數目減半,且去除對此等開關產生互補驅動信號之 須要,因此減少的面積與成本。 -0 可以藉由回到第1圖而看出其他之優點,而此開關12 是直接連接至積體電路之外部連接或接腳。這意味著它們 可以直接曝露於外部應力,例如:ESD脈衝或意外施加之 電流脈衝。此等輸入典型地須要承受至少2kV之脈衝與 ±200mA之注入電流而不會鎖定或受到永久損害。某些應力 15 1261414 可以由從輸入至電力供應之保護二極體吸收但是因為輸入 擴散直接連接至接腳,大約一半應力典型地必須在開關中 吸收。因此,此輸入開關電晶體12較正常之尺寸為大(例 如,從金屬接觸至輸入擴散至電晶體閘極),並且將由護環 5 擴散或其他結構圍繞,而將寄生電流直接轉回供應接腳而 不是由内部電路擷取。其亦須適當低阻抗之金屬連接以承 載連續之200mA誤差電流。因此,此等開關會占用相當大 的石夕面積。 相較之下,第3圖中之第一開關24ca是藉由典型至少 10 lkohm之電阻器23ca而與外部輸入21c隔離。實際上所有應 力現在將由輸入保護二極體(未圖示)吸收,而以小於1毫安 (mA)之電流流經電阻器23ca而進入開關24a之擴散中,因此 對用於開關24ca之護環只須採取最小特殊防護。此進一步 減少所須之1C材料與成本。 15 此等額外電阻器庫之實體尺寸典型地受限於由製造過 程之照像與其他步驟所獲得之匹配公差,其在當一般尺寸 調整時隨著時間快速改進,然而,此所須承受實體應力之 裝置尺寸受限於在過應力情況中可容忍之最大電流密度, 並且在此製程縮減時並無重大改進。因此,執行本實施例 20 所須之增益控制器22之1C材料隨著製造過程中之改善繼續 縮減,而在習知技術中等同選擇開關之尺寸將保持大致相 同。 本實施例之其他優點為:可以獨立設計裝附於各通道 之電阻器率之阻抗位準。例如,可將一通道設計具有高阻 16 1261414 =以適應相當高阻抗之信號來源丨,而可將另—通道, 计1低阻抗位準以改善_訊。而且不同通道可以 不=數目或型紅增麵則。如果將―或多個通道連接至 =已知信號位準之專用來源,以致於通道可以具有固定 曰孤亚且/、具有兩個電随器23與一個開關24。 口口取决於須要多少個輪出,可以如同所示將若干個選擇 器 20i-20ii並聯。 、Vout/Vin = -(R13 c+R 13 d+... +R13 n)/(Rs+R 12 c+R 13 a+R 13 b) Vout=Output signal voltage amplitude at node (Vo) Vm= The input signal voltage amplitude generated by the same signal voltage source (lc) is 5 Rs = the same signal source impedance (lc and 2c) R12c, R14b = the on-resistance of the MOSFET switch (12c and 14b) R13a, R13b = each resistor Ri3a, The resistor of R13b can theoretically turn off one or more of these switches and short the segments of the pen resistor key 13 to provide an additional gain setting. However, it is actually simple to keep the operation by selecting only one off at a time. Figure 2a shows a CMOS transfer gate device comprising a PMOS portion and an NMOS portion, and is typically used in the switch implementation (12, 14) of the selector 1 described above. The device (12, 14) includes a source S and a drain D, 15 connected to other circuit elements in the selector 10, and PMOS and NMOS gates Gp and Gn to control the device _^PM〇s and Conductive of the NM0S part. When the PMOS gate Gp is high (+ve supply, vdd) and the NMOS gate Gn is low (-ve supply, ground), then the nm〇S and PMOS portions of the device are off (off) . When the PMOS gate is at a low potential and the NMOS gate 20 is at a high potential, both portions are turned on. In addition, when :: and the pole D and the source S are grounded (or less than the pmos threshold voltage Vtp above the ground potential) 'this PMOS portion is cut off (0, and when; and the pole and source are At vdd (or less than this +ve power supply from the NMOS threshold voltage Vtn), the NMOS portion is cut. These characteristics are shown in Figure 12 1261414 in Figure 2b. This drain-source conduction changes with the drain and source voltages. In other words, the conduction characteristics of this device vary with the voltage at the drain and source. Please refer to Figure 1 again. The selection of the switching device 12a-12c 5 and the voltage and source voltages change with the input signal, which causes the gain of the circuit to change as the input signal goes higher and lower (as in the above example, R12c is relative to Rs+R13a). +R13b changes), that is, causes voltage gain modulation depending on the signal. This causes distortion of the output signal. Figure 2c shows the characteristics of the CMOS transmission gate used in portable audio equipment, its power supply voltage (Vdd) Typically lower than non-portable devices to protect battery life It can be seen that this device is actually cut off for certain levels of the input voltage, causing more distortion. This problem can be mitigated to some extent by adding some steps to the manufacturing process, which allows for selection. The threshold voltage of the transistor is reduced from a normal 15 800 mV to, for example, 300 mV. This CMOS switch including this low threshold transistor will have some of the improved features as shown in Figure 2d. However, when temperature and normal manufacturing are exceeded Tolerance, this transistor can start conducting or at least a large amount of leakage when Vin (at node 11a, llb...) = Vdd or ground. If a large audio signal is applied to an unselected input, 20 or if accidentally or Deliberately shorting the input to ground or power, or even when the device is attached or removed to an external terminal, this can cause the interfering signal to overlap to the audible penetration of the selected audio signal. This effect does not occur for purely mechanical switches, which is unacceptable to many users. 13 1261414 Figure 3 shows an analogy according to a first embodiment of the invention The selectors 2 are each connected to a number of input sources la-lc via input connections 21a-21c. Each input connection 2la-21c is coupled to a respective gain controller 22 core 22c. Each of the control controls 22a-22c includes each An impedance key having, for example, a 5 dry discrete impedance 23a-23n of the resistor. The gain switches 243-241 are tapped from an impedance chain between the resistors 23a-23n and the inverting input (I) of the operational amplifier 25. The output (Vo) of the operational amplifier 25 is connected to each of the gain controllers 22 (at a 黾 resistance of 23 η) to form a negative feedback loop for input signal amplification. By controlling these gain switches 24, a different number of resistors 23 can be switched into the 1〇 feedback loop and the signal input path to control the input signal gain. This gain is determined as described above with respect to the circuit of Figure 1. By turning all of the gain switches 24 of the gain controller 22 open, this signal from the corresponding input source 1 cannot be transferred to the operational amplifier 25 and is therefore not amplified (zero benefit). By having only one gain controller (22c) 15 with a closed (g〇sed) gain switch (preferably only one of 24ca-24cn), only the corresponding input source 1 (: is transmitted to be used for amplification) The operational amplifier 25. Therefore, the gain switch 24 also functions as an input selector. The plurality of gain controllers 22 are each connected to a respective input source 1 to implement a wheel selection function and a gain control function, and The process of removing the selector or input switch (12) used in the configuration of Figure 1 is removed. This removes the gain depending on the signal level, and thus the associated wheel-out signal distortion. 24 is the inverting wheel (1) connected to the high-impedance op amp. We do not transmit any signal current (except for the negligible amount required to drive the small-input capacitor (1) pf of the large 12 (10). Moreover, since the = switch is connected to the ground, the source voltage and the immersion voltage of each CM 〇 are determined, and are related to the supply voltage and not related to the signal level, so the on-resistance is independent of the signal. Therefore, the CMOS switch Ό, 屯 resistance chain 23 can be used together to occupy a small amount of 1C material. In fact, when the transistor is turned on, the source and the gate voltage are known and valued, so that the CMOS switch can be replaced by the Lancome fine switch, and the omitted transistor and its redundant drive circuit and internal connection are further saved. area. Moreover, if these switches are all NM〇s, the input signal can vibrate the supply voltage exceeding ·1〇, and the parasitic drain _ substrate diode of the PMOS (currently omitted) is not biased. Press to avoid the circuit may be locked or damaged when turned on or off, and allow for large spike input signals. The input switch in Figure 1 usually needs to pass the input signal, which has a voltage from the near positive supply voltage to the near negative supply voltage, so the open- 15 off must be a CMOS switch, ie, parallel nm 〇 ^ pm s. However, the gain selection switch and the short-circuit switch of this embodiment have fixed drain and source voltages, so only NMOS or PMOS is required for each switch, and the number of electric_crystals can be halved, and the switches are removed. The need for complementary drive signals, thus reducing the area and cost. -0 Other advantages can be seen by returning to Figure 1, which is an external connection or pin that is directly connected to the integrated circuit. This means they can be exposed directly to external stresses such as ESD pulses or accidentally applied current pulses. These inputs typically must withstand at least 2 kV pulses and ±200 mA of injection current without locking or permanent damage. Some stresses 15 1261414 can be absorbed by the protective diode from the input to the power supply but because the input diffusion is directly connected to the pin, approximately half of the stress typically must be absorbed in the switch. Thus, the input switch transistor 12 is larger than normal (eg, from metal contact to input diffusion to the transistor gate) and will be surrounded by the guard ring 5 or other structure to turn the parasitic current back directly to the supply. The foot is not drawn by the internal circuitry. It also requires a suitably low-impedance metal connection to carry a continuous 200mA error current. Therefore, these switches will occupy a considerable area of the stone. In contrast, the first switch 24ca in Figure 3 is isolated from the external input 21c by a resistor 23ca that is typically at least 10 lkohms. In fact, all of the stress will now be absorbed by the input protection diode (not shown) and flow through the resistor 23ca at a current of less than 1 milliamp (mA) into the diffusion of the switch 24a, thus protecting the switch 24ca. The ring only requires minimal special protection. This further reduces the amount of 1C material and cost required. 15 The physical dimensions of these additional resistor banks are typically limited by the matching tolerances obtained from the photo and other steps of the manufacturing process, which are rapidly improved over time when general sizing, however, this is subject to the entity The size of the device of the stress is limited by the maximum current density that can be tolerated in the case of overstress, and there is no significant improvement in the reduction of the process. Therefore, the material of the 1C of the gain controller 22 required to perform the present embodiment 20 continues to decrease as the manufacturing process improves, while the size of the equivalent selection switch in the prior art will remain substantially the same. Another advantage of this embodiment is that the impedance level of the resistors attached to each channel can be independently designed. For example, a channel design can have a high impedance 16 1261414 = to accommodate a relatively high impedance signal source, while another channel, a low impedance level can be improved to improve the signal. And different channels can be no = number or type red. If the "or multiple channels" are connected to a dedicated source of known signal levels, the channel can have a fixed 曰 and /, with two electrical followers 23 and one switch 24. The mouth depends on how many rounds are required, and several selectors 20i-20ii can be connected in parallel as shown. ,

此運算放大ϋ25之非反相輸人(+)典型地藉由在運算 放大器25之正與負電力供應之間半途之良好去柄合偏壓I 1〇偏壓。此在過度負載之前在運算放大器輸出之可供使用尖 峰至尖峰電壓振幅最大化。此低壓積體電路中之負的電力 供應在正當情況下接地,因此此偏壓為電源電壓之Γ半 Vdd/2,而Vdd為例如3·3ν之正的供應電壓。The non-inverting input (+) of the operational amplifier ϋ 25 is typically biased by a good handle-to-handle bias I 1 半 between the positive and negative power supplies of the operational amplifier 25. This maximizes the available peak-to-spike voltage amplitude at the op amp output before over-loading. The negative power supply in this low voltage integrated circuit is grounded under proper conditions, so this bias is half the Vdd/2 of the supply voltage, and Vdd is a positive supply voltage of, for example, 3·3 ν.

如果使用此種單-電源設計,則在所選擇輸入叫之靜 15態電壓必須亦為Vdd/2,而輸人信號le通常為靜態接地電 壓。因此,在第3圖中所示之電容器2c須要連接此輸入信 號。然後,此在所選擇輸入21c上之直流偏壓經由電阻器鏈 23c獲得,而無須任何額外的偏壓電阻器。 以靜態之方式此經由各別電阻器鏈23b等而在未經選 20擇輸入21b上之電壓亦為Vdd/2。因此,電容器沘將被預先 充黾,並且g輸入由一通道切換至另一通道時此電容無須 充電。這在當將輸入切換時,幫助避免大的電壓暫態,其 作為非所欲之“波聲”與“呼”聲而可聽到。(請注意第1圖之電 路將須要從各未經選擇通道至Vdd/2之額外開關,其較佳經 17 1261414 由中間電阻(>lkohm),以避免將輸入信號來源短路與過度 負載)。 第4圖顯示類似於第3圖實施例之另一實施例,但額外 地具有裝置36,用於將各阻抗鏈33短路至參考電壓化。為 5 了清楚起見,只將第3圖與第4圖之一些共同元件顯示於第4 圖中。第4圖是針對第3圖先前實施例之潛在問題,而存在 著從運算放大器之輸出經由各別電阻器串,將輸出信號連 接回未經選擇之輸入。例如,如果電阻器串之電阻總共為 60kohm,以及源極阻抗為600〇11111,則所選擇之輸入信號將 1〇在節點21b出現,而只衰減6〇〇〇hm/60.6kohm〜40dB。如果 此在之信號亦經由2〇ii而使用於並聯豸道中(例如,用 於錄製-信號而同時收聽另一個),則其串擾是無法令人接 受的。 15 20 笔壓力典型土也為參考前-實施例之中間·電源 偏壓,而在未經選擇之交流連接輸人上維持正確之靜能帝 壓。^路電路如备是對於各未使用之增益控制器3^ 動""疋用於所有未經使用之輸入或是除了與所選擇輸入 關增贫控制器(32C)以外之所有增益控制器。各短路 6C包括至各別阻抗鏈33a-33c之連接38a-38c,較佳 3 ^之中點。開關37a_37c連接至各連接38a-38c,且其 妾至*考電㈣。此短路開關的、爪)是在各短路電路 / 36bU_(dGsed),其連接至“未經選擇”之增, 制器(32a、32b)。 、评〈日现控 配置滅少在各種選擇器20i、20Π之輸入“通道”之間 18 1261414 串擾之可能性。如果與此未使用阻抗鏈33有關之來源/阻抗 為高,則此輸入連接或鏈η之左端將具有重疊於其上之 小的信號振幅。此可以呈現在已選擇來源1之另-選擇器 20u之輸入’因此造成串擾。此短路電路36動作,而將重疊 5於未使用鏈33上之信號以短路去除以防止串擾。 如果開關37b具有零阻抗,貝u匕在節點鳥之輸出信號 會被70全抑制。然而,小的開關在此節點典型地將仍然具 有例如1 kohm之阻抗。如果此點是沿著60k〇hm串之中點, 且此來源阻抗為600〇hm,則此信號將在381;)衰減大約 10 lkohm/30k〇hm=30dB,並且當它到達31b時會再衰減600 ohm/30kohm= 34dB。這仍然只有總數64犯。 在第5圖中顯示本實施例之進一步加強,其顯示根據第 4圖之述擇杰30,但其中此短路裝置36,包括多個短路開關 37i、J7ii、37iii,此措由降低各短路電路36’之整體阻抗而 15 進一步減少串擾。 例如’如果連接38i、38ii、38iii具有間隔15kohm之分 接頭,則總共衰減是大約: (lkohm/15kohm)*(lkohm/15kohm)*(lkohm/15kohm)*( 600〇hm/15kohm)=23.5+23.5+23.5+28dB=98.5dB,此為可接 20受之位準,與裝置接腳之間或pCB連接之間之其他不可避 免之寄生耦合效應可相比較。 第6圖顯示此可以減少串擾之短路電路36”之另一加強 形式。此最靠近運算放大器輸出(V〇)之短路開關37iii會耗用 彳觀之電流。例如,對於如同以上之lVrms之輸出音訊信衆 19 1261414 與15kohmi氣阻為區段,將會耗用66yArms之信號電流。 如果Vr之來;原阻抗為1〇^以,則所產生之66^vrms之分量將 施加至所有之Vr端子,包括在此與其他選擇器級中之其他 開關與接地。此信號只低K1Vrms位準83dB。如果以所選 5擇之不同信號施加至另一選擇器塊20ii之運算放大器之接 地則即使在1之增盈,此串擾信號會加倍(因為在此結構 中k正的輸入至輸出之增益=2)而產生無法接受之774^之 串擾。當添加更多的選擇器塊時,此串擾甚至會更大。類 似的串擾會由其他通道選擇器塊之出現在短路開關上之類 10似漣波所造成。為了允許在晶片上之金屬線路,其對於1μιη 覓之執道而3典型的為soohm/mm,這在沒有不經濟之過度 寬之金屬執之情況下,難以獲得小於數〇hm之來源阻抗。 第6與7圖之配置減少可能之串擾。第7圖顯示用於供應 麥考私壓乂!*之緩衝電路,此緩衝參考電壓%之運算放大器 15之開啟迴路(〇pen4ooP)輸出阻抗,通常藉由至節點X之大的 負回饋而被減少至幾分之幾歐姆。此在各阻抗鍵%中右手 短路開關或最靠近運算放大器35之短路開關37iii,是由具 有以私阻為yy代表電阻之晶片上(〇}此丨口)金屬連接而接線 回至節」:、έΧ其他的短路開關以及較佳此運算放大器(35)接 20地,疋經由獨立之金屬軌(具有由電阻器XX代表之電阻)而 連接回節點X。此等取決於信號之電流將通過短路開關以及 經由各別的連接電阻(χχ與yy)其可能為數十歐姆而至節點 X %此等屯阻會產生可觀之電壓漣波,但在節點X之電壓 將改艾更夕,因為其阻抗通常為例如O.lohm而非數十歐 20 1261414 姆。取決於:增益選擇器阻抗、串擾規格、通道數目、從 即點X至開關之距離,而須要從各別的通道選擇器獨立地連 接短路開關。 -置藉由將各別金屬軌星形地連接回適當之點(X) )而避免以上的串擾效應。此提供具有小於數歐姆輪出阻抗 之爹考緩衝器,而無須將此連接圍繞晶片,因為此種線路 連接須要非常寬的金屬(會使用大的面積 > 此右手邊短路開 關37111承載大部份電流(由於在運算放大器輸出之lVrms)。 就是此與金屬連接線路阻抗祕之電流,其對參考節歸 · 10生大部份之信號串擾。藉由將此連接各別連接回至去除耦 合之參考緩衝器輸出而降低此效應。此等其他開關通過少 疔多的電流,可以容忍較高阻抗之回歸通道而不會導致可 觀之串擾。 此外,取決於電路佈局與寄生方式,則亦為有利將在 · 15手側開關(37bi’等)之參考電壓端子獨立地連接回星形點 X。這是為了避免經由各電阻器鏈之“最左,,部份將事擾從一 未使用之通道輸人至另-未使用通道之輸人,如^不$ · 的選擇器塊中使用此等通道之第二者,則可察覺此串擾。 如同以上所討論之增益選擇器開關,由於%電壓是電 2 0源之一半,因此對於短路開關無須完整之c M 〇 s傳輪問,並 且-細MOS(或PMOS,但對於相同的尺寸丽〇8典型地會 有較低的電阻)就足夠。 在以上說明中所提及之半導體開關是假設為 MOSFET。此等兀件在標準之低成本廣泛供應之cm⑽矽積 21 1261414 體電路技術中可以獲得,其驅動簡單且提供低的漏電電 流。然而,可以使用其他的半導體裝置例如:加強型JFET 或MESFET。亦可在類似結構中使用雙載子電晶體。 以上所說明之選擇器較佳使用於音訊設備,例如:高 5 傳真放大器或汽車中之立體音響,用於從多個音訊輸入來 源選擇。然而,亦可考慮其他的應用,例如:多頻道精準 測量技術。 本發明亦提供一種操作選擇器之方法,以便選擇特定 的輸入來源,並且以便設定用於該來源之適當增益因數。 10 藉由將與此所選擇輸入來源無關之增益控制器上所有的開 關開啟(open)而選擇此來源。此外,藉由將與所選擇輸入來 源有關之增益控制器上適當之閘關閉(close),而將適當之 增益因數施加至所選擇之輸入來源。 本發明在以上是參考實施例說明。其修正與改變明顯 15 之用意為包括於其範圍中。 I:圖式簡單說明3 第1圖顯示用於類比選擇器之已知結構; 第2a-2d圖顯示CMOS裝置之概要圖式,以及用於此裝 置各種操作情況之導電特徵圖形; 20 第3圖顯示根據本發明實施例用於類比選擇器之結構; 第4圖顯示根據另一實施例之結構; 第5圖顯示第4圖結構之加強; 第6圖顯示第5圖結構之另一加強;以及 第7圖顯示用於參考電壓源之緩衝電路。 22 1261414 【圖式之主要元件代表符號表】 la、lb、lc...輸入來源 2a、2b、2c...電容器 lOi、10ii...選擇器 11a、lib、11c...輸入連接 12、14...裝置 12a、12b、12c...輸入開關 13.. .電阻器鏈 13a...l3n...電阻器 14a··. 14η小·.增益開關 15.. .運算放大器 201.. .選擇器 2011.. .選擇器 21a、21c.··節點 22b、22c...增益控制器 23ba…23kn···離散阻抗 23ca...23cn·.·電阻器 24ba· · .24cn·卜· ·開關 24ca...24cn小..增益開關 25.. .運算放大器 30i、30ii...選擇器 31a、31b、31c··.節點 32b、32c...增益控制器 33ba、34ba...阻抗鏈 33ca、33ca···阻抗 35…運算放大器 36b、36c...短路電路 37a-c...開關 37bi’、37i、37ii、37iii...開關 38a-c...連接 38i、38ii、38iii…節點 D...汲極 S...源極If such a single-supply design is used, the voltage at the selected input must also be Vdd/2, and the input signal le is typically a static ground voltage. Therefore, the capacitor 2c shown in Fig. 3 is required to connect this input signal. This DC bias on selected input 21c is then obtained via resistor chain 23c without any additional bias resistors. In a static manner, the voltage across the unselected input 21b via the respective resistor chain 23b or the like is also Vdd/2. Therefore, the capacitor 沘 will be pre-charged, and the capacitor does not need to be charged when the g input is switched from one channel to another. This helps avoid large voltage transients when switching inputs, which can be heard as undesired "waves" and "calls". (Note that the circuit in Figure 1 will require additional switches from unselected channels to Vdd/2, preferably via intermediate resistors (>lkohm) via 17 1261414 to avoid shorting and excessive loading of the input signal source) . Figure 4 shows another embodiment similar to the embodiment of Figure 3, but additionally having means 36 for shorting each impedance chain 33 to a reference voltage. For the sake of clarity, only some of the common elements of Figures 3 and 4 are shown in Figure 4. Figure 4 is a potential problem with the previous embodiment of Figure 3, with the output from the operational amplifier being connected to the unselected input via separate resistor strings. For example, if the resistor string has a total resistance of 60 kohm and the source impedance is 600 〇 11111, then the selected input signal will appear at node 21b and only attenuate 6 〇〇〇/60.6 ohms to 40 dB. If the signal is also used in parallel ramps via 2〇ii (for example, for recording-signals while listening to the other), the crosstalk is unacceptable. 15 20 The typical pressure of the pen pressure is also the reference to the middle of the pre-example and the power supply bias, while maintaining the correct static pressure on the unselected AC connection. The circuit is provided for each unused gain controller and is used for all unused inputs or all gain controls except the selected input-down lean controller (32C). Device. Each short circuit 6C includes a connection 38a-38c to a respective impedance chain 33a-33c, preferably a midpoint of 3^. Switches 37a-37c are connected to respective connections 38a-38c and are turned to * test (4). The short circuit switch's claws are at each short circuit / 36bU_(dGsed), which is connected to the "unselected" booster (32a, 32b). The evaluation of the daily control configuration between the various selectors 20i, 20Π input "channel" 18 1261414 the possibility of crosstalk. If the source/impedance associated with this unused impedance chain 33 is high, then the left end of this input connection or chain η will have a small signal amplitude superimposed thereon. This can be presented at the input of the other selector 20u of the selected source 1 thus causing crosstalk. This short circuit 36 operates to remove the signal superimposed on the unused chain 33 by short circuit to prevent crosstalk. If switch 37b has zero impedance, the output signal of the node bird will be fully suppressed by 70. However, a small switch will typically still have an impedance of, for example, 1 kohm at this node. If this point is along the midpoint of the 60k〇hm string and the source impedance is 600〇hm, then this signal will decay at 381;) by approximately 10 lkohm/30k〇hm = 30dB, and will return to 31b when it reaches 31b Attenuation 600 ohm / 30kohm = 34dB. This is still only a total of 64 commits. A further enhancement of the present embodiment is shown in FIG. 5, which shows the JJ 30 according to FIG. 4, but wherein the short-circuiting device 36 includes a plurality of short-circuiting switches 37i, J7ii, 37iii, which reduces the short-circuit circuits. 36's overall impedance while 15 further reduces crosstalk. For example, if the connections 38i, 38ii, 38iii have taps spaced 15kohm apart, the total attenuation is approximately: (lkohm/15kohm)*(lkohm/15kohm)*(lkohm/15kohm)*(600〇hm/15kohm)=23.5+ 23.5+23.5+28dB=98.5dB, which is the level that can be connected to 20, which can be compared with other unavoidable parasitic coupling effects between device pins or pCB connections. Figure 6 shows another enhanced version of this short circuit 36" which reduces crosstalk. This short circuit switch 37iii, which is closest to the operational amplifier output (V〇), consumes a considerable amount of current. For example, for an output like the above 1Vrms The audio signal 19 1961414 and the 15kohmi air resistance are segments, which will consume 66yArms signal current. If Vr comes; the original impedance is 1〇^, the resulting 66^vrms component will be applied to all Vr Terminals, including other switches and grounds in this and other selector stages. This signal is only 83dB lower than the K1Vrms level. If the signal selected by the selected 5 is applied to the ground of the op amp of the other selector block 20ii, even At 1% gain, this crosstalk signal is doubled (because k positive input-to-output gain = 2 in this structure) produces an unacceptable crosstalk of 774^. This crosstalk is added when more selector blocks are added. It will be even larger. Similar crosstalk can be caused by 10 like chopping of other channel selector blocks appearing on the shorting switch. To allow metal lines on the wafer, it is exemplified by 1μιη觅3 For so Ohm/mm, which is difficult to obtain source impedances less than a few hms without uneconomical over-extended metal. The configuration of Figures 6 and 7 reduces possible crosstalk. Figure 7 shows the supply for Mai Khao The buffer circuit of the private voltage *!*, the output impedance of the open loop (〇pen4ooP) of the op amp 15 with the buffer reference voltage %, is usually reduced to a few tens of ohms by the large negative feedback to the node X. The right-hand short-circuit switch or the short-circuit switch 37iii closest to the operational amplifier 35 in each of the impedance keys % is wired back to the node by a metal connection on the wafer having a private resistance of yy. The other short-circuiting switches and preferably the operational amplifier (35) are connected to ground 20, and are connected back to node X via a separate metal rail (having a resistor represented by resistor XX). These dependencies on the signal current will pass through the short-circuit switch and via separate connection resistors (χχ and yy) which may be tens of ohms to node X%. This resistance will produce considerable voltage ripple, but at node X. The voltage will change to Ai, because its impedance is usually, for example, O.lohm instead of tens of Euros 20 1261414. Depends on: gain selector impedance, crosstalk specification, number of channels, distance from point X to switch, and separate short-circuit switches from separate channel selectors. - The above crosstalk effect is avoided by star-connecting the individual metal rails back to the appropriate point (X). This provides a reference buffer with less than a few ohms of wheel-out impedance without having to surround the wafer because such a line connection requires a very wide metal (a large area will be used). This right-hand side short-circuit switch 37111 carries most of the The current (due to the lVrms at the op amp output) is the current that is connected to the metal line impedance, which is the most common signal crosstalk for the reference section. By connecting the connections back to the decoupling This effect is reduced by the reference buffer output. These other switches can tolerate higher impedance return channels without causing considerable crosstalk by less than a large amount of current. Also, depending on the circuit layout and parasitic mode, It is advantageous to connect the reference voltage terminals of the 15 hand side switch (37bi', etc.) back to the star point X independently. This is to avoid the "leftmost" part of each resistor chain, and the part will be disturbed from an unused one. The channel is input to another person who does not use the channel, such as the second one of the channels in the selector block of ^ not $ · can detect the crosstalk. As discussed above The selector switch, since the % voltage is one-half of the source of the electric 20, so there is no need for the complete c M 〇s pass-through for the short-circuit switch, and - fine MOS (or PMOS, but for the same size, the Radisson 8 typically has A lower resistance is sufficient. The semiconductor switches mentioned in the above description are assumed to be MOSFETs. These components are available in the standard low-cost wide-ranging cm(10) hoarding 21 1261414 bulk circuit technology, which is easy to drive. And a low leakage current is provided. However, other semiconductor devices such as a reinforced JFET or MESFET can be used. A dual carrier transistor can also be used in a similar structure. The selector described above is preferably used in an audio device, for example : High 5 fax amplifiers or stereos in cars for selecting from multiple audio input sources. However, other applications, such as multi-channel precision measurement techniques, may also be considered. The present invention also provides a method of operating a selector, In order to select a specific input source, and to set the appropriate gain factor for that source. 10 By using the input source selected here The source is selected by turning off all of the switches on the gain controller. In addition, the appropriate gain factor is applied by closing the appropriate gate on the gain controller associated with the selected input source. The selected input source. The present invention has been described above with reference to the embodiments, and the modifications and changes thereof are intended to be included in the scope. I: Simple description of the drawings 3 Figure 1 shows the known analog selectors. Structures; Figures 2a-2d show schematic diagrams of CMOS devices, and conductive features for various operational aspects of the device; 20 Figure 3 shows the structure for an analog selector in accordance with an embodiment of the present invention; The structure according to another embodiment; Fig. 5 shows the enhancement of the structure of Fig. 4; Fig. 6 shows another enhancement of the structure of Fig. 5; and Fig. 7 shows the snubber circuit for the reference voltage source. 22 1261414 [The main component representative symbol table of the drawing] la, lb, lc... input source 2a, 2b, 2c... capacitor lOi, 10ii... selector 11a, lib, 11c... input connection 12 , 14...devices 12a, 12b, 12c... input switch 13.. resistor chain 13a...l3n...resistor 14a··. 14η小·.Gain switch 15. Operational amplifier 201 .. Selector 2011.. Selector 21a, 21c.. Node 22b, 22c... Gain Controller 23ba...23kn···Discrete Impedance 23ca...23cn·.·Resistor 24ba·· .24cn · · · · · 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 ...impedance chain 33ca, 33ca... impedance 35... operational amplifiers 36b, 36c... short circuit 37a-c... switches 37bi', 37i, 37ii, 37iii... switches 38a-c... 38i, 38ii, 38iii...node D...bungee S...source

Gp ·. ·閘極Gp ·. · Gate

Gn...閘極Gn...gate

Vo...輸出電壓Vo...output voltage

Vr...參考電壓Vr... reference voltage

Vtp...臨界電壓 X...節點 XX、yy...電阻器Vtp...critical voltage X...node XX, yy...resistor

23twenty three

Claims (1)

4261414 i 拾、申請專利範圍: 第92126541號申請案申請專利範圍修正本 94〇721. 1· 一種用於從多個類比來源裝置中選擇其一之類比輸入 選擇器,其包括: 5 多個用於連接至該輸入來源之輸入連接,各該輸入 連接是連接至各增益控制H,其具有至少—個零以及一 或多個非-零增益,在該輸入連接與該增益控制器之間 沒有半導體開關。 2.如申料利範圍第1項之類比輸人選擇H,其中各該增 10 讀彻包括—❹辦導㈣關,其各具有源極或汲 極連接至與輸入信號無關之電壓。 .如申,專利乾圍第2項之類比輸入選擇器,其中此與輸 入信號無關之電壓是選擇器之接地。 15 如申%專利範圍第2項之類比輸入選擇器,更包括具有 矜入”輸出之-放大器,並且其中各該增益控制器更包 括阻抗鏈連接介於職A器輸出與各該輸人連接之 °玄開關;I:連接介於該鏈與該放大器輸入之間。 5·如申請專利範圍第3項之類比輸入選擇器,更包括具有 20 輸入14輸出之―放大器,並且其中各該增益控制器更包 括阻抗鏈連接介於該放大器輸出與各該輸入連接之 間’相關是連接介於該鏈與概A器輸入之間。 •如1請專利範圍第4項之類比輸入選擇器,更包括:當 在令&益日寸用於將各該增益控制器短路至參考電壓 裝置。 24 ^1414 K 請專利範圍第5項之類比輪八選擇器,更包括··當 在令增a時用於將各該增益控制 裝置。 器短路至參考電壓 >•如申請專利範圍第6項之類 ^ w . 、輪入廷擇器,其中該短路 ;置包括:連接介於該阻抗鍵與該參考電屡之間之開 9. =申請專利範圍第7項之類比輸入選擇器 衣置匕括.連接介於該®抗鏈與該參考 ’其中該短路 電壓之間之開 10 15 20 :申明專利範圍第8項之類比輪入選擇器,其中該放大 裔具有連接至該參考電塵之第二輸入。 U·:申請專利範圍第9項之類比輸入選擇器,其中該放大 杰具有連接至該參考電塵之第二輸入。 12·如申明專利範圍第8項之類比輸入選擇器,其中該短路 裝置更包括—或多個第二開關,其連接於該阻抗鏈與該 參考電壓之間。 13·如申請專利範圍第9項之類比輸入選擇器,其中該短路 裝置更包括一或多個第二開關,其連接於該阻抗鏈與該 參考電壓之間。 14.如申請專利範圍第1()項之類比輸人選擇器,其中該短路 '置更匕括或夕個弟二開關’其連接於該阻抗鍵與該 參考電壓之間。 1 5·如申#專利範圍第”項之類比輸人選擇器,其中該短路 裝置更包括一或多個第二開關,其連接於該阻抗鏈與該4261414 i Pickup, Patent Application Range: Application No. 92126541, Patent Application Revision No. 94〇721. 1· An analog input selector for selecting one of a plurality of analog source devices, including: 5 And an input connection connected to the input source, each of the input connections being connected to each gain control H having at least one zero and one or more non-zero gains, between the input connection and the gain controller Semiconductor switch. 2. If the input of the scope of the scope of the application is the same as the input selection H, each of which includes the connection (4), each of which has a source or a dipole connected to a voltage independent of the input signal. For example, the patent is the analog input selector of item 2, where the voltage unrelated to the input signal is the ground of the selector. 15 such as the input selector of the second patent scope of the patent, further includes an amplifier having an inductive output, and wherein each of the gain controllers further includes an impedance chain connection between the output of the device and the input of each of the inputs The I switch is between the chain and the input of the amplifier. 5. The analog selector of the third application of the patent scope includes an amplifier with 20 inputs and 14 outputs, and each of the gains The controller further includes an impedance chain connection between the output of the amplifier and each of the input connections. The correlation is between the chain and the input of the device. • For example, the input selector of the fourth item of the patent scope is It also includes: when the &Yi Ri inch is used to short-circuit each of the gain controllers to the reference voltage device. 24 ^1414 K Please refer to the fifth round of the patent range, such as the eight-wheel selector, including a is used to short-circuit each of the gain control devices to the reference voltage > • as in the scope of claim 6 or the like, the round-in device, wherein the short circuit; the inclusion includes: the connection between the impedance keys With the reference Between the opening of the 9. = patent application scope category 7 analog input selector clothing. The connection between the ® anti-chain and the reference 'the short-circuit voltage between the opening 10 15 20: the scope of the patent An analog input of the eighth item, wherein the magnified person has a second input connected to the reference electric dust. U: an analog input selector of claim 9 wherein the magnifying jesus has a connection to the reference The second input of the electric dust. 12. The analog input selector of claim 8 wherein the shorting device further comprises - or a plurality of second switches connected between the impedance chain and the reference voltage. An analog input selector as in claim 9 wherein the shorting device further comprises one or more second switches connected between the impedance chain and the reference voltage. The analog input selector, wherein the short circuit 'sets more or eves the second switch' is connected between the impedance key and the reference voltage. 1 5 · 申# patent scope of the first item Input selector, where the short circuit The device further includes one or more second switches connected to the impedance chain and the 25 J261414 參考電壓之間。 16. 如申請專利範圍第 1卜12、13、14或15項之類比輸入選擇器,其中該等開 關為]ViOSFET。 17. 如申請專利範圍第卜2、3、4、5、6、7、8、9、1〇、 1卜12、13、14或15項之類比輪入選擇器,其中該等開 關為NMOS開關。 18. -種積體電路,其包括如申請專利範圍第卜2、3、4、 5、6、7、8、9、10、u、12、13 10 15 20 入選擇器。 啤15項之類比輪 19. -種音訊設備’其包括如中請專利範圍㈣項之積體電 路0 2〇· 一種積體電路’其包含如申請專利範圍第8、9、10、u、 12 13、14或15項之類比輸入選 一 更包括、擇卩。,其中該短路裝置 更匕括·竭他短路裝置開關獨立無 考電壓之-第三開關。 21· —種音訊設備,其包括如申一 路。 ^專利範圍弟2〇項之積體電 泣如申請專利範圍第18項之積體電路, M0SFET。 寺開關為 關為 23·如申請專利範圍第2〇項之積 — MOSFET。 /、中6亥等開 24·如申請專利範圍第18項之 關為 、月且咆路,其中兮莖P弓 NM0S開關。 τ為寺開 26 ς/(Ι2614ί14 \ 25. 如申請專利範圍第20項之積體電路,其中該等開關為 NMOS開關。 26. —種操作選擇器之方法,此選擇器包括用於連接至多個 輸入來源之多個輸入連接,各該輸入連接是連接至具有 5 至少一個零以及一或多個非零增益之一個別增益控制 器,此方法包括以下步驟:25 J261414 Between reference voltages. 16. For example, the analog input selector of the patent scope No. 1, 12, 13, 14 or 15, wherein the switches are] ViOSFET. 17. For example, if the patent application is in the range of 2, 3, 4, 5, 6, 7, 8, 9, 1 , 1 , 12, 13, 14 or 15, the switch is an NMOS. switch. 18. An integrator circuit comprising a selector as in the patent claims 2, 3, 4, 5, 6, 7, 8, 9, 10, u, 12, 13 10 15 20 . Beer 15 or similar wheel 19. An audio equipment 'which includes the integrated circuit of the patent range (4) of the patent application 0 2〇 · an integrated circuit 'which includes the patent scopes 8, 9, 10, u, 12, 13 or 14 or more than the input of one to include, choose. Among them, the short-circuiting device further includes the third switch which is independent of the voltage of the short-circuiting device switch. 21·—A kind of audio equipment, including, for example, Shen Yilu. ^ Patent scope of the 2nd item of the integrated body sobbing as in the application of the scope of the 18th element of the integrated circuit, M0SFET. The temple switch is closed as 23. MOSFET as the second item of the patent application scope. /, Zhong 6 Hai and so on 24 · If the scope of the application for the scope of the 18th is the month, and the road, which is the stem P bow NM0S switch. τ is the temple open 26 ς / (Ι 2614ί14 \ 25. Such as the application of the scope of the 20th integrated circuit, wherein the switch is an NMOS switch. 26. A method of operating the selector, this selector includes for connecting at most Multiple input connections of input sources, each of which is connected to an individual gain controller having 5 at least one zero and one or more non-zero gains, the method comprising the steps of: 藉由提供該個別增益控制器具有非零增益、以及藉 由提供其他增益控制器各具有零增益,而選擇一個輸入 來源。An input source is selected by providing the individual gain controller with a non-zero gain and by providing other gain controllers each having a zero gain. 2727
TW92126541A 2002-11-26 2003-09-25 An improved analogue selector TWI261414B (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
GB0227553A GB2395849B (en) 2002-11-26 2002-11-26 Improved analogue selector
US10/347,300 US6828845B2 (en) 2002-11-26 2003-01-21 Analogue selector

Publications (2)

Publication Number Publication Date
TW200419904A TW200419904A (en) 2004-10-01
TWI261414B true TWI261414B (en) 2006-09-01

Family

ID=34276801

Family Applications (1)

Application Number Title Priority Date Filing Date
TW92126541A TWI261414B (en) 2002-11-26 2003-09-25 An improved analogue selector

Country Status (2)

Country Link
CN (1) CN100544499C (en)
TW (1) TWI261414B (en)

Also Published As

Publication number Publication date
CN1503598A (en) 2004-06-09
TW200419904A (en) 2004-10-01
CN100544499C (en) 2009-09-23

Similar Documents

Publication Publication Date Title
CN1671040B (en) Low noise operational amplifier
US9190994B2 (en) RF switch branch having improved linearity
US7928794B2 (en) Method and apparatus for a dynamically self-bootstrapped switch
US7675380B2 (en) Integrated digitally controlled linear-in-decibels attenuator
US4683386A (en) Electronic attenuation value control circuit in which switching noise is suppressed
Nagari et al. An 8$\Omega $2.5 W 1%-THD 104 dB (A)-Dynamic-Range Class-D Audio Amplifier With Ultra-Low EMI System and Current Sensing for Speaker Protection
TW200428773A (en) Boost circuit
EP2460271A1 (en) Switches with bias resistors for even voltage distribution
JP2013501410A (en) Switch with improved bias
CN102595279B (en) Circuit with three-stage of power-on sequence used for suppressing the pop noise in audio system
KR20040048988A (en) Operational amplifier with chopped input transistor pair
EP0037406A1 (en) Cmos operational amplifier with reduced power dissipation.
JP2006345190A (en) Distribution type amplifier
US20060006941A1 (en) Class D-amplifier
EP3703258A1 (en) Headset
US6828845B2 (en) Analogue selector
US7068107B2 (en) Variable gain amplifier
US20040239417A1 (en) Amplifiers
US20020060605A1 (en) Amplifiers
US20050206454A1 (en) Low noise op amp
TWI261414B (en) An improved analogue selector
TWI745742B (en) Digital to analog converter
US6380644B1 (en) Switching circuitry providing improved signal performance at high frequencies and method of operation thereof
TW200412025A (en) Circuit apparatus operable under high voltage
CN110771042B (en) Controlling switches in variable impedance elements

Legal Events

Date Code Title Description
MM4A Annulment or lapse of patent due to non-payment of fees