(19)
(11)EP 2 858 246 B1

(12)EUROPEAN PATENT SPECIFICATION

(45)Mention of the grant of the patent:
11.09.2019 Bulletin 2019/37

(21)Application number: 13368037.1

(22)Date of filing:  07.10.2013
(51)Int. Cl.: 
H03K 17/14  (2006.01)
H03K 17/16  (2006.01)

(54)

Capacitance minimization switch

Kapazitätsminimalisierungsschalter

Commutateur de minimisation de capacitance


(84)Designated Contracting States:
AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR

(43)Date of publication of application:
08.04.2015 Bulletin 2015/15

(73)Proprietor: Dialog Semiconductor GmbH
73230 Kirchheim/Teck-Nabern (DE)

(72)Inventor:
  • Morris, Tim
    79539 Loerrach (DE)

(74)Representative: Schuffenecker, Thierry 
120 Chemin de la Maure
06800 Cagnes sur Mer
06800 Cagnes sur Mer (FR)


(56)References cited: : 
US-A1- 2002 036 576
  
      
    Note: Within nine months from the publication of the mention of the grant of the European patent, any person may give notice to the European Patent Office of opposition to the European patent granted. Notice of opposition shall be filed in a written reasoned statement. It shall not be deemed to have been filed until the opposition fee has been paid. (Art. 99(1) European Patent Convention).


    Description

    Technical Field.



    [0001] The present disclosure is related to integrated circuit switches, and in particular with switches designed to minimize effects of parasitic capacitance.

    Background Art



    [0002] Semiconductor switches are utilized in electronic circuits to control when a signal is allowed to propagate from one node to another. An implementation of such a switch in CMOS technology is the transmission gate (examples shown in FIG. 1A and 1B), comprising NMOS and PMOS transistors connected in parallel for the purpose of allowing good conductivity for a wide signal voltage range. There are a number of different ways in which combinations of transistors can be connected together to implement a switch function; however, they all perform the same basic function of connecting two nodes when "on", and isolating the two nodes when "off". For DC signals inside an intended operating range, a switch appears like a resistor, the impedance depending on the architecture used and the dimensions of the transistors. There can also be dependence on the DC level of the signal for most architectures.

    [0003] In FIG. 1A is shown a data switch of prior art comprising an N-channel transistor 10 and a P-channel transistor 11 connected in parallel, where the sources of the two transistors are connected together to form a data input A and the drains of the two transistors are connected to form data output B. An ON signal, a high logic signal, is applied to the gate of the N-channel transistor 10 and through an inverter circuit 12 to the gate of the P-channel transistor 11. To turn off the switch an off signal, which is a low logic signal, is applied to the gate of the N-channel transistor and through the inverter circuit to the gate of the P-channel transistor.

    [0004] In FIG. 1B is a second data switch of prior art where a second parallel connected pair of transistors, an N-channel transistor 15 and a P-channel transistor 16, are connected in tandem 13 with a first pair of transistors 10 and 11. An N-channel transistor 14 is used to ground the tandem connection 13 to insure isolation between A and B.

    [0005] A semiconductor switch presents a resistance to AC signals inside the intended operating range, but parasitic capacitance is also present. The impedance of the parasitic capacitive will present an AC path to the power supplies and chip substrate, and will ultimately cause signal attenuation at higher frequencies. In addition there will be capacitive impedance between the two sides of the switch that will cause some signal propagation when in the "off" state, thus reducing the isolation of the switch against higher frequency signals. A disadvantage of prior-art switch design shown in FIG. 1A is that lower resistive impedance in the "on" state dictates larger transistors. Larger transistors inherently include larger unwanted capacitances; therefore strong switches with low on-resistance also have larger parasitic capacitances, which limit the maximum frequency of a switched signal.

    [0006] The purpose of the switches in FIG. 1A and 1B is to make the voltage at node B equal to the voltage at node A while in the "on" state. Any change in the signal voltage at nodes A and B will require the total sum of all parasitic capacitances to be charged, or discharged, by a charge, ΔQ, equal to the sum of capacitances multiplied by the change in voltage, ΔQ=C·ΔV. Thus the faster the signal frequency, the higher the current required to charge and discharge the parasitic capacitances, i=C·dv/dt.

    [0007] US 2006/0164164 A1 (Rogers et al.) is directed to compensating for parasitic capacitances in a VCA circuit. US 8,400,848 B2 (Fujimura) is directed to a bit line capacitance compensation capacitor, which compensates the capacitance of the bit line, and a peripheral capacitance compensation capacitor, which compensates the peripheral capacitance of the bit line. In US 8,344,808 B2 (Samavedamet al.) a capacitance compensation is directed to using a compensation device coupled to a gain device. US 8,344,802 B2 (Huang et al.) is directed to an operational amplifier device capable of using transmission gates for capacitance compensation. US 6,462,611 B2 (Shigehara et al.) is directed to body effect compensation for electronic switches connected between two terminals. In US 6,150,884 (Fattaruso) an improved operational amplifier circuit with nested transconductance is directed to capacitance compensation.

    [0008] US 2002/0036576 discloses an analog switch and an analog multiplexer which are realized by which electron charges which have been stored in a stray capacitance provided on the output side thereof before a switch is conducted do not give an adverse influence to a level of such an analog input voltage which is subsequently entered after the switch has been switched. An analog switch circuit is arranged by insulating gate type transistors and a voltage follower which is parallel-connected to these insulating type transistors. When the analog switch circuit is turned ON, the voltage follower is firstly brought into an active state, and thereafter, these insulating gate type transistors are brought into conductive conditions.

    Summary of the invention



    [0009] It is objectives of the present disclosure is to sense current diverted from a CMOS transmission gate and inject that current into the input of the transmission gate.

    [0010] It is also an objective of the present disclosure to sense a voltage across a resistor in an RC network to detect parasitic capacitance current and allow output current of the transconductance amplifier to be connected in-phase to the input of the CMOS transmission gate.

    [0011] It is further an objective of the present disclosure to detect gate current of the N-channel and P-channel transistors of the transmission gate and apply that current to the input of the CMOS transmission gate.

    [0012] Common to the prior art is the use of at least one transistor of any type, which can be controlled by a stimulus to make it alternately low resistive (in "on" state) and high resistive (in "off' state). There is also the presence of a stimulus that controls when the transistor(s) is in the "on" or "off' state and the existence of two electrically separate nodes (A & B shown in FIG. 1A, 1B, and 2A), which are connected to either side of the transistor (or network of transistors).
    The present disclosure would not work without the use of some element that senses the rate of change of voltage of the signal being propagated by the switch, (dv/dt). The invention also would not work without the use of some element or device that scales dv/dt with the known parasitic capacitances from the transistor(s), which results in a voltage or current that is proportional to the current that flows through the parasitic capacitances according to the equation, i=C·dv/dt. In the embodiment of FIG. 3a, dv/dt is sensed by a simple RC network and an operational transconductance amplifier (OTA). The measured dv/dt is scaled by the gm of the OTA. In the exemplary circuit of FIG. 3b, these required sensing and scaling functions are performed in one step by measuring the current that flows through the parasitic gate capacitance towards V+ (for the NMOS in the example) or V-(for the PMOS in the example). The current is proportional to both the capacitance and dv/dt.

    [0013] The present disclosure also would not work without converting the measurement of the current in the parasitic capacitances into another current, which is sourced from the local power supply(ies) and is fed into the signal at node A or B, or both A and B, or at any node in-between which has equal electrical potential in the "on " state. In order to reduce the effective capacitance of the switch, the current that is fed back into the system must be of opposite polarity to the current flowing into the system through the parasitic devices. For example, raising the voltage of node A in FIG. 3a requires that current must flow from nodes A and B into the output of the inverter via the gate capacitance of the PMOS transistor, therefore the OTA must deliver some current back into node A.

    Description of the drawings



    [0014] This invention will be described with reference to the accompanying drawings, wherein:

    FIG. 1A is a circuit diagram of a CMOS transmission gate of the prior art;

    FIG. 1B is a circuit diagram of a tandem CMOS transmission gate of the prior art;

    FIG. 2A is a circuit diagram of parasitic capacitance of a CMOS transmission gate of the present disclosure;

    FIG. 2B is an equivalent circuit of the CMOS transmission gate of the present disclosure;

    FIG. 3A is a circuit diagram of a transconductance amplifier detecting parasitic capacitance current to be applied to the input of a CMOS transmission gate;

    FIG. 3B is a circuit diagram of an exemplary circuit of current amplifiers connected to a CMOS transmission gate to detect and feed back parasitic capacitance current; and

    FIG. 3C is a circuit diagram of an exemplary circuit of current amplifiers connected to a tandem CMOS transmission gate to detect and feed back parasitic capacitance current.


    Description of the preferred embodiments



    [0015] The most significant parasitic capacitances present in the simple transmission gate of FIG. 1a are illustrated in FIG. 2A and reduced to a simple model in FIG. 2B. In FIG. 2A is shown the main parasitic capacitances 20 associated with N-channel transistor and 21 associated with the P-channel transistor of a CMOS transmission gate, which affect signals that are switched from A to B. These capacitances include input capacitance from A to the gate of each transistor and from A to the substrate of each capacitor, and output capacitance from B to the gate of each transistor and B to the substrate of each transistor. FIG. 2B shows an equivalent circuit of the switch of FIG. 2A in which the two transistors of the switch are represented by a parallel resistance Rp to signals flowing from A to B, where Rp is a high value when the switch is biased off with a low logic signal, and Rp is a low resistance when the switch is biased on with a high logic signal. The input capacitance Cps is the equivalent of the parallel parasitic capacitance of the two transistors at point A and the output capacitance Cpd is the equivalent of the parallel parasitic capacitance of the two transistors at point B.

    [0016] In FIG. 3A a CMOS transmission gate 31 of the present disclosure is shown. The rate of change of an input signal voltage is sensed through a dedicated RC network. Note that the sense capacitor, C, will add to the parasitic capacitance of the switch. An OTA (operational transconductance amplifier) 30 measures a current, iRC, by measuring the voltage drop, VR, across the resistor R of the RC network. The output of the OTC is applied back to the input of the CMOS transmission gate. The gain, gm, of the OTA (operational transconductance amplifier) must be correctly set to account for the values of sense resistor, R, sense capacitor, C, and total parasitic capacitance, Cpara (which includes sense capacitance, C).

    [0017] Sensed is dv/dt = iRC / C = VR / (R·C) where iRC = VR / R and dv/dt = VR / (R·C), and forced is the current iFB < Cpara·dv/dt where iFB is less than lost current otherwise instability can occur, wherein dv/dt = VR / (R·C); iFB < (Cpara·\/R) / (R·C) > gm·VR ; iFB = gm·VR and.·. gm·VR < (Cpara·VR) / (R·C). Therefore gm < Cpara / (R·C) This method is suitable as long as the parasitic capacitances are known. Positive feedback means instability will occur if loop gain is greater than or equal to one. It also relies on the RC time constant of the sense components being significantly shorter than the rise/fall times of the switching signal.

    [0018] In FIG. 3B the parasitic capacitance does not need to be known because it is sensed directly by current amplifiers 32 and 33. A first input to current amplifier 32 is connected to the gate of the N-channel transistor of the CMOS transmission gate 31 and a first input to current amplifier 33 is connected to the gate of the P-channel transistor. The second input to current amplifier 32 is connected to the CMOS transmission gate on/off input and the second input to current amplifier 33 is connected to the output of the inverter circuit 34, which is connected to the transmission gate on/off input. The most dominate current that is "lost" through the parasitic capacitance surrounding the transistor gates is sensed directly and mirrored back into the signal path. For this reason, the size of the transistors is unimportant and the same compensation circuit can be applied to any size of switch, providing that the maximum current, accuracy, and speed are suitable.

    [0019] The gain of the two current amplifiers 32 and 33 is set to be less than unity to insure circuit stability; therefore, the amount of compensation of the effects of the parasitic capacitance is somewhat less than one hundred percent. However, most of the loss through the parasitic capacitance is compensated.

    [0020] In FIG. 3C the same concept as in FIG. 3B is applied to a transistor switch architecture, which provides better isolation in the "off" state. There are two CMOS
    transmission switches 36 and 37 connected in tandem. An N-channel transistor 39 is used to ground the tandem connection to insure isolation between input A and output B of the tandem switch. The gates of the two N-channel transistors are connected together and connected to a first input of current amplifier 32. The second input to the current amplifier 32 is connected to on/off control input of the tandem transistor switch. The output of amplifier 32 is connected to input A of the tandem transistor switch. The gates of the two P-channel transistors are connected together and connected to a first input to current amplifier 33. The second input to current amplifier 33 is connected to the output of the inverter circuit 34, and the output of current amplifier 33 is connected to input A of the tandem transistor switch. The gain of the current amplifiers 32 and 33 is set to be less than one (1) to maintain stability of the circuitry since the feedback that is accomplished is positive feedback.

    [0021] Current flowing into, or out of, the gates of the transistors that make up the CMOS transmission switches comprises current required by the gates to operate and capacitive current required to charge (discharge) the parasitic capacitance that surrounds the transistors than make up the CMOS transmission switch as shown in FIG. 2A. Therefore, since the signal connected to input A is less than unity, the correction for the effects of the parasitic capacitance is less than total, but most of the effects of the parasitic capacitance is compensated for.

    [0022] When implicitly sensing the current lost through switch parasitic capacitance any type of capacitor might be suitable for emulating the parasitic capacitance before scaling (capacitor C in FIG. 3A), however the most suitable type of capacitor would probably be a MOSCAP (capacitor formed across the gate-oxide of a MOS transistor), since the MOSCAP could be matched to the dimension, orientation, and electrical characteristics of the switch transistor. Such matching would lead to a roughly constant relationship between the sensed current and the total lost current.

    [0023] When explicitly sensing the current lost through the parasitic capacitance of the switch, no additional sensing capacitor is required, although additional capacitors could be utilized to further emulate parasitic capacitances other than the gate-drain and gate-source capacitances. However, to some extent the additional parasitic capacitance present will be proportional to the gate-source and gate-drain capacitances and so increasing the ratio of fed-back current to sensed current would be sufficient.
    In all sensing schemes (implicit and explicit sensing) it is important that the ratio of total fed-back current to total lost current (∑iFB/∑iLOST) does not exceed the value of one (1). It should also be clear that ∑iFB/∑iLOST should not equal the value of one (1) since any mismatch or variation could then cause ∑iFB/∑iLOST to exceed one. If ∑iFB/∑iLOST exceeds the value of one then the circuit wants to inject more current into the signal than would be lost through the parasitic capacitance. This means that any signal change, including that caused by noise falling within the bandwidth of the circuit, could cause the circuit to push exponentially more current into or out of the signal until the circuit saturates. It would then be possible that the circuit overpower the original source of the signal.

    [0024] One possible application for the circuit concept described is for a switch to optionally pass or block USB data signals, depending on whether the switch is chosen to be "on" or "off'. USB data signals (transmitted on two separate wires) should have a differential amplitude between 200mV - 400mV, switching at speeds from 1.5Mb/s (low-speed), to 12Mb/s (full-speed), 480Mb/s (high-speed), and 5Gb/s (Super-Speed). Data lines are terminated by impedances from 18Ω to 45Ω and the data integrity relies on the ability of the data-line to pass signals from the transmitter to the receiver without large overshoots or attenuation. The ideal data line would have zero capacitance and zero impedance, allowing the transmitter and receiver to completely define the signal characteristics as defined in the USB standard specifications. But when multiplexing is required to facilitate the sharing of 2 pins for multiple functions, one of these being USB data communication, then inevitably a switch will be required in the data-line which will add additional capacitance and impedance. To get as close to the ideal data-line as possible, the switch should be low impedance (meaning it will be large), and should introduce minimum capacitance. Because larger switches introduce larger capacitance, the capacitance minimization concept described in this invention can be used.

    [0025] While the invention has been particularly shown and described with reference to preferred embodiments thereof, it will be understood by those skilled in the art that various changes in form and details may be made without departing from the scope of the invention, as defined by the appended claims.


    Claims

    1. A capacitance minimization circuit, comprising:

    a) a CMOS transmission gate (31) comprising signal loss caused by parasitic capacitance; and

    b) an amplifier circuit (30), connected to the input of the CMOS transmission gate, configured to sense a rate of change of an input signal voltage of the CMOS transmission gate in order to detect said signal loss caused by the parasitic capacitances and to feed back a current according to the signal loss caused by the parasitic capacitance to the input of the CMOS transmission gate.


     
    2. The circuit of claim 1, wherein said amplifier circuit (30) is an operational transconductance amplifier (30) configured to sense the rate of change of the input signal voltage via a voltage drop across a resistor (R) of an RC network (R,C), wherein the RC network is connected between the input of the CMOS transmission gate (31) and ground, wherein the operational transconductance amplifier (30) is further configured to provide an output current to the input of the CMOS transmission gate to compensate the signal loss due to the parasitic capacitances.
     
    3. The circuit of claim 2, wherein said operational transconductance amplifier (30) forms a part of a positive feedback loop, wherein the loop gain is set to be less than one.
     
    4. The circuit of claim 2, wherein said transconductance amplifier (30) has a gain set to be less than one.
     
    5. A method of capacitance minimization, comprising the steps of:

    a) forming a CMOS transmission gate (31) with an N-channel transistor connected in parallel with a P-channel transistor between an input and an output; and

    b) connecting an amplifier (30) to the input of the CMOS transmission gate in order to sense a rate of change of an input signal voltage of the CMOS transmission gate indicating a signal loss caused by the parasitic capacitances of the CMOS transmission gate and to feed back a current according to the signal loss caused by the parasitic capacitances of the CMOS transmission gate to an input of the CMOS transmission gate in order to compensate the signal loss.


     
    6. The method of claim 5, wherein the amplifier is a transconductance amplifier (30) capable of sensing a current containing a portion of the current of the parasitic capacitance which can be amplified and applied to the input of the CMOS transmission gate to compensate for parasitic capacitance current loss.
     
    7. The method of claim 5, wherein said operational transconductance amplifier (30) as a gain of less than 1.
     


    Ansprüche

    1. Kapazitätsminimierungsschaltung, umfassend:

    a) ein CMOS-Sendegate (31) mit Signalverlust, der durch eine parasitäre Kapazität verursacht wird; und

    b) eine Verstärkerschaltung (30), die mit dem Eingang des CMOS-Sendegates verbunden ist und ausgelegt ist, um eine Änderungsrate einer Eingangssignalspannung des CMOS-Sendegates zu erfassen, um den durch die parasitären Kapazitäten verursachten Signalverlust zu erfassen und einen Strom entsprechend dem durch die parasitäre Kapazität verursachten Signalverlust an den Eingang des CMOS-Sendegates rückzukoppeln.


     
    2. Schaltung nach Anspruch 1, wobei die Verstärkerschaltung (30) ein operativer Transkonduktanz-Verstärker (30) ist, der ausgelegt ist, um die Änderungsrate der Eingangssignalspannung über einen Spannungsabfall über einen Widerstand (R) eines RC-Netzwerks (R,C) zu erfassen, wobei das RC-Netzwerk zwischen den Eingang des CMOS-Sendegates (31) und der Masse geschaltet ist, wobei der operative Transkonduktanz-Verstärker (30) ferner ausgelegt ist, um dem Eingang des CMOS-Sendegates einen Ausgangsstrom zum Kompensieren des Signalverlustes aufgrund der parasitären Kapazitäten bereitzustellen.
     
    3. Schaltung nach Anspruch 2, wobei der operative Transkonduktanz-Verstärker (30) einen Teil einer positiven Rückkopplungsschleife bildet, wobei die Schleifenverstärkung auf weniger als eins eingestellt ist.
     
    4. Schaltung nach Anspruch 2, wobei der Transkonduktanz-Verstärker (30) eine Verstärkung aufweist, die auf weniger als eins eingestellt ist.
     
    5. Verfahren zur Kapazitätsminimierung, umfassend die Schritte:

    a) Bilden eines CMOS-Sendegates (31) mit einem N-Kanal-Transistor, der parallel zu einem P-Kanal-Transistor zwischen einem Eingang und einem Ausgang geschaltet ist; und

    b) Verbinden eines Verstärkers (30) mit dem Eingang des CMOS-Sendegates, um eine Änderungsrate einer Eingangssignalspannung des CMOS-Sendegates zu erfassen, die einen Signalverlust anzeigt, der durch die parasitären Kapazitäten des CMOS-Sendegates verursacht wird, und um einen Strom entsprechend dem Signalverlust, der durch die parasitären Kapazitäten des CMOS-Sendegates verursacht wird, an einen Eingang des CMOS-Sendegates rückzukoppeln, um den Signalverlust zu kompensieren.


     
    6. Verfahren nach Anspruch 5, wobei der Verstärker ein Transkonduktanz-Verstärker (30) ist, der in der Lage ist, einen Strom zu erfassen, der einen Teil des Stroms der parasitären Kapazität enthält, der verstärkt und an den Eingang des CMOS-Sendegates angelegt werden kann, um den Stromverlust durch parasitäre Kapazitäten zu kompensieren.
     
    7. Verfahren nach Anspruch 5, wobei der operative Transkonduktanz-Verstärker (30) eine Verstärkung von weniger als eins aufweist.
     


    Revendications

    1. Un circuit de minimisation de capacitance, comprenant :

    a) une grille de transmission CMOS (31) comprenant une perte de signal provoquée par une capacité parasite ; et

    b) un circuit d'amplification (30), connecté à l'entrée de la grille de transmission CMOS, configuré pour détecter un taux de changement d'une tension de signal d'entrée de la grille de transmission CMOS afin de détecter la perte de signal provoquée par les capacités parasites et de renvoyer un courant en fonction de la perte de signal provoquée par la capacité parasite à l'entrée de la grille de transmission CMOS.


     
    2. Le circuit de la revendication 1, dans lequel ledit circuit d'amplification (30) est un amplificateur à transconductance opération (30) configuré pour détecter le taux de changement de la tension de signal d'entrée via une chute de tension au travers une résistance (R) d'un réseau RC (R, C), dans lequel le réseau RC est connecté entre l'entrée de la grille de transmission CMOS (31) et la terre, dans lequel l'amplificateur à transconductance opérationnel (30) est en outre configuré pour fournir un courant de sortie à l'entrée de la grille de transmission CMOS pour compenser la perte de signal due aux capacités parasites.
     
    3. Le circuit de la revendication 2, dans lequel ledit amplificateur à transconductance opérationnel (30) forme une partie d'une boucle de rétroaction positive, dans laquelle le gain de boucle est fixé pour être inférieur à 1.
     
    4. Le circuit de la revendication 2, dans lequel ledit amplificateur à transconductance (30) a une gain inférieur à 1.
     
    5. Un procédé de minimisation de capacitance, comprenant les étapes consistant à :

    a) former une grille de transmission CMOS (31) avec un transistor à canal N connecté en parallèle avec un transistor à canal P entre une entrée et une sortie ; et

    b) connecter un amplificateur (30) à l'entrée de la grille de transmission CMOS afin de détecter un taux de changement d'une tension de signal d'entrée de la grille de transmission CMOS indiquant une perte de signal provoquée par les capacités parasites de la grille de transmission CMOS et de renvoyer en retour un courant à l'entrée de la grille de transmission CMOS selon la perte de signal provoquée par les capacités parasites de la grille de transmission CMOS afin de compenser la perte de signal.


     
    6. Le procédé de la revendication 5, dans lequel l'amplificateur est un amplificateur à transconductance (30) capable de détecter un courant contenant une portion de courant de la capacité parasitique qui peut être amplifiée et appliquée à l'entrée de la grille de transmission CMOS pour compenser la perte de courant due aux capacités parasites.
     
    7. Le procédé de la revendication 5, dans lequel ledit amplificateur à transconductance opérationnel (30) a un gain inférieur à 1.
     




    Drawing












    REFERENCES CITED IN THE DESCRIPTION



    This list of references cited by the applicant is for the reader's convenience only. It does not form part of the European patent document. Even though great care has been taken in compiling the references, errors or omissions cannot be excluded and the EPO disclaims all liability in this regard.

    Patent documents cited in the description