The invention broadly relates to a source follower attenuator circuit, to a differential source follower attenuator, and to a method for attenuating a signal.
An attenuator is an electronic device or circuit that reduces the amplitude or power of a signal without appreciably distorting the signal's waveform. Conventional attenuators are passive devices made from resistors. The degree of attenuation may be fixed, continuously adjustable, or incrementally adjustable.
An existing attenuator circuit utilises a switched resistive network, termed as ladder resistive network attenuator. An example of the attenuator is shown in
These disadvantages make the resistive attenuator not suitable for integrated circuit implementation. Hence, in integrated circuits, common source based attenuators are adopted for either fixed or variable attenuators.
An existing common source based source degeneration amplifier (attenuator) is illustrated in
Here, gm is the transconductance of an input transistor and G is the attenuation degree. It can be understood from equation 1 that accuracy of attenuation of the attenuator (200) is affected by the transconductance gm. The value of gm needs to be much larger than the Rdegen to reach acceptable linearity accuracy for Rload over Rdegen, which means that large current consumption is inevitable to boost up the gm. However, even with the large current, the accuracy of attenuation cannot be improved in CMOS based circuits because room for gm improvement is very limited for CMOS based circuits. The situation becomes worse in deep submicron process CMOS, where output impedance of a transistor is small which further reduces attenuation accuracy. As a result, the circuit (200) of
A need therefore exists to provide an attenuator that seeks to address at least one of the above problems.
According to a first aspect of the present invention, there is provided a source follower attenuator circuit comprising a current source; an input transistor with a source connected to the current source and a drain connected to ground; a control transistor with a source connected to the current source and a drain connected to ground; wherein an attenuated output signal across the source and drain of the control transistor is controlled by transconductances, in an on-state, of the input transistor and of the control transistor respectively.
The attenuator circuit may further comprise a RC low-pass filter circuit connected between the gate of the input transistor and the gate of the control transistor for self-biasing the control transistor from an input voltage provided at the gate of the input transistor.
The RC low-pass filter circuit may comprise MOS components.
The attenuator circuit may further comprise a plurality of control transistors, each control transistor with a source connected to the current source and a drain connected to ground and with gates of the respective control transistors are connected in series to a biasing voltage, wherein the attenuated output signal across the source and drain of one of the control transistors is controlled by transconductances, in an on-state, of the input transistor and all of the plurality of control transistors respectively.
The attenuator circuit may further comprise a plurality of control transistors, each control transistor with a source connected to the current source and a drain connected to ground and with gates of the respective control transistors are connected in parallel to a biasing voltage, wherein the attenuated output signal across the source and drain of one of the control transistors is variably controlled by the transconductance, in an on-state, of the input transistor and by transconductances of those of the plurality of control transistors which are in an on state respectively.
The attenuator circuit may further comprise switches for controlling the on state of the respective control transistors.
Each of the switches may comprise two switch elements for preventing signals from feeding through the switches.
According to a second aspect of the present invention, there is provided a differential source follower attenuator comprising a first and a second attenuator circuits as claimed in any one of claims 1 to 5; wherein the current source of the first attenuator circuit comprises a positive current source, and the current source of the second attenuator circuit comprises a negative current source.
The source of each control transistor of the first attenuator circuit may be connected to the negative current source of the second attenuator circuit, and the source of each control transistor of the second attenuator circuit is connected to the positive current source of the first attenuator circuit, for DC offset through cross-coupling of the first and second attenuator circuits.
According to a third aspect of the present invention, there is provided a method for attenuating a signal comprising providing a current source; providing an input transistor with a source connected to the current source and a drain connected to ground; providing a control transistor with a source connected to the current source and a drain connected to ground; obtaining an attenuated output signal across the source and drain of the control transistor such that the attenuating is controlled by transconductances, in an on-state, of the input transistor and of the control transistor respectively.
The invention will now be described with reference to the enclosed drawings, in which:
Referring to
In
The inventors have recognised that an attenuator circuit incorporating a source follower can exploit characteristics of the source follower, such as efficiently handling signals with large amplitude and providing good linearity between input and output with low power consumption, as well as operating at low voltage and provide a wide bandwidth.
In
Since R is normally much higher than 1/gm of the input device M1 (308), the degree of attenuation can be approximated as the ratio of the transconductances (gmM1 and gmM2) of the input transistors M1 (308) and M2 (312) as shown in equation 3 and further in equation 4.
Since transconductances (gm) of PMOS transistors are proportional to the aspect ratios {(W/L)M1 and (W/L)M2} of the transistors (M1 308 and M2 312) (gm∝W/L), the degree of attenuation G can be transformed into equation 5. In equation 5, the DC voltage value of Vin (304) and Vbias (310) are made to be the same. The transconductance of M2 (gmM2) can be regulated by varying Vbias (310), which provides another means of attenuation control, in addition to the voltage settings Vin (304).
In most integrated circuits, differential circuits are often used. A differential circuit based SFA (400) is constructed as shown in
To simplify the circuit of
To reduce DC-offset from inputs, a controlling transistor can be cross-coupled to cancel DC components from input signals. An example of the DC inputs cross-coupled attenuator is shown in
In this implementation, it is assumed that the DC bias is provided by the attenuator (600) itself, which is common in source followers. However, this may present an issue in biasing the attenuator device (600) because Vgs (gate to source voltage) of the input transistors M1p M1n (410, 424) and attenuating transistors M2p M2n (414, 416) need to be the same size to make the relationship of equation 2 and 3 valid. This can be achieved if AC coupling is adopted and biasing voltages (Vbiasp, Vbiasn) (412, 418) of the input transistors (410, 424) can be set by a voltage biasing circuit. Further, the SFA (600) may work better if DC coupling is used with filters. In this case, two low-pass RC filters are employed as shown in
In the above-mentioned attenuators (500, 600), resistors (Rp, Rn) (502, 508) and capacitors (Cp, Cn) (504, 506) are used which is more suitable for discrete circuits instead of integrated circuits. However, both the resistors (Rp, Rn) (502, 508) and the capacitors (Cp, Cn) (504, 506) can be replaced by MOS resistors and MOS capacitors, which is commonly known (i.e., PMOS, NMOS, CMOS configured). Here, resistors (Rp, Rn) (502, 508) and capacitors (Cp, Cn) (504, 506) are used for providing appropriate biasing voltage at the gates of attenuating transistors M2p (414) and M2n (416).
The degree of attenuation can be varied for the SFAs (500, 600) of
Referring to
The second stage of the ladder (752) has similar arrangement to the SFA (600) of
The degree of attenuation of the differential SFAs (700, 800) is given by
Here, switches S1 (S1p1, S1p2, S1n1, S1n2) and S2 (S2p1, S2p2, S2n1, S2n2) have a binary number of either “1” or “0” for obtaining “on” and “off” states. In series switches implementation of
The proposed SFAs (700, 800) provide high linearity without consuming much power. They (700, 800) also provide wide bandwidth without consuming much power. The SFAs' (700, 800) gains can be set by an input transistor aspect ratio, which is highly accurate and independent from the variation of the power supply, process, and temperature. The SFAs (700, 800) operate at low voltage and can handle very large input signal. The SFAs (700, 800) can also be used with both DC/AC coupling. The SFAs (700, 800) both use source follower as gain attenuator and adopt low-pass filter for self-biasing. Degrees of attenuation control of these two SFAs (700, 800) can be varied by switches.
A practical implementation for the SFA (800) of
Referring to Table 1, the input and control transistors are constructed based on a unit transistor with unit aspect ratio of 5.00 μm/0.18 μm.
Referring also to Table 2, which presents summarised pass-band attenuation results, the measured gains of the attenuator (900) match well with the calculated gains. This indicates the expected performance of the proposed SFA (900). The linearity of the SFA (900) is also measured with two-tones in-band signals, which confirms high linearity of the SFA (900). The SFA (900) also achieves very high bandwidth of more than 150 MHz, which is based on driving 2 pF capacitors. Although the SFA (900) achieves high linearity and high bandwidth, it merely consumes 500 μA from 1.8V supply voltage, which includes biasing circuitries. It is worth mentioning that the SFA (900) only occupies very small area, which is only 120×60 μm2. A test chip of the SFA (900) is packaged with SOIC-8 for testing. The magnitude responses of the SFA (900) controlled by the two bits switches (S1, S2) are shown in
Referring to
It will be appreciated by a person skilled in the art that, where a differential SFA is not required, alternative embodiments can provide a single current source SFA based on either the positive or negative current sides of the SFAs (400, 500, 600, 700, 800 and 900). In such embodiments, instead of the cross-coupling connections to avoid DC offset in the SFAs (600, 700, 800, and 900), the sources of the control transistors, i.e. M3p (706) and M2p (414) in
In the above-mentioned embodiments, NMOS transistors may replace PMOS transistors where appropriate.
It will be appreciated by a person skilled in the art that numerous variations and/or modifications may be made to the present invention as shown in the specific embodiments without departing from the spirit or scope of the invention as broadly described. The present embodiments are, therefore, to be considered in all respects to be illustrative and not restrictive.
Filing Document | Filing Date | Country | Kind | 371c Date |
---|---|---|---|---|
PCT/SG06/00361 | 11/24/2006 | WO | 00 | 4/23/2010 |