Reduced-Length Bond Pads for Broadband Power Amplifiers

Abstract

In a transistor formed on a semiconductor die mounted on a substrate, where the transistor output is connected to a circuit on the substrate, a bond pad electrically connected to a transistor drain finger manifold extends less than the full length of the manifold. By controlling the length of the bond pad, the parasitic capacitance it contributes may be controlled. In applications such as a Doherty amplifier, this parasitic capacitance forms part of the quarter-wave transmission line of an impedance inverter, and hence directly impacts amplifier performance. In particular, by reducing the parasitic capacitance contribution from transistor output bond pads, the bandwidth of a Doherty amplifier circuit may be improved. At GHz frequencies and with state of the art transistor device feature sizes, concerns about phase mismatch between drain finger outputs are largely moot.

Claims

1. An amplifier, comprising: a substrate; at least a first semiconductor die mounted on the substrate; at least a first transistor formed on the first semiconductor die, comprising a first plurality of drain fingers; a first manifold electrically connecting the first plurality of drain fingers; and a first bond pad electrically connected to the first manifold, the first bond pad extending a length less than the length of the first manifold; and at least one bond wire electrically connecting the first bond pad to a circuit on the substrate.

2. The amplifier of claim 1, further comprising a second transistor comprising a second plurality of drain fingers; a second manifold electrically connecting the second plurality of drain fingers; and a second bond pad electrically connected to the second manifold.

3. The amplifier of claim 2 wherein the first and second transistors are formed on the same first semiconductor die.

4. The amplifier of claim 2 further comprising a second semiconductor die mounted on the substrate, and wherein the second transistor is formed on the second semiconductor die.

5. The amplifier of claim 1 wherein the length of the first bond pad is determined, at least in part, to control a capacitance presented by the first bond pad.

6. The amplifier of claim 5 wherein the amplifier is a Doherty amplifier, and further comprising: an impedance inverter on the substrate, the impedance inverter connected via bond wires to the first and second bond pads.

7. The amplifier of claim 6 wherein the capacitance presented by the first bond pad affects the operation of the impedance inverter so as to improve the broadband performance of the amplifier.

8. A method of manufacturing an amplifier, comprising: providing a substrate; mounting at least a first semiconductor die on the substrate, the first semiconductor die having at least a first transistor formed thereon, the first transistor comprising a first plurality of drain fingers; a first manifold electrically connecting the first plurality of drain fingers; and a first bond pad electrically connected to the first manifold, the first bond pad extending a length less than the length of the first manifold; and electrically connecting the first bond pad to a circuit on the substrate via at least one bond wire.

9. The method of claim 8, wherein the first semiconductor die further has a second transistor formed thereon, the second transistor comprising: a second plurality of drain fingers; a second manifold electrically connecting the second plurality of drain fingers; and a second bond pad electrically connected to the second manifold.

10. The method of claim 9, further comprising mounting a second semiconductor die on the substrate, the second semiconductor die having at least a second transistor formed thereon, the second transistor comprising: a second plurality of drain fingers; a second manifold electrically connecting the second plurality of drain fingers; and a second bond pad electrically connected to the second manifold.

11. The method of claim 8 wherein the length of the first bond pad is determined, at least in part, to control a capacitance presented by the first bond pad.

12. The method of claim 11 wherein the amplifier is a Doherty amplifier, and further comprising: placing an impedance inverter on the substrate; and connecting the impedance inverter via bond wires to at least the first bond pad.

13. The method of claim 12 wherein the capacitance presented by the first bond pad affects the operation of the impedance inverter so as to improve the broadband performance of the amplifier.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The present invention will now be described more fully hereinafter with reference to the accompanying drawings, in which embodiments of the invention are shown. However, this invention should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art. Like numbers refer to like elements throughout.

[0020] FIG. 1 is a block diagram of a representative, conventional Doherty amplifier circuit.

[0021] FIG. 2(a) is a diagram of a semiconductor die containing two transistors, bonded to a substrate, wherein bond pads on the transistors are coextensive with drain finger manifolds.

[0022] FIG. 2(b) is a diagram of a two semiconductor dice, each containing one transistor, bonded to a substrate, wherein bond pads on the transistors are coextensive with drain finger manifolds.

[0023] FIG. 3 is an equivalent circuit schematic representation of the transmission line of a Doherty amplifier impedance inverter.

[0024] FIG. 4(a) is a diagram of a semiconductor die containing two transistors, bonded to a substrate, wherein bond pads on the transistors are shorter than the drain finger manifolds.

[0025] FIG. 4(b) is a diagram of a two semiconductor dice, each containing one transistor, bonded to a substrate, wherein bond pads on the transistors are shorter than the drain finger manifolds.

[0026] FIGS. 5(a)-5(d) are diagrams depicting various lengths and placements of bond pads along respective manifolds.

[0027] FIG. 6 is a graph of simulate frequency response for amplifiers using conventional bond pads and shortened bond pads according to embodiments of the present invention.

[0028] FIG. 7 is a flow diagram of a method of manufacturing an amplifier.

DETAILED DESCRIPTION

[0029] For simplicity and illustrative purposes, the present invention is described by referring mainly to an exemplary embodiment thereof. In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention. However, it will be readily apparent to one of ordinary skill in the art that the present invention may be practiced without limitation to these specific details. In this description, well known methods and structures have not been described in detail so as not to unnecessarily obscure the present invention.

[0030] FIG. 3 is an equivalent circuit schematic of key elements of the Doherty amplifier circuit of FIG. 1. The first 18a and second 18b amplifiers are modeled as current sources, and the impedance inverter 22, which may be implemented as a quarter-wave transmission line, is modeled as a pi-network, using lumped capacitances C1 and C2, and lumped inductance L. As used herein, a lumped component represents spatially distributed electrical and physical circuit effects as one (or two, when associated with separate amplifiers 18a, 18b) electrical components, such as a resistor, capacitor, inductor, or the like. Such lumping of circuit effects greatly simplifies simulation, and is reasonably accurate over at least some range of operating conditions (power, frequency, etc.). At least some of the lumped inductance L and capacitances C1 and C2 represent parasitic effects. For example, the bond wires 38 contribute to the inductance L. The lumped capacitances C1 and C2 comprise primarily the source to drain capacitances C.sub.DS of the transistor devices 18a, 18b, but also include parasitic capacitance contributed by the large areas of the bonding pads 36. This parasitic capacitance limits the broadband performance of the Doherty amplifier 10.

[0031] According to embodiments of the present invention, the sizes of bond pads connecting to the drain finger manifolds of the first and second transistors are varied, at least in part to control the parasitic capacitance contributed by the bond pads. In this manner, designers can control the resulting capacitances, and utilize them in the design of the quarter-wavelength impedance inverter to optimize the Doherty amplifier's wideband performance.

[0032] FIGS. 4(a) and 4(b) depict reduced-length bond pads 40, according to embodiments of the present invention. FIG. 4(a) depicts the case of two (or more) transistors 18a, 18b formed on a single semiconductor die 26, which is mounted on a substrate 30. FIG. 4(b) depicts the case of each transistor 18a, 18b being formed on a separate semiconductor die 27, 28, both of which are mounted on the substrate 30. In both cases, the transistor 18a, 18b outputs are connected to an impedance inverter 22 formed on the substrate 30, via a plurality of bond wires 38. The bond wires 38 connect, on the die 26, or dice 27, 28, to reduced-length bond pads 40. That is, the bond pads 40 extend along their respective output manifolds 34 less than the full length of the manifolds 34. The specific length of the bond pads 40 may be varied for each specific implementation, to control the parasitic capacitances they generate and contribute to the quarter-wave transmission line of the impedance inverter 22. The lower limit of the bond pad 40 length is determined by the desired capacitance, as well as by ensuring enough area to bond a sufficient number of bond wires 38 to carry the output current.

[0033] Prior art concerns about the phase misalignment of output signals from different drain fingers 32 are largely moot with state of the art circuit feature sizes, and at GHz frequencies. For example, a bond pad 40 according to embodiments of the present invention is on the order of 1 mm; at 2 GHz, the wavelength is 150 mm. Hence, changes in the electrical length due to the bond pad 40 do not have appreciable effect on the signal phases.

[0034] FIGS. 4(a) and 4(b) depict the bond pads 40 aligned with the outer edges of the corresponding manifolds 34, as may be appropriate to accommodate a long impedance inverter circuit 22. However, the size and position of the bond pads 40 in FIGS. 4(a) and 4(b) are not limiting. For example, FIGS. 5(a)-(d) depict other representative sizes and placements of bond pads 40, at least one of which extends less than the full length of the corresponding manifold 34.

[0035] FIG. 5(a) depicts a bond pad 40 on the first (upper) transistor that extends the full length of the corresponding manifold 34only the bond pad 40 on the second (lower) transistor is less than the length of the corresponding manifold 34. Such an arrangement may be appropriate where only a slight reduction in parasitic capacitance is required, and/or a large bond pad 40 area is required to support a given number of wire bonds 38 to the first transistor.

[0036] In FIG. 5(b), both bond pads 40 are shorter than their respective manifolds 34. In this embodiment, the bond pads 40 are both positioned toward the center, minimizing the length of the impedance inverter 22. Additionally, the bond pad 40 on the second (lower) transistor is considerably shorter than the corresponding bond pad 40 depicted in FIG. 5(a)reflecting, in this embodiment, a greater need to reduce the parasitic capacitance presented to the impedance inverter 22.

[0037] In FIG. 5(c), both bond pads 40 are aligned with the tops of the respective manifolds 34, and the bond pad 40 of the second transistor extends more nearly the full length of the corresponding manifold 34.

[0038] In FIG. 5(d), both bond pads 40 are positioned at the center of the respective manifolds 34. This minimizes the maximum electrical length (and hence any potential phase shift) of transistor output signals on the furthest drain fingers from the bond pad 40.

[0039] In general, according to embodiments of the present invention, at least one bond pad 40 may assume any length that is less than the respective manifold 34, and may be placed at any position along the length of the manifold. The length of the bond pad 40 is determined to achieve a required or desired capacitance, such as to optimize the broadband performance of the amplifier 10.

[0040] FIG. 6 depicts a graph of the simulated frequency responses for comparable Doherty amplifiers having conventional bond pads 36 extending the full length of respective manifolds 34 (dashed curve), and bond pads 40 according to embodiments of the present invention, wherein the length of each bond pad 40 is adjusted to be less than the length of the corresponding manifold, to optimize its parasitic capacitance contribution (solid curve). The curves exhibit a similar 3 dB point on the lower end, where m1=3.005 dB at 1.750 GHz. At the higher frequencies, however, the conventional bond pad 36 lengths yield m3=2.947 dB at 2.180 GHz, while the inventive shortened bond pads 40, the lengths of which were optimized based on their capacitance, yield point m2=2.971 dB at 2.240 GHz. Optimizing the bond pad 40 length thus improved the 3 dB bandwidth from 430 MHz to 490 MHzan increase of 14%.

[0041] FIG. 7 depicts the steps in a method 100 of manufacturing an amplifier. A substrate is provided (block 102). At least a first semiconductor die is mounted on the substrate (block 104). The first semiconductor die has at least a first transistor formed thereon. The first transistor comprises a first plurality of drain fingers; a first manifold electrically connecting the first plurality of drain fingers; and a first bond pad electrically connected to the first manifold. The first bond pad extends a length less than the length of the first manifold. The first bond pad is electrically connected to a circuit on the substrate via at least one bond wire (block 106).

[0042] Embodiments of the present invention may be advantageously employed in any amplifier application where one or more transistor outputs are wirebonded to a substrate in a package, and where control of parasitic capacitance is advantageous. Embodiments are particularly well suited to Doherty amplifier configurations in such packaging, as the parasitic capacitance is incorporated into the, e.g., quarter-wave transmission line of an impedance inverter, and directly affects the amplifier operational characteristics, such as its bandwidth. As such, embodiments of the present invention are particularly well suited for wireless communication applications, such as massive MIMO (mMIMO) multi-chip module (MCM) power amplifiers and macro driver power amplifiers.

[0043] Embodiments of the present invention present significant advantages over the prior art. By engineering the length of a bond pad, at a length less than the full extent of a corresponding drain finger manifold, to achieve a required or desired capacitance, the operation of the impedance inverter is optimized, improving the wideband performance of a Doherty amplifier. At high frequency and small feature size, the phase mismatch in output signals from different drain fingers, having different electrical lengths from the drain finger to a bond wire, is negligible.

[0044] The term directly electrically connected or electrically connected or simply connected describes a permanent low-ohmic connection between electrically connected elements, for example a wire connection between the concerned elements. Although such a connection may have parasitic effects, such as the parasitic inductance of a bond wire, no component or element is interposed between the connected elements. By contrast, the term electrically coupled or simply coupled means that one or more intervening element(s) or components, configured to influence the electrical signal in some tangible way, may be (but is not necessarily) provided between the electrically coupled elements. These intervening elements may include active elements, such as transistors or switches, as well as passive elements, such as inductors, capacitors, diodes, resistors, etc.

[0045] Spatially relative terms such as under, below, lower, over, upper and the like, are used for ease of description to explain the positioning of one element relative to another element. These terms are intended to encompass different orientations of the device in addition to different orientations than those depicted in the figures. Further, terms such as first, second, and the like, are also used to describe various elements, regions, sections, etc. and are also not intended to be limiting. Like terms refer to like elements throughout the description.

[0046] As used herein, the terms having, containing, including, comprising and the like are open ended terms that indicate the presence of stated elements or features, but do not preclude additional elements or features. The articles a, an, and the are intended to include the plural as well as the singular, unless the context clearly indicates otherwise.

[0047] The present invention may, of course, be carried out in other ways than those specifically set forth herein without departing from essential characteristics of the invention. The present embodiments are to be considered in all respects as illustrative and not restrictive, and all changes coming within the meaning and equivalency range of the appended claims are intended to be embraced therein.