TECHNICAL FIELD

The present invention relates generally to integrated circuits, and in particular, to circuits for and methods of implementing a gain stage in an integrated circuit.

BACKGROUND

The transfer of data is an important aspect of any integrated circuit device. While data may be transferred between integrated circuit devices through a lossy channel, one important aspect of any integrated circuit device is the efficient processing of the received data within an integrated circuit. As the performance requirements for integrated circuits continue to become more demanding, it is important to provide improved circuits and methods of transferring data. More particularly, there is always pressure to increase the speed of an integrated circuit while reducing both the size of the integrated circuit and power consumed by the integrated circuit. The same pressure is applied to the data links between integrated circuits. As the number of logic elements and memory cells in integrated circuits is increased, the density of data links and their power is also increased.

A received serial data stream experiences post-processing in an analog domain before it converted into the digital domain and de-serialized. This postprocessing is implemented on multiple cascaded high-speed analog circuits amplifying and equalizing a signal, distorted and attenuated in a lossy channel, to the level necessary for error free sampling into the digital domain. Another usage of high-speed analog circuits is precise clocking needed for accurate sampling of the received data stream. All high speed analog circuits require high bandwidth and low power consumption that make it very attractive to use a peaking network based on a passive inductor. However, broad usage of a passive inductor for increasing bandwidth and reducing power consumption of high speed analog circuits is prevented by large inductor footprints that not only consume integrated circuit "real estate," but can also block power and signal distribution. A particular implementation of the inductor may also affect the resistance of the inductor and the magnetic coupling of the inductor, for example. Accordingly, it is difficult to implement an inductor of a high speed analog circuit which achieves many of the objectives of the circuit. SUMMARY

A circuit for implementing a gain stage in an integrated circuit is described. The circuit comprises a first inductor formed in a first plurality of metal layers; a second inductor formed in a second plurality of metal layers, the second inductor coupled to a center tap of the first inductor; and wherein the second inductor has a diameter that is less than a diameter of the first inductor.

In such a circuit, one or more of the following may be true: the integrated circuit can comprise a plurality of metal layers extending from a lower metal layer to an upper metal layer, and the first inductor can be situated above the lower metal layer; the first inductor can comprise a first coil on a bottom layer of the first inductor, a second coil, a third coil, and a fourth coil on a top layer of the first inductor, and the third and fourth coils can have traces of smaller widths than traces of the first and second coil; the second coil can have a plurality of loops having traces of an outer loop with widths that are larger than widths of traces of an inner loop; the first coil can comprise a plurality of metal layers; the gain stage can comprise a current mode logic gain stage, and the first inductor and the second inductor can be coupled to transmit first data of a complementary pair of data; and/or the circuit can further comprise a third inductor comprising the first plurality of metal layers and a fourth inductor comprising the second plurality of metal layers and coupled to a center tap of the third inductor, wherein the fourth inductor has a diameter that is less than a diameter of the third inductor, and wherein the third inductor and the fourth inductor are coupled to transmit second data of the complementary pair of data. Integrated circuits and systems comprises one or more such circuits are also decribed.

According to an alternate arrangement, a circuit for implementing a gain stage in an integrated circuit comprises a first transistor coupled to receive an input signal at a control terminal; and a T-coil inductor coupled to the first transistor, the T-coil inductor comprising a first inductor and a second inductor coupled to a center tap of the first inductor; wherein the second inductor has a diameter that is less than a diameter of the first inductor. A method of implementing a gain stage in an integrated circuit is also described. The method comprises implementing a first inductor in a first plurality of metal layers; implementing a second inductor in a second plurality of metal layers, coupling the second inductor to a center tap of the first inductor; wherein the second inductor has a diameter that is less than a diameter of the first inductor.

Optionally, one or more of the following may apply to the described methods: the integrated circuit can comprise a plurality of metal layers extending from a lower metal layer to an upper metal layer, and implementing a first inductor can comprise forming the first inductor above the lower metal layer; implementing a first inductor can comprise forming a first coil on a bottom layer of the first inductor, a second coil, a third coil, and a fourth coil on a top layer of the first inductor, and the third and fourth coils can be formed having traces of smaller widths than traces of the first and second coils; implementing a first inductor can comprise forming the second coil with a plurality of loops having traces of an outer loop with widths that are larger than widths of traces of an inner loop; implementing a first inductor can comprise forming the first coil of the first inductor in a plurality of metal layers; and/or implementing a second inductor in a second plurality of metal layers can comprise implementing the second inductor outside of the first inductor.

BRIEF DESCRIPTION OF THE DRAWINGS

Fig. 1 is a block diagram of a system for transmitting data;

Fig. 2 is a block diagram of a gain stage circuit which could be

implemented in the system of Fig. 1 ;

Fig. 3 is a cross-sectional view of an integrated circuit having a plurality of metal layers;

Fig. 4 is an expanded view of an inductor of the gain stage circuit of Fig.

2.

Fig. 5 is a top plan view of a first metal layer of the inductor of Fig. 4 implemented in an integrated circuit;

Fig. 6 is a top plan view of a second metal layer of the inductor of Fig. 4 implemented in an integrated circuit; Fig. 7 is a top plan view of a third metal layer of the inductor of Fig. 4 implemented in an integrated circuit;

Fig. 8 is a top plan view of a fourth metal layer of the inductor of Fig. 4 implemented in an integrated circuit;

Fig. 9 is a top plan view of a fifth metal layer of the inductor of Fig. 4 implemented in an integrated circuit;

Fig. 10 is a top plan view of a sixth metal layer of the inductor of Fig. 4 implemented in an integrated circuit;

Fig. 1 1 is a system for producing an integrated circuit; and

Fig. 12 is a flow chart showing a method of implementing an inductor in an integrated circuit.

DETAILED DESCRIPTION

Turning first to Fig. 1 , a block diagram of a system 100 for transmitting data comprising a first transceiver and a second transceiver is shown. The first transceiver 102 comprises a transmitter circuit 106 coupled to a receiver circuit 108 by a pair of transmission lines 1 10 and 1 12. The outputs of the transmitter 106 comprise differential outputs which are coupled by pull-up resistors 1 14 and 1 16 to a first reference voltage, such as VCC. The differential outputs may be a pair of complementary data, for example. The receiver circuit 108 is coupled to receive the differential signal at two inputs which are coupled by pull-down resistors 1 18 and 120 to a second reference voltage, such as ground. While the outputs of the transmitter circuits and the receiver circuit of Fig. 1 are coupled to pull-up and pull-down resistors, respectively, it should be understood that the outputs of the transmitter circuits could instead be coupled to pull-down resistors and that the inputs of the receiver circuits could instead be coupled to pull-up resistors.

In order to provide the bi-directional transfer of data, a transmitter circuit 122 coupled to a receiver circuit 124 by a pair of transmission lines 126 and 128. The outputs of the transmitter 122 also comprise differential outputs which are coupled by pull-up resistors 130 and 132 to the reference voltage VCC. The receiver circuit 124 is coupled to receive the differential signal at two inputs which are coupled by pull-down resistors 134 and 136 to ground. While the system 100 of Fig. 1 is an example of a portion of a system having transceivers, it should be understood that the receiver circuits described in more detail below could be implemented in any type of system having transceivers for transmitting and receiving data.

Turning now to Fig. 2, a block diagram of a gain stage circuit 200 which could be implemented as gain stage 108 or 124 of Fig. 1 for example is shown. In particular, the gain stage circuit 200 comprises a plurality of elements coupled in series for generating a first output (OUTp). The elements coupled in series include a first resistor 202, a second resistor 204, an inductor 206, a resistor 208, and a transistor 210 as shown. The transistor 210 is coupled to receive a first input data signal (INp) of a complementary pair of data signals at its gate, and generates the first output data signal OUTp at its drain.

A second series of elements is coupled to receive a second differential input data signal of the complementary pair of data signals and generate a second differential output data signal. In particular, the elements of a second group of elements coupled in series include a first resistor 216, a second resistor 218, an inductor 220, a resistor 222, and a transistor 224. The transistor 224 is coupled to receive a second input data signal (INn) of the complementary pair of data signals at its gate and generate a second output data signal OUTn at its drain. A control transistor 230 may be coupled to receive an enable signal at its gate to enable the gain stage.

As shown in the expanded view of Fig. 2, the inductor 206 comprises T- coils with a first inductor 232 and a second inductor 234 coupled in series at a center tap 244, with a third inductor 236 coupled to the center tap 244. The second coil 212 extends between a first terminal 246 and a second terminal 248, where the first terminal 246 is coupled to the center tap 238. It should be noted that the inductor 236 is not a physical element in the circuit, but is included to model the negative inductance in the circuit and compensate for the negative inductance (as a result of the mutual coupling of the first inductor 232 and the second inductor 234) by providing positive inductance through inductor 212. It should also be noted that, unlike resistor 202 which is an element implemented in the gain stage circuit 200, the resistors 204 and 208 represent parasitic impedances of the inductor. The inductor 212 may have an inductance which is less than the inductance of inductor 206. As will be described in more detail in Figs. 4-10, the smaller inductance of inductor 212 can be achieved by having a smaller diameter of coils of the inductor 212 compared to the coils of inductor 206. The implementation of the inductors 206 and 212 will be described in more detail in reference to Figs. 4-10. Further, it should be noted that the inductor 220 is implemented in the same way as inductor 206 as set forth above.

Turning now to Fig. 3, a cross-sectional view shows an integrated circuit having a plurality of metal layers. As shown in Fig. 3, metal layers having conductive traces for different signal types are provided. By way of example, a ground trace is indicated in solid black, a signal trace is indicated by cross- hatched lines, and a power voltage trace is indicted by vertical lines. The integrated circuit includes a substrate 302 having circuit elements 304 which are coupled to various interconnect elements formed in a first metal layer M1 on the surface of the substrate 302. Vias, shown for example by 306, enable the connection of various traces to traces in other layers. The metal layers are separated by a dielectric layers, shown here as the white material not occupied by the metal traces or vias formed in the dielectric material of the dielectric layers. External contacts 308 may be implemented as input/output (I/O) contacts on a top surface 310 of the integrated circuit as shown for enabling the input and output of data associated with the circuit elements 304 and providing reference voltages, such as power and ground to the circuit elements.

The integrated circuit of Fig. 3 as shown comprises 12 metal layers M1 -

M12, and 12 corresponding via layers V1 -V12. As will be described in more detail below, the metal layers may have different thicknesses and be used for different types of signals or circuit elements, including the inductors of Fig. 2 as shown and described in more detail below in reference to Figs. 4-10. While 12 metal layers and corresponding via layers are shown, it should be understood that greater or fewer layers could be implemented.

Turning now to Fig. 4, an expanded view shows an inductor of the gain stage circuit of Fig. 2. As shown in Fig. 4, the inductor 206 has 4 coils 402-408 implemented in at least 4 metal layers, where each coil has a plurality of loops. Similarly, the inductor 212 comprises four coils 410-416, where each of the coils has a plurality of loops. The coil 402 extends from a first terminal 240 to a second terminal 420 and has 4 loops as shown. The coil 404 extends from the first terminal 422 to a second terminal 424, and also has 4 loops. The coil 406 extends from a first terminal 426 to a second terminal 428 and has three loops. Finally, the coil 408 extends from a first terminal 430 to the second terminal 242 and has four loops.

The loops of the four coils are formed in different metal layers, and the coils are coupled together by vias between the metal layers. More particularly, the second terminal 420 of the coil 402 is coupled to the first terminal 422 of the coil 404 by a via 434. The second terminal 424 of the coil 404 is coupled to the first terminal 426 of the coil 406 by a via 436, and the second terminal 428 of the coil 406 is coupled to the first terminal 430 of coil 408 by a via 438.

The inductor 212 is coupled to the center tap 244 of the inductor 206 by a via 442 at the terminal 246. A second terminal 446 of the coil 402 is coupled to the first terminal 448 of the coil 412 by a via 460. The second terminal 450 of the coil 412 is coupled to a first terminal 452 of the coil 414 by a via 462, and a second terminal 454 of the coil 414 is coupled to the first terminal 456 of coil 416 by a via 464.

The metal layers used to from each of the inductors 206 and 212 may be consecutive metal layers and may have terminals which are coupled by a via in a single via layer. As will be described in more detail below, a given coil may be formed using a plurality of metal layers. The plurality of metal layers used to form inductor 212 may be the same as or a subset of the plurality of metal layers used to form inductor 206.

Top plan views of various metal layers are shown in Figs. 5-10.

According to one arrangement of metal layers, the bottom coil 408 may be formed from a series of coils in different layers which are connected by vias between the metal layers. As shown in Fig. 5, a coil 502 comprises 4 loops which extend from the terminal 504 to the terminal 506. Conductive elements 508 comprising a plurality of vias 510, shown on each side of each loop, enable connecting the coil 502 to a coil 602. The vias 512 correspond to the vias 438 to couple terminal 430 of coil 408 to terminal 428 of coil 406. The coil 408 as shown is a square structure having a diameter d-i . However, it should be understood that other shapes, such as rectangular or circular loops, could be implemented.

The coil 602, which also extends from a first terminal 604 to a second terminal 606, has substantially the same size and shape as the coil 502.

Similarly, conductive elements 608 comprising a plurality of vias 610, shown on each side of each loop, enable connecting the coil 602 to a coil 702. The coil 702, which also extends from a first terminal 704 to a second terminal 706, has substantially the same size and shape as the coils 502 and 602. The coils 502, 602 and 707 together form the bottom coil 408, where the terminals 504, 604, and 704 are a part of the terminal 430 and the terminals 506, 606, and 706 are a part of the terminal 242. That is, each of the plurality of vias 508, 608 and 708 on each side of the loop for each loop in the metal layers of Figs. 5, 6, and 7 form a single coil with 4 loops having a thickness of the three metal loops and the vias between the metal layers. Also shown in Fig. 7 is the coil 416, which is only made up of a single metal layer. That is, unlike the coil 408 which is made up of traces in 3 metal layers coupled together by vias, coil 416 comprises metal traces only in the metal layer having coil 702. The coil 416 is also a square coil having a diameter d ₂ which is less than the diameter d-i , where d1 could be approximately 13 micrometers and d ₂ could be approximately 5 micrometers.

The remaining coils 406, 404, and 402 of the inductor 206 are also formed in a single metal layer with a corresponding coil 414, 412 and 410 of the inductor 212. More particularly, the coil 406 comprises 3 loops, where the width of traces is generally decreasing from the outer loop to the inner loop. That is, W3 is generally wider than w ₄ , which is generally wider than w ₅ . The resistance of the loop is minimized by having wider routing with gradual reduction from the edge to the center. As shown in Fig. 9, each of the sides of the loops of coil 404 (which has 4 loops) has approximately the same width, which is generally less than the widths of the various traces of coil 406. Finally, each of the sides of the loops of coil 402 (which has 5 loops) has approximately the same width w-i , which is generally less than the widths of the various traces of coil 404. The width w ₂ of coil 404, which has only 4 loops, is generally greater than wi of coil 402, which has 5 loops. The thickness of the metal layers can vary, where the thickness of the lower metal layers starting from M 1 is less than the thickness of the upper metal layer M12. While the coils of the inductors of Fig. 4 could be implemented in any the metal layers, the inductors are preferably implemented in the upper metal layers.

Inductors 402 and 410 may be formed in metal layer M12 and have a thickness t-i , inductors 404 and 412 may be formed in metal layer M1 1 and have a thickness t ₂ , inductors 406 and 414 may be formed in metal layer M10 and have a thickness t ₃ , and inductors 408 and 416 may be formed in metal layer M9. As set forth above, inductor 408 may be formed in a plurality of metal layers to provide and increased thickness U compared to a thickness of the metal layer M9 alone. However, the lower coil 416 of inductor 212 may be formed in the metal layer M9 alone and has a thickness t ₅ . The widths of the traces of coils 410-416 have a value w ₇ , which is approximately equal.

The T-coil implementation with the smaller coil 212 provides up to 3.5 times the bandwidth (BW) extension of internal circuitry (compared to

conventional T-coils) based on a multi-layer inductor 206 having a diameter below 15 micrometers and a multi-layer inductor 212 having a diameter of 5 micrometers. The small T-coil 206 may have high self-resonance frequency, but may suffer from high magnetic coupling between coils when implemented alone. In order to overcome these drawbacks, an additional series inductor 212 coupled to the center tap of the inductor 206 compensates for excessive magnetic coupling. The additional inductor 212 also helps to reduce the effect of T-coil resistance on the operation point and DC gain of the stage. In addition, the arrangement set forth above may simplify the design of a bias circuit for the multiple sequential stage by having different loading, but using the same biasing method. That is, fine tuning of the gain stage can be performed by only changing the dimension of coil 212 to enable keeping DC bias same for all CML stages.

The circuits set forth above may be implemented in a current mode logic (CML) stage, for example. An internal CML stage has less power consumption compare to a conventional receiver input. Current in the driver is in the range of 2-5mA compare to 10mA for a conventional receiver. This allows for the same multi-layer structure to make the footprint 2x-3x smaller than an input T-coil, and therefore have a smaller parasitic capacitance or a larger self-resonance frequency. However, the cost for this performance improvement is additional series resistance from thinner metal lines, and a variation over temperature that is on the order of 0.3-0.4% per K that can reach approximately 100% over PVT.

Another problem related to the small footprint of t-coil 206 is an increase of magnetic coupling between T-coil branches associated with smaller size. The configuration results in magnetic coupling k of approximately 0.6-0.8, while a required magnetic coupling is easily achievable with a larger footprint having a magnetic coupling k of approximately 0.2-0.6. If mutual inductance is large, the effective coupling between coils can be controlled by adding inductance to the center tap. Accordingly, large mutual coupling between the coils is beneficial by allowing larger inductance of the branch, but the smaller interwindings provide smaller parasitic resistance and area.

Turning now to Fig. 1 1 , a system for producing an integrated circuit is shown. The system 1 100 comprises computer aided design (CAD) equipment 1 102, which could be any computer adapted to run CAD software. The CAD equipment 1 102 receives data, such as a master pinlist 1 104, and is coupled by a communication link 1 106 to semiconductor manufacturing equipment 1 1 10. The semiconductor manufacturing equipment 1 1 10 generates a wafer 1 1 12 having a plurality of die, as is well known in the art.

CAD equipment 1 120, which is also coupled to receive the master pinlist 1 104, receives a bonding diagram 1 122 and substrate artwork 1 124. The CAD equipment 1 120 is coupled by a communication link 1 126 to bonding equipment 1 130. The communication links 1 106 and 1 126 could be any wired or wireless communication link. The bonding equipment generally provides wire bonds from a die from the wafer 1 1 12 to a substrate receiving the die, as will be described in more detail in reference to other figures. The die/substrate 1 132 is coupled to packaging equipment 1 134 which generates a finished component 1 136, such as an integrated circuit package. Although the system of Fig. 1 1 provides various elements required for producing an integrated circuit package, it should be understood the elements shown in Fig. 1 1 could be combined, or additional elements could be provided. In any configuration, the system of Fig. 1 1 enables generating the required traces in the metal layers to implement the circuit of Fig. 1 , and particularly the loops of the inductors shown in Figs. 4-10.

Turning now to Fig. 12, a flow chart shows a method of implementing an inductor in an integrated circuit. A first inductor is implemented in a first plurality of metal layers at a step 1202. A second inductor is implemented in a second plurality of metal layers at a step 1204. The second inductor is coupled to a center tap of the first inductor at a step 1206, wherein the second inductor has a diameter which is less than the first inductor. The methods of Fig. 12 could be implemented using the circuits of Figs. 1 -10 as described above, or other suitable circuits. While particular elements are shown in Fig. 12, it should be understood that additional details related to the elements of Fig. 12, or additional elements, can be found in the description of Figs. 1 -10.

It can therefore be appreciated that a new and novel gain stage circuit and method of transmitting data have been described. It will be appreciated by those skilled in the art that numerous alternatives and equivalents will be seen to exist that incorporate the disclosed invention. As a result, the invention is not to be limited by the foregoing exemplary embodiments, but only by the following claims.