Switched-capacitor network packaged with load

This applicationis a reissue application of U.S. application Ser. No. 15/779,238, filed May 25, 2018, now U.S. Pat. No. 10,348,191, whichis the national phase under 35 USC 371 of international application no. PCT/US2016/063533 filed Nov. 23, 2016, which claims the benefit of the Nov. 25, 2015 priority date of U.S. Provisional Application No. 62/259,823, the contents of which are incorporated herein by referencein their entirety.

The invention relates to power converters, and in particular, to power converters that rely on switched-capacitor networks.

Many modern electronic devices rely on a stable source of power. In many cases, a power source, such as a battery, does not provide an appropriate voltage or current. Accordingly, it is useful to provide circuitry that will condition that power so that it is suitable for use by the device. A device that provides such power is often referred to as a “power converter.”

A known power converter includes a regulator in series with a charge pump. An example of such a regulator operates by switching an inductor between two states according to some switch duty cycle. The inductor in this regulator performs two functions. One is to control the output voltage of the converter. The other is to promote adiabatic charge transfer among the capacitors within the charge pump.

Known power converters operating according to the above principles are described in U.S. Pat. Nos. 8,860,396, 8,743,553, 8,503,203, 8,693,224, 8,339,184, 8,619,445, 8,723,491, 8,817,501, 9,497,854, 9,143,037, 8,724,353, 9,502,968, 9,203,299, 9,041,459 and U.S. Publication No. 2013-0229841, U.S. Publication No. 2013-0094157, U.S. Publication No. 2013-0154600, U.S. Publication No. 2015-0311786, U.S. Publication No. 2014-0327479, U.S. Publication No. 2016-0028302, U.S. Publication No. 2014-0266132, U.S. Publication No. 2015-0077175, U.S. Publication No. 2015-0077176, U.S. Publication No. 2016-0197552 and International Publication No. 2014/062279, International Publication No. 2015/138378, 2015/138547, International Publication No. 2016/149063 the contents of which are all herein incorporated by reference in their entirety.

The invention includes novel circuit arrangements in which a switched-capacitor converter is coupled to a regulator.

In one aspect, the invention features a power converter comprising a charge pump and a regulator together with a controller for controlling the charge pump and the regulator.

In some embodiments, the regulator connects to an output of the charge pump. In others, the regulator is coupled to an input of the charge pump and a magnetic filter is coupled to an output of the charge pump.

Some embodiments include a die, such as a processor die. Among these are embodiments in which the power converter is mounted beneath the die. Also among these are embodiments that further include a first active component, a second active component, and corresponding passive components, with the first and second active components being beside the die, and the corresponding passive components being beneath the die.

In some embodiments, a die, active components, and corresponding passive components together define a packaged processor. Among these embodiments are those in which the packaged processor is configured to be mated to a socket and those in which the packaged processor is configured to be soldered directly to a motherboard.

In one aspect, the invention features a power converter that includes a power-converter controller, a charge pump, and a regulator. The power-converter controller includes a regulator-controller that controls the regulator and a charge-pump controller that controls the charge pump.

In some embodiments, the regulator connects to an output of the charge pump.

In other embodiments, the power converter further includes a magnetic filter, the regulator couples to an input of the charge pump, and the magnetic filter couples to an output of the charge pump.

Also among the embodiments are those that include a printed-circuit board having first and second opposed faces, a motherboard, solder bumps, and a first die. The first die, which includes circuitry that consumes power provided by the power converter, protrudes from the first face. Meanwhile, the power converter protrudes from the second face, though not by as much as the solder bumps protrude from the second face. As a result, when the printed-circuit board is mounted to the motherboard with the second face facing the motherboard, a gap formed by the solder bumps accommodates the power converter.

Other embodiments include a printed-circuit board having first and second opposed faces, a motherboard, and a first die that includes circuitry that consumes power provided by the power converter. The first die protrudes from the first face and the power converter protrudes from the second face. When the printed-circuit board is mounted to the motherboard with the second face facing the motherboard, the hole in the motherboard accommodates the power converter.

Yet other embodiments include a printed-circuit board having first and second opposed faces, a motherboard, a socket layer, and a first die. The first die, which includes circuitry that consumes power provided by the power converter, protrudes from the first face, Meanwhile, the power converter protrudes from the second face. When the printed-circuit board is mounted to the socket layer with the second face facing the motherboard, a recess defined by the socket layer on the motherboard accommodates the power converter.

Other embodiments include a printed-circuit board having opposed first and second faces, a motherboard, and a first die. Active elements of the power converter are in a second die. The first die, which includes circuitry that consumes power provided by the power converter, and the second die both protrude from the first face. The power-converter's passive components protrude from the second face. When the printed-circuit board is mounted to the motherboard with the second face facing the motherboard, an accommodation space defined by the motherboard accommodates the passive components.

Yet other embodiments include a printed-circuit board with first and second opposed faces, a motherboard, and a first die. In these embodiments, the power converter includes active components spread across plural second dies and passive components. Both the first die, which includes circuitry that consumes power provided by the power converter, and the second die protrude from the first face. Meanwhile, the passive components protrude from the second face. When the printed-circuit board is mounted to the motherboard with the second face facing the motherboard, an accommodation space defined by the motherboard accommodates the passive components. Among these embodiments are those that also include a heat spreader disposed to be in thermal communication with both the first die and at least one second die.

Also among the embodiments are those that include a printed-circuit board having opposed first and second faces, a first die protruding from the first face, and at least a portion of the power converter protruding from the second face. The printed-circuit board is mounted to be in electrical communication with the motherboard. When the second face faces the motherboard, an accommodation space receives at least a portion of the power converter.

Among the embodiments are those in which the power converter includes a second die, a charge-transfer capacitor, and an inductor. The power-converter controller includes circuitry on the second die. In these embodiments, the charge-transfer capacitor is a constituent of the charge pump and the inductor is a constituent of the regulator. The second die, the inductor, and the charge-transfer capacitor are mounted on the first face of the printed-circuit board.

Also among these embodiments are those in which the power converter includes a package that includes a substrate that defines an upper layer and a lower layer, a second die, a charge-transfer capacitor, and an inductor. The power-converter controller includes circuitry on the second die. In these embodiments, the charge-transfer capacitor is a constituent of the charge pump and the inductor is a constituent of the regulator. The package is mounted on the first face of the printed-circuit board. Within the package, the second die is integrated into the lower layer and the charge-transfer capacitor is integrated into the upper layer. Among these are embodiments in which the power converter further includes first and second interconnect layers within the package, with the second die having a device face that faces the first interconnect layer. The first interconnect layer is connected to the printed-circuit board to provide a connection to the second die and to the second interconnect layer. The second interconnect layer connects to components in the upper layer. Also among these embodiments are those in which the inductor is mounted to the printed-circuit board, those in which the inductor is in the package, those in which the inductor is in the lower layer, those in which the inductor is in the upper layer, and those in which the inductor is embedded in the printed-circuit board.

Also among the foregoing embodiments are those in which the power converter further includes first and second interconnect layers within the package. In these embodiments, the second die has a device face that faces the second interconnect layer. The first interconnect layer connects to the printed-circuit board to provide a connection to the second interconnect layer, and the second interconnect layer connects to components in the upper layer and to the second die. Among these embodiments are those in which the power converter further includes thermally-conducting bumps configured to provide a path for heat to flow from the second die to the printed-circuit board and those in which the power converter further includes a thermally-conducting pad configured to provide a path for heat to flow from the second die to the printed-circuit board.

In any of the embodiments that include a die, the die can take several forms, including, but not limited to, a microprocessor, an array of logic gates and a clock, wherein the array is configured to load and execute instructions in synchrony with a clock signal provided by the clock, and an array of programmable logic gates.

In another aspect, the invention features a charge pump packaged with a load that receives charge provided by the charge pump. The charge pump has a plurality of switches that, when connected to a plurality of capacitors, cause the plurality of capacitors to assume a selected configuration. The switches are configured to cause transitions between the configurations of the capacitors. The charge pump comprises a first terminal and a second terminal and transforms a first voltage at the first terminal into a second voltage at the second terminal.

In some embodiments, the charge pump has a controllable adiabaticity.

Also among the embodiments are those having a communication link between a controller of the charge pump and the load to receive instructions from the load.

Yet other embodiments include those that have either a regulator, a magnetic filter, a switching regulator, or a buck converter connected to the charge pump.

Incorporating the load and the power converter into the same package offers numerous advantages. The output of a real power converter inevitably has some ripple. When a voltage having ripple traverses an extended path, it typically encounters inductance. This inductance will tend to distort the voltage, thus making it more difficult to maintain a precise value of voltage. By eliminating the extended path between the power converter and the load, one can maintain a desired voltage with more accuracy.

However, this is only the beginning of the advantages associated with such integration.

First of all, the ability to more tightly control the output voltage lets loose a cascade of other advantages. For example, as soon as one can control output voltage with great precision, it suddenly becomes possible to offer the load a menu of different voltages that are fairly close together. This is an important feature because there are many applications in which a load may need different voltages at different times.

Eliminating the inductance also eliminates the need to include capacitors that would otherwise be needed to compensate for it. This, in turn, reduces both size and cost of the system.

Yet another advantage is modularity. When a power converter is on the motherboard, the details of the motherboard will depend a great deal on the load. As a result, it is generally not possible to swap a different load into the same motherboard without considerable redesign. Once the power converter and the load are integrated into the same package, the coupling between the power converter and the load becomes encapsulated into the package, and no longer significantly impacts the details of the motherboard. This makes design of the motherboard simpler.

Yet another advantage arises from avoiding the delivery of power at a low voltage. Many common loads, such as microprocessors, require relatively low voltages. Among the tasks of a power converter is to take a relatively high voltage and step it down to a value that the load needs. When the load and power converter are far from each other, power at low voltage and high current must be transferred across a long distance. By placing the power converter and the load in the same package, the path that is traversed by the low voltage power becomes much shorter, thus reducing losses.

Yet another advantage arises from the fact that one can reduce the number of pins when high voltage power is received. This is particularly important for pins that handle power. This is because each pin introduces a certain amount of inductance. Thus, one way to reduce overall inductance is to place a great many pins in parallel. This is particularly important when large currents are present, as would be the case when power is delivered at low voltage. By delivering power at high voltage, one can reduce the current, and hence the number of pins required to carry that current.

1 FIG. 10 1 shows a power converterhaving a first terminal maintained at a first voltage Vby a source. Examples of a source include a power supply, a battery, a stack of batteries, and another power converter.

10 1 2 2 The power convertertransforms this first voltage Vinto a stable second voltage Vat a second terminal thereof. It presents this second voltage Vto a load for consumption. Examples of a load include a processor, a field-programmable gate array, or any of a variety of electronic devices that require one or more stable output voltages.

10 Neither the load nor the source is a part of the power converter. These structures are shown only as an aid to understanding. To further emphasize that the load and source are not part of the power converter, these structures have been shown in phantom with dashed lines.

In the illustrated embodiment, the load is shown as connected to the second terminal and the source is shown as connected to the first terminal. These structures are shown connected this way only as an aid to understanding. There is no particular prohibition against connecting the source to the second terminal and the load to the first terminal.

10 55 55 55 10 10 10 In many applications, the load will require different voltages, possibly at different times. To accommodate such applications, the power converterincludes a control bus. Embodiments include those in which the control busis an analog control bus and those in which it is a digital control bus. The control buspermits a load to communicate with the power converter, thus enabling the load to request the voltage that it needs. Thus, many practical implementations of the power converterwill include several outputs, each connected to different combinations of components so as to be able to provide different output voltages to accommodate different load requirements. Alternatively, it is possible to provide multiple power converters, each tuned to provide a different output voltage, so as to accommodate different load requirements.

10 12 14 16 12 18 20 14 24 26 28 30 16 32 34 The power converterincludes several modules, namely: a regulator, a charge pump, and a magnetic filter. The regulatorincludes a regulator inputand a regulator output. The charge pumpincludes first, second, third, and fourth charge-pump terminals,,,. The magnetic filterincludes a magnetic-filter inputand a magnetic-filter output.

1 FIG. 2 4 FIGS.- 2 FIG. 2 4 FIGS.- 16 The foregoing modules can be connected to each other in a variety of ways. For this reason, no specific connection is shown in. However,show particular combinations of these modules. As is apparent from its absence in, the magnetic filteris an optional feature. Although the embodiments shown inhave one charge pump and one regulator each, this is by no means required. It is possible to mix and match the various components to achieve particular performance specifications. Thus, it is possible to have a power converter that uses two regulators and one charge pump or vice versa.

12 14 36 38 40 36 1 2 36 10 The regulatorand the charge pumpinclude switches that, during operation, open and close at particular times relative to each other. These times, however, cannot always be predicted in advance. As a result, it is useful to provide a control systemhaving a regulator controllerand a charge-pump controller. The control systemensures that various voltages, such as the first voltage Vand/or the second voltage V, remain within a relatively tight tolerance of a desired value. The control systemalso controls the startup and shutdown sequence of the power converterand in the event of a fault condition, such as over-voltage or over-current of a node.

38 40 42 44 14 12 38 40 46 48 10 To carry out their control functions, the regulator controllerand the charge-pump controllerreceive feedback via first and second feedback lines,for providing feedback from the charge pumpand regulator. The regulator controllerand the charge-pump controlleralso rely on third and fourth feedback lines,connected to the first and second terminals of the power converter.

38 56 12 40 58 14 56 58 42 44 46 48 54 50 52 The regulator controllerprovides regulator-control instructionsto the regulator. Similarly, the charge-pump controllerprovides charge-pump-control instructionsto the charge pump. Both the regulator-control instructionsand the charge-pump-control instructionsare based on feedback received from the various feedback lines,,,and synchronized based on an external clock-signaland first and second control signals,.

10 12 Among the embodiments of the power converterare those in which the regulatorcomprises a converter with the ability to regulate the output voltage. Also among the embodiments are those that include a switching regulator. Among these embodiments are those that rely on a buck converter. Such converters are particularly attractive because of their speed and high efficiency. Other suitable converters include boost converters, buck/boost converters, fly-back converters, forward converters, half-bridge converters, full-bridge converters, Cuk converters, resonant converters, and linear regulators. Among the embodiments that use a fly-back converter are those that use a quasi-resonant fly-back converter, those that use an active-clamp fly-back converter, those that use an interleaved fly-back converter, and those that use a two-switch fly-back converter. Among the embodiments that use a forward converter are those that use a multi-resonant forward converter, those that use an active-clamp forward converter, those that use an interleaved forward converter, and those that use a two-switch forward converter. Among the embodiments that use a half-bridge converter are those that use an asymmetric half-bridge converter, those that use a multi-resonant half-bridge converter, and those that use a LLC resonant half-bridge.

14 In some embodiments, the charge pumpcomprises a switched-capacitor network or switching-capacitor network. Among the more useful switched-capacitor topologies are: Ladder, Dickson, Series-Parallel, Fibonacci, and Doubler, all of which can be adiabatically charged and configured into multi-phase networks. A particularly useful switching-capacitor network is an adiabatically charged version of a full-wave cascade multiplier. However, diabatically charged versions can also be used.

2 4 FIGS.- 1 FIG. In any power-converter architecture, size and efficiency can be traded off. For example, a given power converter can be made more efficient by making it larger. Conversely, it can be made smaller by making it less efficient.illustrate three different embodiments of the power converter architecture shown in. Each of these embodiments was devised to improve upon the efficiency-size trade off of traditional power-converter topologies.

2 FIG. 10 1 2 14 1 24 28 18 20 2 shows a first embodiment in which the power converterreceives an input voltage Vat its input terminal and transforms it into an output voltage Vat its output terminal. The first embodiment features a charge pumpthat receives the input voltage Vat its first terminaland outputs an intermediate voltage at its third terminal. This intermediate voltage is provided to the regulator's input, which ultimately maintains, at its output, the power converter's output voltage V. In the illustrated embodiment, the input voltage is nine volts, the output voltage is one volt, and the intermediate voltage is three volts.

28 26 30 In the first embodiment, the charge pump's third terminalconnects to its second terminal, which therefore also receives the intermediate voltage. The charge pump's fourth terminalis grounded.

3 FIG. 14 12 18 1 20 24 14 28 32 28 26 34 2 1 2 shows a second embodiment similar to the first embodiment but with the charge pumpand regulatorhaving swapped locations. In this second embodiment, the regulator's inputreceives the input voltage Vand provides a first intermediate voltage at its output. This intermediate voltage is received by the charge pump's first terminal. The charge pumpthen outputs a second intermediate voltage via its third terminal, which connects to the magnetic filter's input. This third terminalalso connects to the charge pump's second terminal, which therefore also sees the second intermediate voltage. The magnetic filter's outputmaintains the power converter's output V. In this embodiment, the input voltage Vis nine volts, the output voltage Vis one volt, the first intermediate voltage is five volts, and the second intermediate voltage is one volt.

4 FIG. 24 1 26 20 18 34 28 32 1 2 shows a third embodiment in which the charge pump's first terminalreceives an input voltage Vand the charge pump's second terminalreceives an intermediate voltage provided at the regulator's output. The regulator's inputin this embodiment is fed back from a magnetic filter's output. As was the case in the second embodiment, the charge pump's third terminalconnects to the magnetic filter's input. In this embodiment, the input voltage Vis nine volts, the intermediate voltage is four volts, and the output voltage Vis one volt.

14 1 2 1 2 1 2 1 2 3 1 2 3 5 FIG. A typical charge pumpfor this and other embodiments is a switched-capacitor network or switching network as shown in. The charge pump includes a first set of switchesand a second set of switches. These are operated such that, except for a brief dead-time interval, the first set of switchesand the second set of switchesare in opposite states. These switches,selectively interconnect capacitors CA, CA, CA, CB, CB, CB to cause the circuit to transition between two states.

6 FIG. 2 4 FIGS.- 2 4 FIGS.- 5 FIG. 6 FIG. 10 1 12 1 2 3 1 2 3 14 1 2 3 1 2 3 1 2 shows a side-view of an implementation of one of the architectures shown in. The power-converterincludes an inductor Lcorresponding to that in the regulatorofand charge-transfer capacitors CA, CA, CA, CB, CB, CB corresponding to those shown in the charge pumpof. The perspective of, as well as the subsequent figures, limits visibility of the charge-transfer capacitors CA, CA, CA, CB, CB, CB to only first and second charge-transfer capacitors CA, CA.

1 2 3 1 2 3 1 62 60 60 1 2 36 The charge-transfer capacitors CA, CA, CA, CB, CB, CB and the inductor Lare mounted on a printed-circuit boardalong with a first die. The first dieincludes a power-management integrated circuit having circuitry to implement the switches,as well as circuitry to implement the control system.

1 2 60 The two downwardly facing arrows indicate that the actual transistors to implement the switches,are found on the lower face of the first die. This face, will be referred to herein as the “device face.” The pair of arrows shown will identify the device face in the following figures.

64 60 62 64 7 FIG. Solder bumpsprovide connections between the first dieand the printed-circuit board. In the embodiment of, these solder bumpsprovide electrical connections. However, in other embodiments, solder bumps provide an avenue for heat dissipation.

7 FIG. 6 FIG. 7 FIG. 5 FIG. 7 FIG. 6 FIG. 1 62 1 2 3 1 2 3 60 10 60 is a side-view of an implementation similar to that shown in. In, the inductor Lremains mounted on the printed-circuit board. However, the charge-transfer capacitors CA, CA, CA, CB, CB, CB and the first dieare now on separate layers of a two-layer package. This reduces the distance between the switches and capacitors in a network such as that shown in. It also reduces the footprint of the power converter. Since the first diecan be very thin, the overall thickness of the embodiment shown inis not much greater than that of the embodiment shown in.

82 1 2 3 1 2 3 1 2 60 82 A substratewithin the package supports the charge-transfer capacitors CA, CA, CA, CB, CB, CB in the package's upper layer. These charge-transfer capacitors CA, CA are discrete elements that, in some embodiments, are encapsulated in a protection compound such as an epoxy. The first dieis embedded within the substrateto form the package's lower layer.

43 43 82 43 60 43 1 2 3 1 2 3 82 43 43 The package also includes upper and lower interconnect layersA,B on opposite sides of the substrate. The lower interconnect layerA permits connection to the transistors on a lower face of the first die. The upper interconnect layerB provides connections to the charge-transfer capacitors CA, CA, CA, CB, CB, CB in the package's upper layer. A number of conductive vias extend through the substrateto provide electrical communication between the upper and lower interconnect layersA,B.

7 FIG. 6 FIG. 6 FIG. 10 10 An advantage of the configuration shown inis that the overall power-converterwill have a smaller footprint than that shown in the implementation of. On the other hand, the implementation shown inwould be easier to manufacture and would result in a lower profile power-converter.

8 FIG. 7 FIG. 60 65 43 62 1 62 shows an implementation similar to that shown in, but with the device face now being on an upper face of the first die. Electrically conductive bumpsprovide electrical connections between the lower interconnect layerA and the printed circuit board. This permits electrical communication between the packaged components and external components. These external components would include the inductor L, which remains mounted on the printed-circuit board.

43 60 64 62 The lower interconnect layerA also provides an avenue for heat dissipation for the first dievia thermally-conductive bumpsthat connect the first die's lower face to the printed-circuit board.

9 FIG. 8 FIG. 8 FIG. 65 65 64 64 shows an embodiment similar to that shown in, but with an electrically-conductive padB replacing the electrically-conductive bumpsand a single thermally-conductive padB replacing the thermally-conductive bumps. Because all of the electrical connections are on the edges, this type of package can be optically inspected. Instead of having to carry out the more expensive x-ray inspection procedure that would be required by the package shown in, which contains hidden solder bumps.

10 FIG. 8 FIG. 10 FIG. 8 FIG. 1 62 1 shows an embodiment similar to that shown in, but with the inductor Lno longer being on the printed-circuit board. Instead, the inductor Lhas been incorporated into the package's upper layer. Since inductors are typically taller than capacitors, the embodiment shown inembodiment will typically be taller than the one in. However, since all of the components are in the same package, the system can be tested as whole. This ensures higher quality and reliability.

11 FIG. 10 FIG. 1 82 1 43 43 shows an embodiment similar to that of, but with the inductor Lhaving been integrated into the substrate. The inductor Lis implemented as a coil that is formed around a magnetic core using the lower interconnect layerA, the upper interconnect layerB, and conductive vias. This is embodiment is particularly useful for forming a large number of low-inductance inductors.

12 FIG. 11 FIG. 1 62 60 60 36 12 14 12 38 14 40 shows an embodiment similar to that shown in, but with the inductor Lhaving been integrated into the main printed-circuit boardand with the first diehaving been split into two separate dies. There are numerous possibilities for the how the first diecan be split. In some embodiments, one die includes circuitry for implementing control systemand the other die includes switches that form part of the regulatorand the charge pump. In another embodiment, one die includes the regulatorand the regulator controllerwhile the other die includes the charge pumpand the charge-pump controller.

13 17 FIGS.- 6 12 FIGS.- 6 12 FIGS.- 13 17 FIGS.- 6 12 FIGS.- 13 17 FIGS.- 10 62 10 show the power convertersofas part of a larger package. To establish a frame of reference, the printed-circuit boardfromcan be seen in each of. As is apparent, the power convertershown inis oriented upside-down in.

13 17 FIGS.- 18 FIG. 62 10 62 72 10 As can now be seen in, the printed-circuit boardhas a lower face and an upper face. The lower face carries the power converterand numerous interconnecting structures for connecting the printed-circuit board, and all components thereon, to a motherboard. Although only one power converteris shown, this is only an accident of perspective. It is quite possible for there to be multiple power converters, as shown for example in.

66 68 66 10 66 1 FIG. The upper face supports a second dieand a heat spreaderthat drains heat from the second die. The second die has circuitry that is to consume power provided by the power converter. It therefore corresponds to the load shown in. In some embodiments, the power-consuming circuitry within the second dieincludes a microprocessor. In other embodiments, it includes a field-programmable gate array.

10 62 10 10 74 72 76 70 62 72 72 78 76 10 13 FIG. 14 FIG. 15 FIG. Since the power converteris on the lower face of the printed-circuit board, some kind of accommodating space must be provided to receive the power converter. In the embodiment shown in, the power converterhas a low-profile so that it fits in an accommodating space formed by the solder bumps themselves. In the embodiment shown in, a holein the motherboardprovides the accommodating space. In the embodiment shown in, socketsinstead of solder bumpsengage the printed-circuit boardonto a motherboard. In this embodiment, there is no need to perforate the motherboard. Instead, an accommodating spacebetween the socketsreceives the power converter.

60 62 66 60 60 60 72 60 1 2 36 80 72 74 72 68 60 66 16 FIG. In the embodiments discussed thus far, the first dieis mounted on the lower face of the printed-circuit board. However, this need not be the case. In some embodiments, the first die can be mounted on the upper face of the printed-circuit board beside the second die. Moreover, the first diecan be split to form a plurality of first dies, each with different portions of the power-management integrated circuitry. An embodiment having a plurality of first dieson the upper face of the printed-circuit board62can be seen in. The first diesinclude the active components of the power-management integrated circuit, such as the switches,and the control system. The passive componentsremain on the lower face of the printed-circuit board62, where they fit into an accommodating space, such as the holein the motherboard. In this embodiment, the heat spreaderdrains heat from both the first diesand the second die.

17 FIG. 1 2 36 66 80 72 74 72 shows an embodiment in which the active components of the power-management integrated circuit, such as the switches,and the control system, have been integrated into the second die. The passive componentsremain on the lower face of the printed-circuit board62, where they fit into an accommodating space, such as the holein the motherboard.

18 FIG. 14 FIG. 62 62 10 shows a view looking toward the bottom face of the printed-circuit boardin a typical implementation such as that shown in. The printed-circuit boardshows four power convertersadjacent to five coupling capacitors and surrounded by numerous terminals, which include ground terminals, power terminals, and I/O terminals.

In some implementations, a computer accessible storage medium includes a database representative of one or more components of the converter. For example, the database may include data representative of a switching network that has been optimized to promote low-loss operation of a charge pump.

Generally speaking, a computer accessible storage medium may include any non-transitory storage media accessible by a computer during use to provide instructions and/or data to the computer. For example, a computer accessible storage medium may include storage media such as magnetic or optical disks and semiconductor memories.

Generally, a database representative of the system may be a database or other data structure that can be read by a program and used, directly or indirectly, to fabricate the hardware comprising the system. For example, the database may be a behavioral-level description or register-transfer level (RTL) description of the hardware functionality in a high level design language (HDL) such as Verilog or VHDL. The description may be read by a synthesis tool that may synthesize the description to produce a netlist comprising a list of gates from a synthesis library. The netlist comprises a set of gates that also represent the functionality of the hardware comprising the system. The netlist may then be placed and routed to produce a data set describing geometric shapes to be applied to masks. The masks may then be used in various semiconductor fabrication steps to produce a semiconductor circuit or circuits corresponding to the system. In other examples, Alternatively, the database may itself be the netlist (with or without the synthesis library) or the data set.