The design of a system or board’s power distribution network, or power tree, often shifts between centralized and decentralized. It’s an oscillation driven by advances in technology and components as well as design requirements. In situations where saving space for other functions is of primary concern, designers can turn to tiny DC-DC converters that come with other benefits.
The added benefits of these tiny DC-DC units include allowing designers the flexibility to reevaluate their power tree topology, its impact on board layout with fewer constraints, improved performance and higher efficiency, along with savings in overall real estate.
This article will discuss the role of ultra-small DC-DC converters before introducing sample devices and how to best apply them.
Why the shift to ultra-small converters?
The emergence of tiny step-down (buck mode) DC-DC power converters reflects a shift from using larger intermediate bus converters (IBCs) to power relatively large point of load (POL) converters, which then power a relatively large subsystem comprising multiple ICs.
Instead, designers now have the option of using highly distributed, physically tiny converters that can be placed right next to their load, which can be just a single IC and its support components.
There are two reasons for using these highly distributed DC-DC converters. First, new micro components, higher operating frequencies (into the megahertz (MHz) range), advanced fabrication techniques, and enhancements to packaging allow for easy-to-use DC-DC units with impressive performance. Second, delivering power rails this way also brings many primary and secondary benefits to the circuit design, the overall board layout, and the end product.
Further, and although it may seem counterintuitive, using many smaller converters can actually reduce total power subsystem footprint, save on pc board space, and deliver opportunities to add features and functions.
Look at the specifics
It’s interesting to look at the performance and size specifications associated with these converters. For example, the Texas Instruments LMZ10501 “nano” module is a step-down DC-DC converter which can drive up to a 1 amp (A) load (Figure 1).
Figure 1: The Texas Instruments LMZ10501 DC-DC converter can supply up to 1 A with an efficiency of up to 95%. (Image source: Texas Instruments)
Despite this output rating, it is true to its “nano” designation as it comes in an 8 pin, 3.00 millimeter (mm) × 2.60 mm µSIP package, including its inductor (Figure 2).
Figure 2: The LMZ10501 DC-DC regulator comes in 3.00 mm × 2.60 mm µSIP package including its inductor. The bottom view shows its contacts (left); and the top view is dominated by the inductor (right). (Image source: Texas Instruments)
The LMZ10501 is not a bare-bones device as it includes a soft-start function based on an internal current limit, as well as current overload protection and thermal shutdown. In a typical application with basic operation, it requires only an input capacitor, output capacitor, a small VCON filter capacitor, and two resistors (Figure 3). The integral inductor has a DC current rating of 1.2 ADC, supported by a “soft” saturation profile going up to 2 A.
Figure 3: The LMZ10501 requires just three small capacitors and two resistors for operation; the relatively large inductor is part of the IC itself. (Image source: Texas Instruments)
The choice of the external capacitors requires careful consideration. To provide an optimal balance between size, cost, reliability, and performance, both the input and output filters should be low ESR MLCC components. A single 10 microfarad (μF) capacitor (0603 or 0805 size) rated at 6.3 or 10 volts is usually adequate for bypassing of VIN; multiple 4.7 μF or 2.2 μF capacitors can also be used.
Note that choosing a capacitor with too small a value can lead to instability due to lower loop phase margin. In contrast, if the output capacitor is too large, it may prevent the output voltage from reaching the 0.375 volt level required at the end of the startup sequence. There is no significant benefit in using values larger than those recommended.
Look at the size implications
With such a small footprint, designers can reset their thinking and look at new ways to deliver power to the various ICs and other components. Instead of a larger supply located some distance way – such as the corner of the pc board – this µSIP allows the final stage regulation to be done right next to the load. As an added bonus, the devices are fully compatible with standard pick-and-place machines and soldering stations.
How does using this multiplicity of smaller units save space? It does so in both obvious and not-so-obvious ways:
- They reduce the need for the high value, physically larger bulk capacitors at the upstream supply since so much of the regulation is now done local to the load.
- They allow tailoring of the final DC rail (or rails) to the specifics of the load at the upstream DC-DC or AC/DC power unit.
- Since this DC rail is situated close to the load, there’s reduced need for small bypass capacitors on the rails. In effect, the ultra-tiny DC-DC converter at the load not only delivers the needed power, but also can act in the role of some or all of the bypass capacitors.
- Improved transient response due to close location.
- The converters can be individually sized to the operate within their optimum load vs. efficiency window. This increases overall efficiency, spreads their modest dissipation over a wider area, and may eliminate the need for a fan or heatsink.
- The devices are so thin that they can be located on the bottom of a pc board, even when the board is in a closely spaced rack or thin enclosure. Again, this enhances layout flexibly which can lead to a space saving design.
- Crosstalk and noise between a “noisy” IC and sensitive ICs is greatly reduced.
- Although these converters are not electrically isolated, if there is a need for a small isolated converter, it needs to only to be sized for the isolated function.
- Finally, it reduces the need for wider circuit board traces to reduce IR drop and parasitics on the DC rails, which affect load side transient performance.
Note that these tiny DC-DC converters are not limited to loads under 1 A. For example, the TPS82130 MicroSiP™ power module, also from Texas Instruments, provides a 3 A output current from a 3 to 17 volt input, integrating a synchronous step-down converter and an inductor, and delivers an adjustable output voltage between 0.9 and 6 volts (Figure 4).
Figure 4: The TPS82130 DC-DC module from Texas Instruments requires only a few external passive components and can deliver up to 3 A at 0.9 to 6 volts (user adjustable) from a DC input between 3 and 17 volts. (Image source: Texas Instruments)
Don’t be misled by the “module” designation here as this device measures just 3.0 mm × 2.8 mm × 1.5 mm. A look at the appropriate performance curves show that its overall high efficiency peaks at a little over 1 A, and stays high up to the full 3 A rating (Figure 5).
Figure 5: Efficiency of the TPS82130 DC-DC is about 60% or better once it is operating at higher loads and peaks above 1 A, allowing it to be sized to optimally match the load. (Image source: Texas Instruments)
Addressing the relative timing issues
When a system has multiple rails, there are often issues related to their turn-on/turn-off timing with respect to each other. There are three basic types of timing; sequencing, ratiometric, and simultaneous, along with variations of each. Any of these can be implemented by using the Enable (EN) pin and the Soft Start/Tracking (SS/TR) pin on the TPS82130, along with some resistors and/or capacitors (for simplicity, assume just two rails).
In sequential timing, the second device turns on only after the first
device has reached regulation (Figure 6).
Figure 6: Multiple TPS82130 units can be configured for sequential timing, where the left regulator turns on before the right one. Note: while the ICs in the figure are labeled TPS62130, the TPS82130 has improved specifications but the same functionality and pinout. (Image source: Texas Instruments)
In ratiometric timing, both output voltages start at the same time and reach regulation at the same time (Figure 7). It is called “ratiometric” since the two voltages are usually different making their dV/dt slopes differ, but they are related by a constant factor.
Figure 7: In the configuration for ratiometric timing (left), the second voltage rise starts and ends at the same time as the first (right), with a fixed ratio between them. (Image source: Texas Instruments)
Finally, in simultaneous start-up, the slopes of both output voltages are the same, resulting in the voltages reaching regulation at different times (Figure 8).
Figure 8: In the simultaneous mode, both voltages begin their rise at the same time, but they each reach regulation at different times. (Image source: Texas Instruments)
In addition to relative start-up sequencing, there may be concerns with “soft start” (the rate at which the voltage rises on power-on) and the relative tracking of the actual rail voltages with respect to each other. The TPS82130 also makes provision for this using its SS/TR connection.
What to do with this newly freed space?
There are many possibilities for how to use the “now available” space, but choosing the right ones depend on the application priorities. For many designs, the first place to consider is improving both electrical and mechanical ruggedness, areas which are often “cut back” when space is tight.
This may mean adding power supply clamps and crowbars, transient voltage suppressors, and reverse polarity protection on the always vulnerable I/O lines. On the mechanical side, adding extra pc board supports and attachment screws, hold downs, battery clamps, or other structural improvements might be a good use of the space.
Next, it’s time to look at adding extra capabilities or functionality that might be useful. Perhaps there’s room now for a slightly larger battery, or a bigger display and driver IC, more indicator LEDs, or even added user pushbuttons. Perhaps a larger memory option is now feasible, even if it requires a larger IC package. The modest amount of space possibly gained by using these tiny, local DC-DC converters may provide just enough to do more, especially as the board layout now has added flexibility.
Sometimes, less can mean more in unanticipated ways. The availability of ultra-miniature DC-DC buck converters allows placement of the regulation very close to the load, which has a ripple effect on electrical performance, board layout, size and type of upstream power unit, and thermal map.
A derivative effect of using ultra-miniature converters is that they make more board space available within a design’s fixed envelope, thus allowing for other electrical and mechanical improvements, and the addition of new features and functions.
- Texas Instruments, SLVA470A, “Sequencing and Tracking with the TPS621-Family and TPS821-Family”