Understanding the Basics of DC-DC Conversion

Mark Patrick, Mouser Electronics


With the advent of distributed power, DC-DC conversion is more important than ever. This article looks at some of the options available for designers

Figure 1: DC-DC converters can be isolated or non-isolated

When Allesandro Volta made a battery out of copper and zinc disks and brine-soaked cardboard, it was a novelty and the low power DC produced had few practical applications. As electricity became better understood and it was discovered that rotating magnets could induce AC current in wires, the world of power generation opened up. AC was fine for lighting, heating and driving motors and, using transformers, its voltage could be raised and lowered.

DC began to have uses though; unlike with AC, charge could be stored in batteries or capacitors, powering the first telephones for example, before newer designs took power from the line itself. If you wanted a particular DC voltage, you put cells in series. If you did want high power DC, different voltages could be produced using a DC motor-generator set with mechanical gearing in between.

As electronics emerged, different DC levels were needed, such as 150V for vacuum tubes and engineers started thinking about how DC voltages could be electronically converted up and down. The DC-DC converter concept was born.

DC-DC converters today

Today’s applications for DC-DC converters split into different categories: isolated and non-isolated, step-up and step-down or both. Modules are available at all power levels for the function and many have become ‘commodity’ parts.

Let’s first define what we mean by ‘isolation’: if there is an electrical connection, (usually ground), between input and output, the part is non-isolated. Isolated parts have an internal transformer which passes energy between input and output magnetically so the output can ‘float’ with respect to the input (Figure 1). We’ll consider grades of isolation later.

Non-isolated DC-DC converters

The simplest non-isolated DC-DC converter is a series resistor, but that’s not very useful as the voltage dropped varies with load current, so a practical circuit uses a transistor to drop voltage, controlled by feedback to keep that voltage constant. This is a ‘linear’ regulator available typically in a three-terminal TO-220 package. A problem though, is that the voltage can only be dropped, not increased, and the device dissipates significant power, the difference of input and output voltage multiplied by load current. The solution is ‘switched mode’ regulators whose transistor is either fully on or off, in both cases dissipating little power, passing pulses of voltage to the output through an inductor and capacitor which ‘smooth’ the pulses back to DC. The output voltage is controlled by varying the width of the pulses. Efficiency gains over linear circuits are dramatic (Figure 2) and a nice feature is that the switched-mode regulator can also be configured to increase the voltage, becoming a ‘boost’ converter. When it just drops voltage, it’s a ‘buck’ converter.

Click image to enlarge

Figure 2: Linear and switched-mode loss comparison


There are other varieties which can produce negative output voltage (buck-boost) and the ‘SEPIC’ circuit which can produce a positive output voltage above or below a positive input. This is very useful in battery applications where a load voltage needs to be kept constant as a battery is discharging (Figure 3).

Click image to enlarge

Figure 3: SEPIC converter outline


Maximum efficiency with minimum size is a driver in many circuits and DC-DC conversion is critical. Usefully, as you increase the frequency of pulses in a switched-mode converter, the inductor and capacitors shrink. Another effect though is that the transistor dissipates a little power at each switching edge and the more switching edges per second (frequency), the more power is dissipated. Efficiency and size therefore tend to be in opposition. Magnetics technology has changed little to improve things but latest wide band-gap semiconductors such as Silicon Carbide (SiC) and Gallium Nitride GaN) promise less dissipation per switching edge and therefore higher frequency operation and smaller magnetics. This technology is allowing non-isolated DC-DCs to provide efficiencies well over 95% with output currents to 100A and more. An example is the PTH04040Wpart from TI rated at 60A, with its output adjustable from 0.8V-2.5V from an input of 2.95V-5.5V in a package with just 51.94 x 26.54mm footprint.

Modular boost converters are less common but an example might be the ABXS002A3X41-SRZ from GE Critical Power which converts 8-16V input to 16-34V output at 2.3A in a footprint of 27.9 x 11.4mm.

Non-isolated DC-DCs are often called ‘Point of Load’ (PoL) converters as they can be placed right at a typical load of a processor or FPGA, so that the load voltage is as accurate as possible. The most complex types will have many features such as digital control so that performance can be varied ‘on the fly’ to suit load conditions, usually set over an I2C line with a standard protocol such as PMBus. VRMs (Voltage Regulator Modules) are particular types of PoLs which meet IC manufacturers’ specific requirements, for example from Intel.

Isolated DC-DC converters

Isolation can be needed for different reasons; often it is convenient to split input and output ground so that current paths can be separated and not interact. A common use is when powering an RS485 interface – isolated power rails for the driver stop current flowing between grounds on the host and connected equipment (Figure 4).

Click image to enalrge

Figure 4: Isolated RS485 interface



Having a ‘floating’ output also gives the ability to connect the load ground point to either DC-DC output terminal so that, for example, a floating 12V can be configured as -12V by connecting the positive output to ground. Equally, the 12V could be ‘stacked’ on another voltage, the input 12V say, to give 24V total which might be useful for a small motor drive (Figure 5).

Click image to enlarge

Figure 5. ‘Stacking’ an isolated output on another voltage



Coupling power through a transformer can also provide some immunity to EMI, particularly the common-mode type - noise on a ground connection with respect to local earth. 

An important reason for isolation is often safety. You might think that if a DC-DC has low input and output voltages, safety is not an issue. However, the isolation barrier in a DC-DC is often used as part of a wider insulation system to achieve an overall safety rating. An extreme example would be in a medical ‘patient connect’ application (Figure 6). Here, the DC-DC converter must have full medical-rated isolation for the highest system voltage, maybe 230VAC. The rationale is that other equipment connected to the patient may fail applying a dangerous voltage to them. The DC-DC could then provide a path for lethal current to flow back into the source equipment through the ‘unspecified connection’, so it must therefore have a high grade of insulation.

Click image to enlarge

Figure 6: DC-DC requires ‘Two Measures of Patient Protection’


DC-DCs with this specification for medical applications will be noted as having two ’Measures of Patient Protection’ (MOPPs) at the appropriate system voltage such as the THM 30WIseries from Traco Power.

Many DC-DC converters are advertised as having an isolation rating simply quoted as a voltage, perhaps ‘3kVDC’. Users should be careful; this is just the factory ‘test’ voltage and although it’s an indication of barrier robustness, unless a specific safety agency level of isolation is quoted such as ‘basic’, ‘reinforced’ or medical ‘measures of protection’, the DC-DC should not be used as part of a safety isolation system. Even if a standard is quoted such as ‘EN 60950 reinforced’, a system voltage must also be specified – reinforced at 30VAC system voltage is worth nothing as a safety barrier in 230VAC systems.

Isolated DC-DC converters are available in many forms with the simplest low power types often having no regulation – they are ‘ratiometeric’ converters – the output varies in proportion to the input. These parts can be low cost and useful for providing ‘spot’ voltages for interfaces or analogue circuitry. They do often have minimum load requirements though, as their no-load output voltages can be much higher than the nominal value. These parts can be efficient at their rated load but are much less so at light load, perhaps less than 50%. With careful design, these DC-DC types can provide high performance with good agency rated isolation such as the Murata 2W surface-mount NXJ2 series.

DC-DC converters with active regulation can accept varying inputs and loads while maintaining tight output voltage regulation. A 2:1 input variation used to be a standard, such as 18-36V, but parts are now available with very wide inputs of 5:1 and more such as the CM1901-9RG from Bel Power, intended for rail applications with extreme input dips and surges.

DC-DC converter specifications

In selecting a DC-DC converter, often environmental considerations come first: is isolation needed and if so to what safety agency level? What is the maximum ambient temperature and is there airflow available? Quoted operating temperatures are often with derating - a 10W converter may only produce 3W at its maximum temperature of 85°C without significant airflow.

Efficiency is always a headline requirement and ties in with operating temperature. Efficiency variation with load is also important though as most users operate parts at less than 100% rating and efficiency may be much lower at this point.

EMI levels vary with DC-DC converter type. Unlike AC-DC converters, there is no statutory level of emissions and extra external filtering may be necessary. A small low-cost converter may not be so attractive when it needs large expensive filters to keep noise down to acceptable levels.

Now maybe you will think about actual input and output volts and amps but don’t forget protection features; robust converters from quality suppliers bought through Mouser save money in the long run.

Mouser Electronics