When you think about powering medical devices, the first thoughts that come to mind are likely batteries, power supplies, cables, or connectors. But there’s another side to power management—microchips. Based on a range of different semiconductor manufacturing processes, microchips used in medical devices help manage power in everything from both low- and high-power imaging systems to portable and implantable devices. Regardless of the medical device systems for which they are designed, integrated circuits perform several important functions, increasing power density, maximizing power efficiency, and optimizing the balance between static and active energy states.

Increasing Power Density

In such power-hungry medical devices as imaging equipment, power management is critical. While power supplies for x-ray machines tend to range from 2 to 3 kW, high-end systems such as CT scanners can exceed 100 kW. “In such applications, one of the most important objectives for system designers is to increase power density, or the amount of output power that can be generated per unit area,” remarks Edgar Ayerbe, product marketing engineer at Durham, NC–based Cree Power. “For example, some customers’ power supply cabinets have a fixed size,” Ayerbe explains. “Within that space, the goal is to generate as much power as possible. Thus, increasing the output of a power supply by changing the discrete components without having to add more heat sinks or fans is a big plus.”

To accomplish this goal, Cree has developed metal-oxide-semiconductor field-effect transistors (MOSFETs) based on silicon carbide (SiC) that offer distinct advantages over standard silicon insulated-gate bipolar transistors (IGBTs) or MOSFETs. Compared with silicon-based IGBTs and MOSFETs, SiC MOSFET structures achieve much lower losses at the same voltages and even at much higher switching frequencies, according to Ayerbe.

Cree’s C2M0025120D MOSFET structure achieves lower switching and conduction losses than silicon technologies by optimizing the inherent material properties of silicon carbide, Ayerbe says. This functionality enables imaging equipment manufacturers to increase the power of their systems without having to substantially redesign the mechanical and thermal components. For example, using SiC-based technologies, a power engineer can often increase the power density by a factor of two to five. In many cases, this improvement is accomplished simply by replacing the power semiconductors, resulting in more precise images.

“Silicon carbide is advantageous because of its electrical properties,” Ayerbe remarks. “With SiC, it’s possible to use a very thin layer of material and achieve a high blocking voltage. This thin layer also allows for lower ON-resistance per unit area.” Standard silicon technologies, in contrast, achieve higher blocking voltages by incorporating up to five or 10 times more silicon material in the blocking layer. This additional material increases the device’s resistance and switching losses. “The higher the losses, the more resistance the device generates, Ayerbe adds. “Eventually, it acts like an inefficient incandescent light bulb generating a lot of heat from the power losses.”

To protect patients, medical imaging systems typically feature reinforced galvanic isolation to separate the input power source from the output power that the system generates. Employing reinforced isolation usually requires the use of large transformers with sizable footprints. But because the C2M technology can achieve higher switching frequency with low power losses, it can help to increase the power density and thus lead to the use of smaller transformers.

“A challenge going forward is that while we offer a fast and power-efficient chip, many supporting products are not yet available,” Ayerbe notes. “For example, we need faster gate drivers for silicon carbide technologies.” Another challenge is magnetics. Newer magnetic materials at a lower cost are required to take full advantage of the power density capability of silicon carbide. “We’ve demonstrated that we can increase the power density at least fourfold using off-the-shelf products,” Ayerbe says. "If we had better infrastructure in place, including improved magnetic materials, we could perhaps increase power density tenfold.”

Enhancing Power Efficiency

A primary consideration when developing electronic medical devices is how to control leakage currents that may be present from electrical terminals that connect the patient to a system ground, comments Alan Martin, an applications engineer at Dallas-based Texas Instruments (TI). If an external fault occurs in such cases, the patient might get shocked or even electrocuted. To avoid this issue, systems generally feature a redundant layer of insulation in the power-delivery path.

“The historical method used for forcing power across each insulated layer, or isolation boundary, is a switching-mode power supply,” Martin says. “Often, the power levels necessary for patient-connected devices are in the several-watt range. Given such power levels, some sort of power transformer is required as part of a switching-mode power supply, which pushes power to the electronics on the patient-connected side. Generally, this also requires the creation of a feedback path from the secondary side back to the primary side to close the feedback loop and provide regulated voltage for the patient-connected circuitry.”

Nowadays, however, a feedback loop may not be necessary because power can pass through the transformer and undergo postregulation on the secondary side—a configuration that reduces the number of devices bridging the insulated layer. Each device that bridges the insulated layer requires a recognized insulation rating, adds to stray capacitance (the ac leakage current path), and increases costs and PCB real estate, motivating system designers to minimize the number of devices crossing the boundary.

Because of their wide input supply range, TI’s Simple Switcher power modules are suitable as secondary-side postregulators. Combining a dc/dc converter with power MOSFETs, a shielded inductor, and passives, these modules provide the functionality of a synchronous switching regulator and power efficiency greater than 95%. They also eliminate loop compensation design, inductor selection, and complex layout placement challenges. "These switchers are fairly generic," Martin notes. “They can be employed in the types of general-purpose monitoring devices that are used in the OR or critical-care units, including pulse oximetry, end-tidal CO₂, ECG, and blood-pressure monitors.”

“An isolation barrier is necessary in all medical device applications, states Marc Davis-Marsh, another applications engineer at TI. “Once you’ve gone past the isolation, you’re basically looking at what you are powering in the medical device.” But because medical devices perform imaging functions or sense small signals, the primary considerations after passing the isolation barrier is to minimize noise levels in order to avoid affecting imaging or sensor signals.

Traditionally, many designers have used low-dropout regulators (LDOs) to minimize noise, Davis-Marsh says. But because power levels and the complexity of medical device equipment have increased, LDOs are no longer sufficient for reducing power dissipation. And as temperatures have risen in electronic medical equipment, devices have become more susceptible to long-term failures. Thus, instead of using LDOs, designers incorporate switching regulators, which offer higher efficiency than LDOs and help to minimize heat.

“However, despite their utility, switching regulators also come with a tradeoff, Davis-Marsh notes. “They produce their own noise. In general, dealing with this noise comes down to how well laid out the regular is so that it is well bypassed with capacitors and filters that contain the noise within as small a loop area as possible. A small loop area minimizes EMI and the conducted noise, and that’s what allows you to use switching regulators in some lower-noise environments.”

When synchronous switching regulators began to appear on the market, power efficiency approaching 95% became possible, Martin says. In addition to reducing heat, achieving higher efficiency made it possible to extend the battery run time. “Since many medical device products are designed with a single battery to implement an uninterruptable power supply, the use of switching regulators has made it possible to either run batteries longer or decrease their size to maintain the same run time while decreasing the weight of the medical device.”

Optimizing Static and Active Energy States

Introducing an ultra-low-power microcontroller into the design to maximize battery life—that’s the key to managing power in such portable medical diagnostic devices as glucose meters, according to Sean Newton, FAE manager at Austin, TX–based STMicroelectronics. But how do you get there?

“For such systems, the main requirements are to have low-power components, the ability to place the devices in low-power states, having a powerful CPU core to control and perform advanced calculations, having ample nonvolatile memory to store both program images and user data, and having a variety of peripherals to connect different components, whether they be analog or digital,” Newton explains. Coupled with low-power optical sensors, the STM32L1 microcontroller performs these functions, helping product designers to achieve low-power targets in medical device applications.

Achieving ultralow power has been characterized by two main themes, Newton says: “How long can I stay in my lowest power mode before waking to run my application—known as static energy; and how fast can I process the required application before returning to my lowest power mode—known as active power? The total application energy expenditure equals static energy plus active energy.”

A critical consideration is the overall device architecture. For example, the STM32L1 can turn on and off based on sampling demand, so that it does not have to be powered completely during the operational period, Newton remarks. For sensors connected using an inter-integrated circuit (I2C), a serial peripheral interface (SPI), or a universal asynchronous receiver/transmitter (UART), data can be collected, stored, and processed in batches. And depending on the processing required, applications can scale the CPU clock to the highest frequency required, quickly process the data, and then return to a lower power state. Alternatively, an application can scale to the most power-efficient frequency and then return to the lowest power state.

For applications such as glucose meters, the two main states include standby and operational states. Critical to preserving battery life is the ability to keep operational energy usage to a minimum. For a glucose meter, a typical use case is to run five times a day, with around two minutes of operation time for each user interaction, according to Newton. “During the inactive time, the meter is in standby mode, with a real-time clock running for time tracking. During operation, management of operating power is critical. Too much operating energy can overload the ability of the battery to deliver the required energy, shortening its lifetime. Thus, the primary goal is to keep the required operating energy to a minimum.”

Two main objectives affect power consumption: reducing leakage current in the static-power state and ensuring a low runtime current for achieving dynamic operation. Meeting these objectives depends on the fabrication technology used to make the microcontroller. For example, leakage current has increased in ASIC designs as microcontroller features have shrunk below 60 nm. “However, while improved chip fabrication technology is crucial for achieving low static energy consumption, this goal is also a function of all the static components that make up the device, including the size and number of registers in the CPU core, the number of peripheral registers, and the amount of SRAM contained in the part,” Newton comments. “Thus, technology considerations are a major factor in designing future low-power chips.” —Bob Michaels

Bob Michaels is senior technical editor at UBM Canon.

[email protected]