EMBEDDED BOARDS

(click to enlarge)Embedded boards with multiprocessor technologies offer the computing power necessary for many medical applications.

Today, OEMs are busy designing embedded computer board–based devices for a variety of medical applications: medical modality instruments, clinical diagnostic instruments, surgical imaging management systems, point-of-care devices, ultrasound systems, and computer-aided diagnosis and therapy.

So, when faced with the fact that all of these devices must be designed with rigorous requirements, what are the OEM's key considerations for selecting a medical embedded computer board, and thereby, its supplier? This article reviews the seven steps to making the right choice.

Step One: Choosing the Technology

The first modern embedded system was the Apollo Guidance Computer, developed by Charles Stark Draper at the MIT Instrumentation Laboratory. The first mass-produced embedded system was the Autonetics D-17 guidance computer for the Minuteman missile, released in 1961. Since then, embedded systems have come down in price and moved successfully from military to commercial applications and can be found in devices that range from the iPhone and MP3 players to traffic lights to a wide array of medical applications.

Thanks to innovators such as Intel, embedded computer boards have benefited from the latest and greatest microprocessor technologies, and there has been an enormous rise in processing power and functionality. But the choice of technology is still driven by the requirements of the application—e.g., desktop versus laptop, Windows versus Linux, new technology versus legacy systems.

Picking the Processor. Of the two leading cpu design philosophies—reduced instruction set computer (RISC) and complex instruction set computer (CISC)—CISC can be the suitable choice for embedded computer boards that run popular operating systems (OSs) and provide the most versatile functions.

Embedded computer boards equipped with the latest multicore processor technologies, such as Intel's Core 2 Duo (C2D) and Core 2 Quad (C2Q) processors, provide the increased computing muscle necessary for complex medical applications, yet output at a lower power consumption. (This lower power consumption pleases the budget makers who need to keep an eye on costs of operation.) Intel's Q965 Express Chipset, for example, has made a significant contribution to the success of today's integrated embedded application designs. Many board vendors have chosen the Q965 because it supports Intel's C2D and Pentium 4/Celeron D processors in the LGA775 socket, with scalability for future processor innovations. In addition, the C2D's dual-core technology features two independent processor cores in one physical package, running at the same frequency for true parallel processing. The Intel C2Q processor incorporates four processing cores in one package for projects that require even greater multitasking.

Both the C2D and C2Q are suitable for medical OEMs designing applications that require high computing performance and low energy consumption. For example, to take full advantage of today's sophisticated applications, such as surgical imaging management systems and ultrasound systems, computers would need a 40% increase in power such as that provided by multicore processors. At the same time, however, device manufacturers would not pay more for this power. The same multicore processors are engineered to be energy efficient—with as many as 291 million transistors—to consume 40% less power than single-core processors.

Legacy Systems. New processor technologies need to be backward compatible to support the customer's inherent legacy OS or application software. Some applications lend themselves to mobile computing, which requires a fanless construction. Fanless embedded computer boards can also be a consideration for environments such as the OR or endoscopy imaging, where heat and noise can be a distraction. Embedded boards with a nonmobile chip set and processor typically require a heat sink and a fan.

Looking Good. Although high resolution is a de facto standard for many medical applications, some projects may benefit from the use of dual display and dual video achieved through low-voltage differential signaling (LVDS). This is a common feature on embedded computer boards that use the latest video chip. One of the many advantages of supporting two functioning LCD displays on the same cpu is that one display can be dedicated as a primary medical image, and the second can be used for other information or for field service, i.e., more information is available for the healthcare professional, patient, and engineering service professional. On the most up-to-date embedded computer boards, dual-display and dual-video options are offered as a combination of VGA/LVDS, VGA/DVI, DVI/LVDS, or dual LVDS.

Step Two: Finding the Form Factor

Sometimes a single-board computer (SBC) is too large at 13.33 × 4.98 in. (338.58 × 126.39 mm) for a customer's requirements. An OEM should then be able to turn to a range of smaller sizes. For example, many manufacturers offer a selection of embedded computer boards in different form factors. These boards include 3.5- and 5.25-in. ECX, ETX, and half-size ISA SBC. Other options include PC/104 and embedded main boards such as Mini-ITX, Micro-ATX, and ATX. The small-sized boards—typically 4 × 5.75 in.—usually include mobile chip sets and processors with options for dual-video output, fanless configuration, and expansion capabilities through ISA, PCI, or mini-PCI. The large-sized boards, which can range up to 9.6 × 12 in. (243.8 × 304.8 mm), are available with mobile or the latest multicore technologies.

Composite or S-video interfaces are also available on certain boards, which is useful when the medical device uses a TV as the display unit instead of a computer monitor. The ECX form factor provides the small footprint at 3.5 in. but lacks the physical capacity to accommodate the latest multicore technologies. The benefit of bigger boards such as the ATX is that they not only support the latest microprocessors, but they also include more features and expansion slots. The ATX form factor typically offers six expansion slots. The Mini-ITX form factor is, however, probably the most economical. An integrated board can be as small as 170 mm² or 6.7 sq in., which is significantly more compact than the ATX form factor at 115.2 sq in. (2926.1 mm²).

Some manufacturers make an advanced Mini-ITX embedded system board that takes advantage of Intel C2Q technologies. Although they typically support a single expansion slot, these advanced embedded computer boards use a PCI-E × 16 slot to utilize an extra graphic card. Six SATA (serial advanced technology attachment) ports enrich the system capacity, making the advanced Mini-ITX especially suitable for data storage. Its gigabit Ethernet ports provide high performance for networking transmission.

Another economy of the Mini-ITX with mobile chip set and processor is that the board can be passively cooled because its architecture inherently consumes low power. This makes it the ideal choice for applications in which noise can be a distraction.

Compact Carrier Boards. Given that compact size, computing power options, reliability, ease of use, and function expansion are the key design considerations for every embedded application, another innovative form factor is the computer-on-module (COM), Express architecture, which provides modular computing options. Such an approach includes a module of computing core and a carrier board with condensed fundamental computer functions and interfaces for additional functional expansion.

A typical compact carrier board form factor includes a host computer modular board that is connected with the carrier board through a COM Express–defined interconnection. The advantages of this form factor include easy future processor upgrade and quick time-to-market design.

The modular bootable host computer engine is plugged into the carrier board, which connects to the power supply. The application-specific system functions and peripheral expansion are all built onto the carrier board. The concept is simple: install the latest cpu now, and when it is time to upgrade, just replace the cpu module. OEM engineers save time because only the carrier board needs to be designed and implemented. This helps shorten the development time so that the product can be delivered faster. And upgradability and scalability of the modular approach increases the flexibility of the microprocessor supply now and in the future.

Step Three: Expansion and Peripheral Options

Most medical applications are dynamic. Breakthroughs change the course of treatment; technology catches up to comply; the current application or hardware needs to expand; new peripherals need to be added.

Expansion choices on most embedded computer boards may include ISA, PC/104, PCI, PCI-X, or PCI-E. These options are usually based on the add-on card bus standard and throughput because many functional expansion add-on cards are built based on these interfaces.

ISA. Industry Standard Architecture (ISA) was a computer bus standard designed for IBM-compatible computers. Apart from specialized industrial use, ISA is rarely used in today's designs and is often referred to as the legacy bus.

PCI. The peripheral component interconnect, or PCI, standard specifies a computer bus for attaching peripheral devices to a computer motherboard. These devices can take either one of the following forms:

• An integrated circuit fitted onto the motherboard itself, called a planar device in the PCI specification.
• An expansion card that fits into a socket.

The PCI bus is very common in modern PCs, where it has displaced ISA and VESA (Video Electronics Standard Association) local bus as the standard expansion bus, but it also appears in many other computer types. The bus is now succeeded by PCI Express (PCI-E), which is standard in most new computers and other technologies.

PC/104. The PC/104 bus, used in industrial and embedded applications, is a derivative of the ISA bus, utilizing the same signal lines with different connectors. PC/104 is an embedded computer standard controlled by the PC/104 Consortium, which defines both a form factor and computer bus. PC/104 is intended for stackable configuration with different functional modules. This form factor benefits many OEMs that want a customized rugged system that does not require months of design.

Unlike the popular ATX form factor, which uses the PCI bus and is currently used for most PCs, the PC/104 form factor has no backplane. Instead, it allows modules to stack together like building blocks. The stacking of buses is naturally more rugged than typical PC configurations. This is a result of mounting holes in the corner of each module. These holes enable the boards to be fastened with standoffs.

PCI-X and PCI-X 2.0. Peripheral component interconnect extended (PCI-X) is a computer bus and expansion card standard designed to supersede PCI. It is essentially a faster version of PCI, running at twice the speed, and is otherwise similar in physical implementation and basic design. It has itself been replaced in modern designs by PCI-E, which features a very different logical design.

Despite its backward compatibility with PCI-X and PCI devices, PCI-X 2.0 has not been implemented on a large scale. This lack of implementation is primarily because hardware vendors have chosen to integrate the faster PCI-E instead.

(click to enlarge)PCI Express provides flexibility and scalability.

PCI-E is supported primarily by Intel and is intended to be used as a local interconnect only. It was designed to be software compatible with the preexisting PCI standard, making the conversion of PCI cards and systems to PCI-E as simple as replacing the physical layer, without requiring a change to the supporting software. The increased bandwidth on PCI-E has led to unification among embedded board designers because it is fast enough to replace almost all existing internal buses, including PCI.

Gold Finger Boards. For many of the small embedded computer boards where space is at a premium, a strip, or finger, of gold is placed along the outer edge of the board. The gold finger physically connects the embedded boards to the other boards in the medical device.

Step Four: Solid Engineering Support

Once the physical features of the embedded computer board are resolved, OEMs need to seek a suitable supplier. The board specifications can often determine the supplier or vice versa. Many suppliers can provide similar boards. However, OEMs can narrow the choice by selecting the supplier best able to provide solid engineering support for their designs. For example, the supplier should be capable of modifying the BIOS (basic input-output system) to enable the manufacturer to include unique features.

In addition, it is essential that suppliers be able to provide driver support for nonstandard operating systems, e.g., Linux instead of Windows. It is equally important that the supplier be able to help the OEM meet the necessary regulatory requirements.

Step Five: Customization and Flexibility

For fast time to market, OEMs usually pick a standard embedded computer board product—often something out-of-the-box. There are times, however, when a medical device design requires creating a unique feature not met by a standard product or reducing the cost by eliminating unnecessary parts on a standard board. In such cases, it is essential to identify a supplier that can provide customization in-house.

Step Six: Lasting a Lifetime

For medical device OEMs, it is crucial that the life of the embedded computer board parallel the life of the machine it powers. Longevity is not only important for user satisfaction, but in the medical market, it is also necessary because of the time factor associated with seeking and gaining FDA approvals. With this in mind, it is imperative that OEMs select a supplier that works closely with a component manufacturer committed to providing longevity support. Some do. Unfortunately, many don't.

Intel makes a commitment to members of its Intel Communications Alliance (ICA) to ensure the longevity of products in embedded applications that are covered by its embedded road map support. Members can offer their customers a product with at least a five-year life-cycle support.

Step Seven: An FDA-Compliant Supplier

Because an embedded board is often a critical component, particularly when designing devices for an industry with severe approval guidelines, it is important to investigate the supplier's credentials. Choose a supplier whose quality assurance protocols include provisions for FDA standards, and make sure that its manufacturing processes can pass an FDA audit. Seek a supplier that holds, at a minimum, certification to ISO 9001:2000, which includes ANSI/ISO/ASQ Q9001:2000.

Conclusion

There are seven critical steps to identifying the appropriate embedded board and supplier for a given application. The first steps in selecting an embedded computer board for the medical device market are to source a product that matches the OEM's current needs and is backward compatible with any legacy systems currently in use. An embedded computer board should also be able to expand to meet future needs.

Given that today's applications are processor intensive and that device manufacturers are constantly seeking to reduce costs, these boards offer the maximum computing muscle with low power consumption. Moreover, they can provide longevity that mirrors that of the devices they power.

Along with board selection, it is also important to identify the appropriate supplier. Providing devices in such a strictly regulated industry requires flexible support, product longevity, and quality control procedures that parallel FDA standards.

Frank Shen is product marketing director at American Portwell Technology Inc. (Fremont, CA). He can be reached at franks@portwell.com.