You've seen those live-saving defibrillators in offices, shopping malls and even homes, standing at the ready for immediate use in case of a medical emergency.

When the Philips Medical Systems division of Royal Philips Electronics introduced its HeartStart defibrillator in 2002, it was among the first low-cost, mass-market units designed for ordinary, untrained users. It received attention and plaudits due to its small size (a little bigger than a large hardcover book), light weight (2.1 kg/4.6 pounds) and ease of use, with prompts delivered by a soothing human voice. (Watch this video to see the defibrillator unit in action.)

This type of unit--formally known as an automated external defibrillator (AED) to distinguish it from the implanted pacemaker--was soon either mandated or voluntarily installed in tens of thousands of public locations and private homes.

The initial AED design began with a team at Heartstream Inc. (US Patent 5,735,879, April 7, 1998), prior to the company's acquisition by Hewlett-Packard/Agilent (1998), which subsequently sold it to Philips (2001). The product had to meet a maze of constraints with respect to size, weight, performance, regulatory mandates and cost (around $2,000 retail).

It required innovative circuitry based largely on standard components and a custom application-specific integrated circuit (ASIC) processor to arrive at its user-friendly appearance and operation. Despite the inherent complexity of restarting a heart via defibrillation, the designers did not want a graphical user interface, multi-line text display or touchscreen to direct and guide a user, just a voice.

Addressing Three Challenges

The design addresses three traditional analog challenges: 1) generating carefully shaped high-voltage/current pulses from a low-voltage, small-capacity battery pack; 2) enabling analog inputs to make sensitive measurements at the contact points of these pulses; and 3) providing for sophisticated battery/power-management algorithms to account for long periods of idleness followed by sudden, unanticipated operating cycles.
Successful defibrillation depends on multiple factors, including the voltage of the delivered waveform, the energy (joules) of that waveform and the shape of the waveform.

In the early years of defibrillators during the 1940s, the waveform was nothing more than a 50/60-Hz AC line stepped up by a factor of 5× or 10× via a standard transformer. The waveform’s effectiveness was limited while the risk to the cardiac patient was high. Work was done in the 1950's on more complex defibrillation protocols, with the goals of having battery-powered units and providing a more effective, safer waveform.

Research showed that the most effective resuscitation pulse is the biphasic truncated exponential (BTE) waveform, which reaches as high as +1750 V and down to about -500 V. Its period ranges between 5 and 20 msec, with currents reaching 20 A, corresponding to total energy of about 150 J.

In addition to generating its unique waveshape, the Philips device had to drive this bipolar, high-voltage energy pulse into the body's impedance from a small unipolar supply. (For children, for example, the BTE waveform "dose" needs to be far less: up to +600 V and -200 V, spanning 8 to 20 msec, with energy of 50 J.)

The biphasic waveform from Philips takes the well-known biphasic defibrillation waveshape and adds adaptive aspects based on patient ECG and stimulus response. Source: PhilipsHowever, the actual current and voltage delivered cannot be static values set in advance. The parameters, and thus the energy delivered, must be adjusted dynamically and dependent on variations in the actual impedance of the patient. In more basic terms, the delivered power is the product of varying voltage and current delivered into the load impedance, while energy is the time integral of the delivered power. To achieve this delivery, Philips added "intelligence" to the biphasic waveform as part of its proprietary algorithm.

Reading the Patient’s Heart

Knowing the patient's impedance is critical to managing the continuous interplay of voltage, current and energy while delivering the BTE waveform shape. So, before the defibrillator generates its first pulse, it reads the patient's electrocardiogram to determine if shock is even appropriate. It also determines the impedance of the patient's chest by injecting a low-level AC current at the contact electrodes and measuring the voltage drop that results. Only then does it generate a pulse with the appropriate specific BTE waveform values. After each jolt, the impedance is re-measured and the pulse set-up modified as needed. Impedance can change, for example, due to sweat and other factors.

Because of the complex BTE waveform shape, the design team considered using direct digital synthesis (DDS) for its generation. This was rejected due to complexity, cost and consistency issues associated with system transients. The team also looked at using a supercapacitor to store and then release battery energy for the expected high-rate, high-energy discharge/recharge cycle. This was overruled due to size and long-term performance concerns. Instead, the team chose to use a standard, high-quality 100-μF film capacitor.

Circuitry is managed by a custom ASIC and requires analog circuitry for ECG monitoring, high-energy waveform generation and battery power management. Source: PhilipsThe pulse circuit manages the discharge of this capacitor via a bank of (silicon-controlled rectifiers SCRs) and an insulated-gate bipolar transistor (IGBT). These, in turn, are controlled by the system central processing unit, an application-specific integrated circuit (ASIC). The design needed special attention to control the current's di/dt slewing, which occurs as the circuit reverses the waveform polarity via the SCRs to create the biphasic (bipolar) swing. It does this by generating the positive-going waveform, turning the waveform generator off for a few hundred microseconds to allow transients to settle, then generating the negative-going portion.

Battery Pack

The battery pack is comprised of nine standard CR123 lithium batteries rated for nominal 9 V/4.2 AHr capacity (the type also used in many cameras). Philips decided to provide the batteries as a sealed, custom pack to avoid the possibility that users might insert a cell incorrectly, use non-lithium cells with the same form factor but different ratings, or have lower-quality or even counterfeit third-party replacement cells.

Simulating the ASIC's performance during its approximately 1.5-second operating cycle took 11 days on a 2002-vintage high-end PC. The mock-up prototypes used no fancy technologies. Instead, corrugated cardboard, duct tape and glue were the primary rough-fit mockups, with more formal enclosures built closer to the final decision stage.

Although it would have required less memory, the calm, reassuring voice that guides the user was not generated via computer-voice digital synthesis, as this would sound harsh and artificial. Instead, a human narrator recorded the phrases in a relaxed and clear voice to be digitized, concatenated, decoded, amplified and sent to the unit's loudspeaker.

One concern is that, unlike a defibrillator in an ambulance or emergency vehicle, the AED likely will remain unused and unchecked for long periods. To ensure that the battery pack and system remain viable despite months and perhaps years of inactivity, the unit runs daily self-test procedures. These evaluate battery condition, internal circuitry, the overall waveform-generation and delivery functions, the calibration of key components and the high-voltage circuitry.

Prototype Evaluation

As with any product, two questions must be answered: does the product work to the design's target specifications, and are those specifications correct?

The Philips team answered the first question with tests and measurements using industry-standard patient simulators from Symbio and Dynatech-Nevada (subsequently part of Fluke Biomedical). The team answered the second question via detailed analysis of the array of fully vetted medical literature which reported on tests on animals, as well as data analyzing performance of existing non-portable defibrillators.

In just over a decade, the portable, easy-to-use defibrillator has become almost a common fixture in many locales. Its engineering design simplicity and effectiveness didn't come easy, but required a focus on what was needed, what was allowed and what was achievable in this life-or-death application.

Question or comment about this article? Contact the editor: engineering360editors@ihs.com