The Evolution of Smart Power Relays

Since the mid-1980s, semiconductor technology has promised to provide smart power products for solid-state relays that could carry high current and provide unique functions for customers. However, electromechanical relays — with all their shortcomings — continue to be the choice for many applications, due to their cost, ease of use, or efficiency advantage. While both technologies have improved, semiconductor technology has narrowed or closed the gap on these and other application parameters.

Today, technology developments required to make smart power relays today's choice come from advances within the last 15 years in power MOSFETs, mixed signal analog ICs, and power semiconductor packaging. Let's look at these advancements to see what the next five years could bring.

Electromechanical vs. SSRs

Solid-state relays (SSRs) for the lower current applications have existed for many years. In many cases, these are optically isolated products. A variety of solid-state relays exist — most of which are low (<5A) current. Semiconductor manufacturers have been making subassemblies for these products with a variety of technologies, including optoisolators, MOSFETs, thyristors, and power ICs. Previously, it would have been difficult to compare a semiconductor SSR to an electromechanical relay on an apple- to-apple basis. However, the table summarizes how these products currently compare. As shown in Fig. 1, the electromechanical relay usually needs the help of a few semiconductors to meet the requirements of the application — especially if the application is automotive.

The areas where electromechanical relays maintained an edge over the solid-state version are the places that semiconductor manufacturers have made the most improvements. As noted, in several cases, the solid-state version is approaching parity with the electromechanical relay, particularly when the additional system requirements, such as the control IC for protection, diagnostics and interfacing to an MCU, and the cost of the additional fuse, are taken into account.

One of the more recent and very small electromechanical relays has a 90.6 mm footprint but is more than 11 mm in height. It's rated for 14V, can handle 25A with a surge rating of 35A, and dissipates 800mW. Although the contacts are at the milliOhm level, the coil consumes more than 50mA. The acoustic noise level is 60 dB at 5 cm, and is rated for 100,000-cycle operating life. Pricing for 10,000 is $1.50.

In comparison, a recently introduced solid-state relay, Motorola's MC33982, has 12 × 12 = 144 mm footprint but is only 2.2 mm in height. The solid-state relay is rated for a maximum of 27V and can handle 60A (as long as the maximum junction temperature isn't exceeded). The on-resistance is 2mΩ, and the operating supply current is 20mA. With no moving parts, acoustic noise isn't an issue.

In a pulse width modulation (PWM) application operating at 4 kHz, the solid-state relay will have switched 100,000 times in 25 sec. It has been in qualification for over 1000 hr (100,000 miles) of operation in an automotive environment. Pricing for 10,000 is $3.90. This pricing includes the capability to replace the fuse that the electromechanical relay requires, as well as including additional circuitry to allow customers to configure the relay to meet their specific application requirements for current limit, controlled turn on, and fault detection and prevention. For the electromechanical unit to meet this current level, two would be required. The cost parity at these volumes is obvious. The size would double for electromechanical relay and be prohibitively high in some applications.

SSR Progress

Since the high current solid-state relay has taken years to get to a head-to-head comparison with electromechanical relays, the progress made for the three key pieces of the solid-state relay must be understood.

The most fundamental piece of the relay is the switch portion. The electromechanical relay contacts have always presented an infinite impedance in the open state and only a milliOhm or so loss in the closed state. As shown in Fig. 2, the power MOSFET portion of a solid-state relay had considerably higher on-resistance in the late 1980s and has made remarkable improvements. Since the earliest power ICs were monolithic devices, the area available for the power portion was significantly reduced, and 100mΩ in a TO-220 package was a more typical value. Today's capability approaches 2mΩ in the industry-standard TO-220 package. Improvements in MOSFET avalanche capability allow a 40V unit to be used today where 50V was previously required.

Since the relay is either open or closed, it has few critical switching parameters. For the MOSFET, on-resistance isn't the only critical parameter. Other changes that have occurred include increased capability to withstand overvoltage (improved avalanche), which provides increased ruggedness and reduced gate charge for easier switching.

As mentioned earlier, the first smart power ICs were monolithic, single output devices with key functions such as short-circuit protection, simple diagnostics, and some ease of interface aspect, such as the drive circuitry for a high side switch. Fig. 3, on page 31, shows Motorola's SMARTMOS^™ technology evolution after the initial (SMOS1 — not shown) introductions of the mid- to late 1980s. The die shrinks of over an order of magnitude in the critical dimensions since the first process has allowed far more functionality in a given area, especially for the CMOS portion of the mixed signal process. Recent processing changes have even added nonvolatile memory.

The newest process technology — the 0.25-micron SMARTMOS 8 deep trench technology — allows silicon designers to dramatically reduce the analog portion of new designs for extremely efficient layouts. Deep trench technology allows the isolation of logic from the noise generated by switching power devices in applications such as motor drivers. Deep trench also allows designers to pack more features into a chip and still have it fit into existing packaging. One of the critical performance metrics in highly integrated power is substrate injection. SMARTMOS 8 has reduced substrate injection to several orders of magnitude lower than previously available products.

Packaging SSRs

The most obvious change in high current solid-state relays or a smart power product that forms the basis of a solid-state relay is the packaging. Fig. 4 shows that the earliest smart power IC packages (circa 1985) were typically a 5-pin TO-220 or 11-pin or 15-pin single in-line package (SIP) that weren't surface-mount technology. Designed to house a single chip, the employed silicon process offered required isolation to separate the power device from the circuitry. Pin-count was a limiting factor to adding more functionality. The thermal capability of the TO-220 is currently 1°C/W, but was higher in the 1980s due to die size limitations and plastic packaging capability.

In the 1990s, several versions of packages with multiple pin-counts were introduced by different suppliers. Surface-mount packages became common, but a single die attachment/heat spreader was still a common characteristic of these packages. This required monolithic silicon design or complex assembly processes to isolate multiple die products. Thermal capability of these packages is typically 2°C/W.

The latest leadless packaging is strictly for surface-mount assembly. The overmold construction allows new versions to be easily defined and implemented within a given (12 mm × 12 mm) outline. A critical design feature is the ability to isolate multiple die from each other, allowing different semiconductor technologies and different circuit topologies to be designed into the package. For example, in a single switch solid-state relay, the mixed signal circuitry can be on one isolated island and the power MOSFET on another, with additional isolated islands used to provide connection points for multiple source wires. The thermal capability of the QFPN is typically 1°C/W.

Target Applications

An application where the bar has been raised is portable power tools. Customers attach value to extended operating time without recharging — especially commercial users. Fig. 5 shows a simple circuit used for applying power, speed control and reverse operation. Since the only load on the battery is the motor, a double-throw, double-pole, center-off switch is used. A thermal switch is added to protect the motor. Replacing the electromechanical relay with a solid-state relay that can be PWM controlled can eliminate the loss through potentiometer and provide extended battery life. Using the same switch, an H-bridge isn't required, and a single solid-state relay can replace the electromechanical unit and provide the power, speed control and more. The solid-state relay can also include the overcurrent and overvoltage protection to protect the motor and battery. This is an efficient means of providing the speed and torque control and long life on a single battery charge that customers need for portable drills, saws, sanders, and other portable tools.

Solid-state relays can also address the humidity, salt water, and vibration-related failures of electromechanical relay contacts that are common in marine applications. Suppliers of relays are expanding their portfolio to include more solid-state offerings to provide products that will survive for several years — even in ocean-going crafts.

These applications are similar to those encountered in automotive applications, with the additional environmental requirement to survive higher levels of humidity, salt water, and vibration. Since the solid-state unit doesn't have contacts and is sealed, it's a natural for motor, lighting, and heating loads. The additional protection, diagnostic, and user-selectable switching features built into the solid-state unit provide the same value to the marine application as they do in automotive applications.

As anyone who follows semiconductor technology knows, history has a way of repeating itself — and with a high degree of predictability. Past trends are good indicators of where technology will be within the next five to 10 years. The progress noted in the three key areas of a solid-state relay will be extended on a similar trajectory. It's expected that we'll see on-resistance that's a fraction of today's values, features that are easily doubled, and inevitably smaller packages with even more pin-outs in future generations of solid-state relays.