Tag : ASICs

ASIC and FET innovations will dominate power device technology

Innovations in power devices in the 1990s will be in the areas of power application-specific integrated circuits (ASICs) and cost-effective, high-voltage field effect transistors (FETs) with low on-state resistance. Power ASICs will make it possible to integrate solid state relays (SSRs) with temperature sensors, or to put special timing or decision-making logic on an SSR.

FETS will be used in devices where standard SSRs cannot be used because of a lack of room for a heat sink or other cooling device. They will also be used wherever power dissipation is a major concern. The 1990s will also see the development of closer relationships between designers and device suppliers, opening a way to solving a number of long-standing performance problems.

As a manufacturer of solid-state and time-delay relays, the most important┬átechnological innovations I expect to see in the ’90s will be in the areas of ASICs and FETs. Innovations in ASICs will enable us to do things we couldn’t do before, and FET innovations will make it possible for our devices to fit into areas that were previously closed to solid-state power switches.

Equally important will be the benefits that designers will realize from the closer relationships they must develop with their device suppliers. This trend can be seen throughout the industry as electronic technology becomes more complex–and more potent. Unless designers develop these relationships, much of the available technology may not translate into products that help solve their design problems.

As power ASICs become a reality–that is, as it becomes possible to develop semiconductor devices for specific power applications at acceptable cost–manufacturers similar to ourselves will be able to offer functions that previously were never considered. For example, if a maker of heat pumps typically uses solid-state relays (SSRs) in combination with temperature sensors, we could possibly combine the two into one device, integrating a temperature sensor into an SSR. Or we could put special timing or decision-making logic into an SSR, if it seems to be sensible.

The question is: How are we going to know whether that’s a sensible thing to do? And if it is, how are we to know exactly what type of timing or decision-making circuitry to include?

There’s only one source for that information–the design engineers who use our products. Unfortunately, those designers aren’t in the habit of discussing the details of their designs with component manufacturers. The traditional way of buying power switching devices, such as SSRs, is to choose them from a catalog. Rarely do designers sit down with a relay manufacturer to discuss how, where, and why they’re using that manufacturer’s products. Yet, unless designers start doing that, the full benefits of power ASIC technology will go unrealized.

What sort of device improvements might come out of such improved communication? One improvement is a means for dealing with a common objection to the use of SSRs–when they fail, they fail closed. Users would prefer that they fail open and leave their loads unenergized.

It’s possible to build circuitry into an SSR to detect when the output switch has failed. Such circuitry would compare the state of the output with the state of the input. Then, if there were an output without an input, the circuitry would give a failure indication and also open a particular one-shot electromechanical device say, a fuse, to shut the system down. The details are less important at this point than the fact that until now, such devices would have been too large and too costly to be worth serious consideration. Now we know that they can be made small enough, and we’re fairly sure they’ll be economical in next few years.

I look forward to the advent of the cost-effective, high-voltage FET with a low on-state resistance. Presently, typical SSRs that switch ac loads utilize thyristors as their switching elements. Those devices, whether they’re back-to-back SCRs or triacs, typically have an on-state drop of 1.0 to 1.5 V, which means they dissipate some 20 to 30 W with even a moderate 20-A load.

A thermal load of 20 to 30 W can’t be ignored and requires either a heat sink or some form of cooling. The need to dissipate significant power militates against using SSRs in many applications where there simply isn’t enough room for a heat sink or other cooling apparatus.

Today, high blocking voltage and low on-state resistance are almost a contradiction in terms for FETs–at least at affordable prices. Nevertheless, I believe that high-voltage FETs with very low on-state resistances–a few milliohms–will become available at acceptable prices during the 90s, making it possible to build high-voltage, high-current SSRs that dissipate very little power.

Those relays will be used in close quarters where SSRs can’t be used today. In addition, designers will prefer them over today’s devices even in applications where there’s sufficient room for proper cooling, simply because it’s always a good idea to minimize power dissipation.

But, in my opinion, the interesting effects of technological change in my area of interest in the ’90s will be on “how” we do business with each other, rather than on the details of “what” we do.