Spare the Juice—SWaP is a Hot Topic


  • Page 1 of 1
    Bookmark and Share

Over the past couple of years there has been a marked growth in the concern for minimizing power consumption in all kinds of devices as well ast in life in general. This has not previously been the norm and still appears somewhat unique, especially in a nation that still has places like the Las Vegas strip. But it is happening. Possibly the proliferation of portable, battery-powered devices has been a big impetus. Concern from sectors like the military for size, weight and power (SWaP) optimization has certainly played a role. The realization in the industrial sector of the amount of power consumed by electric motors—about 65 percent of all electricity used in factories—has spurred cost-saving efforts aimed at better motor control.

Recently I chaired a panel at the AdvancedTCA Summit on infrastructure where the presentations centered on power management and cooling issues, among related topics. It came out that in a typical data center, depending on a number of variables, somewhere between 30 and 40 percent of the total power budget is devoted to simply forcing air through the facility to dissipate the heat produced by all the silicon on all the boards in all the rackmount chassis in such a building. How do we bring those numbers down?

Well, more efficient cooling technologies might help, but the main way is to reduce the power dissipation of the silicon in those systems. This example explains why there is such a concentration in the semiconductor industry on low power and power management for all systems, not just the handheld and mobile devices. For example, in this issue of RTC, we have an article from ZeroWatt on a compression technology for A-to-D converters that analyzes the signal characteristics to minimize power consumption. Power consumption at the component level is also a major concern for advanced storage systems and can also be seen in the article on 6 Gbit/s SAS in this issue as well. Attacking the cumulative consumption of all silicon devices is needed to make significant advances in optimizing SWaP.

We are seeing innovations in software development and analysis tools aimed at getting a better picture of the details of power consumption. These include such things as the ability to predict how long a device may be placed into one of several sleep modes and still be reliably available for the next scheduled task. In addition, several vendors are introducing probes that can help correlate power usage with the code being executed.

In the example of the data center, reduction in silicon power consumption reduces the heat that must be dissipated (i.e., power wasted) and as a consequence, the power that must be consumed to move the air thus saving electricity and—not inconsequentially—dollars. Such a two-tier example could be extended to the factory floor where heat must also be carried away from equipment, but also in terms of using computer intelligence to reduce the power consumption of those motors that are responsible for so much of the electrical bill. 

Such things as open- and closed-loop variable speed drives are becoming more economical and more widely used thanks to the miniaturization and cost drops in the computer intelligence needed. Motor control is also moving into more intelligent controllers that can dynamically monitor the motor’s load as the load changes and calculate the amount of power needed at any time for the required RPM and torque. Such techniques can not only save on power consumption, they can also increase the life span of the motor.

Increasingly, we see the application of the sort of intelligence used to manage the power in silicon and then in integrated computer systems, spreading out into the aforementioned “life in general.” Computer intelligence is well known to be widely used in hybrid vehicles, but is now also showing up to reduce the fuel consumption of more traditional internal combustion engines. Likewise, we are all now familiar with the efforts to build out the Smart Grid—an effort that is dependent on extensive data networks, intelligent monitoring of conventional and newer alternative sources of energy, and which also includes as a necessary element, intelligence in everyday household appliances. All of these are rich opportunities to apply the low-power devices and the expertise that has gone into developing them to the issues of more efficient power consumption in the world at large.