MIT & TI Team Develop Chip up to 10 Times More Energy-Efficient than Current Chips

From MIT News comes what seems to be one of a long string of disruptive technologies in the semi-conductor industry. The potential impact of these chips is unimaginable at this time.  Can you imaginge a pacemaker or other implantable medical device that requires no batteries?

Researchers at MIT and Texas Instruments have unveiled a new chip design for portable electronics that can be up to 10 times more energy-efficient than present technology. The design could lead to cell phones, implantable medical devices and sensors that last far longer when running from a battery.

The key to the improvement in energy efficiency was to find ways of making the circuits on the chip work at a voltage level much lower than usual, Anantha  Chandrakasan, the the Joseph F. and Nancy P. Keithley Professor of Electrical Engineering, explains. While most current chips operate at around one volt, the new design works at just 0.3 volts.

Reducing the operating voltage, however, is not as simple as it might sound, because existing microchips have been optimized for many years to operate at the higher standard-voltage level. "Memory and logic circuits have to be redesigned to operate at very low power supply voltages," Chandrakasan says.

One key to the new design, he says, was to build a high-efficiency DC-to-DC converter--which reduces the voltage to the lower level--right on the same chip, reducing the number of separate components. The redesigned memory and logic, along with the DC-to-DC converter, are all integrated to realize a complete system-on-a-chip solution.

One of the biggest problems the team had to overcome was the variability that occurs in typical chip manufacturing. At lower voltage levels, variations and imperfections in the silicon chip become more problematic. "Designing the chip to minimize its vulnerability to such variations is a big part of our strategy," Chandrakasan says.

So far the new chip is a proof of concept. Commercial applications could become available "in five years, maybe even sooner, in a number of exciting areas," Chandrakasan says. For example, portable and implantable medical devices, portable communications devices and networking devices could be based on such chips, and thus have greatly increased operating times. There may also be a variety of military applications in the production of tiny, self-contained sensor networks that could be dispersed in a battlefield.

In some applications, such as implantable medical devices, the goal is to make the power requirements so low that they could be powered by "ambient energy," Chandrakasan says--using the body's own heat or movement to provide all the needed power. In addition, the technology could be suitable for body area networks or wirelessly enabled body sensor networks.

A liitle far away from this blogs usual energy effiency topics, but I hope it is of interest to many of you.

Comments

[Tom Konrad]

Interesting to me.. while it may be targeted at medical and military apps in 2-3 years, it could easly be in computers and consumer-electronics in a decade.

Just think about the energy savings of low-voltage server farms!

[al]

Though I'm skeptical about high tech saving us in fossil scarce, I'm not thrilled in living in 1900AD either

[Charlie]

I really doubt that this new low-voltage chip technology could improve cellphone "talk-time" performance. The phone still needs to transmit radio signals with enough power to have an acceptable S/N ratio at the cellular network antennas.

[Will]

They say this is 10 times more energy efficient than current chips.

Current chips 'work at around 1 volt' while their tech works at '0.3 volts'. That does not appear to add up to a 10% betterment.

I don't want to do this advance down, as the improvements are still mighty, and will surely make a huge difference. But I would like them to make their figures match up in their press releases.

If they do, could someone tell me why?

[eric]

Power consumption of a circuit goes as the square of the voltage (assuming that the resistance remains the same). This a 3x reduction in voltage would lead to perhaps a 10x reduction in power.

[Carl Hage]

There is a huge opportunity to reduce power consumption of electronics. The vast majority of power consumed is frivolously wasted. Besides new circuit technology, powering down a circuit when not needed could save, and using circuit techniques that don't waste power.

If electronics were a car, in some cases it would be as if they wanted to simplify the design to have only a brake. The engine runs at full power, then the brake is used to control speed, dumping unused energy as heat.

Consider a typical computer. Nowadays they have multiple cores (processors), but usually everything is powered on, except when the system is "off", and then a quarter of the electronics is really on (to wake on network or key press). There is only a small difference in power between idle and compressing a video, though sometimes the clock rate is adjusted to save a little power.

It's technically possible (with today's technology) to make a computer that uses 1/100 the power, and is instant on (within 2 seconds). Many new computers come with multiple cores. Instead of having 2-4 always on, turn them off, then add another slow ultra low-power processor to run when 99% idle (most of the time). Move idle RAM to flash when sleeping and power down disks. When a computer is powered off, the router, modem, and printer could see there is nothing connected and go to standby. Instead of using the regular power supply and powered up motherboards for standby, disconnect from the wall, and use a small battery or capacitor to power a timer, network sensor, radio signal detector, infrared remote, etc.

[While grabbing coffee I see an article on an x86 processor that is .6W for the handheld market. Compare that to the AMD Quad FX at 250W. Use the 60W processor when needed, otherwise, power it down and use the .6W processor, and power everything down when really idle.]

If a watch runs off a button cell for 7 years, a wireless IR motion detector runs years on AA cells, there is no reason for the TV or VCR to burn 11W just to run the remote ON or timer.

The real problem is that consumers have no idea about power consumption or effiency of electronics, so the marketing depts tell the designers not to worry about power consumption or total cost-- just the most GHz or megapixels.

The solution is to require a yellow sticker on all electronics-- just like a refrigerator. Measure energy in $ over 5-10 years, if always on, always off, and extra for hour/day. People will buy the computer than uses $50 of electricity instead of $500, and designers will make them with little extra cost.

Another idea is to have a competition to see which computers (or server clusters) have the lowest power that meets a typical use performance critera, e.g. display a web page in .5s. For servers, it could be the lowest power for a typical profile of hits, e.g. for a typical site or large site like cnc or yahoo.

If reviewers measured things like energy cost, time to power up/down, open a web browser and show a page, etc., instead of peak gaming speed or GHz, we could choose between things that matter instead of irrelevant statistics.

[bigTom]

Low power electronics and computers are two very different markets/technology. This new stuff is more of an anabler for low lost very low poer devives like sensors. It will have little direct impact on the energy business.

technofossil is correct, that not a lot of attention to power consumption has gone into computers, at least not to sleep modes (although the battery powered laptop area has always emphasized power savings, as a way to extend battery life). An important way to make a system green, is to run a slower clock rate. Power scales as roughly the third power of the clock rate. A quad-core chip running at 85% of the clock of a single core chip should consume about the same power as a single core chip running at 100% (and do 2.7times as much work). Having two supported clock modes say slow/fast would be a good way to go, and the switching time between modes is on the order of a microsecond, so say 50% speed might make a decent sleep mode.

What is really needed is green consumers, to start pushing the manufacturers. If Intel got word that overall power consumption was an important to distinquish their chips from the competition, with a couple of years more efficent products would be on the market.

[merchant accounts]

The flip side of a high risk merchant account is that it functions increases the operating costs of a business thus reducing their bottom line.

[damer]

I really doubt that this new low-voltage chip technology could improve cellphone "talk-time" performance.

[folding campers]

The key to the improvement in energy efficiency was to find ways of making the circuits on the chip work at a voltage level much lower than usual, Chandrakasan explains. While most current chips operate at around 1 volt, the new design works at just 0.3 volts.

[Diamond Drill Bits]

I think it is one of the wonderful blog i have gone through ever..It gave many information regarding the energy..I think it would help electrical people more..

[Furniture Stores in Los Angeles]

This is SO interesting, thanks for posting it! It still has to do with energy, lol.

[Filipino Car Lease Broker Los Angeles]

Technology such as this could literally change the world.

[Dentist West Hollywood]

Wow, a pacemaker or device with no batteries would be amazing.

[backup camera]

Is this being used in anything now?

[seo services]

MIT always has such great ideas coming out of it, I wish I had been smart enough to go there! Lol.

[Microsoft Office]

The new President will have to embrace this exact plan if the United States is to avoid economic catastrophe.