Power management is ever-more critical as consumer applications eclipse computer applications and the whole dc-dc converter arena vibrates with innovation. Meanwhile, two trends should have a considerable impact on 2005.
First, ultracapacitors will find more applications as prices drop and engineers understand their capabilities relative to batteries. Second, fuel cells will appear in consumer products, from notebooks to active headsets. Instead of plugging in a product to recharge it, you'll squirt it full of methanol.
The Sept. 29, 2004 Electronic Design Update (ED Online 8878) featured a story about a bridge-power UPS the size of a bread box packed with 22 2600-F electric double-layer capacitors. Designed by Maxwell, this device targets telcos, data centers, and hospitals. While the announcement stirred up lots of reader interest, Maxwell hasn't been resting on its laurels.
The company's 1.6- and 2.3-kW PowerCache products are the first devices to come online when the grid fails. They provide 48-V power during startup or transfer to generators, microturbines, fuel cells, or other alternative primary and backup energy sources. Before ultracaps, bridge-power units used mechanical flywheels to store energy. (In fact, most of them still do.)
To build an ultracapacitor, designers need an electrode with enormous surface area and an electrolyte with tiny ions. Maxwell solved the electrode problem in the early 1990s with a form of finely divided carbon that provides 2000 m2 of surface area per gram and an electrolyte with ions small enough to penetrate the nanoscale porosity of that carbon. Maxwell's initial soda-bottle-size 3-kF ultracapacitors cost tens of thousands of dollars to make. But now, 2.7-kF ultracaps cost around $50, and prices continue to fall.
Ultracaps store about 5 W-h (18 kJ) of energy per kg and can discharge currents up to 1000 A at 2.5 V. A good battery might store 20 or 30 W-h/kg but could never return it to the system at high currents. Also, unlike batteries, an ultracap can be repeatedly charged and discharged without damage.
The latest news from Maxwell concerns 15-V, 58-F Boostcap ultracapacitor packs and modules that contain six standard D-cell form-factor ultracaps. Intended for regenerative braking, energy storage, wind-turbine blade-pitch control, and smaller fail-over systems, the pack and the module both provide a simple solid-state solution to buffering short-term mismatches between the power available and the power required.
The packs are shrinkwrapped, and the modules have an aluminum enclosure. Dimensions are roughly 9 by 3 in. Both weigh about a pound. Packs cost $127 and modules go for $200 in small quantities.
Supposedly, 2005 will be the year of the direct methanol fuel cell (DMFC) for portable consumer gear. DMFCs for notebook computers will be able to power a machine for up to 10 hours on around 100 cm3 of methanol. Smaller DMFCs without even a pump or fan could be used for personal media players and smaller devices.
Last summer, Toshiba announced the prototype of a thumb-size DMFC that can be integrated into devices as small as digital audio players and wireless headsets for mobile phones (see the figure). It puts out 100 mW, enough to power an MP3 music player. It will last up to 20 hours on a single 2-cm3 charge of highly concentrated (99.5%) methanol. Tanked up, it weighs less than 10 g.
Toshiba's prototype is passive. While active DMFCs use a pump and fan to feed methanol and oxygen into a cell stack, passives rely on the concentration gradient to deliver and circulate methanol and oxygen in the stack. Actives are larger and better suited to notebooks.
The company had to expend considerable engineering effort on the structure of the fuel cell's electrodes and polymer electrolyte membrane so it could work with nearly undiluted methanol. Generally, fuel cells work best with a methanol-to-water concentration of less than 10%, but that would require far too large a fuel tank. Hitachi and NEC also have demonstrated DMFCs.
Also, fuel cells aren't about to leave lithium batteries in the dust. This is good for chip makers who have been working on charging circuits and "gas-gauge" ICs for rechargeables.
Just last month, Maxim introduced a fuel-gauge chip that calculates the remaining capacity for lithium-ion (Li-ion) and Li-ion polymer batteries. The chip integrates its own fuel-gauging algorithms, taking some burden off the host. (The chip will still send low-battery warnings to the host, of course.)
Earlier in the year, Texas Instruments announced a two-chip gas-gauge set that promises to make battery-life readings more realistic. The problem is that battery self-discharge rates vary with temperature, how long the battery sits between uses, and whether the battery is fully discharged before recharging or not.
TI's proprietary algorithm exploits longer periods of inactivity to recalibrate the "starting position" for state of charge, also eliminating self-discharge effects. It continually updates full-charge capacity by comparing state of charge before and after the system load is applied.