Tech Talk

For four blissful years, the exchanges that trade in bitcoin operated within a cloud of legal uncertainty, awaiting the day when the regulatory beast would awaken to its new opponent. Now, that day has come. This week, the Department of Homeland Security took a quick and hard strike at MT Gox, the largest online exchange, serving its payment processor Dwolla with a warrant (later obtained by ars technica) to seize the MT Gox account. Dwolla is one of the preferred ways of getting government currencies in and out of MT Gox and the news caused temporary tremors throughout the Bitcoin community. Trading volume spiked and the exchange rate bobbled down to $106 before climbing back up.

These exchanges are the supply lines for Bitcoin, which has steadily increased in value over the last year. With one supply line down in the U.S., many people will be looking for alternatives. And soon they will find them.

Two brothers, Mark and John Russell are scheduled to unveil a new automated exchange kiosk this weekend at a Bitcoin conference in San Jose, CA. They're calling the machine Robocoin. It will provide a physical place for people to convert their dollars to bitcoins and vice versa.

As has been reported this week, the Kepler planet hunting space telescope, may have to end its mission earlier than hoped, due to the failure of the system that keeps it pointed in the right direction. That system consists of four reaction wheels, which are basically electric motors attached to fly wheels. By speeding up or slowing down, they transfer angular moment to the satellite, rotating it around its center of mass.

Kepler’s mission is find exoplanets by staring, unmoving, at small patches of space and look for periodic dips in the brightness of the stars there. Those dips could mean the presence of planets. But without at least three working reaction wheels—Kepler is down to two—the satellite can’t steer it’s gaze or keep it from gently drifting in the solar wind.

According to David Cooper, CEO of Microsat Systems Canada Inc., in Ottawa, Ont., a provider of reaction wheels for small satellites, there are two main classes of things that can go wrong with reaction wheels—mechanical and electrical. And that means Kepler's pointing system was probably damaged either by the shock of launch or in space by radiation.

I dream a dream of perfect calibration: A single chip embodying the “metrological triangle”—with built-in, reproducible, quantum standards for the volt, ohm, and ampere, completely defined by just two universal constants, Planck’s and the electron charge.

We’re two-thirds of the way there: Thanks to quantum Hall resistance and Josephson voltage measurements, the ohm and volt can be practicably defined within 10 parts per billion. Both, however, depend on empirical measures of current, typically via watt-balance measurements that are accurate to only 100 ppb. (Watt balances, in their turn, depend on the definition of the kilogram, which is still evolving, related initiatives like a Compton-wave definitions of mass.)

That leaves the amp, waiting for a way to produce exquisitely accurate currents.

Single-electron pumps (SEPs—not related to the “Someone Else’s Problem” invisibility field invented by Douglas Adams) produce extremely precise currents, sending electrons leaping one at a time from quantum dot to quantum dot across a series of potential barriers. It’s something like a line of backpackers—each bearing just one electron charge—picking their way across a stream single file by hopping from stepping-stone to stepping-stone.

Researchers use oscillating voltages to drive the current, but there are built-in challenges. Lower frequency pumps—metallic fixed-tunnel barrier systems and normal/superconducting hybrid turnstiles—work in the megahertz range. These can move only a few million electrons per second through the pump, and produce picoampere (pA) currents that are difficult to detect. Semiconductor-based tunable barrier pumps operate at gigahertz frequencies, producing a thousand times as much current—but electrons need some finite time to step from stepping stone to stepping stone. Like the backpackers trying to cross the stream too fast, the jostle and bump and some fall into the stream, disrupting the regular flow of current.

Researchers at Britain’s National Physical Laboratory have tackled the problem with a graphene-based, nanoscale device (described in Nature Nanotechnology) that produces very precise current flows from about 20 MHz to 400 MHz.

Malcolm Connolly (with colleagues at NPL and Cambridge’s Cavendish lab and collaborators from Lancaster University) constructed the pump from shaped monolayers of graphene on silicon dioxide. The heart of the pump is an archipelago of quantum dots stretching between two graphene peninsulas. The pump path is flanked by graphene electrodes connected to a gigahertz radio frequency generator and plunger gate circuits that control the flow of electrons into and across the pump.

The accuracy of NPL’s graphene SEP does fall off a bit a higher frequencies—victim of the same collision of shortening driver cycles and electron transfer times that affects other gigahertz approaches. But, say the researchers, the effects are smaller and the measured accuracy is much higher at a given frequency—about an order of magnitude better than metallic pumps can manage. As a result, the researchers predict an error rate in the range of 10 ppb at 90 MHz—on par with the accuracies achievable for resistance and potential standards. Ten pumps operating in parallel would deliver 100 pA “with metrological accuracy.”

Thus, say Connolly et al., “a realization of the quantum metrological triangle in a single graphene device is also now within sight.”

Image: Malcolm Connolly, NPL/Cambridge

After four successful years in space, the Kepler planet hunting space telescope is in serious trouble. A key component that keeps the spacecraft pointing at the right patch of stars, a reaction wheel, has failed. Kepler went into space with four of these and needs three, but this new failure leaves it with just two. Even so, at least one Kepler expert thinks there may be a way to save the satellite.

The US $600 million telescope hunts for exoplanets in our own galaxy. It uses a 95-megapixel camera to register slight dips in stellar brightness that signal a planet's passage across its host star. So far the mission has found more than 2700 candidate exoplanets, several of them in the habitable zone of their stars. To find these it must continually stare at a patch of sky containing some 4.5 million stars.

It’s this staring that’s in danger with the loss of the reaction wheel. The device is used to gently point the telescope in the right direction, using other patches of stars as a reference. Reaction wheels are electric motors attached to fly wheels. By speeding up or slowing down, they transfer angular moment to the satellite, rotating it around its center of mass. Kepler’s have to be pretty good ones. According to a report at the 2011 IEEE Aerospace Conference, the telescope must be able to stare for more than 15 minutes at a time with a stability of 0.009 arc seconds for each axis of rotation. (By comparison, a comma in an Apollo mission manual left on the moon is about 0.001 arc seconds as seen from Earth.)

Kepler’s first reaction wheel failure was in July 2012. Earlier this month another one started to go wonky, registering signs of friction. I’ll give a more detailed description of how reaction wheels fail and what can be done about it tomorrow, but for now, here’s Stanford University’s News service interviewing Scott Hubbard, a consulting professor of aeronautics and astronautics about saving Kepler:

Q: How might NASA engineers go about getting Kepler functional again?

A: There are two possible ways to salvage the spacecraft that I’m aware of. One is that they could try turning back on the reaction wheel that they shut off a year ago. It was putting metal on metal, and the friction was interfering with its operation, so you could see if the lubricant that is in there, having sat quietly, has redistributed itself, and maybe it will work.

The other scheme, and this has never been tried, involves using thrusters and the solar pressure exerted on the solar panels to try and act as a third reaction wheel and provide additional pointing stability. I haven’t investigated it, but my impression is that it would require sending a lot more operational commands to the spacecraft.

The mission was set to continue through 2016. Kepler’s loss could be a blow to other instruments such as HARPS-N at the Telescopio Nazionale Galileo in the Canary Islands. HARPS-N, which IEEE Spectrum’s Rachel Courtland visited in 2011, is used to confirm the exoplanet status of objects Kepler spies.

PHOTO: NASA

A new version of D-Wave's supposed quantum computers could help NASA hunt for alien worlds or enhance Google's mammoth search engine before the end of the year. The U.S. space agency and Internet search giant have joined a growing list of high-profile customers using the latest D-Wave machine despite lingering skepticism from quantum computing experts.

Photo: D-Wave Systems

It’s launch season in Silicon Valley; birdies leave their nests and nascent companies leave their incubators, accelerators, or classrooms and announce their products to the world. Until recently, software, particularly apps, dominated launch season; credit cards and parental funding can't take a hardware product from idea to manufacturing. But in recent years, with 3-D printers and other rapid prototyping tools readily available, and funding within reach, thanks to Kickstarter, Indiegogo, et al., a lot more would-be entrepreneurs are daring to build things out of plastic and circuits as well as code. And they’re making launch season a lot more interesting.

This week Haxlr8r, one of the first hardware-only accelerator programs, launched its second class. Haxlr8r is especially interesting for its international approach—it requires its teams to spend most of the three months they spend under Haxlr8r’s wings in Shenzhen, China. Explained Zach Hoeken Smith, program director for Haxlr8r: “Shenzhen is the best place in the world to design and build products.” Being there, he says, allows entrepreneurs to quickly change their choices of components, boards, and other parts because they can get their hands on just about anything they might put in a final product within minutes or hours, rather than days or weeks. Entrepreneurs there can also quickly find the right manufacturing partner, and sit down with experts to refine their designs to maximize their manufacturability in the early stages, not after prototypes are complete.

“We believe,” says Smith, “that you don’t wait until you get a million bucks on Kickstarter before you work with manufacturing.”

In a first, an unmanned plane today successfully took off from a U.S. Navy aircraft carrier. In another first, a small, (mostly) unmanned jetliner recently flew through British commercial airspace.

The navy trial involved a Northrop Grumman X-47B plane, one of just two test models built as part of a US $1.8 billion program. As it is the policy to test each element of the plane's autonomous systems independently, today's flight concentrated on the catapult-assisted takeoff. It went without a hitch. This time, however, the flight was controlled remotely by a human, and the plane landed on a conventional runway, that is, on dry land.

Future flights will test the X-47B's ability to fly itself and land on a carrier. Overlapping optical and other sensors, together with GPS connections and internal maps, are designed to allow the plane to navigate and avoid mid-air collisions by itself. Human judgment takes over only when the plane is taxiing--a tricky shell game, played on a deck crowded with moving vehicles.

The plane's sleek, batlike airframe is designed to elude most radars, carry heavy bomb loads, and travel about twice as far as most manned fighters. That's just what the doctor ordered, because it would allow the Navy's floating islands, multi-billion-dollar behemoths all, to lurk at a healthy distance from shore-based missiles and other high-tech weaponry. Next year the Navy plans to test the unmanned plane's ability to refuel in mid-air, which it must do to manage truly long-distance flights.

There are many advantages to going pilotless. For one, it saves humans from getting killed or captured. For another, it allows a plane to make hairpin turns and other maneuvers that generate g-forces high enough to drain the blood from any head, right stuff or not. Finally, it saves on the weight of the many systems that protect the pilot.

Britain's less operatic but perhaps equally consequential feat took place last month, but was announced yesterday. A Jetstream airliner flew 800 kilometers with no intervention from the human pilot except during the takeoff and landing phases--all in commercial airspace. That British air-safety regulators allowed the test suggests that they think unmanned airliners are close enough to justify serious planning.

The flight was organized by a joint project of the U.K. government and local aerospace vendors called the Autonomous Systems Technology Related Airborne Evaluation and Assessment, or Astraea, which happens to be the name of a Greek goddess (the Brits have always had a lot of such coincidences).

Back in December, Lambert Dopping-Hepenstal, the director of Astraea, told Spectrum that the twin-turboprop, 18-seat Jetstream served as a "flying laboratory," with engineers on board to monitor everything. He said that the test vehicle hadn't been specially configured to take off and land autonomously, but that it could do so--as can any modern airliner.

"Here, though, you’ve taken the pilot away, and preprogrammed it to fly a route," he said. "In the event of comm failure, it will look after itself and follow the rules of the air in avoiding conflicting traffic."

Complete autonomy would require quite a lot of advances in the ability of a plane to sense and avoid danger in the air and on the ground,  Dopping-Hepenstal added.  "I’m not quite sure you’re rushing to an economic solution here, but rather doing things that are now difficult or impossible to do with manned aircraft--long endurance, extreme environmental environments. I’m not convinced by the short-term economic argument today."

An airline passenger advocacy group has jumped into the fray over Boeing's 787 -- calling for the FAA to scale back the Dreamliner's airworthiness clearance and for an investigation into Boeing's FAA-approved fix to its battery fire problems.

FlyersRights.org, calling itself "the largest airline passenger organization," on Wednesday submitted a formal petition to FAA chief Michael Huerta and Secretary of Transportation Ray LaHood. The group cited expert testimony that questions Boeing's battery fix, now being implemented around the world to get 787s back in the air within weeks. FlyersRights also questioned FAA's giving Boeing the ability to create some of its own regulatory tests for the Dreamliner's lithium ion batteries.

The FAA, FlyersRights president Paul Hudson says, "gave Boeing essentially all the authority to approve their own batteries. And it's backfired. Now there's an interest by all parties to make the best out of that situation. But the public isn't well served by that."

At NTSB hearings  last month, Boeing Vice President and top engineer of the 787 program Mike Sinnett testified that Dreamliner's FAA certification process "was the most extensive effort in our history." However, he also later noted of some of the 787's battery tests, "In retrospect I believe we don't feel that it was conservative enough."

Included with FlyersRights' petition was an independent assessment of Boeing's battery fix by David Zuckerbrod, head of the Baltimore-based battery consulting firm Electrochemical Solutions.

Zuckerbrod says makers of lithium ion batteries work within some of the most stringent quality control standards in industry today. Even still, he says, "They make cells by the billions, and yet the failures are in the hundreds or thousands every year."

D-Wave's supposed quantum computers have attracted plenty of skepticism alongside some serious interest from huge corporations such as Google and Lockheed Martin. Now recent testing has shown that D-Wave's machine can indeed beat standard computers head-to-head in solving certain problems.

The D-Wave computer performed up to 3600 times faster than a high-performance machine running IBM software while solving an optimization problem, according to the New York Times. D-Wave's machine only proved slightly faster than the standard computing on two other optimization problem tests, but the results still seem encouraging for the company's future prospects.

“Ours is the first paper to my knowledge that compares the quantum approach to conventional methods using the same set of problems,” Catherine McGeoch, the Beitzel professor in technology and society at Amherst College in Massachusetts, says in a press release.

McGeoch, a founder of "experimental algorithmics" in computer science, was enlisted by D-Wave as an outside consultant to devise tests for comparing the company's machines with conventional computers. She plans to present a co-authored paper on the test results at the 2013 Association for Computing Machinery (ACM) International Conference on Computing Frontiers in Ischia, Italy on May 15.

The latest success of D-Wave's computer does not mean that it or quantum computers in general will replace consumer laptops anytime soon. Instead, D-Wave's machine has proven good at tackling specific optimization problems with one best solution—puzzles similar to the "traveling salesperson" problem that asks for the shortest possible route to visit a list of cities exactly once before returning to the original city.

Lots of folks have been trying for years to figure out how to get today’s kids interested in Science, Technology, Engineering, and Math (STEM) careers. A pair of serial entrepreneurs—Brent Bushnell, profiled in Spectrum's 2012 Dream Jobs Special Report, and Eric Gradman, whose latest venture is Two-Bit Circus—think efforts so far have been missing something—art. So they’re mixing in art with engineering to create a new acronym, STEAM, and a new venture, the STEAM Carnival.

Bushnell and Gradman envision STEAM Carnival as taking classic carnival games, which have always appealed to kids, and updating them with technology, particularly the kinds designed to wow youngsters, like lasers, tesla coils, motion capture systems, robots, and shooting flames. An unexpected combination? Not really, Brent Bushnell’s father, Nolan Bushnell, worked as a carnival barker before launching the video game industry by founding Atari. (A story I hold near and dear to my heart, having worked as a barker for a very low-tech carnival game myself before becoming a technology journalist.)