Hizook

I ran across this on Gizmodo:

Japanese boffins are now making animations by creating small plasma balls in mid-air. The technology doesn't use vapor or strange gases, just lasers to heat up oxygen and nitrogen molecules: up to 1,000 brilliant dots per second, which makes smooth motion possible. They could be used as street signs, advertising or to create giant plasma monsters to destroy entire cities. Maybe.

My complaints about most blogs is depicted perfectly in this post... In a brilliant scientific reporting style, they managed distill an otherwise ingenious technology into half-witted "appeal-to-the-masses" quips. To make matters worse, their only reference source is a link to a Google translated page (from Japanese) from the National Institute of Advanced Industrial Science and Technology (AIST) in Japan. Normally, that wouldn't be a problem, except for the fact that the translation quality is lacking.

Oh well, enough of my rant about popular blogging style; they did introduce me to a cool technology of which I was previously unaware. Anyway, I scanned through some of the older, English press releases and found a coherently written one in English:

Most of the 3D displays reported until now draw pseudo-3D images on 2D planes by utilizing the human binocular disparity. However, many problems occur, e.g., the limitation of the visual field, and the physiological displeasure due to the misidentification of virtual images.

We have developed a 3D display, which utilizes the plasma emission phenomenon near a focal point of focused laser light. By controlling positions of the focal points in three directions of X-, Y-, and Z-axes, real 3D-images constructed by dot arrays were displayed in the air (3D-space).

Ah, much better description! From what I can gather from the Google translated page (referenced by Gimodo), they are using a solid-state femtosecond laser (with peak power in the sub-tera-Watt range -- around 0.1 TWatt) firing at 1kHz to produce 1000 plasma "dots" per second -- effectively allowing them to display continuous motion. Overall, they mention an average power consumption of 200 Watts.

This system differs in concept from Holographic 3D Video (shown below left) and what the Japanese release called "human binocular disparity" systems, such as the Actuality Perspecta (two images, below right) from Actuality Systems. I've talked briefly about these in an older post here.

This laser-induced plasma display is compelling. However, there are obstacles with it becoming as useful as competing 3D video systems -- namely resolution and color. The resolution is limited by the ability to generate high-power laser pulses. Current systems just can't handle too many tera-Watt pulses before overheating. Also, the color of the plasma is related to the excited gas composition and temperature. The temperature is easy to vary (just change the laser output power or duration). Unfortunately, it is not so trivial to change the gas composition, especially at the location of individual 3D pixels at 30-60 Hz for seamless video (at least, this seems difficult from my perspective).

Still, cool stuff! (Bonus points for anyone who can find a video...)

UPDATE 8/15/2007:   I found an older English press release that describes their earlier system with a nano-second laser and 100 pulses per second.  The diagram below shows how the laser produces the plasma "ball."

My buddy Francesco works at GA Tech on this project.

Scientists at the Georgia Electronic Design Center (GEDC) at Georgia Tech are investigating the use of extremely high radio frequencies (RF) to achieve broad bandwidth and high data transmission rates over short distances.

The research focuses on RF frequencies around 60 gigahertz (GHz), which are currently unlicensed—free for anyone to use—in the United States. GEDC researchers have already achieved wireless data-transfer rates of 15 gigabits per second (Gbps) at a distance of 1 meter, 10 Gbps at 2 meters and 5 Gbps at 5 meters.

“Our work represents a huge leap in available throughput,” Pinel said. “At 10 Gbps, you could download a DVD from a kiosk to your cell phone in five seconds, or you could quickly synchronize two laptops or two iPods.”

I don't really have any cool pictures to contribute (maybe Francesco can help me take a few -- maybe one of the latest chip die?), so this one from the press release will have to do.

So, perhaps the cooler application (well, other than downloading DVDs to your cell phone) would be the ability to stream video from a Mac Book Pro to dual 22-inch motors... Since Dual-Link DVI is only 7.4 Gbps (wikipedia), this technology should let you stream all that video data in real-time... Awesome!

Oh, and some people online have been complaining about the poor range... Well, I do believe they are mis-informed. This is actually done intentionally. The reason that portion of the spectrum is unlicensed is (in large part) because of the following graph (from here).

If you look closely, you'll see that there is a large amount of RF attenuation around 60 GHz due to oxygen in the atmosphere. This means that signal strength will fall off quickly, making it bad for long-range commercial transmission (which is probably why it is unlicensed).

I believe this frequency is even used for satellite-to-satellite communication for a similar reason -- to prevent ground-based interference and interception. So yeah... I suppose the small transmission range is good for personal privacy...

Cool technology, but thank goodness I don't have to work with Cadence for chip design!!!

I've been thinking a bit about the use of computer learning (and classification) algorithms to design mechanical and electrical systems (and also computer science systems).

My curiosity was first piqued when I read about some of Jason Lohn's work at NASA's Evolvable Systems Group. The electrical engineering projects were interesting for me, particularly the evolved antenna designs (because of their aesthetics).

The two left antennas are the ST5-3-10 and ST5-4W-03. They were designed for the Space Technology 5 Project (ST5) by using evolution/genetic algorithms. It is curious since the computer-generated designs have higher gain, fewer matching requirements, more uniform coverage, and a 40% shorter design cycle compared to conventionally designed antennas for the project! The two right antennas were designed using evolutionary algorithms when mission requirements were altered. Conventionally, the entire design cycle would have to be repeated, at substantial cost and delay (basically double the initial efforts). However, the evolutionary algorithms resulted in substantial savings in both.

This curious UHF helical antenna was designed for the Mars Odyssey Spacecraft. Its evolutionary design resulted in a product that had the same technical specifications of the conventionally designed antenna, but occupying 1/4 the volume (a big deal on spacecraft). You can read more the evolved antenna designs from the group's website, or a pair of papers here and here (both local copies). Antenna design has largely been a "black-art," trial-and-error process (at least from my perspective), so this is a very cool application!

The NASA group also uses evolution algorithms to design MEMS resonators, fault-tolerant FPGA circuits (such as a fault-tolerant 3-bit multiplier), and even analog circuits.

The circuit above shows a circuit for an evolved 85 dB amplifier. Basically, they use a genetic algorithm (with SPICE as the evaluator) to successively evolve a better circuit.

All of these developments are great achievements, and interesting applications of "computer learning algorithms." But how much faith can we really have in the designs?

Now in general, I figure "computer learning algorithms" include genetic algorithms, cellular automata, hidden Markov models, neural networks, among many, many others. These algorithms are applied to a number of applications across a many disciplines. But the problem (at least from my perspective) is that they lack determinism. By the very nature of their design (at least in most instances), the human observers remain unaware of the underlying low-level functioning of the system (particularly at intermediate steps). Rather, we rely on known evaluation functions/simulators (such as the SPICE circuit simulator, or RF simulations) to evolve/learn the design. Then we test an actual prototype to verify the design. We may even subsequently analyze the design and learn some new deterministic insight (hey, why not).

So why the skepticism with lack of determinism -- especially if the design passes simulation and physical testing? Well, the first reason is that it is often impossible to perform comprehensive testing for large systems (both 'good' system states as well as spurious ones such as noise, feedback, brown-outs, etc). Now, I will concede that most of these problems can be handled by a human expert analyzing the system (in a deterministic way) after the learning algorithm. But I think the major problem are the number of non-experts who do not perform the post-learning analysis, but rather use the algorithms in a "black-box" fashion. These programmers see the tools as a "quick" solution where you throw data in and an answer magically pops out.

Well, what happens when non-experts apply learning algorithms to critical (meaning lives at stake) applications (such as bridge or aircraft design)? Yikes.

I suppose I'm being a bit rough on the machine learning algorithms -- they are a great tool. They produce interesting, thought-provoking results that teach us a lot. However (particularly for a non-expert such as myself), I think they should be used cautiously... I believe I'm going to try to focus on deterministic methods henceforth.

**The inspiration for this rant began when Matt Reynolds asked me "Why use hidden Markov models?" in regard to my recent publication of Hambone. The most direct answer, of course, is that Dr. Thad Starner wanted us to use GT²K, a hidden Markov model toolkit, and that the related work authors used a hidden Markov model to perform their classification. Since I am not a machine learning person (though some of the co-authors are), the "black-box" approach (again, from my perspective and given my background) didn't necessarily sit well with me. Of course, I would have preferred to use PCA (principle component analysis) to determine a more deterministic classification method, but time just wasn't permitting... And such is life...

UPDATE (July 9, 2007): I found a copy of a preliminary paper by the investigators of this technology. You can find a copy online here, or locally here. It is a pretty interesting read. They are using a lithium niobate (LiNbO₃) guided-wave acousto-optic modulator (at a higher bandwidth) instead of more typical surface acoustic wave (SAW) modulators. Apparently their lithium niobate modulator can also diffract along two axes and rotate polarization (for later selective filtering). I has assumed they were using "normal" SAW modulators, so this technique is "new" to me... To tell the truth, all of this stuff is still a bit over my head, but I'm learning quickly... This stuff is fascinating!

There was an article on MIT's Technology Review entitled "Holographic Video for Your Home." It discuses, at length, developments by Michael Bove (et. al.) to create smaller holographic displays using novel modulators and semiconductor lasers to generate dynamic holograms. First, lets take a look at a few of the images (and descriptions) from the article:

	Bove's team has developed a high-bandwidth, multi­channel light modulator (left) that converts a one-gigahertz electrical signal into a holographic video. The signal makes the clear crystal in the center of the device vibrate at specific frequencies.
	When laser light shines into the crystal, as seen in this image, the vibrations change the directions and intensities of the emitted light, creating diffraction patterns--the basis of a hologram
	An earlier holographic-video system (left) required racks of equipment to drive the modulators and moving mirrors.
	A new modulator decreases the number of optical components needed (left). The optics of Mark III will eventually fit into a box half a meter long.

Now, a few choice quotes:

The Media Lab's video holograms appear to float above a piece of frosted glass. An electronic device behind the glass, called a light modu­lator, reproduces interference patterns that encode information about the pictured object. Laser light striking the modulator scatters just as it would if it were reflecting off the object at different angles.

The Media Lab's video holograms appear to float above a piece of frosted glass. An electronic device behind the glass, called a light modu­lator, reproduces interference patterns that encode information about the pictured object. Laser light striking the modulator scatters just as it would if it were reflecting off the object at different angles.

Aware that this sort of display wouldn't cut it in consumer applications, Bove and his team have laid out plans for the next generation of the system, Mark IV. Mark IV will use a set of powerful red, blue, and green semiconductor lasers to shine full-color videos onto a screen the size of a computer monitor. A prototype could be ready within the next couple of years.

Cool stuff. It mentions in the article that early holographic video systems were developed all the way back in the 1980's. This stuff isn't new per-se, but it is being driven by the "shrinking" size of optical components. I have a number of colleagues who work in this area, and I tend to share their belief that revolutionary breakthroughs are eminent. Among the potential breakthroughs: very small and sensitive sensors, semiconductor-laser based TVs and projectors, contact-lens displays, optical computers, etc. Micro-optical components are really at the point that MEMS was during the early 1990's -- on the verge of a major break-out.

I should probably mention that I like the holographic video system better than other 3D video generation techniques -- mostly due to the lack of moving parts. There are a number of systems that use persistence of vision (or advanced projection techniques) via a spinning mechanism. For example, the Actuality Perspecta system from Actuality Systems.

The Perspecta Spatial 3-D System v1.9 creates 10”-diameter three-dimensional imagery.

Of course, I suppose there is some benefit to having a commercially available unit already in production -- unlike the holographic video systems being developed at MIT. For example, check out the medical image from Actuality.

Either way, these "spinning" displays have severe limitations: they're not physically interactive (you can't touch the images); their refresh rates are limited by mechanical inertia (they don't scale well); they're prone to mechanical breakdown (moving parts); and finally, if they fail or break there will be a lot of kinetic energy to dissipate (notice the thick glass to protect it)!

Anyway, that's just my 2¢.

Roland Piquepaille had an interesting write-up on an Israeli teams' efforts to create a miniature, inductive-powered robot to explore the inside of the human body (blood vessels in particular).

At first, I thought the above image was the actual prototype, in which case I would have been very impressed. Then, I read on...

You can see above "an artist's rendition of what the tiny submarine robot would look like." (Credit: Unknown, via the Jerusalem Post)

"Artist's rendition," indeed.

The researchers stress that the project is an "interesting development, but it has a long way to go before it is used in medicine." Solomon says that the tiny robot could be controlled for an unlimited amount of time to carry out any necessary medical procedure. The power source is an external magnetic field created near the patient that does not cause any harm to humans but supplies an endless supply of power for it to function. The robot's special structure enables it to move while being controlled by the operator using the magnetic field.

Yep, makes sense to use inductive/magnetic coupling for power, create the smallest such robot (that I know of), and claim great aspirations about fighting cancer and malignant tumors, but where is the functional prototype? "A long way to go," no doubt. I think there needs to be a little more substantiation when making such extreme claims. Publications are always good, as are active project websites.

Anyway, I've have seen a fair number of actual gastrointestinal robots for endoscopies (aka, "inside the human body") that use a similar idea. Most use peristaltic (undulating) motions to prevent tearing, lacerations, and punctures. However, some designs still use "gripping" feet mechanisms. One great example is a hybrid design from Carnegie Mellon (project homepage here and press release here). Most importantly, they are already performing testing in plastic tubing and pig intestines.

	This prototype shows a "six-legged" (actually, they apply pressure to the gastrointestinal tract walls) robot.
	This is another prototype (apparently 3-legged?) robot in an actual capsule.
	An overview of the CMU approach. Click on the image to enlarge.

Another great example comes from a New Scientist article entitled "Worm-inspired robot crawls through intestines." It features work performed by a multi-national European research team, and again, this research has produced working prototypes (with videos)!

	A prototype crawling through a pig intestine. Be sure to check out the video here.   (Local copy of the video here)
	A more recent prototype being held in a researcher's hand. Check out the video of the legs actuating here.   (Local copy of the video here)

I'd like to take a moment to rant about what seems to be "bad" journalism. First, I'd like to pardon Roland Piquepaille, as he represents a fair amount of skepticism in his article. However, the remainder of the blogosphere is taking this "advancement" on dogmatic faith as an absolute truth. Maybe the Israeli team's work is a huge advancement, but maybe it isn't (there just isn't enough data at this point to make a clear distinction). Either way, the blogosphere isn't helping science by disseminating (mis)information. This is just plain wrong, and I hypothesize (along with some colleagues) that this may be true of several other recent "advancements" featured on many blogs.

I make a pledge to represent anything I post as accurately as possible, and to the best of my knowledge (what little I have). Hopefully the other blogs will too.

It is always interesting to see "common" electronic components used in fun, novel ways. Often, these methods have been well-known for quite some time and may even be the premise behind dedicated, more expensive sensors. However, the methods have a tendency to become vogue every few years. The decision to use these "alternative" components can be boiled down to a few key factors: cost, space, novelty, and usually the most important, availability (it's a nice alternative when you don't have the dedicated sensor sitting on your desk). I'm going to discuss two examples; can you think of any others?

As a first example, let's look at one of the most visible (and often annoying) components that everyone is familiar with -- the light-emitting diode, or LED. To get a handle on what I'm referring to, watch this video (local copy here) of an LED array being used as a touch-sensitive panel.

This project comes from New York University, and as you can see, they're using the LEDs to emit light while simultaneously using them as touch sensors (actually, reflective light sensors). You see, in the "normal" operating mode (forward-biased), LEDs consume power by emitting light. However, LEDs can be hooked up backwards (reverse-biased), where they will exhibit light-dependent characteristics similar to photodiodes. This means that the LED can be used to both transmit and detect light simultaneously! Of course, the LED can't actually be sensing and transmitting light simultaneously -- the effects are mutually exclusive since they use different operating points. However, if you're toggling the LED between modes fast enough, it will have the appearance of simultaneous emission and detection.

Now, if a single LED is hooked to a microcontroller as shown below, it can be used to both transmit (emit light) and receive (sense light) using just two I/O lines (by using the RC characteristics of the microcontroller I/O lines, no analog-to-digital converter is required). This means that virtually any LED can serve as a bi-directional communications port! This was explored in a paper (local copy here) entitled "Very Low-Cost Sensing and Communication Using Bidirectional LEDs" from Mitsubishi Electric Research Lab.

Of course, this capability has also been known for quite some time, as illustrated by a patent entitled "Proximity sensor with a light-emitting diode which simultaneously emits and detects light" (Patent #4564756) from 1984.

My second example of a "clever" sensing device using a cheap component comes from µblog.  Basically, Nick couldn't find a thermistor (a cheapish part that changes resistance with temperature), so he used a run-of-the-mill diode instead.  Basically, the reverse bias leakage current of a diode varies with temperature, as shown in the graph below. 

Thus, the diode can be used as a temperature-dependent resistor (or essentially a thermistor) when held at at a constant reverse-bias voltage.  Nick used a custom analog sensing and control circuit (to control a CPU fan), but consider a circuit similar to the one above (with "reverse bias" and "discharge").  You could use the same circuit (again, knowing the RC characteristics of the microcontroller I/O pins) with the regular diode in the place of the LED to sense temperature.

Again, this technique isn't exactly "new."  The same concept is empolyed in many higher-end microcontrollers with internal temperature sensors. 

So, I've listed two examples of clever sensing techniques using cheap, common components.  Can you think of any others? 

I'm on a quest to build the ultimate home electronics & computer lab. I do a ton of work from home: from school assignments, PhD research, to hobbyist electronics design and testing. In the process, I've accumulated a lot of useful tools. Most have been purchased with personal funds (no sponsored or donated equipment), but some pieces are on loan from friends/colleagues. Be sure to help me out with recommendations. Most importantly, what are the contents of your "home lab?"

Let's get started with a picture of the "main" part of my lab. This picture shows the bulk of my equipment.

Here's a list of the hardware in my computer/electronics lab.

	This is my Fujitsu T4210. It is actually a custom build with 1GB RAM, 60GB Hard-drive, Bluetooth, WiFi, DVD/CDRW and Spare Battery bay, Intel Dual Core Duo T2300E (1.66Ghz) processor. With the extra battery in the place of the DVD/CDRW drive (my normal configuration), I get a healthy 6+ hours of battery life -- enough for a solid day of classes.
	Dual Samsung SyncMaster 226BW 22-inch monitors. These wide-screen monitors are hooked up to the Shuttle PC sitting behind (on the right). The best part is the combined resolution of 3360x1050 -- that's enough for about 4 full-sized applications to be running simultaneously!
	This is a Shuttle SN25P10. It was acquired on Ebay, and it came with dual Western Digital WD360 (36.7GB) Raptor Hard-drives and a NVidia GeForce 6600GT. It currently runs Ubuntu Feisty (7.04) with RAID-1 on the two WD360s (across 3 partitions). It is also connected (via a 7-port USB hub) to the stack of external hard-drives sitting next to it.
	On the far right, you can see a stack of 5 Western Digital WD2500JB 250GB IDE Hard-drives sitting inside some Addonics USB enclosures. These are hooked up to the shuttle PC via a 7-port USB hub. They form a RAID-5 array with a 1TB capacity. The RAID-5 is (obviously) done in software via Ubuntu.
	This is my awesome oscilloscope, the Tektronix TDS2024B. This is a 4-channel, 200Mhz, 2 Giga-samples/second digital oscilloscope. It has a USB port on the front for a flash drive, and can connect to a PC (for control and data) over another USB port on the back. It has FFT capabilities built-in, so that's a great plus! I still haven't had the time to play with all the cool functions of this scope, but I look forward to tinkering some more.
	This is the Weller WESD51 Soldering Iron. I went through about 3 crappy Radio Shack soldering irons before indulging in a good, digital, temperature-conrolled iron from Weller. Being able to switch iron tips from very fine (for SMT parts) to very fat (for 14-guage wire) is great.
	This is the Kepco Bipolar Operational Power Supply/Amplifier (Model BOP 1000M). It has a range of ±1000V at ±40mA, and a bandwidth of about 1.5kHz. This isn't actually a piece of my personal equipment; it is on loan from a professor at Georgia Institute of Technology, and is being used in my current research.
	This is the Meterman LCR55. It is a glorified resistance-inductance-capacitance meter. The great thing about this meter is its range. It measures resistances from 20MΩ to 20Ω, capacitances of 2000μF to 200pF, and inductances from 2000H to 200μH. The one in the photo is actually a loaner from Matt Reynolds, as mine is currently in transit (after borrowing his, I had to get one of my own).
	This is a Fluke 112 multimeter. It is your run-of-the-mill multimeter from Fluke. It works great.
	These are function generators. The one in the picture is the Instek SFG-2110 Function Generator. I'm borrowing this one from Matt Reynolds for my research. Since this is such a fundamental piece of equipment, I've ordered my own: the EZ Digital FG-7002C (which should arrive any day now).
	My trusty Samsung S730 Digital Camera. Its function is self-explanatory.
	A Fluke 80K-40 High Voltage Probe. It can handle voltage up to 40kV (though it has a relatively small bandwidth of a few kHz).

So those are the main components of my lab. I have a number of other small components, including a Dremel, PIC Development tools (I have two ICD2's), hack saw, a wet lab with chemicals, tons of electromechanical components like motors and solenoids, discrete electrical components (two large "dressers" full), power-supplies and wall-warts, etc. There is also a wireless router, cable modem, two old computers (with CRT monitors), two old laptops, an old oscilloscope, etc.

Can you think of anything crucial I'm missing...?

Back in Fall 2004 and Spring 2005, I took the Computer Design I and II courses in Computer and Electronics Engineering at the University of Nebraska (at the Peter Kiewit Institute).  The goal of these classes was really cool, even if somewhat antiquated:  Build a functioning 8086-based computer by hand, on your own!  Well, here was the final product...

In the picture above, you can see the fruits of my labor.  On the left is an LCD with keypad and 8279 keypad/display controller.  On the right, you can see the "motherboard" (if you want to call it that).  Starting from the top:  passives, 8284 clock generator, 8086 microprocessor, 74HCT573 address/data latches, 74HCT245 buffers, GAL22V10 programmable address decoder and chip select generator (this is really more modern; in the past, this would have been done with discrete logic), 74HCT573 latches and associated LED bar displays,  two 5256 32K-by-8bit RAM (I think), DS1287 real-time clock, another GAL22V10 decoder for chip selects, 8259 Interrupt Controller, 16550 Serial UART, SP233 TTL-to-RS232 level converter, DB29 as serial connector.  The two ZIF sockets held ROM chips, but they have long since been removed and erased.  This device could be controlled by either the keypad or over the serial port via a terminal on a modern PC (either hyperterminal on Windows or minicom on Linux).  The whole thing was fully functional.  We also built an 8051 microcontroller board with USB, but it was relatively uninteresting.

So, this post is about wire-wrapping (the process of doing point-to-point wires to make electrical connections).  Well, take a look at the back of the boards...

This is the reason it took several months to build!  Obviously, modern circuit design is much more efficient (and can handle much higher frequency signals).  With modern PCBs, you design the boards with CAD software, send the design to a fab-house, get them back, and then just solder down the components.  With wire-wrapping, you must make each individual connection by hand.  It is a rather painstaking, tedious process that is prone to error (it took about two nights to wire-wrap this whole board, then many more to debug and add functionality).

Even though my 8086-based computer was an impressive piece of work, it still pales in comparison to John Pultorak's Apollo Guidance Computer. 

This report describes my successful project to build a working reproduction of the 1964 prototype for the Block I Apollo Guidance Computer. The AGC is the flight computer for the Apollo moon landings, with one unit in the command module and one in the LEM.

I built it in my basement. It took me 4 years.

If you like, you can build one too. It will take you less time, and yours will be better than mine. 

You, sir, are one crazy (good) engineer, but I think I'll pass on building an AGC of my own...  It's amazing that a computer like this got us to the moon.  Returning should be easier now, right?

The image above should illustrate the point of this device to any electrical engineer. You'll recognize the basic topology of a "typical" MOSFET (Metal-Oxide Semiconductor Field-Effect Transistor), only the channel is not a semiconductor material, but rather ions contained in an aqueous solution. The operation of the device relies on the electrowetting effect (see Wikipedia). Basically, the hydrophobic (yellow) layer normally (with no potential applied) attracts the oil. When you apply a potential, the electrowetting effect "attracts" the aqueous solution (particularly the charged particles in it) to the channel where they can conduct electricity. Beautifully simple.

For those a bit more versed in "techese" the operation is best discussed by the inventors (Kim and Steckl):

Control of current flow in the LiquiFET is obtained by controlling the presence in the channel of one of two fluids: a conducting, fluid, aqueous electrolyte or an insulating, fluid, non-polar oil. Fluid location in the channel region is determined by the electric field applied to the gate through the competitive electrowetting effect. We used a water droplet with KCl as the conducting medium and an oil film with a non-polar dye for visualization purposes.

The device structure consists of the glass substrate, a dielectric-covered transparent ground electrode, source and drain metal contacts, a hydrophobic insulator layer of amorphous fluoro-polymer, a hydrophobic/hydrophilic grid, the two fluids (electrolyte & oil), and the top gate electrode (Au wire).

For zero gate bias, the low surface tension oil preferentially covers the low surface energy hydrophobic insulator, forming a thin film that excludes the high surface tension polar electrolyte solution. When a negative bias is applied to the gate, the resulting field across the hydrophobic insulator effectively increases its surface energy and reduces its hydrophobicity, attracting the polar water molecules and electrolyte anions to the insulator surface. The water increasingly displaces the oil layer with increasing bias and forces the displaced oil to the side regions of the structure. When the gate bias is removed the oil returns to its original position due to the capillary force acting to minimize the energy of the system.

Anyway, the FET characteristic is unmistakable in the Vds-Id curves...

This is pretty cool stuff -- very novel. You can find their paper here (or here locally). I'd like to thank New Scientist for bringing this technology to my attention.

Imagine having sunglasses which can change color or opacity on-demand with the flick of a switch! This is the goal of Chunye Xu (and team) at the University of Washington. In the image below, the lenses block 55% of incident light (on the left) and 95% (on the right). There is also a video of the lenses in action (locally here).

The sunglasses use a thin-film of electrochromatic material. Wikipedia has a decent page that discusses the effect. It is also described in the news release as follows.

Researchers made the glasses using electrochromic materials that change transparency depending on the electric current. Many groups, including the UW, are developing such materials for so-called "smart windows" that could soon be used in energy-efficient homes and offices. Most smart windows use liquid-crystal technology or inorganic oxides. Those materials are expensive to produce and require a constant or frequent injection of power to hold their tint. The UW glasses are based on a new type of smart window using organic, rather than inorganic, oxides. These are cheaper to manufacture and require less power.

The prototype glasses are powered by a watch battery that attaches to the glasses frame, and the wearer spins a tiny dial on the arm of the glasses to change color or shade. The lenses were created by sandwiching a gel between two layers of electrochromic material. Applying a small voltage moves charged particles from one layer to another, and changes the transparency. Once the glasses are a certain tint they will stay that way without power for about 30 days. A single watch battery is able to power thousands of transitions, Xu said.

Also, the prototype shown above only produced blue hues. This isn't exactly "new," as viologens (another electrochromatic material) produces a bluish hue. However, the big breakthrough comes in the ability to create red and green hues, which Xu and colleagues have done, as discussed in their academic paper (or here locally). This opens the technology to some new applications (from another news release).

By combining the polymers of different colors into multiple layers and supplying different levels of current from the batteries in the sunglasses, a wide variety of different colors can be produced in the lenses, Xu says.

If the power consumption is low enough, this technology could be used (instead of the magnetic bead technology) for the next generation of electronic-ink (or E-Ink) and electronic-paper. Electronic-paper is already being touted as one of the "next big things" for smart materials. In fact, we've already seen commercial products, such as the Sony Reader pictured. Electrochromatic electronic-paper would allow for color rendering yet still boast the very low power consumption required for prolonged operation. Further, the electronic-paper would be read (much the same as "normal" paper) in lit conditions. If thin-film solar cells can be integrated into the electrochromatic layers, the electronic-paper could be entirely self-powered!

So a question that could be posed is "How is this better than the lenses that change transparency when you go from indoors to outdoors?" Well, those lenses use the photochromatic effect, where the lenses change color/opacity based on the amount of incident light. In this case, the lenses are sensitive to UltraViolet (UV) light. When outdoors the lenses become darker in the presence of UV, and when indoors the lenses become lighter as the UV is no longer present. There are 2 or 3 problems with photochromatic lenses (compared to these new electrochromatic lenses).

• When you're in the car, the car's windshield blocks much of the UV light, causing the lenses to become more transparent. Since people spend most of their outdoor-time inside their car (well, at least I do), the lenses' ability to turn dark is useless.

• You can't adjust the lens settings to be brighter or darker based on your personal lighting preference. My threshold for bright light is much less than most other peoples'. This might have something to do with too much time in front of a computer monitor...

• For those who care about fashion, you can't change the color of photochromatic lenses to match your attire. (I personally don't care about this though.)

Hopefully people aren't naive enough to fall for a "number of cores" competition, much like the Intel/AMD "number of Mega/Giga Hertz" competition a while back. Intel and AMD are now planning quad-core and 8-core chips for future desktop PCs, and I'm sure this will usher in a new era of performance computing. This is all well-and-good, but why a 1000 core processor?!?

OK, clearly a 1000-core chip isn't going to be used for general purpose computing (at least not at this time). But power-performance ratios are really the future of embedded and special-purpose computing. Consider the Asynchronous Array of Simple Processors (AsAP) project, which implements 36 cores (6x6 array) of 32-bit processors (9-stage pipeline, 54 32-bit instructions, and 16-bit datapath).

AsAP processor operates at 475 MHz; and each processor dissipates 32 mW while executing applications, 84 mW while 100% active, and 144 mW worst-case at 1.8 V. Most of AsAP's area (66%) is for the core which is a high area utilization. Each processor occupies 0.66 mm², which is more than 20 times smaller than the other traditional processors such as ARM. AsAP processor also achieves more than 5 times higher performance density and energy efficiency compared with others, as shown at below.

One really unique thing about the AsAP (at least from my perspective) is its asynchronous FIFO buffers that allow each core to talk to its surrounding ones despite the differences in clock rates/latencies (see this paper -- or locally here). In this fashion, a core is only active (consuming power) when it has data in its buffer to process. Each processor is programmed with a small code snippet (in C or assembly), which can be assigned (or auto-mapped) into the AsAP. Now for the really cool part. With 36 cores each with a small instruction set (54 possible instructions), 64 words of RAM, 128 words of ROM, you can do some really amazing things!

• FIR Filters
• Signal Convolution
• Sorting
• CORDIC sin, cos, arcsin, arccos, arctan
• Pseudo Random Number Generators (LFSR)
• CRC Calculations
• Huffman Encoding
• FFTs
• JPEG Encoders (9 cores, 224mW @ 300 MHz) (shown below)
• A complete 802.11a/g wireless LAN base-band transmitter (22 cores, 407mW @ 300 MHz)

All of these applications offer huge power and performance savings compared to other, traditional solutions (as explained in this presentation -- or locally here). The fully-functional, complete 802.11a/g wireless LAN base-band transmitter implemented in 22 cores at a mere 407 mW is perhaps the most impressive application to date! Check out the diagram below to see what each core is doing!

But wait, the title of this post mentions a 1000-core processor, not 36. Well, that is all just a matter of scaling. Additional cores can be (easily?) added into the array.

We have designed a 0.18 μm CMOS chip that was fabricated during the summer of 2005 (the 36-core chip). Early testing in the fall of 2005 has shown it is fully functional! We believe it is the highest clock rate fabricated processor designed in any university. A 13 mm x 13 mm chip utilizing the exact same design in 90 nm CMOS would contain more than 1000 processors and be capable of more than 1 Tera-op/sec peak performance.

Just imagine what would be possible with 1000 cores! (see the image below) One could argue, "This can be done with FPGAs." And you'd be correct (in fact, they already have the AsAP architecture loaded into FPGAs for testing). However, having the processors in silicon gives you additional benefits.

• Performance increases (FPGAs are limited because of general-purpose nature)
• Power decreases (FPGAs logic is always active, the AsAP only runs when it has data)
• Easy functional decomposition and mapping
• Programming each core (and then auto-mapping) in C or assembly instead of Verilog and VHDL

I hope they work with a chip-fab to make some of these chips available to hobbyists. These chips would be really valuable in robotics, as you could activate and deactivate the different cores depending on your power budget (not to mention the power savings inherent with the AsAP design).

As a side-note, I almost worked with Dr. Baas on these project(s) for my PhD. Ultimately, I decided chip design wasn't my thing (at least not for a PhD, but it would have been great for a MS). However, my good friend Toney is doing some work on the project. Last I knew, he was working with the test setup, shown below. It should be obvious that this group at UC Davis does some amazing work. Good job guys!

You can find out more about the AsAP project here.

OK, this post is a bit long, but that's because the topic is really revolutionary (and fits nicely into several of my personal interests). First, I'll go over an abridged version with pictures and videos, then get into a little more scientific discussion.

Basically, the idea is that you place a diode in a liquid (or on its surface), and then apply an alternating current (simple AC) signal across the liquid. The diode will then be propelled through the liquid. Don't believe me, check out this video. You can then hook up the diodes in a large ring, and it will spin like a gear (video). Use a LED and the circuit can light up. Use a zener diode, and you get constant velocity. Use a photodiode, and you get light-controlled velocity. Make the diodes stationary, and you can pump the liquid (even at the micro scale). Use multiple diodes to make a (micro) mixer. Use lots of diodes to make micro robots (not done yet, but conceivable). It's all quite interesting, and it can be done in a fish tank!

   

OK, now for the advanced section. The principle works using the parasitic capacitance of the diode. During half the AC cycle, the diode will be forward biased, and nothing happens. During the reverse biased half-cycle, a parasitic capacitance will cause a charge to be built up across the diode terminals. The electric field from the capacitor causes fluid to flow via electroosmosis (basically, charged ions in the fluid move in response to an electric field). The motion of the fluid around the diode causes an "equal and opposite" force on the diode, propelling it forward. Of course, there are many factors that affect the rate (such as fluid pH). The whole process is described (along with the applications previously mentioned) in the article.

This all begs the question, "What sort of propulsion can you get, and what other devices can be used?" Can we cause rotation using one diode? Can we inhibit or enhance the effect for better control? Can it be controlled by a microcontroller? Can it be done over large distances using RF coupling instead? What if a bipolar transistor (NPN or PNP) is used? What sort of motions are produced by them? Obviously, a parallel-plate capacitor would have a zero net electroosmotic flow, but can the geometry be altered so that even a capacitor can exhibit this effect? All very good questions, and I intend to build a fish tank to test them out!

Anyway, the article also includes a "supplementary material" section that has a whole bunch of movies, which are described below.

All movies are real time except Movie M7 (double-speed) and in WMV format. The semiconductor devices on Movies M1-M6 float in a large Petri dish full of 1e-6 M NaCl solution. Two AC-powered electrodes from thin wire are placed above and below (or left and right) of the scale. E(ext) was 120 V/cm and the frequency was 1 kHz. The scale on the back is spaced at 1 cm in movies M1 to M3, M6 and at 0.5 cm in movies M4 to M5.

• Supplementary Movie M1: A 1-mm long semiconductor diode propels to the right when the field is turned on. Field direction is horizontal.

• Supplementary Movie M2: A low magnification view of a miniature diode propelling. The diode (small speck near the bottom of the first frame) covers a distance of almost 5 cm in about 43 s. Field direction is vertical.

• Supplementary Movie M3: Two LEDs orient vertically when the field is turned on, light up and begin propelling. Note that the diodes move in opposite directions because they are oriented oppositely with respect to their electrode polarity.

• Supplementary Movie M4: "Diode-powered gear" begins rotating when the field is applied due to the directional propellant force of the diodes attached around the O-ring.

• Supplementary Movie M5: LED-powered gear without external illumination. The diodes rotate the gear and light up - note that the lit diodes are always to the left and right of the gear as they are the ones receiving most power from the surrounding vertical field.

• Supplementary Movie M6: Motility of photodiodes controlled by exposure to light (from laser pointer). The photodiode velocity drops sharply when they are illuminated and is restored when the light is turned off.

• Supplementary Movie M7: Separation of two types of particles within the channel of the microfluidic device in Figure 1. At first, only an AC field is applied and both types of particles rapidly move to the right by diode pump driven flow. However, under the simultaneous action of balanced AC + DC external fields, the small particle (1 µm amidine-stabilized latex) very slowly moves to the right, while the large particle (2 µm sulfate-stabilized latex) begins moving to the left. Compare with Figure 5.

I've also saved local copies of the movies (M1, M2, M3, M4, M5, M6, and M7). I originally found out about this article from New Scientist.