SemiSerious

After my longest absence since starting SemiSerious, I thought it was about time to get posting again. For the handful of people actually reading this blog, I feel terrible about not keeping it going through the early summer. I hope to ramp up again in August with weekly posts appearing regularly by the time September rolls around.

Unfortunately, I have nothing semiconductor-related to post today. Instead, I thought I would just give an update on where I’ve been. My father passed away in June. After travelling back to Saskatchewan with my wife and two daughters for the funeral and other things, we took an extended family vacation. It just seemed like the right time for the kids to be with their maternal grandparents.

Now that I think of it, maybe a post about Dad is connected to semiconductors. I guess you could say he was there on the ground as consumer products transitioned to solid-state technology. Around the time I was born in the mid-sixties, my father was off work from the seed cleaning plant because of a workplace injury. He decided to use his unproductive time to enroll in a home self-study course in electronics - radio and television technology and repair. That led to a moonlighting business fixing TV’s (in our basement), then to a job at the local electronic parts distibutor, and eventually to his owning and operating that same parts business.

It was watching Dad in the shop that ignited my own interest in electronics. He always enjoyed having his kids around, so there were lots of opporunities to see him working. Dad taught me the resistor color code and Ohm’s Law.

In a lot of ways, it was difficult for Dad to watch the industry migrate from vacuum tubes to semiconductors. He was fascinated by the transistor and the IC, but the ever-decreasing cost of electronic products unsettled him. Cheaper TV’s meant fewer would ever be repaired. But it was not just about the loss of business for himself and his colleagues. It meant that more things were just thrown away and piled up in landfills.

My Dad was born in Southern Saskatchewan to immigrant parents (first generation North Dakotans) and spoke only Norwegian until picking up English at school where he completed the eigth grade. He was 83.

OmniVision made the first announcement of a backside illuminated CMOS image sensor intended for the mass market two weeks ago. It seems their lead (at least in the PR world) did not last long as Sony made a similar claim this week. Gizmodo has front-side versus backside captured images along with a neat animated graphic comparing the two sensor structures.

As one news site pointed out, Sony is the “800 pound gorilla” in the image sensor business, so their announcement adds credence to the BSI concept. So there are now two manufacturers claiming to have developed a backside technology. And these are not just middling players. OmniVision and Sony are both “number ones.” OmniVision took over the lead in CMOS image sensors from Micron (now Aptina) last year while Sony is the world’s largest supplier of image sensors when you factor in both CIS and CCD.

Another news article from the past week reminds us that backside illumination is not a new concept in imagers built for scientific applications. NASA’s Mars Reconnaissance Orbiter is presently capturing digital images with an e2V CCD system. The probe’s High Resolution Imaging Science Experiment (HiRISE) instrument employs image sensors from the British company.

But the OmniVision and Sony sensors will be built for consumer applications. I’m guessing we might not see too many of these on the market for at least a year. Beyond that, though, it might be only a matter of another couple of years before the majority of imagers shipped “swap ends” for getting the light into the photodiodes.

At the Intertech-Pira Image Sensor Conference a few weeks ago, I suggested that Google Street View might create a post-cameraphone demand boom for imagers. My reasoning (if it’s fair to call it that) was that the folks at Google have given us a useful virtual tourism concept simply by driving camera trucks around about 40 major US cities. But it struck me that Google’s approach is neither time nor location scalable. If you want fresh images, you need trucks rolling 24-7 just to achieve something like annual updating. Perhaps Google (who has the money for such an undertaking) will replace its current rolling camera approach with a vast array of permanent camera installations. Street View could then be “live.”

What about new locations? More trucks more time…I think you get the picture. With apologies to Little Rock and Milwaukee, I am interested in virtually touring around some places outside the good old USA. Well Google now has a competitor, and although you can only view three locations today, one is outside the US. MapJack offers Sausalito and San Francisco, California along with Chiang Mai, Thailand. Even with by giving us the first location beyond the borders of the US, it would hardly be fair to say that MapJack has leap-frogged Google Street View. However, MapJack does provide a better user experience and a convenient additional search window to browse for shopping or dining. I first heard of MapJack from Scott Bourne on This Week in Photography.

I don’t understand any of the nuts and bolts holding MapJack together, but it appears to depend on Google’s own Map grids and location-based information like dining and other entertainment options for anyone who intends to take a step beyond virtual touring. It will be interesting if MapJack gets large enough for Google to think about shutting down access to its map grid information.

Surely part of virtual tourism is a way to maintain your carbon-neutrality. But that won’t be the case if Google and MapJack are spewing truck exhaust to capture images. Vast networks of low-cost IP surveillance cameras are the only way to go.

The expression “turn off your clothes” might soon be applied more literally than it once was for the geeky guy wearing the bright yellow Hawaiian shirt. MIT Technology Review recently reported on a technology that might some day extract power from the clothes we wear. It’s all part of a new movement to harvest energy from the various sources of kinetic energy that surround us.

Georgia Tech materials science professor Zhong Lin Wang and his team have adapted their zinc oxide nanowire technology to a new way to generate power. They have grown the nanowires on top of a polymer surface a bit like hair. The flexing and movement of the hairs can generate electricity from the movement or even the breath or heartbeat of the person donning the new fabric. According the the MIT report, as much as 80mW are available per square metre of the fabric. The energy conversion ability of the nanowires has been known for some time, but their use in clothing is a new development.

One other source of power that has attracted the interest of the US military generates power from the motion of a person walking. Simon Fraser University professor, Max Donelan, is the lead researcher for this presently untapped source of portable power. For a 21st centruy soldier in the field, a few extra minutes of battery life in a GPS receiver might actually save his life.

The zinc oxide nanopartilce clothes and the SFU Biomechanical Energy Harvester represent the larger-scale devices within the broader field of environmental energy harvesters. I have previously posted (see “More Buzz” and “Nanotechnology from Gigatools”) about nanogeneration of power, and I believe this will be the enabling technology for a wide range of devices from wireless sensor networks to self-powering drug delivery robots and diagnostic devices inside our bodies.

In a somewhat related development, researchers are proposing self-cleaning clothes that harness the power of the sun to do the dirty work. The secret is in the titanium dioxide coats the textile fibers. According to this work, the titanium oxide spontaneously combusts, vaporizing the stain on top of it. Perhaps future developments will allow some of the heat of that combustion to be recycled so that the sloppiest among us will be able to get the best charge out of our clothes.

The Mobile World Congress has wrapped up, and there were plenty of interesting things to see and new announcements made.  But unfortunately I was stuck at SI’s booth fighting off the hordes of people clamoring to learn more about our semiconductor analyses   I did, however, make it out of the booth long enough to present with my Portelligent colleague, Jeff Brown, about the “Evolution of the Smart Phone - Systems to Semiconductors”.  While it started a bit slow with only two non-SI/Portelligent people there at the beginning, it was standing room only in a hot, noisy room by the time we concluded, which was fabulous!  A lot of questions were asked as well, which is always a good sign that people liked what you had to say instead of just wanting to bolt for the doors once you stopped speaking.
 
The presentation examined three smart phones from HTC (Universal, TyTN, and TyTN 2) and compared/contrasted them with three smart phones from Nokia (N90, N93, and N95).  Jeff covered the system level side of the presentation where he talked about the different feature sets and overall board sizes.  One of the interesting areas that he discussed was how HTC came into the smart phone market as a PDA vendor, so they had a full blown (and large) first offering, and then had to figure out how to meet the phone market criteria, the big one being size.  On the flip side, Nokia entered looking at a phone and had to grow to compete in the smart phone arena with the successive offerings, mostly in terms of functionality.  In the end, the two latest offerings from both companies are quite comparable from a system level point of view.
 
This is where I took over and looked at some of the components that made the phones functional.  One thing that surprised me was how loyal both companies were to the component manufacturers.  Many people think that cost is king in cell phones, and this may be true for entry- and mid-tier offerings, but other than changing the Bluetooth from TI to CSR, the major components for each company remained consistent.  Looking at the different chips that were used, my conclusion was that HTC had a very efficient design where they removed duplicate functionality from devices and were able to dramatically reduce the number of chips used in their phones while still increasing the functionality.  Nokia’s phones resulted in a very effective design, taking advantage of the economies of scale that a company of that size has at its power and they have a great opportunity to optimize in the future.
 
A copy of the slides can be downloaded at www.semiconductor.com/mwcpres.

- Gregory A. Quirk

Yesterday marked the dawn of a new era of scaling for CMOS devices. I hope that statement fits with all the hype around the launch of the 45nm microprocessor from Intel. In fairness, though, switching to metal gates and high-k dielectrics represents an important milestone in semiconductor technology. Gordon Moore’s well-worn comments are appropriate:

“The implementation of high-k and metal materials marks the biggest change in transistor technology since the introduction of polysilicon gate MOS transistors in the late 1960s.”

Intel released TEM images of its 45nm PMOS transistor. The embedded SiGe source / drains are evident, but we’ve all seen those before. For his keynote, Paul Otellini seemed confident that we could not tell what Intel’s secret sauce in the dielectric and workfunction metals was, but they had something to hide on top of the gate stack. We can’t see the CMP surface, but that’s a no-brainer for a replacement gate (aka gate last) technique. I think they want to keep the capping material on top of the gate electrode hidden until they present at IEDM (Paper 10.2). At about 9:30 on the morning of December 11 in Washington, DC, Kaizad Mistry will open the trench coat on the 45nm HKMG process. Considering the secrecy Intel has been able to maintain on this process, I think the trench coat is a fair reference. Intel deserves full marks for keeping its employees and vendors quiet for so long. Steve Jobs is jealous, I’m sure.

Or maybe he’s not. SI will be opening the kimono on the Penryn this Friday to all our clients participating in the analysis. Intel will still generate lots of excitement at IEDM (at least if you aren’t analyzing 45nm or buying one of our reports).

There’s been much hype and there will be much more.  The Penryn MPU hit Time magazine’s list of best inventions of 2007 along with the iPhone cover girl. Intel may covet the iPhone socket and may well win it for next generation devices, but there may could be another connection to Time’s list. NASA’s methane powered rocket might one day look to the Intel marketing machine as a steady source of fuel.

If you have attempted to post comments to articles on this site and are wondering what happened to them, please accept my apologies. Technical difficulties have left many recent comments in limbo. I promise that your patience will be rewarded next week when you will, at last, see these on SemiSerious. I think many of these will spark some meaningful debate on technology.

Thanks for all your insighful comments on our blog posts and please keep them coming. If for no other reason at all, it is very rewarding to know that our posts are being read.

A couple weeks ago I scratched rather shallowly into the topic of turning motion energy into useful electricity (check it out). I was impressed with the development of a technology that could convert power line frequency mechanical vibrations into about 40mW from a device fitting inside a one centimetre cube. This, I argued, would start to put wireless sensor networks on the map.

Once again through Technology Review, I have now discovered that someone has taken the miniaturization of the concept to a much smaller level. Technology Review recognized Xudong Wang with its TR35 award for young innovators. His energy conversion device is small enough for implant into the body to power biosensors. I have not been able to confirm the power density or typical dimensions, but one nanoamp was reported in Technology Review. The device produces its electrical current by converting ultrasonic frequencies. I cannot say for sure, but I assume the intention is to externally provide ulstrasonic energy to the biological entity containing the device.

So unlike the devices discussed previously in SemiSerious, Xudong’s invention may not be intended for harvesting wasted background energy from power line vibrations so unavoidable in our electrified world. On the other hand, optimization of the process used to produce the ZnO nanowire arrays appears to be leading to huge efficiency improvements. It isn’t a stretch to think that a small device on a white water river adventure in your blood stream or in your gut could provide enough power for many sensors and related circuits.

On a loosely related note, Georgia Tech appears to be upwardly mobile, at least in university rankings which puts them at number seven for US public institutions. In fact, they are tops in nanotech.

I will not be the one to draw any connections, but the White House decision to support the ITC ruling to embargo smart phones containing Qualcomm chipsets could lead to a dumbing down of the U.S. (at least in consumer technology). Enter the blue jean and black turtleneck bedecked superhero with the iPhone. Steve would argue - and would receive support from many users and gadget reviewers - that the iPhone’s predecessors were not really smart anyway. Fortunately for the technical intelligence of American consumers, Apple has designed alternatives into their smart phone. The iPhone uses an Infineon baseband processor and Samsung applications processor.

Followers of the iPhone - especially here in Canada where it is not yet available - will know that hackers are still working at unlocking the functionality for use on service providers other than AT&T. According to hackers, the toughest nut to crack in the iPhone has been the Infineon S-Gold 2 baseband processor. The S-Gold 2 chip contains an ARM926 core and EDGE modem. Since Apple has been very stringent about control of the software and use of the iPhone so far, perhaps this was part of the decision to go with Infineon in the baseband socket.

Unfortunately for Apple, there seems to be an effective hardware hack for the phone using the TurboSIM from Bladox.

So look for increased iPhone sales in the short term and for the second generation device if it comes before the end of the year.

Any discussion of Major League Baseball that includes the word “green” usually means one thing - money. In the SF Giants case, though, it means using solar energy. A solar farm incorporating 590 Sharp Corp. panels was installed before San Francisco’s AT&T park hosted the All-Star game. The installation was reported to cost about $1.5 million to be funded by 15 million PG&E customers since the panels are on their grid. The total power available is about 120kW.

Obviously, Mark LaPedus was onto this story a long time ago. But I am adopting the excuse that Barry Bonds will be breaking Hank Aaron’s home run record imminently. Therefore, this news is still very topical.

 So even if Barry Bonds isn’t clean, a portion of the power for the Giants’ ballpark is.