SemiWiki – Page 791 – The Open Forum for Semiconductor Professionals

November 8, 2014

Xiaomi Already #3 in Smartphones Behind Samsung and Apple

Xiaomi Already #3 in Smartphones Behind Samsung and Apple
by Paul McLellan on 11-08-2014 at 7:00 am
Categories: Mobile

The Q3 results are mostly in now and there are two big stories. The iPhone 6 of course, which started shipping just before the end of the quarter. The other story is China, now the #1 mobile market in the world. These have led to very different stories for different companies.

Samsung is basically flat with Q2, but that was a big drop. Smartphones are such a large part of Samsung’s business it has affected overall corporate earnings. They are down a full 50% on revenues that are down around 20%. “its smartphone business lost ground in an intensely competitive global market”. They are still #1 with almost 25% market share (but they were over 30% recently) shipping around 78M smartphones.

Apple is still #2 of course, and they have said that they are shipping everything they can build. They sold 10M phones in just the first week of the iPhone 6 launch. Since they launched on 19th September that was all in this quarter. In total they shipped nearly 50M phones and very slightly gained market share. Q4 will probably be up since it will be the first full quarter with iPhone 6 and then, assuming Apple stick to making only one new phone per year, will creep down until this time next year.

The most amazing result is that #3 is Xiaomi, shipping 18M phones for close to 6% market share. Considering how recently they were created, this is amazing. All those phones sold in Asia, mostly China. They are eating Samsung’s lunch.

Chinese manufacturers also hold #4 #5 #7 #8 positions. Lenovo is #4 (and this is still without the Motorola Mobility acquisition from Google, which hasn’t closed yet). If it closes in Q4 and the numbers are all merged I assume they will be #3 in Q4, but Xiaomi is growing fast. Lenovo shipped 17M phones.

Huawei is #5 shipping nearly 17M phones. In fact some analysts have them ahead of Lenovo since the numbers are so close. And almost exactly the same number is LG (the second Korean entry into the table). Then Coolpad and ZTE.

Bringing up the rear are Sony, Microsoft and Motorola Mobility all with around 9M phones shipped, more or less.

What I think is happening is that at the high end of the market in the US, Samsung and Apple are doing very well. You certainly see a lot of Galaxy and iPhones around, although it is always wise not to assume that what you see in silicon valley generalizes to middle America. In US Samsung has 36% market share and Apple 26%. And as late as this time last year, Apple and Samsung’s profits together were 109% of the entire mobile industry, everyone else in aggregate losing money.

But Samsung used to be #1 in China and now Xiaomi is, at a much lower price point. The iPhone 6 didn’t launch in China until October so it has no impact on the Q3 numbers but Apple is an aspirational brand that everyone wants if they can afford it. Apple is wildly profitable, with Samsung struggling. Some estimates of Apple’s profits are as high as 85% of all the profit for the whole industry, but its challenge is going to be to keep growing as fast as the market so that its share stays relevant to 3rd parties (it is currently 12%). Android has close to 85% market share (across all the manufacturers) so Apple need to keep iOS desirable. Today, people seem to develop first for iOS and then port to Android, and Apple would like to keep it that way.

It doesn’t look as if Samsung’s position is likely to change very fast either. Kim Hyun-joon, senior executive at its mobile division, said:
the outlook for the fourth quarter in smartphones “wasn’t looking good” as competition will get fiercer. “That will probably hurt our profitability.”

Another small story is Microsoft (the old Nokia handset division) which had dropped out of the top 10 to fall behind Motorola but snuck past them again this quarter. But I still find it hard to believe that Microsoft, and thus WindowsPhone operating system, can get out of the basement at around 3%. As well as Microsoft, Samsung and ZTE do ship some phones with the WindowsPhone OS but less than a million units together. Blackberry (what used to be called RIM) is even lower at under 1% and also has to be doomed, although with their strong position in government and some corporations they presumably are an acquisition target at some point.

Oh, and Amazon? They took a $170M writeoff on their smartphone inventory. The Fire just didn’t seem to light. Amazon don’t publish numbers but some analysts reckon they only sold 35,000 in the first 25 days. Not quite up there with Apple selling 10M in a week. I reckon that means Apple sold as much as Amazon in the first 40 minutes.

Bluetooth Smart Crashes IoT Party

Bluetooth Smart Crashes IoT Party
by Tom Simon on 11-07-2014 at 7:00 pm
Categories: Arm, IP

When you hear the term Internet of Things the first image that comes to mind is, well, a lot of things talking over TCP/IP, maybe using WiFi or god forbid Ethernet. But upon closer examination it seems that Bluetooth is crashing the IoT party. This was driven home for me when my girlfriend gave me a FitBit Flex. For those unfamiliar, it is a tiny capsule of electronics that goes onto an inconspicuous wristband. It can help you track your activity and your sleep. Its accuracy and efficacy is probably fodder for another discussion. But it provides gratifying feedback.

FitBit has done some very clever things. It is dead simple to use, and it links directly to your phone or PC via a small dongle they provide. The real service goes through their website, where your data is aggregated and viewable. Though your phone can show you this info, the app must be signed into their site to work properly. So the FitBit is clearly a ‘thing’. And the internet is involved, but this device relies heavily on Bluetooth Smart, formerly BLE. Thankfully the Bluetooth SIG decided to get away from 3 letter acronyms for this technology.

Being a bit of a hardware hacker I was immediately curious about what was inside. Fortunately ifixit.com posted a teardown showing the device’s insides. It’s mostly what you’d expect: LiPo battery, ST’s ST32L151C6 ARM cortex M3, Nordic Semiconductor’s nRF8001 Bluetooth chip, an accelerometer, haptic feedback device and LEDs, etc.

Space is clearly at a premium. The whole thing is very compact and it looks like lots of 0402 and smaller SMT devices are crammed in. After seeing this I researched Bluetooth chips a bit. The growing trend is for the Bluetooth radio to be combined with the MCU. This is a very interesting development. Of course you need a software stack to run the radio, and certainly there will be application code for whatever the ‘thing’ is going to be doing. In one fell swoop you can a single device that has most everything you need to build a ‘thing’. You can use I2C to read sensors, you have GPIO’s to operate LEDs or other output devices. And, it can talk to your phone or computer!

Each company has a different name for it: Bluetooth MCU, Bluetooth SOC, etc. But now processor choice is becoming an important part of Bluetooth chip selection. And there are definitely different camps. TI unsurprisingly uses the 8051. Nordic uses ARM – mostly Cortex MO.Broadcom also uses ARM and you can get a Cortex M3. CSR is using something called Xap. And Intel seems to not have one yet. Witness that their Edison board has a separate WiFi/ Bluetooth combo chip. If they do have one it is not easy to find on their website. But when they do have one, I expect it will be Quark or Atom based.

So why did FitBit use a separate radio and MCU? They could have saved a lot of area on their very small board. I can think of a few reasons. Perhaps the application team and the RF team were separate and did not coordinate, as was suggest in an online interview I read recently. Or perhaps they needed the compute power of the Cortex M3 and when they were doing development the Broadcom chip was not available. FitBit is quite loyal to Nordic. Maybe they did not want to switch radios. They certainly have switched processors from product to product. Their initial products used a Freescale ARM core, instead of the ST ARM chip in the newer Flex. Interestingly though they use the TI Bluetooth SOC with an 8051 MCU for their PC dongle.

It will be interesting to see what comes in future products from the combining of microprocessors and radios in one chip. It is already making ripples in the maker community and allowing much easier product development and prototyping in small footprints.

November 7, 2014April 10, 2025

Amorphous Silicon and TFTs

Amorphous Silicon and TFTs
by Daniel Payne on 11-07-2014 at 7:00 am
Categories: EDA

Most ICs are fabricated with crystalline silicon (c-Si), which is a tetrahedral structure forming a well-ordered crystal lattice. There’s another form of semiconductor material called amorphous silicon (a-Si) which has no long-range periodic order. It turns out that a-Si is a great material for the active layer in thin-film transistors (TFTs), used in most of our liquid-crystal displays (LCD) because it allows the manufacture over a large-area.

You can use either a planar or staggered structure for an a-Si TFT transistor:

The gate of a TFT device can either be structured on top or on the bottom:

The a-Si TFT device is sensitive to light, so shielding from incident light must be added:

If you’re interested to learn more about a-Si TFT devices and how to use TCAD simulations, then consider attending a webinar on November 18th, set for 10AM-11AM, Korea local time. It’s during the TCAD simulation phase that process engineers can start to predict critical electrical performance like leakage currents. At this webinar you can expect to learn about:

Basic concepts of a-Si TFT TCAD simulation
- Basic equations
- Physical parameters
- Density of states model
TCAD calibration procedure example
- Process and Device simulations for IV curve generation
- Understanding and use of density of states models
- Understanding and use of probability of occupation function
- Correlation between density of states and current density

Related – SiC and Si Power Devices

Presenter
Silvaco is producing this webinar live for the Korean market, however they will also be archiving the webinar on their web site for people to view. The presenter is Nam-Kyun Tak, a Senior TCAD Application Engineer from Silvaco. Nam-Kyun has been at Silvaco since 2010, where he worked on TFT technology development. Prior to Silvaco, he worked at Samsung Electronics for 6 years doing DRAM development.

Nam-Kyun Tak received B.S. and M.S. degrees in electrical engineering and computer science from Kyungpook National University, Daegu, Korea.

November 6, 2014

What does the Ford Mustang and Intel’s Gordon Moore Have in Common with Local Motors?

What does the Ford Mustang and Intel’s Gordon Moore Have in Common with Local Motors?
by Charles DiLisio on 11-06-2014 at 11:00 pm
Categories: Foundries

1964 Vision, Volume and Moore’s Law

The 1964 New York World’s Fair saw Lee Iacocca, then a young 40 year old General Manager, introduce a car that inspired “total performance” and was for a “young America out to have a good time.” This young America would become the baby boomer generation. The Mustang was revolutionary in its affordability, which was obvious in its sales. Ford estimated it would sell 100,000 Mustangs during that first year; in fact, it would sell more than 400,000. In 1964 big corporations (Ford, GM, US Steel), had big factories and drove down costs through greater volume. The mantra that greater volume spurs lower costs that creates greater demand was observed in 1936 by Theodore Paul Wright describing the effect of learning on production costs in the aircraft industry.

1964 was also when a young engineer, Gordon Moore, then at Fairchild wrote in his journal what would become Moore’s Law to be published in Electronics on April 19, 1965. Dr. Moore observed that in an integrated circuit (IC) that transistors doubles approximately every two years. As a result we have seen an explosion of electronics, computing and communication growth as a result of larger wafers —> smaller geometries —> greater integration —>lower prices —> greater electronic product volumes.

Enter the Experience Curve

In 1966, Bruce Henderson of the Boston Consulting Group (BCG) introduced the “experience curve,” which resulted from work done for Texas Instruments. It holds that unit costs go down as a company gains production experience. The experience curve concept allowed Texas Instruments to underbid rivals by postulating falling unit costs. Ultimately, the concept was to drive costs down while becoming the dominant market player.

The “experience curve” observation (mega manufacturing —> volume — > lower cost —> leading market share) has led companies not only in the semiconductor industry, but also manufacturing companies and even countries to seek volumes and lower prices to gain market prowess and success.

Today, the “experience curve” and its corollary Moore’s Law are being challenged based on the very things they created. Miniaturization, computing power and ubiquitous communication allows a disruptive business model to emerge even in traditional automobile industries where scale and volume have been barriers to entry.

The effects on the automobile industry will reverberate into the semiconductor industry and we need to be aware of this disruption. I suggest you read my previous article on SemiWiKi titled: Viva The New Industrial Revolution! What Etsy, 3D Printing, and Kickstarter Means to Semiconductor Companies?

Local Motors — Maker Revolution leads to an Open Sourced Car

Today, manufacturing does not necessarily need to be based on large volumes, driving down costs. Crowd sourcing, 3D printing, low-cost collaborative design tools, and cloud factories, complex manufactured products can go from idea to production quickly and at reasonable costs. It’s the Maker Revolution combined with Internet physics.

Local Motors is changing the way automobiles are designed, build and sold. Its using, crowd sourcing design, combining this with community tools like Siemens Solid Edge Design 1, and utilizing micro-factories to build interesting, niche vehicles that are C.O.O.L (Community, Open-source, Ownership Experience, and Local). As a result, Local Motors is able to go from idea, design and build in a period of five months, at a lower cost. The implication is very similar to what happened in the steel industry where big steel (US Steel) was challenged by mini-mills (Nucor) resulting in a new, profitable business model for steel production (see “The Rise of Mini-Mills”, New York Times, 1981)

As the semiconductor industry moves from the smartphone platform to the Internet of Things (IoT), it may be good to reevaluate the premise of Moore’s Law and the experience curve.

Look at the XC2V FlypMode Military Assault Vehicle. This DARPA project designed to see how a military vehicle idea could move from idea to production. In a matter of five months the vehicle went from military input to design, manufacturing and vehicle delivery to the military.

Semiconductor Companies — Systems Thinking vs. Moore’s Law

What can semiconductor companies learn from Local Motors? Four observations:

Co-creation or crowd sourcing or 3[SUP]rd[/SUP] party collaboration for IC design— necessary to meet time to market (TTM) issues in the future. Can semiconductor designers be open to crowd sourcing design with customers and non-customers?

More than Moore’s Law — focusing religiously on smaller geometries, greater levels of IC integration that require expensive, large capital expenditures may not be necessary to be successful in the future. Can you rethink integration using programmability (FPGAs, MCUs, MPUs) or utilizing 2.5D or 3D packing?

Think Systems vs. Devices — transitioning to the IoT world, you will need to think systems and be able to develop intelligent sensors which incorporate MEMS sensors, MPU/MCU, memory and some RF.

Value, not Volume — most sensor systems will have to be done faster, designed for very specific applications and at moderate volumes.

This New Industrial Revolution is a revolution of Makers, custom products and regionalization. It isn’t about volume or the learning curve as we know it. We are moving from the homogenous (baby boomer), one size fits all market, to the heterogeneous (millennial), highly fragmented one. Passionate design, rapid time to market, and uniqueness will be paramount as technology is becoming ubiquitous.

If you haven’t watched the videos below on Local Motors, you should. Not just for what they are doing but the message behind the company. This message will drive significant change in products and business models of all categories.

See:

November 6, 2014

Lucio and the Kaufman Award

Lucio and the Kaufman Award
by Paul McLellan on 11-06-2014 at 4:30 pm
Categories: EDA, Events

Tuesday was the Kaufman award dinner. This year it was awarded to Lucio Lanza. Last week I wrote about how Lucio ended up in EDA, although that was not where he finished up. He is currently a venture capitalist running Lanza Technology Ventures, one of the few VCs to make any investments in the EDA/IP/semiconductor space. Also, unlike most Kaufman award honorees, he actually worked with Phil Kaufman earlier in his career.

See How Lucio Got Into EDA

Lucio has invested in many EDA and IP companies, served on their boards and mentored their CEOs. Two that were especially significant were Artisan Systems and ARM. Mark Templeton, the founding CEO of Artisan, and Simon Segars, the current CEO of ARM presented the award. Or as they were introduced, they provided “the entertainment.” Since it was election night they pretended to have a debate format. “How’s that independence thing going?” Simon asked. In case you don’t know, Simon is English (me too!) although he lives mostly in the US since he has kids in high school (actually I think he lives mostly on British Airways). Artisan was eventually acquired by ARM and Lucio was on the board for a time (although Simon was not CEO in that era).

Mark started by giving a little bit of the history I covered in my earlier post about Lucio starting in Milan, going to Olivetti, coming to the US to work for Intel and then leaving to Daisy. He then ended up in the somewhat strange position of working for Cadence as a consultant in the Joe Costello years and also working at US Venture Partners (USVP). At Cadence they made many acquisitions and he also started an IP business (which for some reason never survived his departure). Eventually the conflict of interest between making acquisitions and making investments in companies to be acquired became too much and so he left Cadence to be full-time at USVP. And then he left USVP to found his own venture capital company, Lanzo Technology Ventures. It is a one-man show run out of a little office on University Avenue in Palo Alto. I doesn’t appear to even have a website since everyone already knows Lucio.

Simon talked about many of the companies Lucio had been involved with. There were too many to fit on one slide and many of the CEOs of those companies were there at the dinner. One I talked to during the networking session before the dinner was Dave Stewart of CriticalBlue, which has had Lucio on its board since the beginning. Somehow I know him through the Edinburgh connection where they are based.

Mark talked about Lucio showing up at Artisan’s door with a whole lot of pizza. At the time he was not even an investor. He gave Mark some advice. “Double your prices.” Mark did and…nothing happened except they made more money. “I’ll tell you what is going to happen,” Lucio said. “In 15 years or so the IP market will be bigger than EDA.” In 14 years the lines crossed.

Mark and Simon presented the award to Lucio to a standing ovation and then it was Lucio’s turn to talk. He gave more color on the early years of his career before he switched to the challenges facing the fabless ecosystem. He pointed out that SoCs are not systems-on-chip, they are hardware on chip since the EDA industry has ducked the challenge of the software component and that has to change or the cost of design will become prohibitive. Indeed, the aforementioned CriticalBlue, started as an EDA company, has completely refocused their technology on optimizing embedded software.

In the future, Lucio said, “chips” (or whatever technology turns out to be the future) will have hundreds, then thousands then hundreds of thousands of processors. This is what I call “core’s law” since it is exponential but not really obvious yet since we are on the flat part of the curve still. But just like in the 1970s you could predict chips with tens of thousands of transistors even though there were just 128 or something at the time, you can run the numbers out a decade. Who is going to make designing such a system feasible and cost-effective? Lucio finished by throwing this down as a gauntlet for the whole industry for the future, a challenge and an opportunity. And when companies come along that solve a piece of the puzzle, guess who will turn out to be an advisor, investor, board member. Lucio.

Semiconductor Safety

Semiconductor Safety
by Daniel Nenni on 11-06-2014 at 7:00 am
Categories: Cadence, EDA

Semiconductors and automotive are now like peanut butter and jelly. Certainly you can have one without the other but why would you? I remember when a car first talked to me telling me that the door was ajar. It sounded more like, “the door is a jar” but I got the point. Now my car tells me just about everything including what is wrong with the engine or transmission. Not every semiconductor device designed into our cars are considered mission critical but I beg to differ.

Have you ever driven cross country with four children when the entertainment system failed? Talk about mission critical. Less than a year after I bought my car a light went on saying I should “return to workshop immediately” (it’s German). One of the sensors in the transmission intermittently failed so they replaced the entire transmission ($10,000). Thankfully it was covered under factory warranty because that would have certainly been mission critical to my wallet!

Also Read: ISO 26262 driving away from mobile SoCs

Automotive electronics will change dramatically over the next few years but given the 7 year design cycle (which includes 2 years of test) we are not really talking about bleeding edge or even leading edge technology here. Even so, safety verification is an important part of the automotive semiconductor design cycle which is why I took this briefing. My overall concern is that with mobile devices dominating the semiconductor industry we have lost sight of MTBF, failure in time, risk of undetected failures, etc… Seriously, what do you do when your smartphone drops a call? Or an app crashes? Or your phone just quits? How about when your anti-lock brakes fail or your airbags make an unscheduled deployment?

The Incisive Functional Safety Simulator and Functional Safety Analysis technologies are part of theCadence System Development Suite (SDS), addressing the largest and most complex verification and hardware-software co-development challenges faced by semiconductor and system companies:

HIGHLIGHTS:

Solution automates three elements of ISO 26262 compliance including traceability, safety verification and tool confidence level
New Incisive Functional Safety Simulator delivers up to 10x simulation performance versus traditional solutions
New functional safety regression analysis capability in Incisive vManager automatically generates regression profiles and results, enabling the traceable audit trail for compliance

“Addressing functional safety challenges, particularly in automotive electronics, is critical for the success of system and semiconductor companies,” said Charlie Huang, executive vice president, Worldwide Field Operations and System & Verification Group at Cadence. “By partnering with companies like Melexis who embrace functional safety today, Cadence is delivering a solution that enables engineers to more efficiently address one of the key requirements to proliferate fault-tolerant electronics in the automotive industry where the safety of consumers is paramount.”

Cadence enables global electronic design innovation and plays an essential role in the creation of today’s integrated circuits and electronics. Customers use Cadence software, hardware, IP, and services to design and verify advanced semiconductors, consumer electronics, networking and telecommunications equipment, and computer systems. The company is headquartered in San Jose, Calif., with sales offices, design centers, and research facilities around the world to serve the global electronics industry. More information about the company, its products, and services is available at www.cadence.com.

Daylight Savings Time and the IoT

Daylight Savings Time and the IoT
by Daniel Payne on 11-05-2014 at 6:00 pm
Categories: IoT

On Sunday in the USA we changed our clocks back one hour to account for Daylight Savings Time and I was reminded of how far we have to go in getting all of our devices to understand and automatically account for the time. Despite all of the talk about IoT and how it has the promise to automate our lives, we still have to manually set the time. Here’s my experience with DST around the home:

[TABLE] style=”width: 500px”
|-
| Device
| Type
| Connectivity
| Automatically Changed Time
|-
| MacBook Pro
| Laptop
| WiFi
| Yes
|-
| iPad 3
| Tablet
| WiFi
| Yes
|-
| iPad Air
| Tablet
| WiFi
| Yes
|-
| Galaxy Note 2
| Android Phone
| WiFi, GSM
| Yes
|-
| Kindle Paperwhite
| e-book Reader
| WiFi
| No
|-
| Honeywell Thermostat
| Thermostat
| –
| No
|-
| AM/FM Alarm
| Clock
| –
| No
|-
| Microwave
| Appliance
| –
| No
|-
| Oven
| Appliance
| –
| No
|-
| 1998 Acura RL
| Sedan
| –
| No
|-
| 2001 Honda Odyssey
| Minivan
| –
| No
|-
| Cateye Stealth50
| Cyclocomputer
| GPS,
Ethernet
| No
|-
| Insignia
| Set top box
| Antenna
| Yes
|-
| Sony DVD-Bluray
| DVD, Bluray
| WiFi
| Yes
|-

All of our devices using Android, iOS, Mac OS X and Windows understood and made the time change automatically.

The Kindle Paperwhite kind of surprised me because it didn’t update the time in spite of being connected to WiFi, even after a Sync operation it still required me to manually change the time of day. The 3G model of the Kindle Paperwhite does automatically change the time, but ironically the WiFi version doesn’t. Amazon has the more popular Kindle Fire devices, and they do understand how to update the time automatically.

Amazon Kindle Paperwhite

For home automation I could upgrade my thermostat to a WiFi-enabled device from Honeywell or Nest, but I’m not sold on the cost savings so will hold off for a while on that purchase.

Honeywell Thermostat

Newer AM/FM alarm clocks are WiFi enabled and the consumer electronics category of WiFi-driven radios is booming.

I’m not sure that I want my Microwave to be WiFi enabled, because it has a tendency to knock-out WiFi signals enough that my daughter’s iPhone disconnects while our Android phones and Windows laptops stay connected.

Our autos are old enough that they have no networking features, although there’s an interesting group called the Open Automotive Alliance that plans to bring the Android platform to cars starting this year.

Apple has something called CarPlay for infotainment with partners like: Alfa Romeo, Audi, BMW, Chevrolet, Chrysler, Citroen, Dodge, Ferrari, Fiat, Ford, Honda, Hyundai, Jaguar, Jeep, Kia, Land Rover, Mazda, Mercedes-Benz, Mitsubishi, Nissan, Peugeot, Subaru, Suzuki, Toyota and Volvo.

Apple CarPlay

For cycling my Cateye Stealth 50 cyclocomputer does connect to both GPS and Ethernet, although I still had to manually select DST off.

Our home audio-visual equipment is new enough that it connects to WiFi and auto-sets the date and time.

Slowly, but surely I am having to manually reset fewer clocks in our home each year, but we still have a ways to go before the IoT fully automates the time change. How does the time change effect your digital life?

November 5, 2014

Samsung 14nm is the one delayed!

Samsung 14nm is the one delayed!
by Daniel Nenni on 11-05-2014 at 6:00 am
Categories: Foundries

As you may have read, the CEO of Ultratech made some unfortunate statements on the recent quarterly conference call in regards to FinFET Yield. As a result there has been a lot of speculation about the who, what, and why. I blogged about it because it interested me personally plus I wanted to collect more data on the subject. Some of the resulting data was posted publicly but most was emailed to me in private, discussed in person, or on the phone.

Also Read: Let the FinFET Yield Controversy Begin!

One of the many benefits of blogging on SemiWiki is that we have access to an incredible amount of semiconductor related data that passes through the site and our private email. It also gives us bloggers access to the semiconductor rank and file, more so than journalists as we are working semiconductor professionals and travel in the same circles. The bad news is that we get our fair share of hate mail. And when I say “we” it’s mostly me because let’s face it I don’t always play well with others, just check my grade school report cards.

While I speculate and make guesses in my blogs I generally know the truth before going out on a limb. On this occasion however I was wrong. Using deductive reasoning I expertly guessed that it was UMC that pushed out the equipment orders referenced in the conference call. It was not and I apologize. It was Samsung. From what I know, that specific market segment (annealing) is mostly shared between Ultratech and Mattson Technology. Here are clips from the recent conference calls:

UTEK CEO: As we have discussed on past conference calls, the difficult implementation of 3D FinFET microprocessors to high production manufacturing. Once again a major logic manufacturer delayed their FinFET ramp. We had then requested to prepare LSA tools for shipment for the end of the third quarter which was delayed. These LSA shipments for the most part caused our third quarter revenue to be less than projected. These LSA systems have been rescheduled for shipment in the fourth quarter. Due to the continued low yield in FinFET devices for the past two years, we have seen a reduction in new LSA bookings in subsequent shipments…

MTSN CEO:The adjustments of our customers’ production capacity plan caused a slight pause in Mattson’s quarter-to-quarter growth, which is consistent with the industry trend. However, we continue to expect our total 2014 revenue to grow over 40%, it can be achieved 2013. As a recovery of wafer fab equipment spending begins in the coming quarters...

Read the entire transcripts and tell me which one of these companies would you trust your business with if you were a major semiconductor manufacturer? Better yet, which one of these companies would you trust your hard earned investment dollars with?

And just because Samsung allegedly pushed out an equipment order does not mean that FinFETs are in trouble or according to Motley Fool’s “Senior” Technology Specialist: “that neither TSMC nor Samsung quite has the FinFET transistor structure (which promises higher performing transistors at lower power) figured out.”

I was also told that Ultratech lost the Intel and TSMC FinFET annealing business to Mattson which I understand completely since Ultratech clearly “does not play well with others.” If someone wants to briefly explain where annealing fits in with the semiconductor manufacturing process in the comment section it would be greatly appreciated.

FD-SOI, an Opportunity for China?

FD-SOI, an Opportunity for China?
by Paul McLellan on 11-04-2014 at 11:00 pm
Categories: FD-SOI, Foundries, STMicroelectronics

Last month in Shanghai was a meeting of the FD-SOI consortium. The focus of the meeting was largely on the suitability of using FD-SOI to serve the Chinese market. The fabs in China are not right on the bleeding edge and are very cost-sensitive so 28nm is probably as advanced as they will get for a long time if not indefinitely. China has a goal that by 2020 they will manufacture 40% of the semiconductors used in China within the country. This is a big goal since China is actually the largest market for semiconductors in the world. This year the Chinese market will be $161B with just 8.9% being manufactured in Chinese fabs. Of course a lot of those semiconductors are re-exported inside finished goods but a lot are consumed in China too, after all it is also the largest market for smartphones.

The opening keynote was given by Handel Jones of IBS. He pointed out that FD-SOI is a great bridge process between 28nm planar and FinFET. It has a lot of the same power efficiency, low leakage etc but at a price that is roughly the same as planar. Handel still believes that 14nm/16nm will not ramp until Q4/2016 or Q1/2017, which seems very late based on everything else we have been hearing. Handel also had estimates of 28nm wafer volume going out to the middle of the next decade. 28nm is clearly going to be a very long-lasting node which also means that there is potentially a very large market for 28nm FD-SOI.

Handel also had some wafer pricing data. I assume that this is based on a model rather than actual quoted prices since those are usually too commercially sensitive. A 28nm FD-SOI wafer in 2015 is $2400, compared to a 14/16nm FinFET at $4800 and 14/16nm FD-SOI at $3600. If those predictions hold out then that is quite a price differential between FD-SOI and FinFET at the 14/16nm node. By 2017 IBS reckon that the cost of 100M gates will be just 90c in 28nm FD-SOI compared to $1.57 in 14/16nm FinFET, so nearly half as much again for the FinFET.