It is unbelievable but reasonable. The next killer should be PCM(Phase Change Memory), will can replace, NAND, Nor, and even RAM. Spansion walks in a different way from Intel, ST. But they all want to find the next generation memory killer.
____________________________________________
Spansion eyes DRAM market
David Roman(05/17/2007 12:59 PM EDT)U
RL: http://www.eetimes.com/showArticle.jhtml?articleID=199601545
NEW YORK — Spansion Inc. has it's eye on the $30 billion DRAM market.
Spansion, which supplies only NOR flash memory and hasn't reported a profit in 10 quarters due to intense competition and pressure on ASPs, plans to position its proprietary MirrorBit technology as a competitor to DRAM, which will open the market to high-end phones, PDAs and communication handsets.
"We are currently designing a DRAM-like device using MirrorBit technology," said Bertrand Cambou, president and CEO, at Spansion's first analyst briefing held at the Nasdaq Marketsite in Times Square here. Cambou did not say when the product would be ready for market.
"People say the competition for NOR is NAND. This statement is wrong." The competition is DRAM, he said. "We have the opportunity to go after DRAM," which he called "a kind of dirty memory" due to its use of a capacitor and its need for a refresh every millisecond.
MirrorBit could be used in place of DRAM and would provide a lower power, lower priced solution, he said. It would also allow Spansion to play in the market for high-end handsets, lifting it from the mainstream NOR market where compeition is fierce.
"DRAM is a $30 billion market today," Cambou said. "If we take 10 percent of that we would double the size of our company." Spansion reported $2.6 billion in sales in 2006 and was the world's biggest NOR flash supplier last year, according to market watcher iSuppli Corp.
"Right now, 166 MHz is what you need to hit to take the DRAM space," Cambou said. "Today we are doing 133 MHz. If we go 20 percent faster, we can be there."
Cambou said MirrorBit is superior to floating gate technology used by Intel, STMicroelectronics and other NOR competitors. Some analysts question the ability of floating gate to scale to sub-45-nm geometries.
The long-rumored alliance of Intel Corp. and STMicroelectronics, the NOR market's No. 2 and 3 suppliers, will also create an opportunity for Spansion, Cambou said. "Floating gate is losing steam. Going forward we think consolidation is in motion."
Smaller NOR players will not be a factor because they are not investing in 300-mm capacity or 45-nm process technology with Spansion's intensity. "We see ourselves as being the leader in NOR and will redefine the industry bvy making customer-centric solutions with a laser focus on cost."
Added Cambou: "Some of our competition is not excited by our product, but we are. We think it has a lot of potential. We think it is an exciting field. We have the technology it takes to flourish as a company."
Thursday, May 17, 2007
Tuesday, May 15, 2007
TI takes two approaches to IC manufacturing
Fab-lite is a trend in semiconductor industry.
_____________________________
TI takes two approaches to IC manufacturing
Chip maker revs in-house analog, but shifts logic to foundries
Mark LaPedus(05/13/2007 9:00 AM EDT)
URL: http://www.eetimes.com/showArticle.jhtml?articleID=199500883
Dallas — Moving to remain competitive in what has become a brave new world of IC manufacturing, Texas Instruments Inc. last week disclosed the details of its revised "hybrid" fab strategy. The chip maker is bolstering its in-house efforts in analog production, but it is also shifting more of its logic-based IC work and process flow to the foundries.
As a result of the shift, TI has pushed out the production ramp date for its new--and still unequipped--300-mm fab in Texas by about 18 months.
Nearly half of TI's logic chip production is outsourced to the foundries today, but that figure could jump to 70 percent over time, according to analysts. TI says it has no intention of going fabless for logic. But the chip maker's foundry partners for the 45-nanometer node--TSMC, UMC and a yet-to-be-determined vendor--will play a much bigger role than TI has afforded foundries in the past. And by 32 nm, TI will co-develop its processes at the foundries, whereas traditionally it has done that work in-house.
By turning over more control of logic process development to its foundry partners, TI is also offloading some of the risk. It's a calculated move as the chip industry approaches a new inflection point in manufacturing, with complex technologies introduced for the 45-nm node and beyond.
Those technologies include immersion lithography, ultralow-k and, to a lesser degree, high-k dielectrics.
The foundries "will meet TI's requirements," said Kevin Ritchie, senior vice president for the company's Technology and Manufacturing Group.
At TI's financial analysts' meeting here last week, Ritchie revealed two of its foundry partners for the 45-nm node: Taiwan Semiconductor Manufacturing Co. Ltd. and United Microelectronics Corp. Meanwhile, Singapore's Chartered Semiconductor Manufacturing Pte. Ltd. appears to be on the outside looking in, although TI has not yet identified the third partner that will split the 45-nm business with TSMC and UMC.
Another open question is the future of TI's own processor foundry business. For years, the company has manufactured leading-edge Sparc processors for Sun Microsystems Inc. According to Ritchie, UMC will begin making Sparc processors on a foundry basis at the 45-nm node. Sun officials, however, insisted that Sun is still exploring its foundry options and has not selected a new outsourcing partner.
In any case, the new chip-manufacturing strategy is a dramatic departure from the philosophy TI was professing only a year ago, said Robert Lineback, an analyst with IC Insights Inc. Back then, "they were talking about the need to develop technology," Lineback said. "The fear factor of foundries' not being able to quickly ramp leading-edge digital processes was part of the reason for TI's strategy to do its own technology. But TSMC and UMC proved they could do it as good or better than TI's own R&D team and fabs."
Given the soaring costs of fabs and process development, TI's new foundry push makes fiscal sense, said David Lammers, an analyst with VLSI Research Inc. But Lammers believes TI "will not go fabless" in the foreseeable future. With the huge volumes anticipated for its single-chip cell phone products, the company will require both internal and external production, he said.
During a presentation at the financial analysts' meeting, Richard Templeton, the Dallas chip giant's president and chief executive officer, said the manufacturing strategy frees up resources to focus more on design. But that is not to say TI has stopped investing in fabs, Templeton said. On the contrary, capital spending and resources for analog production are "going up."
TI's capital spending is targeted at $900 million for 2007.
Different world
For years, TI prided itself in developing its own leading-edge processes for analog and logic applications. The chip maker also built logic fabs to keep pace with IBM, Intel and other leading-edge integrated device manufacturers (IDMs).
As the 300-mm era arrived early in the decade, however, fab costs soared, prompting some IDMs to adopt "fab lite" strategies. At the same time--indeed, even before the start of the 300-mm era--the top silicon foundries had begun to amass huge sums of capital in order to develop leading-edge processes and fabs.
That changed everything for IDMs and their fabless rivals. For example, by using leading-edge foundries, fabless Qualcomm Inc. has been able to close the manufacturing gap with rival TI.
In January, TI dropped a bombshell when it announced it would withdraw from the costly business of internal digital-logic process development after the 45-nm node, instead relying on foundry-supplied processes at the 32-nm node and thereafter. The company also closed its Kilby wafer fab in Dallas, a move that affected 500 employees, including 200 process development engineers.
The decision was "an emotional change for TI," Ritchie said. After the announcement, he said, "one employee choked backed tears and said: 'I believe it's the right decision, but it hurts.' "
The events fueled speculation among analysts that TI would sell its next-generation 300-mm fab in Richardson, dubbed RFab. Announced in 2003, the fab was targeted for production in the 2009 or 2010 time frame.
TI has finished the shell on RFab but has yet to equip the plant. The company has no plans to sell RFab but is pushing out the production ramp date by a year and a half, Ritchie said. In the meantime, TI will expand production at an existing 300-mm fab in Dallas. R&D wafer lines at DMOS6 will be converted to production, raising the overall capacity to 26,000 wafers per month from 17,000.
Current production at DMOS6 is on a 65-nm process. While TI plans to install its 45-nm process at the fab, with production slated for the second quarter of 2008, foundries are expected to lead TI's production ramp at that node.
Three-pronged approach
TI's foundry strategy is partitioned by product category: wireless, DSP and Sparc processors. At the 65-nm node, TI has three foundry partners for its wireless chips: Chartered, TSMC and UMC. For wireless chips at 45 nm, TI will continue to use UMC and TSMC, Ritchie said. "Disruptive pricing" is a key criterion for the sought-after third partner, he said.
Within its own logic fabs, TI develops processes and makes its own 65-nm high-performance digital signal processors. At the next node, TI and TSMC will manufacture DSPs.
A foundry will also take over production of Sparc processors for Sun at 45 nm. Ritchie said, but Sun has not confirmed, that the foundry will be UMC "for the first iteration."
At 32 nm, TI plans to bring its hybrid strategy to new heights. In the past, process development efforts required three distinct R&D activities. TI developed its own process; foundry partners independently did the same. A third R&D effort brought the separate processes into the production phase.
Starting at the 32-nm node, TI and its foundry partners will jointly work on process technology. The process will be defined and developed at the foundry, instead of within TI, and then be "copied" and transferred or fanned back to TI's own fabs. In a sense, that will make TI's fabs complementary to the foundries, VLSI Research's Lammers said.
Even as it shifts responsibilities to the foundries for digital processes, TI is investing in its own analog processes and fabs--and for good reason. Nearly half of the company's total sales revolve around analog, and the fragmented sector is growing faster than the overall semiconductor industry, said Art George, senior vice president of high-performance analog at the company.
TI is already the world's largest analog supplier. "I believe that high-performance analog could be the core growth engine for TI," Templeton said during his presentation. "We are building our skills and competencies around analog."
TI has 200-mm analog fabs in Japan, Germany and Dallas. In 2008, the company may expand its analog fab capacity, perhaps leveraging the used tools from its shuttered Kilby fab.
Over the past year, TI has quietly rolled out and ramped up four analog processes. BiCom3HV is a 36-volt bipolar silicon-germanium process. HPA07 is a high-precision analog technology.
A035 is a high-density analog CMOS formula. LBC7 is a high-power BiCMOS technology, based on 0.25-micron technology. TI has dropped hints that it will roll out LBC8, an 0.18-micron version of LBC7, in 2008.
_____________________________
TI takes two approaches to IC manufacturing
Chip maker revs in-house analog, but shifts logic to foundries
Mark LaPedus(05/13/2007 9:00 AM EDT)
URL: http://www.eetimes.com/showArticle.jhtml?articleID=199500883
Dallas — Moving to remain competitive in what has become a brave new world of IC manufacturing, Texas Instruments Inc. last week disclosed the details of its revised "hybrid" fab strategy. The chip maker is bolstering its in-house efforts in analog production, but it is also shifting more of its logic-based IC work and process flow to the foundries.
As a result of the shift, TI has pushed out the production ramp date for its new--and still unequipped--300-mm fab in Texas by about 18 months.
Nearly half of TI's logic chip production is outsourced to the foundries today, but that figure could jump to 70 percent over time, according to analysts. TI says it has no intention of going fabless for logic. But the chip maker's foundry partners for the 45-nanometer node--TSMC, UMC and a yet-to-be-determined vendor--will play a much bigger role than TI has afforded foundries in the past. And by 32 nm, TI will co-develop its processes at the foundries, whereas traditionally it has done that work in-house.
By turning over more control of logic process development to its foundry partners, TI is also offloading some of the risk. It's a calculated move as the chip industry approaches a new inflection point in manufacturing, with complex technologies introduced for the 45-nm node and beyond.
Those technologies include immersion lithography, ultralow-k and, to a lesser degree, high-k dielectrics.
The foundries "will meet TI's requirements," said Kevin Ritchie, senior vice president for the company's Technology and Manufacturing Group.
At TI's financial analysts' meeting here last week, Ritchie revealed two of its foundry partners for the 45-nm node: Taiwan Semiconductor Manufacturing Co. Ltd. and United Microelectronics Corp. Meanwhile, Singapore's Chartered Semiconductor Manufacturing Pte. Ltd. appears to be on the outside looking in, although TI has not yet identified the third partner that will split the 45-nm business with TSMC and UMC.
Another open question is the future of TI's own processor foundry business. For years, the company has manufactured leading-edge Sparc processors for Sun Microsystems Inc. According to Ritchie, UMC will begin making Sparc processors on a foundry basis at the 45-nm node. Sun officials, however, insisted that Sun is still exploring its foundry options and has not selected a new outsourcing partner.
In any case, the new chip-manufacturing strategy is a dramatic departure from the philosophy TI was professing only a year ago, said Robert Lineback, an analyst with IC Insights Inc. Back then, "they were talking about the need to develop technology," Lineback said. "The fear factor of foundries' not being able to quickly ramp leading-edge digital processes was part of the reason for TI's strategy to do its own technology. But TSMC and UMC proved they could do it as good or better than TI's own R&D team and fabs."
Given the soaring costs of fabs and process development, TI's new foundry push makes fiscal sense, said David Lammers, an analyst with VLSI Research Inc. But Lammers believes TI "will not go fabless" in the foreseeable future. With the huge volumes anticipated for its single-chip cell phone products, the company will require both internal and external production, he said.
During a presentation at the financial analysts' meeting, Richard Templeton, the Dallas chip giant's president and chief executive officer, said the manufacturing strategy frees up resources to focus more on design. But that is not to say TI has stopped investing in fabs, Templeton said. On the contrary, capital spending and resources for analog production are "going up."
TI's capital spending is targeted at $900 million for 2007.
Different world
For years, TI prided itself in developing its own leading-edge processes for analog and logic applications. The chip maker also built logic fabs to keep pace with IBM, Intel and other leading-edge integrated device manufacturers (IDMs).
As the 300-mm era arrived early in the decade, however, fab costs soared, prompting some IDMs to adopt "fab lite" strategies. At the same time--indeed, even before the start of the 300-mm era--the top silicon foundries had begun to amass huge sums of capital in order to develop leading-edge processes and fabs.
That changed everything for IDMs and their fabless rivals. For example, by using leading-edge foundries, fabless Qualcomm Inc. has been able to close the manufacturing gap with rival TI.
In January, TI dropped a bombshell when it announced it would withdraw from the costly business of internal digital-logic process development after the 45-nm node, instead relying on foundry-supplied processes at the 32-nm node and thereafter. The company also closed its Kilby wafer fab in Dallas, a move that affected 500 employees, including 200 process development engineers.
The decision was "an emotional change for TI," Ritchie said. After the announcement, he said, "one employee choked backed tears and said: 'I believe it's the right decision, but it hurts.' "
The events fueled speculation among analysts that TI would sell its next-generation 300-mm fab in Richardson, dubbed RFab. Announced in 2003, the fab was targeted for production in the 2009 or 2010 time frame.
TI has finished the shell on RFab but has yet to equip the plant. The company has no plans to sell RFab but is pushing out the production ramp date by a year and a half, Ritchie said. In the meantime, TI will expand production at an existing 300-mm fab in Dallas. R&D wafer lines at DMOS6 will be converted to production, raising the overall capacity to 26,000 wafers per month from 17,000.
Current production at DMOS6 is on a 65-nm process. While TI plans to install its 45-nm process at the fab, with production slated for the second quarter of 2008, foundries are expected to lead TI's production ramp at that node.
Three-pronged approach
TI's foundry strategy is partitioned by product category: wireless, DSP and Sparc processors. At the 65-nm node, TI has three foundry partners for its wireless chips: Chartered, TSMC and UMC. For wireless chips at 45 nm, TI will continue to use UMC and TSMC, Ritchie said. "Disruptive pricing" is a key criterion for the sought-after third partner, he said.
Within its own logic fabs, TI develops processes and makes its own 65-nm high-performance digital signal processors. At the next node, TI and TSMC will manufacture DSPs.
A foundry will also take over production of Sparc processors for Sun at 45 nm. Ritchie said, but Sun has not confirmed, that the foundry will be UMC "for the first iteration."
At 32 nm, TI plans to bring its hybrid strategy to new heights. In the past, process development efforts required three distinct R&D activities. TI developed its own process; foundry partners independently did the same. A third R&D effort brought the separate processes into the production phase.
Starting at the 32-nm node, TI and its foundry partners will jointly work on process technology. The process will be defined and developed at the foundry, instead of within TI, and then be "copied" and transferred or fanned back to TI's own fabs. In a sense, that will make TI's fabs complementary to the foundries, VLSI Research's Lammers said.
Even as it shifts responsibilities to the foundries for digital processes, TI is investing in its own analog processes and fabs--and for good reason. Nearly half of the company's total sales revolve around analog, and the fragmented sector is growing faster than the overall semiconductor industry, said Art George, senior vice president of high-performance analog at the company.
TI is already the world's largest analog supplier. "I believe that high-performance analog could be the core growth engine for TI," Templeton said during his presentation. "We are building our skills and competencies around analog."
TI has 200-mm analog fabs in Japan, Germany and Dallas. In 2008, the company may expand its analog fab capacity, perhaps leveraging the used tools from its shuttered Kilby fab.
Over the past year, TI has quietly rolled out and ramped up four analog processes. BiCom3HV is a 36-volt bipolar silicon-germanium process. HPA07 is a high-precision analog technology.
A035 is a high-density analog CMOS formula. LBC7 is a high-power BiCMOS technology, based on 0.25-micron technology. TI has dropped hints that it will roll out LBC8, an 0.18-micron version of LBC7, in 2008.
Atmel memory embeds crypto for counterfeit busting
Crypto is a trend, but now few function is implemented for most application.
______________________________________
Atmel memory embeds crypto for counterfeit busting
Loring Wirbel(05/15/2007 12:11 PM EDT)
URL: http://www.eetimes.com/showArticle.jhtml?articleID=199600358
COLORADO SPRINGS, Colo. — Atmel Corp. is offering a development kit for a nonvolatile memory surrounded by a 64-bit embedded cryptographic processor. The CryptoMemory can be used for mutual authentication in applications such as smart cards, as well as in anti-counterfeit apps as an alternative to bar code systems.
The memory, offered in densities of 1 to 256 kbits, can be divided into as many as 16 configuration and user sections, each with a defined level of read and write access. Four levels of security can be implemented: free access, password protection using eight sets of read-write passwords, mutual authentication using a choice of four sets of 64-bit keys, and full data encryption. Thus one
______________________________________
Atmel memory embeds crypto for counterfeit busting
Loring Wirbel(05/15/2007 12:11 PM EDT)
URL: http://www.eetimes.com/showArticle.jhtml?articleID=199600358
COLORADO SPRINGS, Colo. — Atmel Corp. is offering a development kit for a nonvolatile memory surrounded by a 64-bit embedded cryptographic processor. The CryptoMemory can be used for mutual authentication in applications such as smart cards, as well as in anti-counterfeit apps as an alternative to bar code systems.
The memory, offered in densities of 1 to 256 kbits, can be divided into as many as 16 configuration and user sections, each with a defined level of read and write access. Four levels of security can be implemented: free access, password protection using eight sets of read-write passwords, mutual authentication using a choice of four sets of 64-bit keys, and full data encryption. Thus one
Spansion turns to TSMC for flash at 40-nm and below
Fab-light is a trend for all traditional semiconductor venders...
_______________________________________
Spansion turns to TSMC for flash at 40-nm and below
Mike Clendenin(05/15/2007 6:06 AM EDT)
URL: http://www.eetimes.com/showArticle.jhtml?articleID=199600158
TAIPEI, Taiwan — Spansion Inc is once again turning to Taiwan Semiconductor Manufacturing Co. to increase its NOR flash capacity, signing a deal with the foundry to produce MirrorBit at 40 nanometer and below.
Terms of the deal, which was signed Tuesday, were kept secret.
The agreement is an extension of earlier work between the two companies. TSMC also makes MirrorBit at the 110-nm and 90-nm process nodes. It began ramping up the 110nm parts in the second quarter of 2006 and production of 90-nm parts is expected to begin within months.
The two companies did not say when they expect to begin production of 40-nm parts, but it isn't likely to happen before 2009. TSMC isn't expected to introduce its 45-nm process until 2008. Meanwhile, Spansion won't ramp its internal 45-nm parts until mid-2008.
In a bid to diversify, TSMC rival United Microelectronics Corp. is also looking to increase its flash business and is targeting NAND production on its 65- and 45-nm technology nodes.
_______________________________________
Spansion turns to TSMC for flash at 40-nm and below
Mike Clendenin(05/15/2007 6:06 AM EDT)
URL: http://www.eetimes.com/showArticle.jhtml?articleID=199600158
TAIPEI, Taiwan — Spansion Inc is once again turning to Taiwan Semiconductor Manufacturing Co. to increase its NOR flash capacity, signing a deal with the foundry to produce MirrorBit at 40 nanometer and below.
Terms of the deal, which was signed Tuesday, were kept secret.
The agreement is an extension of earlier work between the two companies. TSMC also makes MirrorBit at the 110-nm and 90-nm process nodes. It began ramping up the 110nm parts in the second quarter of 2006 and production of 90-nm parts is expected to begin within months.
The two companies did not say when they expect to begin production of 40-nm parts, but it isn't likely to happen before 2009. TSMC isn't expected to introduce its 45-nm process until 2008. Meanwhile, Spansion won't ramp its internal 45-nm parts until mid-2008.
In a bid to diversify, TSMC rival United Microelectronics Corp. is also looking to increase its flash business and is targeting NAND production on its 65- and 45-nm technology nodes.
Cypress sells off PSRAM unit to Taiwan firm
Nor flash market ~ PSRAM market
Cypress sells off PSRAM unit to Taiwan firm
Mike Clendenin(05/15/2007 4:48 AM EDT)
URL: http://www.eetimes.com/showArticle.jhtml?articleID=199600119
TAIPEI, Taiwan — Cypress Semiconductor is exiting the market for pseudo SRAM, selling off the business to Taiwan's Elite Semiconductor Memory Technology Inc. (ESMT) for an undisclosed sum.
Cypress offered little explanation as to why they were exiting the business, describing the deal as "consistent with Cypress's new strategic direction." The assumption is that the business just wasn't performing well enough to remain a part of the company's portfolio. The company said ESMT will continue all delivery and support to Cypress's current PSRAM customer base.
ESMT primarily focuses on low-density memory chips, ranging from 1Mbit to 256Mbit, that find their way into products such as DVD players, LCD monitors, MP3 players, set-top boxes, and video game systems.
The company is eager to build up its handset business, but a hole in its product offering has been pseudo SRAM. The company said the deal with Cypress will "significantly" boost its revenue and immediately put it on the radar of handset makers. "It will enable the company to quickly cut into the supply chain of major cellular phone makers, since Cypress is one of the world's top-tier PSRAM suppliers," said Hsing-Hai Chen, chairman at ESMT.
By using a pseudo SRAM architecture that employs a single-transistor DRAM cell rather than a six-transistor SRAM cell, pseudo SRAM helps GPRS and UMTS phone makers add more bells and whistles to their units while maintaining system speed, a compact form and perhaps most importantly, relatively competitive pricing.
Pseudo SRAM is pin-compatible with SRAM, allowing chips to be used as drop-in replacements for the asynchronous SRAM currently used in many cell phone designs. The devices can also include a burst read and write mode that emulates a flash-memory interface.
Numerous other companies offer pseudo SRAM, usually under other names, though. For instance, a number of memory companies have founded the CellularRAM work group, consisting of Cypress, Micron Technology, Qimonda, Renesas Technology, Etron Technologies, Hynix Semiconductor, Silicon 7 and Winbond Electronics. These firms have agreed on a common spec for pseudo SRAM as they develop it under the CellularRAM name.
Cypress was a key member of the CellularRAM camp. ESMT did not immediately respond to requests to clarify whether it will remain part of the CellularRAM alliance and collaborate with the companies on product roadmaps that may require specification changes.
ESMT's existing low power SDRAM product line includes 16-Mbit, 32-Mbit, 64-Mbit and 128-Mbit devices for the handheld market. With the PSRAM product transfer from Cypress, ESMT's Mobile RAM product line will expand to cover lower-density, low-power memory products including 2-Mbit, 4-Mbit, 8-Mbit, 16-Mbit and 32-Mbit PSRAM devices.
Cypress sells off PSRAM unit to Taiwan firm
Mike Clendenin(05/15/2007 4:48 AM EDT)
URL: http://www.eetimes.com/showArticle.jhtml?articleID=199600119
TAIPEI, Taiwan — Cypress Semiconductor is exiting the market for pseudo SRAM, selling off the business to Taiwan's Elite Semiconductor Memory Technology Inc. (ESMT) for an undisclosed sum.
Cypress offered little explanation as to why they were exiting the business, describing the deal as "consistent with Cypress's new strategic direction." The assumption is that the business just wasn't performing well enough to remain a part of the company's portfolio. The company said ESMT will continue all delivery and support to Cypress's current PSRAM customer base.
ESMT primarily focuses on low-density memory chips, ranging from 1Mbit to 256Mbit, that find their way into products such as DVD players, LCD monitors, MP3 players, set-top boxes, and video game systems.
The company is eager to build up its handset business, but a hole in its product offering has been pseudo SRAM. The company said the deal with Cypress will "significantly" boost its revenue and immediately put it on the radar of handset makers. "It will enable the company to quickly cut into the supply chain of major cellular phone makers, since Cypress is one of the world's top-tier PSRAM suppliers," said Hsing-Hai Chen, chairman at ESMT.
By using a pseudo SRAM architecture that employs a single-transistor DRAM cell rather than a six-transistor SRAM cell, pseudo SRAM helps GPRS and UMTS phone makers add more bells and whistles to their units while maintaining system speed, a compact form and perhaps most importantly, relatively competitive pricing.
Pseudo SRAM is pin-compatible with SRAM, allowing chips to be used as drop-in replacements for the asynchronous SRAM currently used in many cell phone designs. The devices can also include a burst read and write mode that emulates a flash-memory interface.
Numerous other companies offer pseudo SRAM, usually under other names, though. For instance, a number of memory companies have founded the CellularRAM work group, consisting of Cypress, Micron Technology, Qimonda, Renesas Technology, Etron Technologies, Hynix Semiconductor, Silicon 7 and Winbond Electronics. These firms have agreed on a common spec for pseudo SRAM as they develop it under the CellularRAM name.
Cypress was a key member of the CellularRAM camp. ESMT did not immediately respond to requests to clarify whether it will remain part of the CellularRAM alliance and collaborate with the companies on product roadmaps that may require specification changes.
ESMT's existing low power SDRAM product line includes 16-Mbit, 32-Mbit, 64-Mbit and 128-Mbit devices for the handheld market. With the PSRAM product transfer from Cypress, ESMT's Mobile RAM product line will expand to cover lower-density, low-power memory products including 2-Mbit, 4-Mbit, 8-Mbit, 16-Mbit and 32-Mbit PSRAM devices.
Subscribe to:
Posts (Atom)
