LONDON – Engineering consultancy UBM TechInsights has announced it has found a phase-change memory die inside a multi-chip package inside a mobile handset.
TechInsights (Ottawa, Ontario), part of the same group that publishes EE Times, has found a NOR flash memory compatible phase-change RAM, from Samsung Electronics Co. Ltd. in a mobile handset but declined to reveal the model of handset citing reasons of customer confidentiality. "We will be able to reveal it in the near future once our work with this client is done," a spokesman for UBM TechInsights said.
Samsung had said in April 2010 it would be shipping a multi-chip package memory, including a 512-Mbit phase-change memory die in the second quarter of 2010 (see Samsung to ship MCP with phase-change). At the time Samsung was not clear about the process technology in use describing it as "equivalent to 40-nm-class NOR flash memory." The process is thought to be at about 65- to 60-nm minimum dimensions. The half-pitch memory distance shown in the microphotograph below (8 cells per micron) bears this out.
Numonyx, now Micron Technologies Inc., released a 90-nm 128-Mbit phase-change memory in 2008, which it formalized as the Omneo range of serial and parallel access memories in April 2010. But the company has not said anything about design wins or volume production. Numonyx has been developing a 1-Gbit phase-change memory in a 45-nm process, but there has been no news of sampling or volume production for a device that was expected to appear early in 2010.
UBM TechInsights has confirmed the commercial availability of a 512-Mbit PRAM die, labeled KPS1N15EZA, and packaged with a Samsung 128Mbit UtRAM die in a multi-chip package in a mobile phone. The Samsung PRAM die is comprised of four aluminum interconnect layers with the memory elements and the top and bottom electrode contacts built between the aluminum metal 1 and the silicon substrate.
"Recent debate on the scaling limitation of phase change memory technologies, combined with production delays from Numonyx, had left many wondering if PRAM would ever realize its promise as a next-generation memory," said Young Choi, senior consultant for professional services at UBM TechInsights, in a statement. "The discovery of this product in a mobile application is a clear sign that designers are now open to using this promising technology."
Cross-section of Samsung 512-Mbit phase-change RAM.
Packaged MCP that includes phase-change memory die.
Mobile phone main board including phase-change memory MCP
new2coding: It took Micron 10 years to admit that PCM does not scale, after having signed an exclusive, royalty-bearing license for the PCM patents in 1994. Intel started their PCM technoPonzi in 2000, and even demoed the abomination they called the "world's first cell phone with PCM" in 2008, culminating with this year's complete failure (Numonyx failed to deliver a 1Gbit 45nm chip, and has gained absolutely no traction with its 90nm 128Mbit chip). Samsung started in 2003. So you should expect all kinds of funny things from them until they admit by 2013, at the latest, that PCM will never be commercialized in volume, because it does not scale, is too slow and power-consuming in write, and too costly to make. Are things clearer now?
Well, I am now a proud owner of a brand-new, commercial Samsung E2550 bought in Singapore. It contains NOR Flash, of course, and no absolutely no PRAM MCP. For those on the hunt for the non-commercial Samsung e2550 handsets that allegedly contain the PRAM MCP, here is a method to quickly identify an "interesting phone" - open the battery compartment, remove the battery, and look above the SIM card placeholder - if the solder bridge is next to the E2550 mark, then this is the regular commercial Samsung phone with NOR Flash in it, if it is next to the E2550L mark, then you MAY have a handset with PRAM MCP. Good hunting! Although, I am now convinced that Samsung simply planted only a handful non-commercial units for a marketing stunt.
The comments on this column have made for about the most entertaining technology read I can remember. Great job bringing these out Peter.
I don't think there is much more to add to the comments especially when people like RG Neale have posted here (not to mention his comprehensive series of articles on PCM scaling). But I did add my own two cents in my last column:
If you want to help me with entertainment value, please spread some of your love over there.
Somobody just brought to my attention an interesting document prepared by Samsung on October 4th, 2010:
Page 3 states: "Samsung’s 512Gb PRAM is combined with Mobile DRAM to deliver performance three times faster than NOR-based MCPs, making it ideal to quickly process large-size multimedia
The "three times faster" is obviously a lie, but hard to prove given that the only PRAM chip in a cell phone has been destroyed.
However, it is obvious that Samsung lied about PRAM being 512Gb. It is just 512Mb. Samsung exaggerated by a factor of 1024x.
And, apparently, nobody has noticed so far. People must be busy installing those chips into fake phones instead of reading marketing materials.
Great job, Samsung!
I agree, the memory market is huge and there is a lot at stake. It's not always about technical merit as we know ;-)
@Volatile Memory, keep your posts coming, I find them informative and entertaining :-)
When a Huge hedgefund And competing companies have have tons of money to loose if PCM suceeds of course you will see all of these"Experts" come out to beat it up and spread the lies ..
good luck ecd.fan keep it up .ooops i mean "virtual memory "
the truth shall set you free
Peter--this article was picked up by the Wall Street Journal's "Best of the Web Today" column, but alas, not for its content, but for the word-play of the headline. See:
Go about 3/4 down,under the heading:
"It's Always in the Last Place You Look"
I wonder who sent it in to them--has to besomeone who both reads EET AND sends things to this popular daily column.--Bill S
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.