SI
SI
discoversearch

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor.  We ask that you disable ad blocking while on Silicon Investor in the best interests of our community.  If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.
Technology Stocks : WDC/Sandisk Corporation
WDC 163.00-0.4%Nov 7 9:30 AM EST

 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
To: Ausdauer who wrote (12703)7/9/2000 1:06:18 AM
From: wily  Read Replies (1) of 60323
 
Ausdauer,

The last SSTI c.c. was informative, but I left with the understanding that 16Mbit was their maximum flash density.

I don't know how to resolve this -- maybe I'll try calling the company on Monday. My experience is that this information is closely guarded. If we had some semi-processing engineers or company employees on this thread we could get some good guesses or better -- like they have on the AMD and Intel threads.

Here's a great picture of the insides of a CF card from the SST website:
ssti.com

A lot less of the interior space is used for the chips (I'm guessing the square-shaped chip is a multi-chip package) than I had imagined.

The most specific information that I've seen regarding number of chips inside a CF card is from the Hitachi PR that said they were making a 448MB card using stacked 256Mbit chips. That comes out to 14 chips or 7 pairs of stacked chips. I think this is a very good reference point (well, it works for me).

A 96MB card with 7 chips would average 110Mb per chip, so they are probably stacking 64Mbit chips. In their 10K they said they are developing 32 and 64Mbit chips, so it can't be a 128Mbit chip in the card. At 64Mbits, it would take 6 stacked pairs.

OTOH, if you go to this page:
ssti.com
and download the "Die Sales Specifications" pdf, you will see on page 5 the dimensions for one of their chips. The footprint is 2.63mm X 2.98mm which equals 7.8mm2.

If you look at the previous picture of the insides of a CF card, you will see that the chip package is square and is about 3/8 the length of the card, which makes the package
(.375*43.2mm)^2 = 262.2mm2

So you could fit about 33 of those chips in the package, or 66 of them if they were stacked.

If they were 16Mbit chips, it would take 48 of them to make a 96MB card. Seems like it's physically possible, but I don't know if it would be practical from an engineering standpoint.

I wonder whether they are working on a prototype through a foundry partner such as Intel or NSM

I assume this is how they develop all their products. ??

I was surprised that there already is pricing mentioned.

Looks to me like they've definitely already got products (CF cards, etc.) of that size.

The need to incorporate flash within the SDMC controller is apparently critical and SanDisk may not have the technology to carry this out successfully on its own. Eli hinted that the Saifun technology had certain advantages over SuperFlash as long as endurance was not a mission critical feature.

I guess they need a very small amount of memory to hold an ID number and maybe some more to hold some programming code? I wonder why this memory has to be embedded on the controller? Maybe because the memory chips will be NAND type and thus not able to store code?

Also the claim that embedding flash at $110 per chip (presumably 64MB capacity) adds value to a device like a digital camera is hogwash

I don't think the article claimed that -- it was talking mostly about set-top boxes. SST also makes CF cards. OTOH, I don't see what the advantage is to making the memory non-removable.

Samples of the 5-V SST58SDxxx and 3.3-V SST58LDxxx are available in 8-, 16-, 24-, 32-, 48-, and 64-Mbyte densities, with production scheduled to begin in August.
ebnonline.com

wily
Report TOU ViolationShare This Post
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext