REGISTER | LOGIN
Breaking News
News & Analysis

True 3D Chips Harness Nanotubes

12/19/2014 10:44 AM EST
14 comments
NO RATINGS
More Related Links
View Comments: Newest First | Oldest First | Threaded View
Page 1 / 2   >   >>
wilber_xbox
User Rank
Author
Re: Does the chip work?
wilber_xbox   12/30/2014 12:17:47 PM
NO RATINGS
To me it sounds an expensive method of stacking chips and the problem when one of the chips is not working or got damaged during the whole fabrication process is more than using the TSV to stack multiple chips. I agee that using TSV we cannot get the kind of density and leakage that we can get in this chip.

resistion
User Rank
Author
Re: Does the chip work?
resistion   12/23/2014 8:36:24 AM
NO RATINGS
I noticed their CNFET Ion/Ioff is 5000, that means they could support an array of 256Kb maybe, but the RRAM ratio has to be significantly less than that, I think that margin is tight...

R_Colin_Johnson
User Rank
Author
Re: Does the chip work?
R_Colin_Johnson   12/23/2014 8:27:01 AM
NO RATINGS
resistion: RRAM is risky and CNTs are risky, it's even more risky to put them together.

So true! It will take years of deveopment and optimization to see if this solution can be commercialized. Thanks for pointing this out.

R_Colin_Johnson
User Rank
Author
Re: System-on-Stacks could be answer to Moore's law SOS
R_Colin_Johnson   12/23/2014 8:24:35 AM
NO RATINGS
resistion: are CNTs manufacturable

Good point. These researchers claim to have found a new was of "bundling" them for reliability and higher current carrying capabilities, but they are still many years from commercialization I believe. Thanks for the comment.

resistion
User Rank
Author
Re: Does the chip work?
resistion   12/22/2014 11:39:37 PM
NO RATINGS
RRAM is risky and CNTs are risky, it's even more risky to put them together.

resistion
User Rank
Author
Re: System-on-Stacks could be answer to Moore's law SOS
resistion   12/22/2014 11:34:36 PM
NO RATINGS
The companies know how to stack with vias already, so the question would be are CNTs manufacturable.

Astronut0
User Rank
Author
Re: System-on-Stacks could be answer to Moore's law SOS
Astronut0   12/22/2014 4:35:44 PM
NO RATINGS
They might have used some kind of repair mechanism.  The memory layers would be a good candidate for that.

R_Colin_Johnson
User Rank
Author
Re: System-on-Stacks could be answer to Moore's law SOS
R_Colin_Johnson   12/22/2014 4:32:35 PM
NO RATINGS
msporer: Yes you are right here. I guess that's one of the advantages of the stacking die using through silicon vias (TSVs) approach. You only stack die that have already tested good.

R_Colin_Johnson
User Rank
Author
Re: Does the chip work?
R_Colin_Johnson   12/22/2014 4:24:02 PM
NO RATINGS
Astronut0: The chip they built is just an initial prototype, but I'll ask them your questions.

msporer
User Rank
Author
Re: System-on-Stacks could be answer to Moore's law SOS
msporer   12/22/2014 3:52:54 PM
NO RATINGS
yield = (layer yield)^(number of layers) ?

Page 1 / 2   >   >>
Like Us on Facebook
EE Times on Twitter
EE Times Twitter Feed