×

Welcome to the Slashdot Beta site -- learn more here. Use the link in the footer or click here to return to the Classic version of Slashdot.

Thank you!

Before you choose to head back to the Classic look of the site, we'd appreciate it if you share your thoughts on the Beta; your feedback is what drives our ongoing development.

Beta is different and we value you taking the time to try it out. Please take a look at the changes we've made in Beta and  learn more about it. Thanks for reading, and for making the site better!

Building Chips Like LEGO

Hemos posted more than 7 years ago | from the i-need-the-yellow-capacitor dept.

Hardware Hacking 50

MattSparkes writes "It seems that 3D silicon chips, allowing designers to fit more components into a smaller space, could soon be made far easier to create with a little inspiration from a classic children's toy. "Silicon wafers covered with matching patterns of Lego-like teeth and holes could aid the development of 3D electronics, say UK researchers." Crucially, this technique can make use of existing machinery."

cancel ×
This is a preview of your comment

No Comment Title Entered

Anonymous Coward 1 minute ago

No Comment Entered

50 comments

Patents? (1, Interesting)

phil-trick (24853) | more than 7 years ago | (#17613918)

Well, Lego (TM) could be on to a nice earner with this one....

Re:Patents? (1)

LiquidCoooled (634315) | more than 7 years ago | (#17613964)

Not really, this is more like baking a cake and pre-shaping the pastry first.

The units are not hand stacked for home use, they are stacked and welded, just the initial dents and hills assist the alignment process.

Re:Patents? (1)

UbuntuDupe (970646) | more than 7 years ago | (#17614196)

I think the parent was saying that since the parts fit by using the same interface shapes as Lego's, they'd violate the patent because they use that specific shape.

Of course, that's wrong, since Lego's patent has long since expired. That doesn't stop the company from threatening imitators though...

Re:Patents? (4, Informative)

TheRaven64 (641858) | more than 7 years ago | (#17614528)

The patents on Lego expired a few decades ago. They recently tried using trademark law as a work-around, by trademarking the arrangement of dots on the surface of their bricks, but it didn't stand up in court.

Re:Patents? (4, Interesting)

sunderland56 (621843) | more than 7 years ago | (#17615646)

Never mind the patents - Lego is a trademark. The Danish toy manufacturer very strongly defends that trademark.


If you have silicon chips that fit together like little plastic children's toy blocks, that's perfectly fine. But if you mention the word Lego - even in internal company documents - you'll have a swarm of lawyers knocking at your door. (Yes, this has happened before).

Re:Patents? (1)

LordSnooty (853791) | more than 7 years ago | (#17634330)

But if you mention the word Lego - even in internal company documents
Have to call this assertion into question (Yes, I used to work for them, at the old UK office in Wrexham). Sure, you might've got a light roasting if you didn't use the preferred LEGO capitalisation, but to suggest that staff could not mention the company's name in documents is absurd. Alternatively if you're telling the truth, it's a recent development and the company sounds in a sorry state.

Coooling (5, Interesting)

LiquidCoooled (634315) | more than 7 years ago | (#17613934)

I hope the central portions of these chips have enough space to allow cooling to be achieved.

If the stack is open, then could the cooling actually be better than a single over the top method.
This could work like the fins inside double layered home radiators.

Re:Coooling (3, Interesting)

arachnoprobe (945081) | more than 7 years ago | (#17614002)

<blockquote>I hope the central portions of these chips have enough space to allow cooling to be achieved.</blockquote>

One could use special cooling-building blocks. These would even be placed right next to the most-heat producing parts of the overall system and enhance the overall cooling effect.

Re:Coooling (0)

Anonymous Coward | more than 7 years ago | (#17614188)

If you can have more transistors on a chip, then you can possibly let them run at a somewhat slower pace, such that cooling becomes a non-issue. Perhaps.

Re:Coooling (1)

arachnoprobe (945081) | more than 7 years ago | (#17614286)

If you can have more transistors on a chip, then you can possibly let them run at a somewhat slower pace, such that cooling becomes a non-issue. Perhaps.
Like thats ever happened before..... But yes, I agree, That would be good. But for the "normal" customer, pure speed is still the main selection criterium. Disclaimer: I'm not using my room-heating-system since I got my G5 Dual ;)

Re:Coooling (0)

Anonymous Coward | more than 7 years ago | (#17615424)

I hope the central portions of these chips have enough space to allow cooling to be achieved.


Maybe they should make them like Meccano [wikipedia.org] instead?

Re:Coooling (1)

The Lerneaen Hydra (885793) | more than 7 years ago | (#17615654)

The type of space you're talking about needs to be in the range of mm for air to flow through easily, and seeing how this was in the range of tens of nanometers (10^-9), not a chance. in fact you'd have trouble trying to force much at all through that channel, even if you were to actively pump water through it you'd have issues (not to mention the possible galvanic problems).

Low power components (3, Interesting)

Calinous (985536) | more than 7 years ago | (#17614106)

You could stack some low power components for this - while current performance microprocessors won't be made this way (too much power to dissipate), there are plenty of other microelectronics which uses power in the hundreds of milliwatts range. These, you could stack 10 high.
      Also, this could reduce the cost even more in the low cost market - instead of needing a PCB with soldered connections, just put all the components on top each other.

Re:Low power components (1)

arivanov (12034) | more than 7 years ago | (#17614268)

You can still do something like the Current Intel pseudo-4-core CPUs efficiently with this if you change the chip packaging. Cooling both top and bottom simultaneously may be able to compensate for the stacking of 2 layers. This will require complete redesign of the socket as we know it, but the gains may actually be worth it for some applications.

Re:Low power components (1)

Calinous (985536) | more than 7 years ago | (#17614496)

I think you would use water cooling before doing this two-sides cooling contraption (not that it isn't capable of working).
      As a side note, you could mount processors on both sides of the board, and use a hole in the middle of the PCB for the connection from one to the other (using some lego-like block)

Re:Low power components (2, Interesting)

stevesliva (648202) | more than 7 years ago | (#17614534)

while current performance microprocessors won't be made this way (too much power to dissipate), there are plenty of other microelectronics which uses power in the hundreds of milliwatts range. These, you could stack 10 high.
Given that current Intel dice appear to be about 50% cache, it would be nice to be able to slice the cache off and put it under the processor. Might be able to make the access path from the different memory banks more symmetric. Or, since DRAM and flash processes are tailored differently than logic processes, you could replace the SRAM cache with DRAM or flash from wafers processed entirely differently than the processor wafers.

Bones are crunchie (1)

Nathgar (995959) | more than 7 years ago | (#17614202)

"Silicon wafers covered with matching patterns of Lego-like teeth and holes." Great... we're adding teeth and holes to our future grey goo doomsday device.

Re:Bones are crunchie (1)

somersault (912633) | more than 7 years ago | (#17614486)

Don't worry.. if it has holes, we can fill it.

Re:Bones are crunchie (1)

metlin (258108) | more than 7 years ago | (#17614786)

Don't worry.. if it has holes, we can fill it.

I *shudder* to think of it. :-\

Must clean mind's eye. ARGH!

Re:Bones are crunchie (1)

somersault (912633) | more than 7 years ago | (#17615202)

Hah *shudders too* dirty mind you have there o_0

Re:Bones are crunchie (1)

metlin (258108) | more than 7 years ago | (#17615476)

Hey, I stop at organic life forms. Of the feminine human variety, I must add.

Oooh, baby. Check out those transistors. Let me jingle my electron around your hole.

Sorry, couldn't resist. =)

Extra-nuclear humour... (1)

avronius (689343) | more than 7 years ago | (#17616102)

Two atoms are walking down the street.

The first one stops suddenly and says, "DAMN! I think I lost an electron!"

The seconds looks at him (quite concerned) and asks, "Gee man, are you sure?"

"Yeah," says the first one, "I'm positive..."

Big deal (1)

robinvanleeuwen (1009809) | more than 7 years ago | (#17614338)

You want to tell me that in 40 or 50 years of space travel nobody
ever cut up a picture or used photoshop/gimp to apply existing ideas
to next generation space ships?

What i remeber from my Lego days is that i ended up with (what i
personally think) a good desing, with all the wrong colours, but
with about a gazillion of unused Lego blocks. And missing Lego
blocks offcourse.

just my 2 cts.

Re:Big deal (0)

Anonymous Coward | more than 7 years ago | (#17614406)

> You want to tell me that in 40 or 50 years of space travel nobody
> ever cut up a picture or used photoshop/gimp to apply existing ideas
> to next generation space ships?

Well... definitely not gimp.

Cost? (1)

ilovegeorgebush (923173) | more than 7 years ago | (#17614348)

Is this cheap to do? What about capacity? Is it similar to the perpendicular discs theory that Toshiba brought to light a while ago? Could this be applied to storage solutions such as USB keys etc? Or is it more 'static electronics' such as DVD players?

Slow news day? (0, Offtopic)

chiefer (1050460) | more than 7 years ago | (#17614478)

Wow, must be a slow news day. As we all know, noone has EVER stacked components in a microprocessor before..[a href = "http://en.wikipedia.org/wiki/Integrated_circuit#S SI.2C_MSI.2C_LSI"]linky [/a]

Recommend 3D Circuit Layout Software? (0)

Anonymous Coward | more than 7 years ago | (#17614598)

I'm working on an electronic art project that involves circuitry distributed within 3D structures. I'd like to try different what-if scenarios given that I have some constraints to live with. My circiuts are small and simple when compared to microprocessors, but they're complex enough to be difficult on paper or in your head. Can anybody recommend 3D circuit layout software for me? OSS, or low cost are the best for me. Thanks!

Lego Everything (1)

qaz20 (264928) | more than 7 years ago | (#17614664)

What I want to know is why can't you have
lego everything? House, furniture, computer.
It could all fit together and be the ultimate
in reconfigurability.
Imagine using this for cubicles!!

Re:Lego Everything (1)

FusionDragon2099 (799857) | more than 7 years ago | (#17620994)

I actually thought about having a house built out of Lego once. Then I realised walking on the floor barefoot would be really, really painful.

a benifit beond the obvius (2, Interesting)

ZivZoolander (964472) | more than 7 years ago | (#17614776)

sure better configurations could be made by stacking, but there is a manufactoring benifit that would reduce the costs of making chips by a very large factor. From what i remember from a tour in national semi. I was told that on a waffer of silicon they would etch several hundred numebr of chips, of which a certin precentage of of chips will have a defect(which are detected by laser optical scanners), due to the qty of transisotrs they are trying to squeeze in a square inch. so if one transistor is bad they just lost the entire chip for that transitor. (and if to many chips are lost they just toss the whole waffer) with the stacking method you dont have to make an the entire chip on the same peice of silicon. now you can partition the chip onto two seperate dyes then "lego" assemble working halfs. you could see a yeild increase as much as 25% depending on sition(if not more) because your only tossing defective partitions. thus reducing overal costs (probably enough to offset the amount to change to the lego method)

Re:a benifit beond the obvius (0)

Anonymous Coward | more than 7 years ago | (#17616364)

people have been doing that to improve yields for decades already(multi-chip-modules).... the problem with it is you loose some performance as you go from chip to chip, so you really have to limit the interconnect between the Die (and architect it well). it seems like this article is mostly aimed at MEMS, but this might be useful for some of those stacked DRAM modules that companies like samsung have made...

Re:a benifit beond the obvius (1)

ChrisMaple (607946) | more than 7 years ago | (#17617414)

The advantage of vertical stacking of bare dice is that the interconnect can be extremely short. I'd guess this could be in the 10's of microns, best case. Multichip modules are 2D, and minimum interconnects are thousands of microns. Vertical stacking thus promises substantial improvement.

Re:a benifit beond the obvius (0)

Anonymous Coward | more than 7 years ago | (#17616392)

Cool idea. Would it possible to get something like that working in 2d?

Re:a benifit beond the obvius (1)

Manchot (847225) | more than 7 years ago | (#17617212)

I really doubt that chip manufacturers could decrease their costs doing what you're describing. The biggest problem is that every planar processing step is incredibly expensive. For one thing, the fabs themselves cost billions of dollars to construct, and it cost millions more to maintain a clean environment. Secondly, you're not going to find your typical factory worker in a fab: pretty much every worker there will have (at least) a bachelor's degree in electrical engineering or materials science. Basically, every redundant step will cost you millions of dollars in operations costs, so the fewer steps you can do, the better. This is why the size of the wafer used has crept upwards as time has gone on: you can fit more chips on a wafer, and reduce the number of steps. What you're talking about would multiply the number of processing steps, greatly increasing your costs.

talking about this for 25 years (1)

peter303 (12292) | more than 7 years ago | (#17614818)

I've been hearing 3D proposals since the 1980s, but not much has come to market. Even kludges like multiple 2D chips bonded together are more common.

Re:talking about this for 25 years (1)

JazzHarper (745403) | more than 7 years ago | (#17615214)

Aside from the obvious cooling and interconnect problems (Google for "hairy smoking golfball"), testing and yield are sigificant obstacles. Testing the entire stack is much more difficult than testing an individual chip. Once you bond together a stack of chips, a flaw in any one of them means you have to throw away the whole stack.

Heathkit (3, Interesting)

sharp-bang (311928) | more than 7 years ago | (#17616614)

When I was a lad, Heathkit marketed an educational analog circuit-building kit wherein the circuit elements (resistors, capacitors, transistors, etc.) were encased in Lego-like bricks and connected on a Lego-like board instead of a breadboard. It was great fun - my brother and I built every circuit in the book, and then some - but unfortunately the kit interoperated a little too well, using the exact same dot-matrix as real Legos. We could sanp real Legos right into the circuits. The kit came off the market very quickly and my understanding is that the settlement with Lego contributed to Heathkit's eventual demise. Oh well.

(If anyone out there has the kit and wants to sell it, drop me a line.)

wait, what? (0)

Anonymous Coward | more than 7 years ago | (#17618032)

which building chips like LEGO, and how do LEGOs chip??

I'm not understanding the headline here and I'm not RTFA.

Nice Idea, but *maybe* not practical (1)

WillRobinson (159226) | more than 7 years ago | (#17618532)

This work has been going on for a long time in the market, in so far as die stacking. The problem I see with this implementation is unequal heat growth will tear them apart. (on large surface area dies) maybe not in the 10 to 20 mill range, but thinking about large areas like processor size dies, this will be a problem.
Check for New Comments
Slashdot Account

Need an Account?

Forgot your password?

Don't worry, we never post anything without your permission.

Submission Text Formatting Tips

We support a small subset of HTML, namely these tags:

  • b
  • i
  • p
  • br
  • a
  • ol
  • ul
  • li
  • dl
  • dt
  • dd
  • em
  • strong
  • tt
  • blockquote
  • div
  • quote
  • ecode

"ecode" can be used for code snippets, for example:

<ecode>    while(1) { do_something(); } </ecode>
Sign up for Slashdot Newsletters
Create a Slashdot Account

Loading...