Electromechanical Switches Could Reduce Future Computers' Cooling Needs 95
Earthquake Retrofit writes "Science Daily is reporting that researchers at Case Western Reserve University have taken the first step to building a computer capable of operating in extreme heat. Te-Hao Lee, Swarup Bhunia and Mehran Mehregany have made electromechanical switches — building blocks of circuits — that can take twice the heat that would render electronic transistors useless. 'The group used electron beam lithography and sulfur hexafluoride gas to etch the switches, just a few hundred nanometers in size, out of silicon carbide. The result is a switch that has no discernable leakage and no loss of power in testing at 500 degrees Celsius. A pair of switches were used to make an inverter, which was able to switch on and off 500,000 times per second, performing computation each cycle. The switches, however, began to break down after 2 billion cycles and in a manner the researchers do not yet fully understand. ... Whether they can reach the point of competing with faster transistors for office and home and even supercomputing, remains to be seen. The researchers point out that with the ability to handle much higher heat, the need for costly and space-consuming cooling systems would be eliminated.'"
Steampunk (Score:4, Interesting)
Miniaturized relays are interesting, but an inverter which operates at 0.0005 Ghz is less interesting. Somehow I don't think we'll be seeing this replace electronics anytime soon. (well, except in lithium battery microcontrollers :-) ). Although it would be interesting technology for a steampunk novel.
Re: (Score:3, Funny)
Oops - gotta go - the device has a lifetime expectancy of only 1 hour, 7 minutes even at this slow speed.
Next time, I'll boot up that old 4.77 mhz PCjr.
Re: (Score:2)
It can be interesting technology for certain peripheral devices, such as encoding sensor data in those 500 C places, as long as the data rates are not too fast.
Diamond Age (Score:1)
Kind of like the "rod logic" in Diamond Age by Neal Stephenson.
Re:Steampunk (Score:4, Insightful)
As you alluded to, there are billions of microcontrollers out there running at less than 0.5 MHz. Heck, the good ol' Mostek 6502 ran at 1 MHz, and started an industry (KIM-1, Commodore PET, Apple ][, etc.) This is still in the research stage, but even at the current speeds, useful processors could be built (but apparently not last very long), especially since this would open new markets where traditional semiconductor gates won't function.
But, even short of a full processor, there could be uses for logical applications (gate arrays).
Re:Slow processors (Score:3, Informative)
The venerable 6502 was also the heart of the VIC 20. A slightly modified/improved version was used in what was the most ubiquitous personal computer, the Commodore 64. Although running at only 1Mhz, most of its instructions executed in less than 3 clock cycles, making for some pretty efficient and fast ML code.
Re: (Score:1)
The venerable 6502 was also the heart of the VIC 20. A slightly modified/improved version was used in what was the most ubiquitous personal computer, the Commodore 64. Although running at only 1Mhz, most of its instructions executed in less than 3 clock cycles, making for some pretty efficient and fast ML code.
So? Today even simple processors can finish several instructions per clock cycle (even though each instruction could take longer). A processor with these mini-relays could still run circles around the 6502 even with a slower clock.
Re: (Score:2, Interesting)
Just to be clear, processors don't run at the toggle frequency (f-sub-t) of an inverter. At best, a CPU will run somewhere south of 1/10th of the inverter toggle frequency. So we're talking more like 50 kHz, probably slower.
Not saying that there aren't applications for a really slow CPU running in a 500 degree C environment. (Like my DVR? Chuckle...)
Re: (Score:2)
Re: (Score:2)
Miniaturized relays are interesting, but an inverter which operates at 0.0005 Ghz is less interesting.
Also less interesting? That if we were to get it up to 1 GHz, then the processor would exhaust itself in two seconds.
Re: (Score:2)
Miniaturized relays are interesting, but an inverter which operates at 0.0005 Ghz is less interesting.
Also less interesting? That if we were to get it up to 1 GHz, then the processor would exhaust itself in two seconds.
Good enough for the flight controller in a smart missile.
Re: (Score:2)
Miniaturized relays are interesting, but an inverter which operates at 0.0005 Ghz is less interesting.
Also less interesting? That if we were to get it up to 1 GHz, then the processor would exhaust itself in two seconds.
Good enough for the flight controller in a smart missile.
I think smart missiles usually travel at least a few minutes away...
Re: (Score:2)
Not to mention, they're heat-resistant, but they don't really have a heat problem, because they don't have any leakage! Plus, they can't be manufactured at any great scale with e-beam lithography, break down in about an hour even at their current slow speed, and are quite a bit larger than existing transistors. They've got some work to do before these get interesting, and I think there's a lot of other potential future technologies that are far ahead of them in development.
Re: (Score:2)
Well if they get the life up to a couple of hours, they can try selling dongles made with them. Combine the worst of old and new tech. Make it so you drive to the video store to buy a dongle to watch a streaming video.
If these things are really mechanical, they probably vibrate some. Why not send vibrations through the water and gas mains to carry data? Or the sewer... yikes... crap speeds?
Time to watch the movie Brazil again... something about all those ducts and tubes and the information ministry.
Re: (Score:3, Interesting)
Miniaturized relays are interesting, but an inverter which operates at 0.0005 Ghz is less interesting. Somehow I don't think...
This is an incorrect and unfair assertion -- unfair by stating the switching rate in GHz.
In the real world, DC-DC inverters run below 1 MHz. From Wikipedia:
"Unlike a linear power supply, the pass transistor of a switching mode supply switches very quickly (typically between 50 kHz and 1 MHz) between full-on and full-off states, which minimizes wasted energy."
http://en.wikipedia.org/wiki/Switched-mode_power_supply/ [wikipedia.org]
Re: (Score:2)
Uh, TFA is referring to a logic inverter, not that kind of inverter.
Although a miniaturized electromechanical SMPS would be interesting.
Re: (Score:2)
Except this has nothing to do with power supplied, but with inverters. You know NOT-gates.
Re: (Score:2)
- Dan.
Re: (Score:2)
Miniaturized relays are interesting, but an inverter which operates at 0.0005 Ghz is less interesting. Somehow I don't think we'll be seeing this replace electronics anytime soon. (well, except in lithium battery microcontrollers :-) ). Although it would be interesting technology for a steampunk novel.
Of course, even if the thing does run at high speeds for an acceptable lifetime, there are a lot of other components in a computer system that would also have to have increased thermal tolerance. It's not just the CPU you have to worry about.
Fire code (Score:2)
Re: (Score:3, Interesting)
We had the opposite problem.
The sprinkler activation system was disconnected at all times (excepting inspection). That made me feel safe.
Eventually -I- figured out what was wrong with the system and fixed it, and with the alarms cleared they felt safe enough to put the activators back on.
Re: (Score:2)
They might be tightfisted, but when nobody allows their equipment to be housed there, they'll lighten up a bit.
Re: (Score:1)
At 500 degrees Celsius, the water will evaporate instantly.
But... (Score:2)
Too big, too slow, too fragile (Score:2, Insightful)
Re: (Score:2)
Within an order of magnitude for a brand-new technology is pretty good. I know people on Slashdot don't usually realize this, but there's a lot of engineering work that goes on between the development of a new technology and when that new technology outpaces the old technology.
Re: (Score:2)
Relays are back! (Score:4, Informative)
I tried to tell them that tubes and transistors were just a fad. Relays were good enough for the Z4 and they're good enough for us. These kids and their newfangled gadgets...
Re: (Score:2)
Remember the Starship Enterprise, NCC1701? (No bloody A, B, C, or D)
It's computer used relays! You could hear them every time someone told the computer to do something! ;)
Re: (Score:2)
With this device we may be on our way back to the vacuum tube, just smaller.
http://www.radioremembered.org/vpwrsup.htm [radioremembered.org]
Re: (Score:3, Informative)
There are STILL applications where vacuum tubes are better than transistors. Guitar amps and other pieces of audio equipment is a prime example. I wouldn't consider an amp that wasn't tube, nor would most serious guitarists. Many audiophiles also swear by tubes, for dynamic range and warmth. If you want a two way radio system to be able to work during a nuclear blast, only tubes will do, as the EMF will render transistors useless.
I would not be shocked to see some new application for tubes or tube-like
Re: (Score:1)
If you want a two way radio system to be able to work during a nuclear blast, only tubes will do, as the EMF will render transistors useless.
Come on... exactly *what* would you be transmitting during a nuclear blast? Except possibly "Ouch. Bye."
Re: (Score:2)
Come on... exactly *what* would you be transmitting during a nuclear blast? Except possibly "Ouch. Bye."
"Roger, right on target, returning to base. Over". is one example.
Also, if you are the one being bombed, you might want to actually tell someone from inside your deep earth bunker, since the hardwire has now been cut.
Re: (Score:2)
EMF hardening of semiconductors is neither new nor difficult, it's just a damn nuisance. And FWIW tubes don't work during an EMF pulse, either.
The problem is that active devices turn on as hard as they can during an EM pulse. If there's enough energy in capacitors attached to the active devices to destroy the active devices, that's what happens. Tubes can't turn on very hard, are inherently resistant to quick destruction by high temperatures, and have enough mass that they're not going to get heated much by
Re: (Score:2)
As far as guitar amps is concerned, semiconductor circuits can be made to perform in the manner that tubes do, and more reliably, but there's just too much myth and momentum for tubes for them to be obsoleted as they should be.
There is a lot more to guitar tone than clarity. I have tested more than a few dozen digital solutions and while it might be hard to quantify the difference to someone who hasn't spent 40 years playing, it is easy to pick in a blind test. It isn't about fidelity (where tubes come u
Re: (Score:2)
Is there really some effect that you couldn't replicate with 192 kHz sampling and a digital effect? Really? That people can actually hear in blind tests?
Yes. And I notice that it is never musicians who actually question this, only those with no experience. It has nothing to do with sampling rate. It isn't clarity, or any of those things that "on paper" would mean "better". It has nothing to do with fidelity. If you don't understand what tube pump and compression are, you will never get it. It has to d
Re: (Score:2)
Actually, tube amps and vinyl sound better especially in the past decade and a half. Tubes have a much better overdrive handling characteristic than solid state transistors do. Ove
Re: (Score:2)
Many audiophiles also swear by tubes, for dynamic range and warmth.
You mean the sort of people who buy these: http://www.wired.com/gadgetlab/2008/06/snake-oil-alert/ [wired.com] ?
Re: (Score:2)
No I mean almost every professional guitarist I have seen on stage. From Stevie Ray Vaughn to Junior Brown, Eddie Van Halen to Chet Atkins. Literally every serious professional guitarist, with very few exceptions. So they are all idiots?
It helps to actually understand what you are talking about if you are going to compare guitar amp technology, or call tube amps "snake oil". Every once in a while, I hear someone say exactly what you are saying, which I find humorous since it is always from someone who i
Re: (Score:2)
Impervious to electromagnetic radiation (Score:4, Interesting)
Both attributes that the military would like.
Re: (Score:1)
In addition to the military, I could see this technology having possible uses in space. If it can perform to standards in the radiation of space this could be interesting. At 500c these chips could also reduce the need for thermal management systems in un-manned spacecraft. That is IF these are producible, IF chips made from these can compete with existing options, and IF they get through that whole 2 billion deal......
Re: (Score:2)
You've hit it exactly. Consider that this is basically pre-alpha technology and it is already in the ballpark for performance while showing exceptional thermal and mechanical advantages. Provided it can be built to withstand ionizing radiation, you've got a perfect match to spacecraft components. Gyroscopes, accelerometers, stellar orientation, sensor polling and alerts, pressure sensors, heat sensors, atmosphere sensors, all of these things require at least some processing. If all of these things could be
Or Mercury (Score:2)
As the article states, these would be ideal for processors that operated on the surfaces of Mercury and Venus. While a probe on the surface of Mercury could probably get away with good (but very large and heavy) heat shielding, Venus especially has always been a problem for probes and the Soviet Venera probes that delivered images of Venus' surface lasted only somehwere around 57 minutes before failing. A probe based on this technology, i.e. a working high temperature design, could last much longer.
Re: (Score:1, Insightful)
Hum, these relays are still working with electrons AFAIK, so they would still generate (susceptible to eavesdrop) and be susceptible to EMF. Pure mechanical switches are not, but thes are electromechanical, thus exploiting mechanics to switch on and off a current
Re: (Score:3, Insightful)
If there is current, there is electromagnetic radiation. And the EMP could create a current transient high enough to cause an arc across the switch contact, which could have temperatures far far in excess of 500 C and melt or fuse the tiny switch. There may be arcs already at the normal operating current wearing it out.
Re: (Score:2, Insightful)
Re:Impervious to electromagnetic radiation (Score:4, Informative)
How prescient ;-)
No, when they mention radiation in the article it's because these devices are radiation-hard, i.e. they will last a long time in a radioactive environment such as many satellites fly in. Standard silicon CMOS devices on the other hand degrade very quickly because charged particles get trapped in the gate oxide changing the gate threshold, degrading performance, and then eventually killing the device. The silicon crystal itself is more vulnerable to defects from radiation, which increases channel resistance, again degrading performance and killing the device.
The SiC MEMs devices are more robust because SiC is more robust at high temperatures and radiation filled environments, plus as a primarily mechanical system rather than electrical, it will probably be more tolerant to crystal defects.
In a computer this will emit just as much electromagnetic radiation as a silicon chip because the radiation comes from the flow of current being turned on and off. It doesn't matter one bit if you do it with a vacuum tube, BJT, MOSFET, or MEMs device, you will get electromagnetic radiation.
It may be more robust to EMP than a Si-CMOS device, but it will still be vulnerable to contact degradation when an EMP causes breakdown of the air or vacuum dielectric.
The military does love the idea of MEMs switches, more so for radar/comms, but they got burned bad after DARPA and DOD agencies funded 10's of millions, or probably more, in R&D with no useful results. The main problem with MEMs switches has always been reliability, which you will see is also a problem in the MEMs devices being promoted in this article.
Re: (Score:2)
Being EMP-safe means all semi-conductors (not just the transistor substitutes) would have to also be EMP-safe. Kinda pointless to use relays for that if an EMP fries the power supply capacitors, for example.
And conventional tempest protection easily outweighs the performance/size/weight compromise of using relays.
We'll need something sexier than this to remove the yawn tag from the OP.
Not obvious how it works (Score:2)
This is what I thought (Score:2)
Two billion cycles (Score:3, Funny)
... is easily explained. There is a flaw in the design. Someone apparently made the "wear out expiration" register signed instead of unsigned.
Re: (Score:2)
Energy conversion to heat still exists (Score:4, Interesting)
There are two sources of heat in modern semiconductor CPU's.
One is leakage, the heat generated by current times resistance squared in transistors that are off.
The higher current that is related to the clock speed is the heat generated by transistors that are turned on using the same current times resistance squared.
To keep the on current at a bare minimum, transistors are paired with one on and one off so the current through the pair should be zero except for leakage. The current flows when they are clocked and the capacitance (stored voltage) of the wire between transistors and gate capacitance of the MOSFET it drives supplies current during switching.
How does this no heat switch avoid the current of switching the capacitance between the switches. From what I can tell is this part is able to handle higher temperatures. I do not see it as a no power (no heat generated) device.
Silicon Nitride has much higher resistance than most metals. Due to the resistance and temperature resistance, it is often used as hot surface ignition in gas appliances. Current through the switches will create heat. It is unavoidable.
At it's current speed of 0.000.5 GHZ clock speed, I can believe the current power consumption is very low. How does this stack up to an Atom CPU clocked at 0.000.5 GHZ?
Re: (Score:2)
The point is not that these relays generate no heat, but that they can survive up to 500C, while modern CPUs survive about 100C, but start working incorrectly. Discrete transistors and low density logic chips (like the 4xxx series) survive up to 125C. Assuming the power dissipation and ambient temperature is the same, you need much smaller heatsink to keep the device at 400C than you need to keep it at 60C.
Atom at 500kHz would probably use less power, it still would melt if the ambient temperature was 150C.
Yet another "breakthrough" (Score:2, Insightful)
This is a research idea that MAY be useful, the demise of CMOS silicon has been highly exaggerated.
From the summary:
"an inverter, which was able to switch on and off 500,000 times per second" -> 500kHz is not so great
"however, began to break down after 2 billion cycles" or about 1 second at current processor speeds. That increases to 4000 seconds at 500kHz, or a little more than an hour.
Also, we can put billions of error free transistors on a chip for a few dollars. THAT is the real hurdle that nothing e
Re: (Score:2)
For specialty needs, if your device needs to function at 500 degrees, then a very costly 500khz processor may be quite ok, only the longevity needs to be fixed.
You'll anyway write specialized, tiny, optimized code to run on it and the hardware controlled by the chip may easily be valuable enough.
Think small chips on spacecraft, where the sunny side may get very hot; or tiny controllers attached to the burning part of missiles (from ICBM's to simple air-to-air) - the temperature capabilities may easily allow
Works at 500Khz and lasts 2 billion cycles.. Oh. (Score:2)
So in other words it'll last 4000 seconds. Little over an hour. Perhaps notalot of use just yet except perhaps in
short lived weapons - missiles and suchlike. But they don't need heat tolerance.
computers will still make lots of heat (Score:1, Interesting)
A gate input acts like a capacitor. Something in the output that feeds it has to limit the inrush current. Whatever that is generates the heat. If the resistance isn't supplied by a transistor, it will be supplied by the wires. The basic formula that describes the heat generated by a gate doesn't change.
What may change is that the circuit may become less damped. That could lead to problems with ringing.
So, I'm not convinced that these mechanical switches will have any advantage on heat generation and t
Great (Score:2)
Re: (Score:2)
Ah, but not if they use molten salt [wikipedia.org] as an electrolyte. I can imagine an organization saving so much on "cooling costs" as they send the server admin into hell on earth to swap out a component or fiddle with the cables.
Quit talking about gigahertz (Score:5, Interesting)
Nobody's going to use this for desktop CPUs. The whole point is that the switches work at 500 degrees C, where silicon doesn't. This technology would be used for embedded control in extremely hostile environments, where 500 kHz would be just fine. The article names the inside of a jet engine and the surface of Venus as examples.
Re: (Score:2)
OK, so my webcam on Venus will have to be 320 x 240 at 6 frames per second. No big deal.
Re: (Score:2)
What, you don't keep your house 500 C? You wouldn't believe how much I save on my A/C bill each month.
These would be great for high-end test equipment (Score:5, Interesting)
Re: (Score:2)
MEMS switches have been around for a while now in the ATE and RF markets:
http://www.memsindustrygroup.org/i4a/pages/index.cfm?pageid=3759 [memsindustrygroup.org]
Coming soon! (Score:1)
So what are they going to call it? (Score:1)
silicon carbide is already used for transistors (Score:1)
for precisely the reasons outlined here. they operate very well at very high temperatures.
anything operating on a mechanical basis will have a finite lifetime. a billion cycles sounds like a lot, but not when everything is switching 10 million times a second.
so these things will never work as a replacement for transistors.
making things smaller will always result in greater leakage. there might be salvation in quantum transistors. however, see tunneling.
also see mems.
Re: (Score:2)
You must be kidding (Score:2)
Mechanical devices are more durable then silicon? Who would have thought.. other then Captain Obvious of cousre.
Contact wear (Score:1)
one of the main reasons transistors were sought out is to eliminate constant maintenance on electromechanical devices
so how long is one of these going to last at 500 degrees C running at ~500khz ?
Re: (Score:1)
It says in the summary these things are breaking after 2 billion cycles. Is it the math of it "all" you are having problems with or is it the reading of the *entire* summary that is causing you trouble?
Babbage? (Score:1)
The engineers took their cue from English inventor Charles Babbage, who built a steam-driven machine to calculate mathematical tables in the 1830s. The group applied nanotechnology to make switches fit today's ever-smaller computing platforms.
So they want to build an computer with electromechanical switches - and they take their inspiration from Babbages's steam-driven machines instead of the computers with electromechanical switches [wikipedia.org] that Konrad Zuse build?
2 billion ops? (Score:3, Interesting)
First we have flash memory can that only be written to N number of times, and now they're building a cpu that can only do N computations?
Re: (Score:2, Interesting)
NASA (Score:4, Interesting)
Seems like something interesting for planetary exploration where standard CPUs on a probe would be rendered useless in a matter of hours. Much as the equipment sent to Venus.
Re: (Score:1)
That's exactly what I was thinking, Venus probes are notoriously hard to cool. Also, solar explorers, or anywhere else in the solar system where there is high heat, would be good targets for high temp electronics. With some improvement on temp and longevity this might be just what is needed. The current record for Venus is 107 minutes.
Not just for inverters... (Score:1, Interesting)
Not yet mentioned is the opportunity to use MEM switches for filters, modulators, phase comparators and a few other useful devices. The basic principle is synchronous rectification, whereby switches of a bridge open and close in sequence to rectify an incoming signal. A key advantage is elimination of L-C components and the ability to generate and filter arbitrary waveforms.
but tubes sound better :-) (Score:2)
Zuse would like it... (Score:1)