IBM probably processes less than 100,000 wafers, mixed 200mm and 300mm, a month. If they are yielding 80% to devices that is 20,000 wafers a month available for solar applications. And this is a very aggressive estimate. It isn't a lot compared to the needs of the photovoltaic people, who now by more silicon that the semiconductor companies.
But recycling is good, I guess, given the cost of making silicon in the first place.
After cooking silicon dioxide (sand) with carbon, the silicon has to be reacted with hydrogen chloride to form dichlorosilane. The dichlorosilane is distilled to purify it. This process is not too energy intensive as the boiling point of dichlorosilane is near room temperature. Then the dichlorosilane is decomposed at high temperature onto huge mandrels, another highly energy intensive step. The polysilicon this produces is melted again and the single crystal silicon is pulled from the melt. That's three 'meltings' each requiring scads of energy. Then there is zone refining for purification, which is equivalent to melting the whole boule several more times.
Can this be done cheaper? Well, if it could, it would be. There would be huge incentives. At the moment you need a hydroelectric plant or some other cheap power next to your facility to make silicon cost effectively, the same as if you were making aluminum. There is a company called Evergreen Solar that uses a different method of making silicon into flat sheets, that may or may not save some of the cost, but I'm betting less than 30%.
Other solar materials such as CIGS may be cheaper to make in the long run as they are deposited as thin films (about a micron or less) on less expensive substrates. There are an enormous number of startups working on this: e.g. Heliovolt and Solyndra.
I have lots of money, because I work for a business. And because I work for a business, I need to check my business e-mail and calendar. I think the iPhone is a great product, and I am looking forward to getting one when ActiveSync is available. I can't fathom what Apple were thinking about when they left this feature out. Maybe people who get a lot of mail at work aren't cool enough to have one?
Has been around years but is still cool
on
Magnetic Fluids
·
· Score: 3, Informative
This has been around >15 years. In fact some of the tools that made the chips in your PC probably had ferrofluidic bearings.
Because these liquids can be held in place by magnets, you can make a feedthrough into a vaccum chamber that can be rotated. The fluids have low vapor pressure so you can have a high vacuum system with a rotating shaft entering it - that's very difficult normally. The fluid seals between the shaft and the sleeve, where the air would normally leak in. Good down to about 10^-10 of an atmosphere.
Try www.ferrofluidics.com.
Regards, John the semiconductor capital equipment designer.
You need SSB radio. It's a non profit service. You aren't going to be able to surf with it, but it works for e-mail. Stan Honey, a fairly legendary ocean racer and tech head, runs it.
I must misunderstand SMP. A lot of scientific programs, that aren't written to be parallelized or don't parallelize well, never fork the running process. So I don't see where the OS can come in and allocate part of the task to another processor. In fact for a single computation running on the box, in a non-parallelized program, the only advantage I can see is that any OS housekeeping would take place on the second processor. With a good OS I can't think this would lead to a huge improvement. If more than one case of the program runs on the same box there will be an improvement but consider that the processors now have to share the memory and IO so the actual performance would depend on a lot of things to do with caching, how much IO is needed, etc. While if one just buys two cheap boxes instead this isn't an issue. Cheers John.
I think Cliff headlined this one wrong. Read what the guy asked, it doesn't mention realtime anywhere. He wants to know how to do cheap scientific computing. Cheers John
I recently put together a cluster of eight Gateway 2000 boxes with Athlon 800MHz proccessors. We loaded Red Hat 6.1 and are using one of the commercial Fortran compilers to do plasma simulations. The boxen are in a very basic Beowulf-type configuration.
Our previous boxes were 600MHz DEC alpha stations running Dec's UNIX, OS/F or whatever it is. We find that the AThlon boxes, which are 32 bit of course as opposed to the 64 bit Alphas, are about as much faster as the clock speed would indicate, i.e. about 30%. As a result we increased our computation resources by a factor of four for less than $20K. We are very happy.
I'm not sure SMP can be justified in this kind of case, as boxen that support it are typically way more expensive than our cheapo Gateways, and SMP generally does not increase speed by the factor you would think. However I'd be interested to hear any results to the contrary. When problems can be split between processors, performance per buck is what matters.
The best solution really has to be tailored to the specific problem being solved; so for example 384MB or PC100 RAM was ample in our case but in a big 3D finite element case it gets to be a problem. As an example we also have an electromagnetic package that runs on NT; because of the software licensing we can't run more than one case at once, and that has to go on the huskiest system I can get the budget for, currently dual Pentium III 733s with 2GB RAMBUS memory. It is way less cost effective than our cluster system.
"I mean, if Mills flies up to him in on of his anti-gravity flying saucers and gives him a rust-resistant suit made of magnetic plastic, will he still say "It's all bunk. Gimme my 11 dimensional strings back!" or what?" Nice comment, you have a way with words there. If you showed all this to me I'd still want proof that anti-gravity was how it worked. And there are rust-resistant suits and magnetic plastic already. Boring of me, huh?
But you make two points here, and the one doesn't follow from the other. If I give you a regular rechargable battery and say I charged it from a power supply driven by a dilithium crystal, but I can't show the crystal to you, there's no reason for you to believe me. I assume there must be bug free code somewhere so your example with the blender doesn't seem to prove much.
What I'm saying is that the explanation for what he's produced is probably mundane. As for the stuff/effects being exotic, as they may be, do you think he'd get many investors if he handed over a glass of water and said he'd made it with hydrions? Now, the Voice article mentions UV light twice but I didn't see any actual evidence that it had been detected by anyone. One guy mentions that they bought some cells and 'got the same results'. Including UV emission? Who can tell? I really look forward to seeing these guys publish their work with a full description of what they did. In these cases you generally get companion publications from other researchers supporting the findings and that'll be nice too. But I don't think i'll hold my breath while I wait.
"I can't be wrong" isn't something you'll hear much in most labs I've worked in, but what you will find is a strong presumption that the scientific work of the past hundred or so years is not going to be unexpectedly turned on its head. It could happen, of course. It just isn't very likely. If you want to stake some money on it, nobody's stopping you but I recommend the lottery instead.
Slashdot Mirror
IBM probably processes less than 100,000 wafers, mixed 200mm and 300mm, a month. If they are yielding 80% to devices that is 20,000 wafers a month available for solar applications. And this is a very aggressive estimate. It isn't a lot compared to the needs of the photovoltaic people, who now by more silicon that the semiconductor companies.
But recycling is good, I guess, given the cost of making silicon in the first place.
It is actually much worse than that.
After cooking silicon dioxide (sand) with carbon, the silicon has to be reacted with hydrogen chloride to form dichlorosilane. The dichlorosilane is distilled to purify it. This process is not too energy intensive as the boiling point of dichlorosilane is near room temperature. Then the dichlorosilane is decomposed at high temperature onto huge mandrels, another highly energy intensive step. The polysilicon this produces is melted again and the single crystal silicon is pulled from the melt. That's three 'meltings' each requiring scads of energy. Then there is zone refining for purification, which is equivalent to melting the whole boule several more times.
Can this be done cheaper? Well, if it could, it would be. There would be huge incentives. At the moment you need a hydroelectric plant or some other cheap power next to your facility to make silicon cost effectively, the same as if you were making aluminum. There is a company called Evergreen Solar that uses a different method of making silicon into flat sheets, that may or may not save some of the cost, but I'm betting less than 30%.
Other solar materials such as CIGS may be cheaper to make in the long run as they are deposited as thin films (about a micron or less) on less expensive substrates. There are an enormous number of startups working on this: e.g. Heliovolt and Solyndra.
I have lots of money, because I work for a business. And because I work for a business, I need to check my business e-mail and calendar. I think the iPhone is a great product, and I am looking forward to getting one when ActiveSync is available. I can't fathom what Apple were thinking about when they left this feature out. Maybe people who get a lot of mail at work aren't cool enough to have one?
This has been around >15 years. In fact some of the tools that made the chips in your PC probably had ferrofluidic bearings. Because these liquids can be held in place by magnets, you can make a feedthrough into a vaccum chamber that can be rotated. The fluids have low vapor pressure so you can have a high vacuum system with a rotating shaft entering it - that's very difficult normally. The fluid seals between the shaft and the sleeve, where the air would normally leak in. Good down to about 10^-10 of an atmosphere. Try www.ferrofluidics.com .
Regards, John the semiconductor capital equipment designer.
Try www.sailmail.com.
You need SSB radio. It's a non profit service. You aren't going to be able to surf with it, but it works for e-mail. Stan Honey, a fairly legendary ocean racer and tech head, runs it.
John.
I must misunderstand SMP. A lot of scientific programs, that aren't written to be parallelized or don't parallelize well, never fork the running process. So I don't see where the OS can come in and allocate part of the task to another processor. In fact for a single computation running on the box, in a non-parallelized program, the only advantage I can see is that any OS housekeeping would take place on the second processor. With a good OS I can't think this would lead to a huge improvement. If more than one case of the program runs on the same box there will be an improvement but consider that the processors now have to share the memory and IO so the actual performance would depend on a lot of things to do with caching, how much IO is needed, etc. While if one just buys two cheap boxes instead this isn't an issue. Cheers John.
I think Cliff headlined this one wrong. Read what the guy asked, it doesn't mention realtime anywhere. He wants to know how to do cheap scientific computing. Cheers John
I recently put together a cluster of eight Gateway 2000 boxes with Athlon 800MHz proccessors. We loaded Red Hat 6.1 and are using one of the commercial Fortran compilers to do plasma simulations. The boxen are in a very basic Beowulf-type configuration.
Our previous boxes were 600MHz DEC alpha stations running Dec's UNIX, OS/F or whatever it is. We find that the AThlon boxes, which are 32 bit of course as opposed to the 64 bit Alphas, are about as much faster as the clock speed would indicate, i.e. about 30%. As a result we increased our computation resources by a factor of four for less than $20K. We are very happy.
I'm not sure SMP can be justified in this kind of case, as boxen that support it are typically way more expensive than our cheapo Gateways, and SMP generally does not increase speed by the factor you would think. However I'd be interested to hear any results to the contrary. When problems can be split between processors, performance per buck is what matters.
The best solution really has to be tailored to the specific problem being solved; so for example 384MB or PC100 RAM was ample in our case but in a big 3D finite element case it gets to be a problem. As an example we also have an electromagnetic package that runs on NT; because of the software licensing we can't run more than one case at once, and that has to go on the huskiest system I can get the budget for, currently dual Pentium III 733s with 2GB RAMBUS memory. It is way less cost effective than our cluster system.
Hope this helps.
John
"I mean, if Mills flies up to him in on of his anti-gravity flying saucers and gives him a rust-resistant suit made of magnetic plastic, will he still say "It's all bunk. Gimme my 11 dimensional strings back!" or what?"
Nice comment, you have a way with words there.
If you showed all this to me I'd still want proof that anti-gravity was how it worked. And there are rust-resistant suits and magnetic plastic already.
Boring of me, huh?
But you make two points here, and the one doesn't follow from the other. If I give you a regular rechargable battery and say I charged it from a power supply driven by a dilithium crystal, but I can't show the crystal to you, there's no reason for you to believe me. I assume there must be bug free code somewhere so your example with the blender doesn't seem to prove much.
What I'm saying is that the explanation for what he's produced is probably mundane. As for the stuff/effects being exotic, as they may be, do you think he'd get many investors if he handed over a glass of water and said he'd made it with hydrions?
Now, the Voice article mentions UV light twice but I didn't see any actual evidence that it had been detected by anyone. One guy mentions that they bought some cells and 'got the same results'. Including UV emission? Who can tell? I really look forward to seeing these guys publish their work with a full description of what they did. In these cases you generally get companion publications from other researchers supporting the findings and that'll be nice too. But I don't think i'll hold my breath while I wait.
"I can't be wrong" isn't something you'll hear much in most labs I've worked in, but what you will find is a strong presumption that the scientific work of the past hundred or so years is not going to be unexpectedly turned on its head. It could happen, of course. It just isn't very likely. If you want to stake some money on it, nobody's stopping you but I recommend the lottery instead.
John