View Full Version : Cryogenics Cooling Mods (OC Thread)
Hi everyone, I am now asking this forum (the second forum I have brought this problem to..) for some advice and help. I think that the subject we are conversing on may be better understood by looking at the forum post at HardForum.com. In any matter, please post your replies to this forum if you are a member here. This way it is easier to respond to questions/comments/offers. I will properly credit individuals that help. Thanks a lot!
Regards,
Prescott Paulin, VP
Research & Development
300 Below, Inc.
My Original Post To The HardForum.com Forums-->
Original Post (http://www.hardforum.com/showthread.php?s=&threadid=248365): http://www.hardforum.com/showthread.php?s=&threadid=248365
Hello everyone. I am the Vice President of Research and Development at 300 Below, Inc. headquartered in Decatur, IL. 300 Below is the world's largest and oldest cryogenic processing company. (300below.com) The reason I am writing this forum is to ask for some help. I believe that cryogenics may very well apply to computers, mostly in the sense of cooling so that they may be overclocked. Another use is for future processors that generate so much heat that they will not sufficiently work with current cooling methods.
So far I have only been hitting dead ends. I am not sure if the cooling effects of liquid nitrogen will damage any parts of a computer by way of condensation. We are able to custom build processing units, and do so for NASA and various other government institutions on a regular basis, but are not sure that building a processing unit to encase a computer is the best option. Do you think the best way to proceed would be to completely process the entire machine and hold it as low as -300F or would a system using, for instance, a peltier effect on the processor via tubing and other equipment (like the water coolers) be more advisable?
I would do the testing myself in the department, but the idea is sketchy so I am not able to allocate funds to this project. If anyone has any ideas I would appreciate your help. If the idea is of use, we can hopefully develop an inexpensive product that will benefit all computer users. Also, if anyone works for a microprocessor manufacturer, case manufacturer, motherboard manufacturer, or system integrator and thinks that their company may wish to do joint testing with 300 Below, Inc. (normally with no charge at all) please let me know. I am very interested in finding some solutions for this industry. So far we have found that many things, from sporting goods to disposable razors to musical instruments to engines, have responded to the process. It is my job and my interest to find a way to improve the methods currently available for cooling.
I think the general agreement around here is that anything below -50C will slowly destroy anything silicon. When you're getting that cold the silicon contracts so much that it starts pulling apart. Might work for a couple of days at -300F (if you know what you're doing and are DAMN lucky).
As far as cooling the whole board to around, say, -30C... I'm not really sure how that would affect some components like capacitors. Overall this sounds like it would cost about as much as a computer and be a pain to work with. Vapochill cases cost about $650 and they keep your cpu just below freezing... that's about as cold as I've seen done with any degree of practicality.
cowanrg
10-29-01, 12:23 PM
Well, here is a little background information for you to use. Computer cannot operate at VERY low temps. I have seen computers that were liquid nitrogen cooled, but compenents such as the motherboard and expansinon cards failed. They just weren't made to be frozen like that. They cannot operate. (plus, as you know, extreme cooling can impede electricity. NiCD batteries for instance, have 30% of their life when in cold conditions.)
Watercooling with peltier would be best. However, as you also know, condensation is an issue. The ideal would be to have a system that could regulate the peltiers so that would always hover right above freezing, so condensation would not form.
You mentioned processing the whole system, and holding it at a low temperature. This could work, but you have to remember, most people want computers that they can move around fairly easy. You would have to insulate (or make external) most peripherals. for instance, if you put a cd into a cd-rom drive that is at even -100F, the cd will maybe crack, or at least, fog up becuase it was at room temp. So, some componenets will have to be outside of the case, thus making a bulky, and complex setup.
My suggestion, as mentioned above, is to make a cooling setup specific to the components you are trying to cool (chip, graphics card, chipset), and make the system regulate this. temperature probes in sync with power control would yield not only great temperatures, but also little worries about condensation, etc.
Send me a private message if you would like to discuss this further. (i will also be in springfield, IL for about a week, and would love to see what you have going on there!). I also have my email on my profile. Hope this helps a little.
Right now, as we have discussed in the HardForums, we are looking to NOT cool the CPU, but rather treat the chip with our normal process. This means that we take the item down to around -300F and bring it back up slowly. This will hopefully permanently change the molecular structure for the better, resulting in better properties of the unit when used normally. Basically we take the chip, process it, send it back, and you overclock/test/etc... We would prefer to test it ourselves to gain more accurate results though, however we do not have the equipment nor the complex knowledge of the computer parts to do the testing ourselves. This is where we need to partner with, say, Intel or AMD to do the testing. Many people have volunteered to get their chips frozen, but I have to figure out what the whole process will entail. Thanks for your feedback. And by the way, it is not a problem to arrange a tour of the facility. We must have some advance notice, and I can probably take care of that for you.
Regards,
Prescott Paulin, VP
Research & Development
300 Below, Inc. (Decatur, IL)
If you would like to email me, please send correpondence to presky@300below.com. Everyone that would like to participate in a test using their chips, please email this address as well. Include the specs of the item you wish to volunteer. Thanks again.
-Prescott
cowanrg
10-29-01, 12:47 PM
OH, that makes a little more sense. You would most likely need to speak with Intel or Amd as you mentioned. However, I wouldn't be surpirsed if they didn't know how the chip would react either. Do you have any data sheets on basic processes on cryogenic freezing? (how it benefits materials?). Have you tried just tossing a chip into the tank? I'm sure there are numerous of us in here that could help you diagnose the chip on a technical level. Let me know how far you have gone into this.
Wow. I've never seen a corporation jump into something like this- 300Below.com is "takin' it to the streets!" so to speak.
Imagine if more companies got involved like this......the OC world would be a better place!!
Prescott, I have a stack of P2 233 socket chips I might donate to the cause. Also, my main chip- a very happy Celeron 566 doing 995mhz. If you can share any more information about what you are actually going to do to the chips, please do. I'd be more inclined to let you work magic on my CPUs if I was sure I'd get it back again-working!:D
I have used these services many times, we used to treat teeth for our grinder this way. I also am aware nascar uses these services to harden certain parts. Mainly what you are doing is just tempering the item. Works great on metals, I can attest to that. Silicon? Heck toss one in. I had our rep treat my pocketknife once and It worked great on it. I am sure someone would be happy to send you a test chip. Then analyze it when it returned.
We do many various motorsports items for various drivers. For example, Nascar and Indy 500 drivers. We have non-disclosure agreements though, so we cannot release the names of these individuals/groups. They find it it work very well for these applications, and if you notice on our website, we build our own custom hum-vees that utilize cryogenically treated parts. Tooling is our main user base, which can explain for this customer attesting to our performance. (BTW- I am curious to know who you are, so if you don't mind, please email me.)
As for the experimental side, I cannot guarantee how the chips turn out. I would expect that there be no damage, and rather an increase of performance. Thanks again everyone for your help!
Sincerely,
Prescott Paulin, VP
Research & Development
300 Below, Inc.
flounder43
10-29-01, 06:54 PM
Sniff, sniff, something smells funny here. Why wouldn't you go to a manufacturer like Intel or AMD?
BTW, can I dip my trumpet in your cryogenic freezer?
flounder43
10-29-01, 07:13 PM
Originally posted by presky
We do many various motorsports items for various drivers. For example, Nascar and Indy 500 drivers. We have non-disclosure agreements though, so we cannot release the names of these individuals/groups.
Strange, every other company that I have ever heard of wouldn't hide, but rather would require their name to be advertised as a provider of parts for race cars. And what specifically do you do for NASCAR? I know enough to know that their are strict rules about altering those cars... Is it a SECRET?
Maybe you could clarify some of your "various . . . items" for "various drivers".
Finally, how much to dip my trumpet?
There's nothing funny going on here... I can't proceed to contact a manufacturer like Intel or AMD because I don't have enough proof that this could work, which would be wasting their time. In the future, if that happened, they would not waste their time again looking at the idea, even if it turned out to be valid.
I am not in the Cryo-Motorsports division so I can't tell you why we do this, or who the people are if we could. The fact is that I know we have non-disclosure agreements signed because the racing teams requested them. This doesn't mean we can't freeze for other people, it's just basically taking away our right to say, "We treated their stuff." They most likely do this because there is such a competitive advantage that they do not want their opponents to find out about the process. It is not illegal to cryogenically process anything in Nascar. When we treat something it neither adds nor takes away weight, but rather just changes the molecular structure. Also, since I am not in that division, I can't tell you what parts we treat. You are welcome to call and ask though.
The price for treating your trumpet will vary. You are more than welcome to call our main office. (800) 550-CRYO There will be someone there to help answer your questions regarding the treatment of your trumpet. I have my saxophone treated and it works great. And although I am not a professional player I can't say what specifically works better, but I can say the tone is much smoother.
Whatever the matter, thanks. Good luck with getting your trumpet treated. We have a quick turnaround for that, as well as developing the original process to treat instruments.
Regards,
Prescott Paulin, VP
Research & Development
300 Below, Inc. (Decatur, IL)
A few points.
First, it would seem that the problem with silicon contraction would occur even with a one time treatment. This should be easily testable. Treat a chip and then put it back in the computer. If you can run the computer for a few weeks thereafter, you're probably safe. Durons are cheap these days and this should give you a first approximation to whether the idea is at all workable.
Second, your description of things so far suggests that the primary benefits acrue to properties such as hardness or rigidity. Are you currently treating anything in order to improve it's electrical properties?
Third, heat dissipation is an increasing problem in computers. Have you considered whether cryogenically treating a copper heatsink might improve it's thermal transfer properties?
Fourth, I'm just up the road from you. If there's anything that would be helpful with, drop me a line.
nihili
-=HN=- Wild9
10-29-01, 08:44 PM
i just have a question, not to sound rude or anything at all just a simple inquiry.
how would cryogenically freezing something then bringing it back up to room temp very slowly change its physical properties in any way?
i mean when you cool something the molecules move more slowly until a point 0 K or i think its -273 C when the molecules stop moving all together, then bringing it back to room temp very slowly would only allow the molecules to begin moving at their normal speed again, well unless (with certain elements or compounds) the molecules rearrange themselves in a neat way, like frozen water molecules, then being able to keep them that way, well like water for example, water molecules arrange themselves in such a way that when water freezes it becomes less dense than its liquid form, if there was a way of bringing it back to room temp, with the molecules staying the same way well you just got yourself a nobel prize,
but you can do that in the opposite direction, take a steel rod and heat it until it glows coool it very slowly and the molecules stay arranged they way they were when heated and i think that process is called tempering.
well just my 2 cents and feel free to corrent me if im wrong on anything, constructive critisicm brings out the flaws in everyone ya know.
flounder43
10-29-01, 08:49 PM
I did some reseach.
300 Below is owned by your father, as is shown by your own web page: http://members.aol.com/lcd2/friends.html
The Illinois Sec. of State confirms: http://www.cyberdriveillinois.com/cgi-bin/business_services/corpone.s?ID=56744894&CTYPE=01
I also know that there is a lot of controversy about "cryo" treatment of various things, including musical instruments. I am skeptical.
Why don't you, and your "company" buy some processors and test them yourselves? You could conduct a scientific test that would confirm your theories about changes to the molecular structure of things.
I am not trying to me harsh or mean. However, I cannot betray my nature to get more of an explanation when such a fantastic proposition is conveyed...
I smell snake oil.
It's a newer process (cold tempering) molecules supposedly get "denser" on Nascar they us it for pistons, cranks and some other things. It really does work on metals, I had tested the life of some tool steel bits v/s treated and got at least 45% better life. But now silicon is a different story.
heh heh, snake oil, heh heh, yeah!
I don't smell snake oil, nor have I ever smelled it in our business operations. You know, you really ought to do some more research before you try to put a false image on our company. Yes, my dad did start the majority of the industry spawning copycats along the way. It's funny to see you are based in Minneapolis where one of our "competitors" is based. I'm not going to draw any conclusions on this, but I will say this. We are a company that does millions of pounds annually, does business with major companies and organizations on a regular basis (like NASA for instance) and has been featured on the Discovery Channel's "Next Step". Matter of fact, the viewer response was an all time high for the show and the ran it over again many subsequent times. So before you try and create a false image and get other people to think we are out to get them, think again. We have more experience and years of operation as well as a size factor under our belt than any other commercial cryogenic processor. If there was a project budget allocation to this particular project I would be happy to do that. The fact of the matter is that we are geared toward more industrial operations such as dies and drill bits that smaller projects such as this don't get any funding. I'm sorry if you have a problem with that, but then again, you don't run the business. Also, regarding the comment about "Me and my 'company'", it's not mine. It's my father's. Good day.
"You could conduct a scientific test that would confirm your theories about changes to the molecular structure of things."
Reply>> Hello! That's why I am writing this forum; to try and figure out how to proceed with testing.
You also state that this is a fantastic proposition. How is that? This is not fantastic. This is boring research that might lead to the development of new products. It's what I do.
I'm sorry if you are confused.
Regards,
Prescott Paulin, VP
Research & Development
300 Below, Inc.
snyper1982
10-29-01, 11:34 PM
now im not trying to be rude or anything, but if your company is so huge, then a 50 dollar duron processor should be small frys to them. Oh sorry, i mean your fathers company, you could actualy buy a few yourself if you were that serious about it, i mean come on your the VP. So what if they get messed up thats what research and development is all about, trial and error, am i right?
Yes, you are correct, but I didn't get any allocation for funding on this particular project. If it was rock solid I would by pass the forums and make a proposal to a chip manufacturer or a systems integrator. You are correct about the research and development, and if this happens then its the test participants that are at risk. We don't lose any money, but then again, if it works the test participants are credited. I would love to be serious about it, and as far as things currently go I am, but there are other people above me that would rather perform R&D for the industrial applications, thus I am stuck with no funding. Sorry, but this is the way the business world works. I don't know why I have to explain operating procedures to people who are complaining. If you don't want to do anything, you don't have to. I don't have to persue the development of an application to this industry, but I think it would be nice to improve the industry if something works. This is all in the name of science and computers. If this works I may end up creating something that all computer users can benefit from, and it may even pave the way to faster and/or more complex processors.
I am not trying to be ingrateful, as if it were my processor that cost me good money, then I would be wary of trusting it to someone who just suddenly popped up on a forum. However, we have been in business for a number of years. On our normal processing we offer 100% money back (no questions asked) guarantee. We have had an extremely low rate of requests for a refund, as our service is top quality. Bottom line: what we currently do works. I don't think anyone should have a problem dealing with us, as we never leave a customer in the dust. I especially will not alienate a customer. Also, look at the previous posts for someone who attests to the improvement of an object when cryogenically treated. You will find a happy customer. Thanks again everyone for your questions and comments. I look forward to answering them to the best of my ability.
Sincerely,
Prescott Paulin, VP
Research & Development
300 Below, Inc.
Thelemac
10-30-01, 12:21 AM
Ok, guys, let's not start flaming.
Thank you for giving us a chance to respond to this idea here, Prescott, and for enduring the slightly harsh treatment.
I am curious as to what it is that you're proposing. Are you suggesting that we treat our processors and then run them like that, in hopes of increasing the thermal transfer properties? I guess I don't quite understand what it is that you think that it will do. I could see using your process to cool the chips (though as previously stated, that isn't feasible due to the whole pulling apart thing), but I don't understand how treating it helps. I admit that I really don't know a whole lot about how the whole cryogenic thing affects metals...if possibly their life is better because they can transfer heat better, or if they're just plain harder.
On another note, I don't know that taking the chip to the temperature is really safe. As it does have problems at ~50C (will run for a short while, but it is pulling itself apart the whole time it's at that temp) I don't really think that cooling the chip that far will result in a chip that actually works. The best way to test that would be to try it on a system that you know works (you could build a test system for no more than $100) and then test it after you treat it. That will give you an idea of whether or not its feasible to spend much time on. I am, of course, assuming that you could get access to the faciliity without much of a problem. If that doesn't work, then you have your answer about treating currently manufactured CPUs.
However, there is a chance that it would help the manufacturing process to use your system, or something similar. For that, of course, you would have to contact a chip maker to find out if using your type of cooling *during* the manufacturing process would help. I think that it would, because you could get the silicon to root itself much closer together so that you could cool the chip more (at least, that's kind of how I think it works) and that would be good. That would probably also help with the overall speed and power consumption, heat release and so forth.
I hope that this has helped you. :)
Thelemac
If, and I use the word loosely. This process worked, what would a regular everyday joe expect to be charged to get a Cryobaked chip?
As well I think one way of eleviating some of the worry that people on this board have of the credability of your idea. Would be to look a little further into making it beneficial to a small select group of volunteers. In otherwords set something up with a few people. They will LEND you their chip for you to process and in the case of failure you would allow them some form of compensation. It's only fair. This way it could be a matter of only needing a few hundred dollars set aside in the case of failure. As a Senior member of a high-tech company, you probably make a lot more then the 20,000 a year that I make. Making a personal sacrifice of a few hundred dollars (that if this works you will not have to use.) would actually be beneficial to your cause. Now personally I think thats the main worry of a lot of the people reading this thread. What my friend are YOU willing to do for them if they help you.
Now I understand that you feel it COULD prove to be a great method of enhancing these chips. Cool, but that's more a bonus to you then it is to the guy that get a few more Mhz out of his processor.
You mention that you are doing this of your own free will not the will or desire of the company, but that if the idea becomes feasable you may be able to push it into an actual area of R&D. Very cool again for you.
I personally feel you have a rather unique concept, something that I personally would love to see turn into a good thing. But people are pesemisitic in nature and need some proof. Setting up some sort of controlled experiment with a few people would allow for that to be settled for the common fellow.
Now I hope you take my words in there inteneded reading. I do not hold any malace towards you or your idea, in fact I wish it well. If you wish to discuss other methods of structuring a seed test for R&D of this idea feel free to contact me as I have a good eye for figrin out that kind of thing :)
Orion
All spelling errors, gramer problems and other assorted problems with my English are all the fault of my highschool teachers as they failed to keep me awake to learn. :D
Prescott, don't be too bothered by the skepticism. We here have heard of many, many magical cooling methods that promised the world. As you can imagine, they usually end up bogus. :D
IMHO, I don't see how your process could help, but then again I know jack squat about the technology you are talking about. I sure hope it turns out to work! Anything that helps you OC!
If I understand what you are talking about, I guess you would need a pretty large sample of chips in order to really understand how well it worked, right? That would definately be cost prohibitive, but try this- you can score working 486s these days for almost nothing. Dunking them in the magic cold sauce would give you insight as to how more modern (and expensive) equipment might behave- if 10 of 10 dunked systems shoot sparks at turn-on, you might need to rethink.:)
Personally I think it's pretty cool you are experimenting like this- it sounds like an idea worth getting into.
Also- for some reason I had a brain flash on trying this process on bare circuit boards before having components added. Like, build a vid card only with the parts that can survive -300 on it, do the process, then finish building the card. Just a thought.
r0ckstarbob
10-30-01, 03:19 AM
presky
i recognize your post here as the same over at the [H]ard fourms... interesting to see the different reactions from people on both threads.
i'd definately recommend emailing Dave Smith, the author of this EXCELLENT article on thermodynamic theory and practice in regards to computer cooling.
http://www.amdmb.com/article-display.php?ArticleID=105
go to the link and his email address is attached to it.
if theres anyone who can fully understand the entirety of what you're proposing, it will be him. and if not, he'll know somebody who will.
good luck, and please keep us posted. try not to mind the skepticism. flounders not really on the witch hunt warpath like he's comming off... we HAVE had alot of folks with some cranked ideas that they were passing off as fact drop onto the forums here. (this idea of yours is absolutely NOT one of those ideas in my opinion btw)
it's something of a trial by fire sort of thing i guess - and in the end, often times the only thing we have to seperate the wheat from the chaff.
peaz
RSB
flounder43
10-30-01, 11:47 AM
Well put, RSB.
It is important to remember that I was not flaming or witchhunting, but was rather trying to find out more about what I felt didn't make sense. So I was merely looking for some answers.
I would have been more receptive if Presky had just laid it all out at first. However, when things don't seem quite right, or things are stated in an overly ambiguous way, there is usually a reason.
Nevertheless, best wishes to you, Presky.
cowanrg
10-30-01, 12:08 PM
I think ROH posted hit on a very poignant topic... price. This is something that should be taken into consideration above all others. If this process runs anywhere more than $50, i doubt response would be that high... as cheap as chips are, it's just not worth it.
do you think you could give us a rough estimate on how much this thing could cost to do? then the question could be asked "if this works to make your processor 30% cooler, would you spend $XX.xx to have it done?" then you can see if you want to move forward on this matter.
Right now I am trying to find if this is a viable idea to take to a system integrator or manufacturer or computer components. From what I hear it is. I am not sure that we would make this process available to the public unless there is demand, and then it would probably be no more than 15 to 20 bucks for a processor and heatsink including s/h. Once again, I am not looking to sell the process on the forums, but rather figure out if this is a viable idea that will benefit users in the future. I think the idea is proceeding smoothly, and we will find newer ways to implement the process with other components. I will let everyone know what turns out.
Regards,
Prescott Paulin, VP
Research & Development
300 Below, Inc.
cowanrg
10-30-01, 02:00 PM
yeah, i understand that you aren't selling anything, but in research and development, the first question is always "if we can do it, can we sell it."
just because it can be done doesnt make it a good idea :) but in that price range, i will grab my pen and ask "where do i sign up?"
looks like you have something.
foxmulder
10-30-01, 04:22 PM
Just a thought on my part but have any of you thought about why "Presky" might not want to talk to the big companies about this? If it does turn out to have merit AMD or Intel would have their own R&D Departments all over it. Especially a company the size of Intel would have no problem getting patents (real fast) and making all the money themselves. If I were him I'd do just this, Hell I'd bet that this board alone has more OCing experts than Intels entire R&D department. The one thing I do find a little strange is that the VP of R&D of a tech company isn't willing to throw $50 of his own at such a project. Aside from that : way to go. I'll be interested to see where this goes.
I know this sounds strange, but my $50 has to go into our industrial projects. All of our research in this area has resulted in great things. This is why I don't put my own money into this project, although I hope it takes off so I can get some funding to experiment more.
The major chip manufacturers probably would not do this, even if tests showed it netted some improvement. I've followed 300 below for several years from the perspective of treating firearms. Whether the process would add some improvement to copper heatsink performance, or a processor is probably moot because, as I understand it, the process is too time consuming. Remember, mass production equals profits. I had the opportunity to use a razor that was treated with this process and yes, it stayed sharper a lot longer, but it was a $.35 piece of merchandise. That, having been said, I am very interested in learning what, if any improvement could be realized, both in copper heatsinks and CPUs. The whole idea titillates my imagination. I watch a lot of TLC and TDC.
Some great person once said, "Our reach should always exceed our grasp."
Hoot
FrozenInHI
10-30-01, 10:28 PM
Originally posted by presky
I know this sounds strange, but my $50 has to go into our industrial projects. All of our research in this area has resulted in great things. This is why I don't put my own money into this project, although I hope it takes off so I can get some funding to experiment more.
your $50 has to go into your industrial projects? if it's resulted in great things, and you seem to believe that this experiment will result in great things, then why wouldn't you invest 50 bones to get a great thing going? I am kinda confused, I mean you want me to send you my 150 dollar chip to you to try in your "Demo Man" freezer, then send it back to me and "HOPE" it works? Just a thought, but i bet you don't get many takers here! Try it for yourself first, then when you find results, post here, and maybe you'll get a few takers. But it seems that a VP of a company would be a bit more resourceful than to just post a thread asking for CPUs to be sent his way so he can "hopefully not destroy them" and return them in better condition than when they left. What would your contingency plan be? What would happen if I sent you my chip and you returned me a Mr. Wizard liquid nitrogen frozen banana that you shattered with a hammer? Would you replace my chip or chalk it up to R & D of our partnership? Also, what's in it for us? If we were to give up some CPUs to you to try out, and it in fact did increase performance without jeapordising lifespan, what would you be willing to do for us?
My main intention of writing this forum is not to get chips to test from people. It is rather to gather information from people on what they think will happen using the process. If my intention was to get chips I would have already posted an address. The process would not differ in any way from our normal treatment procedures if we were to test this. Right now the idea seems plausible, but we are looking into the possible effects on the processor without treating it right now. Thanks again everyone.
Sincerely,
Prescott Paulin, VP
Research & Development
300 Below, Inc. (Decatur, IL)
r0ckstarbob
10-31-01, 02:42 AM
Originally posted by FrozenInHI
your $50 has to go into your industrial projects? if it's resulted in great things, and you seem to believe that this experiment will result in great things, then why wouldn't you invest 50 bones to get a great thing going? I am kinda confused, I mean you want me to send you my 150 dollar chip to you to try in your "Demo Man" freezer, then send it back to me and "HOPE" it works? Just a thought, but i bet you don't get many takers here! Try it for yourself first, then when you find results, post here, and maybe you'll get a few takers. But it seems that a VP of a company would be a bit more resourceful than to just post a thread asking for CPUs to be sent his way so he can "hopefully not destroy them" and return them in better condition than when they left. What would your contingency plan be? What would happen if I sent you my chip and you returned me a Mr. Wizard liquid nitrogen frozen banana that you shattered with a hammer? Would you replace my chip or chalk it up to R & D of our partnership? Also, what's in it for us? If we were to give up some CPUs to you to try out, and it in fact did increase performance without jeapordising lifespan, what would you be willing to do for us?
we're not talking about one chip, we're talking about 10 or 50 or even 100 test chips in as many computers to get even the most distantly accurate predictions to expected performance gain before the production of such a system even BEGINS development. it seems to me mr paulin is doing his job... that is seeking out new and innovative ways to expand his department using the techniques at his disposal... the forums being one of them.
i think it's important to note that unlike us, this isn't a personal hobby for mr paulin, this is his job. it's how he feeds his kids and makes his car payments. would you throw even 50 dollars into your place of buisness as an employee into something like this? no, of course not. at least not without expecting some compensation from your company for the funds you've invested in it on their behalf. and if you did it unsolicited, you'd be lucky to get even that back. it's unreasonable to expect him to, and for his purposes, futile.
lets take a step back and look at it from a business point of view. lets assume that just developing the treatment process to standard to yield the most performance increase per chip, training the personnel, and then designing and building the machines that will actually make this all happen will cost the company in the range of say $150,000 dollars to instigate. (and thats CHEAP and not counting the marketing and promotion that will have to go into letting the general public know of it's existance).
thats $150,000 the company will have to invest in startup fees just to develop and market the process to the general public, at which point the company (like any other company providing a brand new service) has to pray that their product is going to sell and that they will make those dividends up. thats quite a bit of money to risk. it's not necessarily alot to invest, but it's alot to risk. in the grand scheme of the companys overall gross, perhaps not that much, but it's still a hefty chunk of change.
now before this company is going to invest or risk $150,000 dollars and commit a staff of personnel to this task, it's going to want some assurances ahead of time that
they will be able to markedly improve the performance of the chip with a process that they can develop and market as a service that they provide
that this will be a process/product that will likely appeal to the consumer market or that there's even a market in the first place
that they will be able to provide this service at a cost that will be reasonable and reachable to the general public
and providing that the above statements are true, that they can reasonably expect to see a return in their investment at some point, allowing them to at least break even
so whats the first step? research. who's job is it to explore new ideas that could potentially benefit the company and their investors? usually the upper management. in this case, mr paulin has adopted this role.
this isn't bad or evil. this is the way it works. he's researching whether or not theres a market for something like this, he's asking the folks who have the most knowlege and combined practical experience in the arena their expert opinions as to overall feasability, and at some point he's going to need test chips and public product testers. this isnt' snake oil, and it's not BS guys (i'm talking to everyone expressing skepticism in this thread) - this is the process of integrating a new, untested thing into an old idea to make it better. this is totally new territory for mr paulin, his company, and even US - the overclocking/hardware enthusiast market. there is absolutely nothing for him to reference anything against ahead of time to guage whether or not his idea has any merit, so he's come to ask the target market themselves. when it's all said and done, thats what has to happen in the end.
mr paulin wouldn't have to just toss 50 bucks into it of his own money to test out the feasability of it and do just one chip, he'd need at least 10 to 20 chips to get an overall average; thats 10 to 20 seperate computers. and for every computer to get tested under "normal" or even "extreme" work loads and conditions for say just a week... lets say each test computer (hand built) is going to be 500 dollars - thats 5 to $10,000 dollars just in hardware...
and if each computer gets tested for say 96 hours prior-to and after treatment to get the direct performance comparison data he'll need, thats 1,920 to 3,840 hours of testing...
if mr paulin were to do it all by himself, he would need to invest $10,000 of his own money and an entire YEAR of 11 hour days just in testing to compile his data... not counting the time it takes to assemble computers for testing, install OS's, treat the chips, etc etc etc....
all this so he could provide some hard facts at the next board meeting justifying the cause of cryogenically treating chips so he could convince his company to go ahead and invest (or risk) $150,000 of their own money on what he's trying to demonstrate is "a good bet" to set the whole thing in motion.
and all the board has to do is say "NO", and all that time, energy and money is utterly wasted.
providing it worked in the first place. and it's STILL just labrotory data after its all done and it still doesn't have the rigors of public testing or configuration differences. its the same reason they do test showing to the public of movies before they're released.
now does it make ANY sense whatsoever for mr paulin to throw "50 bucks" into the pot any more? no. of course not. does it make sense for us to give the man 20 bucks to do an experiment that has every possibility of increasing the performance of our chips? well, it certainly wouldn't be out of character for us now would it? :)
If AMD or Intel asked you to potentially donate your chips to an experiment like this you'd have ripped your CPU out of your computer just after scribbling the mailing info on whatever hard surface happened to be nearest as quickly as you could (computer screen or not) and sent your chip via carrier pigeon if you had to in order to participate AND YOU ALL KNOW IT!!! :). and the funny part is that they're not even cryogenic experts.
try to see this thing from ALL the angles before you condemn it. he's not asking for volunteers (yet). he's merely getting some public opinion from the folks who know what they're talking about and who, in the end, this idea will make the most impact upon.
thats us. the bad, the beautiful...
...us overclocker nerds.
long live the clock cycle. and your mom. ;)
peazzout
RSB
PS whats he willing to do for us? providing your a tester, it sounds like he's giving you the opportunity to participate in an experiment that will quite possibly give you a chip that performs better then it did when you bought it, at a marginal price, and you'll have it LONG before anyone in the public market has one. it sounds fair to me and it sounds like quite a bit in return. we've got enough data between us to guestimate within reason that there will be little or no damage to the chip after it's been warmed up that we can tell (this from others who have done it under far less controlled circumstances), so we know the risks, where as mr paulin, doesn't. to me, the risks sound pretty decent. i'd certainly give it a shot if the opportunity was presented to me.
Originally posted by Monster of Rock
Dunking them in the magic cold sauce . . .
Magic cold sauce . . Thats GREAT!!! I needed a good laugh! LOL!
------------------------------------------
Hmmm . . . It sounds very interesting, alright
I fired you off an email about it. I'd be interested in testing, but unfortunately, I live just a little bit more down the road from you: Ontario, Canada.
If you were testing heatsinks, I'd send you off my FOP38 in a heartbeat to be dipped in your magic cold sauce :). It could only reduce my temperature! But with my processor, thats a different story. I payed a lot of money for that thing back in the day ($200 CDN) and it hasn't let me down yet. I even chipped the corner off and it still goes (see sig). That, and while its gone I'd have to find a replacement Duron or something to hold me over until it got shipped back, and even then this process might make the die even MORE brittle, and it might shatter like glass under the pressure from my FOP38. And have you considered what might happen to the porcilain (or the stuff that the XP is made out of) when its dipped in your secret sauce? That might shatter as well.
Processor dies are based on a very very VERY refined architechure. Doing this might damage this architecture; throw it out of synch. Then you have one dead processor on your hands, and all you can say is that it was dipped in magic cold sauce when your friends ask you why you have a processor for a keychain.
Personally, I think its a very good idea. I mean, nobody that I know of has thought of this before. But what you're proposing doesn't really have a niche. The people who would require incredibly cool processors use their computers all day/every day. When they send their processors out to be baptised, they can't use their computers unless they buy a replacement processor. Sounds moot, but its not. Strike 1.
People who would do this wouldn't know the outcome until they get their processors back in a few days/weeks/months. That adds a high level of doubt in their minds. A long time to think about wether or not they will see their beloved AXIA or AXIAR or AYHJA again. They also don't know wether or not this process will help or hinder them. It might make their processor run 500 or 600 mhz over stock @ a phoninally low voltage w/ incredibly low temperatures, but then again, it might make them run lower than stock or only at stock, which would be a very bad thing for most people, or they might not even work at all. Strike 2.
Don't get me wrong, I'm not trying to find fault w/ your research, I'm just outlining that the people you're dealing with are essentially, the computational elite. Thats why you came to us. But the fact is, that in doing this, many of us can't use our computers again till we get out processors back, and even after that it might not work. But on the flip side, they might work better than we had ever anticipated. It is that kind of doubt that would keep anybody from doing this. We would need to see hard evidence that this process wouldn't damage our processors i.e.: they would still work when we got them back (not even overclock, just work).
You're playing to a very tough crowd here. Some proof that it wouldn't damage our chips would be nice. Technical data etc.
I hope your research goes well
foxmulder
10-31-01, 05:13 AM
Well from reading the sigs around here it seems that there are quite a few people with older less valuable chips laying around. If I really understood how this process works and why its supposed to increase the performance of my chip I'd give it a whirl. I have a backup Duron 750 that I could spare and I know a lot of other people have multiple systems and processors. Only problem is: I'm in Austria.
As far as feasability I believe that there are a LOt of techno geeks out there who would pay $20-30 for even a 10% increase in performance when/if the effect has been proven.
If I understand this process correctly, it improves the long range order and makes the materials more crystalline like. This would improve the properties of the treated materials. However CPUs are made on crystalline silicon wafers, I dont think much improvement can be made on that. But the buss lines may show some improvement. Inorder to treat a CPU it must not be mounted in a package. The difference in expansion rates would destroy the chip. However the heat sink should show some improvement from this treatment and heat sinks are cheap and easy to test before and after. I say try the heat sink first, less problems in testing.
If it didn't shatter the cpu core from the extreem cold. The very least it would do is make the core less fragile, which would be a good thing for us AMD owners. I have seen cryo treated alluminum blocks that are almost as tough as steel. This process definatly works and works well, at least on metals.
This works well for metals that are not in crystalline form, however this is crystalline silicon. It is not a metal, it is a semiconductor with an almost perfect crystal lattice. I think there would be little or no improvement for the silicon, maybe some for the buss lines. The gate insulators would most likely have problems and the CPU would not function. The temp is reduced slowly so shattering is not a problem, but thermal stress on the die is.
Actually thermal stress only occurs if the temperature is reduced or increased too fast. Hope this helps... -Prescott
Prescott's point about thermal stress is true. It's also largely irrelevant.
The problem with cooling a chip is that you are working with materials that have different coefficents of expansion and very tight tolerances. A chip is not a uniform structure. Each of the materials in the chip will shrink at different rates at the chip cools. While this is true of a saxophone to a certain extent, a saxophone has considerably more leeway for different expansion rates. In a chip, if you're off as much as .2 microns it's a disaster. This makes cooling a chip far more difficult than cooling a saxophone or a tie rod.
If you fill a bottle full of water and cool it to -300, the bottle will break. I don't care if you take three years to cool it, your average bottle just can't withstand the increased pressure as the water freezes. This has nothing to do with thermal stress.
Now it's possible that the materials within the chip are sufficiently malleable that they can compensate for the differences in the coefficents of expansion. I don't know. I suspect the only real way to find out is to cool a chip down and see if it still works. I have a 486 and a P120 I'd be happy to donate to science just to see if it works. Cool the chip down very slowly, I'll pop it back in the machine and see if it still runs. If it doesn't, you can alter the procedure and have a go at it with the second chip. If it does work, then you can use a chip that has been thourougly tested for performance in someone's rig, cool it, and see if it runs better. But with the risk of destruction at least somewhat abated, you should find more takers.
nihili
Originally posted by nihili
If you fill a bottle full of water and cool it to -300, the bottle will break. I don't care if you take three years to cool it, your average bottle just can't withstand the increased pressure as the water freezes. This has nothing to do with thermal stress.
The bottle would only break because when you freeze water, the molicules re-arrange themselves differently. Into an octagonal pattern I believe. Thus increasing the size of the water and therefore the pressure.
Originally posted by Oni
The bottle would only break because when you freeze water, the molicules re-arrange themselves differently. Into an octagonal pattern I believe. Thus increasing the size of the water and therefore the pressure.
Precisely. It's not a failure due to thermal stress. It a failure because the water gets bigger at a certain rate when you freeze it and the bottle doesn't.
I'll admit that this isn't really a case of differing coefficients, but it was the clearest example of something breaking even during very slow freezing.
nihili
If I am correct, isn't water the only thing that expands when it is cooled? This is just off the top of my mind at 11:00, so pardon me if its a stupid question.... If I am right that might mean that the processor would most likely not break in any way during the freezing process...
Regards,
Prescott Paulin, VP
Research & Development
300 Below, Inc.
BTW- Thanks a lot for your previous post RSB... That was really nice of you to take the time to write that.
Thelemac
11-01-01, 12:16 AM
It's not so much the water expanding (there are others that have polar bonds like water has and does that too) but the rate at which different materials shrink could have the same effect.
kevin_bouchard
11-01-01, 12:17 AM
I think they were refering to the fact that when freezing something with multiple parts, that some may shrink/expand more than others, and since cpus are fabricated extremely small, the slightest change could damage it, cracked connections,so connections may short.
Also when testing cpus, older ones probably have a greater chance of servival since the have bigger transistors.
Oh I haven't anything in this thread that suggests that cryogenic freezing promotes heat transfer,electrical conductivity or heat compacity. Does anyone have data about these effects.
Durability wouldn't be a bad thing for those small and crackable cores of Amds cpus :D
r0ckstarbob
11-01-01, 01:12 AM
Originally posted by presky
If I am correct, isn't water the only thing that expands when it is cooled? This is just off the top of my mind at 11:00, so pardon me if its a stupid question.... If I am right that might mean that the processor would most likely not break in any way during the freezing process...
Regards,
Prescott Paulin, VP
Research & Development
300 Below, Inc.
BTW- Thanks a lot for your previous post RSB... That was really nice of you to take the time to write that.
*nods head* no sweat
i believe you're correct in that water is the only thing that expands when frozen, though i'm not entirely certain. i think the point being made is that there is the possibility of damage occuring due to the fact that theres no reason to assume the rate and the measure of contraction within a single CPU is uniform. i believe it's this potential lack of uniformity within the chip (and the risk of subsequent damage) combined with the lack of factual data to support your idea of a beneficial CPU treatment, that is causing hesitation among the members of this forum.
the facts are
we don't know what exactly will happen when a chip is subjected to that kind of treatment process, which makes it a risky venture. we can speculate, but we don't know. in fact the most knowlegeable people you will be able to talk to on the subject are going to be the techs in your own labs. they're going to be the number one experts on it, guarateed.
The CPU is the most important part of our boxes so we get really sketchy when trying to do something directly to it. even AMD's new XP chip has got some folks weirded-out because you have to cut into the chip in order to unlock the clock multiplier so we can overclock the thing. folks here barely trust themselves to do it (and with good reason - some of us can barely make clean 120mm holes in the side of our cases ;) )
trusting someone else to do surgery on our CPU elevates the risk into an entirely new category of weird for the most part, especially if theres no guarantee the patient will live or not. It just makes us skittish - not entirely unwilling though.
among overclockers, although we're technically "risking" our gear every day by overclocking them - there is a certain amount of doctrine behind what we do to let us more or less accurately predict the outcome of our overclocking. its this doctrine that allows us to deal with it adequately ahead of time. there is risk, and there are ways to mitigate that risk to acceptable levels or even stop if we choose within the doctrine. i believe the root of the problem with many of us here is that if selected to be a tester, we're taking a risk, with little or no doctrine to back it up, with no control over the process itself even, and we have to commit to it. once it's done, it's done... theres no undoing it. it's the risk of permanant damage that concerns us most i think.
i believe if you had some doctrine more closely related to the subject at hand to support your idea up, the overall tone of the feedback you are recieving would certainly change. baseball bats and pistons are quite a bit different from delicate electronics with microscopic archetecture. there is concern to be had, thats for certain and i think that concern is certainly worth note.
in the end, you WILL have to say "chuck it all, here goes nothin" and just try to do it and see what happens. forgive me for not having parused your website as fully as i might have, but curiously, has your company ever treated any kind of electronics and gotten an increase in performance? has cryogenically treating electronic components ever lowered the operational temperature or increased the speed? your engineers on staff should be able to (relatively) accurately predict the contraction rates and affects, as they'll have the numbers and experience at hand. i mean you guys are cryo-nerds right? shrinkage is your business. it would HAVE to be.
now...
if the process worked? yup. i bet you'd have a DAMN solid market, at least among we overclockers and hardware enthusiasts... and theres alot of us. 20 bucks for a performance increase? oh yeah. i've blown more then that in arctic silver II or even in case fans just to grab a couple of degrees so i could run a little cooler and grab a few more clock cycles out of my CPU. so has every single kid on this fourm i might add.
are there folks willing to be the guinnea pig for this testing? sure. not everyone, but more then enough to gather your data (for better or for worse). there are folks, both here and over at the Hardforums, more then willing to do product testing and donate a couple of older chips to help you try this idea out. nobodys gonna donate their one and only pride and joy CPU for treatment without some hard data behind whats happening to give them the reassurance they'll need however.
if and when you DO decide to do some product testing, and providing the results and feedback are positive, you may consider giving the testers the option on a 2 for 1 deal. they send you an older chip and and their 20 bucks, you treat and sent the CPU back to them. they test the chip for a couple of weeks, provide you with the data you need to get the green light on the project, and then they - if they so choose - would have the option open to them to send you their pride and joy super-dooper CPU to you to receive the same cryogenic treatment, free of charge.
Example: homeboy sends you a 700mhz Duron and 20 bucks for you to treat and them to test. at the end of the test, they would then have the option to send you a second chip (say, their primary chip - 1.33Ghz CPU for example) for the same treatment, at no cost to them. i mean lets face it, even a treated 700Mhz duron isn't going to compete with an untreated 1.33Ghz Thunderbird, but the incentive to them is to have that 1.33Ghz chip going at full blast, not blowing 20 bones and 96 testing hours on a chip that they don't even use anymore, however well treated.
The other option is to start out treating heat sinks. i'm DEFINATELY behind that one. if it works and works well, you'll have already gotten your foot in the door in the computer market for cryogenically treating computer parts. it would be a small step for both your target market (the hardware enthusiasts) AND your board members to make in providing you with the support that you'd need after successfully treating heat sinks to start treating CPU's (and maybe even entire motherboards if that was an option).
i dunno. was just thinkin bout stuff. this has certainly piqued my interest. thanks for your time and i'm sorry for the long winded posts. i type 90 words per minute but my brain only goes about 40 wpm so there tends to be alot of slack. funny how that works.
RSB
Just a quick question?
Ok, does the material become more dense or just more structured when cooled?
If the proccess of super cooling restructures the material in a more uniform fashion then I would have to say that it could greatly increase the CPU's ability. It would allow for electrical and thermal conductivity through the portions of the CPU now even well fabricated Silicon can have flaws or distortions in it's crystaline structure. Could this not then partially correct these problems???
If this is also the case then it would probably help heatsinks as well. It would allow the heatsink better thermal conductivity due to the restructuring of the metal. Take a look at Pyrolytic Graphite for example. Graphite in a more natural state is rather useless for use as a thermal conductor. Yet when properly structured like it is in Pyrolytic Graphite it is one of the best thermal conductors out there.
just my 2am thought on the subject.
Orion
I think a few missed my points about cryo treating a CPU. I have some experience in semiconductor packaging. If the die is in the package and bonded, the different rates of expansion between the package and die will cause the die to break. The gate insulators are silicon oxide, different rate of expansion for silicon or the metals used in the buss lines. The silicon wafer is a mono crystal, near perfect crystal lattice, the gains would be almost zero for the silicon. maybe some improvement on the buss lines. If a CPU does survive this treatment, I dont think you will see any performance improvement. The heat sink is the first best target.
Originally posted by docJ
I think a few missed my points about cryo treating a CPU. I have some experience in semiconductor packaging. If the die is in the package and bonded, the different rates of expansion between the package and die will cause the die to break. The gate insulators are silicon oxide, different rate of expansion for silicon or the metals used in the buss lines. The silicon wafer is a mono crystal, near perfect crystal lattice, the gains would be almost zero for the silicon. maybe some improvement on the buss lines. If a CPU does survive this treatment, I dont think you will see any performance improvement. The heat sink is the first best target.
I agree. I thnk the chance of successfully cryogenically treating a CPU is very slim. As I've said though, I'd be willing to donate a 486 and a P120 to the attempt. I doubt they'll survive, but I'm not using them anyway.
But if you could improve the thermal transfer properties of the heatsink, you'd have plenty of business. Anyone know what the coefficients of expansion are for copper and aluminum? I'm wondering whether you could successfully treat an aluminum sink with a copper base like the 8045.
nihili
I'd think they would be rather close, A better idea would be to treat the material first then bond the copper and alluminum. That way you would not have to worry about their coefficients of expansion.
But, your right it could be a bit of a mess if the coefficient is to great between the two.
presky: Have you people ever done any work with multi-metal items? I think I saw some titanium covered bits on your web page. Does it have much effect on any bonds between the metals? The soldering of joints etc????
I would probably be willing to send you a heatsink for testing, maybe even a few of them if you decided to look into the benefits for that kind of thing. :)
Orion
BTW: I personally think any difference between the rates for different metals would be almost negligable when dealing with heatsinks, it's once you get down to the small level that CPU's have that I would tend to agree it could be more a problem with multiple substances in a single freeze session.
Yes, we do multi metal and multi coated items with great results. I think that if we can freeze an engine (car, truck, racing, r/c, etc...) whole, then we can probably freeze a CPU whole. -Prescott
Originally posted by presky
Yes, we do multi metal and multi coated items with great results. I think that if we can freeze an engine (car, truck, racing, r/c, etc...) whole, then we can probably freeze a CPU whole. -Prescott
Actually, your experience with freezing whole engines provides an excellent argument that multi metal heatsinks would survive. But it's not a very good basis for concluding anything about CPUs. An engine is made of large parts and engineered specifically to survive lots of physical stress. A CPU is made of microscopic parts and is designed to run in a protected environment free of mechanical stress. I'll admit that I don't know much about racing engines, but I'll wager there aren't many pieces in wich a .2 micron shift during the cooling process would render the engine inoperable.
In any case, whenever you feel like having a go at it, let me know and I'll provide you a working CPU just to see if it will survive. It will even be an old 486 which should be much more durable for this sort of thing.
nihili
vBulletin® v3.8.7, Copyright ©2000-2012, vBulletin Solutions, Inc.