EXTREME Overclocking Forums
Home | Reviews | Forums | Downloads | RealTime Pricing Engine | Folding Stats Contact Us


Go Back   EXTREME Overclocking Forums > General CPU Cooling > Phase Change, Dry Ice, LN2, Other Crazy Stuff

Welcome Guest Visitor! Please Register, It's Free and Fun To Participate!
The EXTREME Overclocking Forums are a place for people to learn how to overclock and tweak their PC's components like the CPU, memory (RAM), or video card in order to gain the maximum performance out of their system. There are lots of discussions about new processors, graphics cards, cooling products, power supplies, cases, and so much more!

You are currently viewing our boards as a "guest" which gives you limited access to view most discussions. You need to register before you can post: click the register link to proceed. Before you register, please read the forum rules. By joining our free community you will have access to post topics, communicate privately with other members (PM), respond to polls, upload your own pictures, and access many other special features. Registration is fast, simple, and absolutely free! To start viewing messages, select the forum that you want to visit from the selection below.

After you have registered and read the forum rules, you can check out the FAQ for more information on using the forum. We hope you enjoy your stay here!

Note To Spammers: We do not allow unsolicited advertising! Spam is usually reported & deleted within minutes of it being posted, so don't waste your time (or ours)!


Please Register to Post a Reply
 
Thread Tools
Old 10-20-2013, 01:43 AM   #61
technogiant
Overclocker
technogiant's Avatar
Senior Member
 
Posts: 335
Last Seen: 09-15-2014
iTrader: 0 / 0%
Thanks Orthello, yeah I'm really pleased how it has worked out....not bragging but considering I've worked this all out myself as I believe it is probably a world first then for it to work first time just as I planned was tbh dumbfounding.

I don't know atm where I'm going with this....even though it has cost a lot of money and effort I'm thinking its very much a proof of concept build.

I just don't think I will ever feel safe using it with the butane in there....it's strange...I was having a discussion with someone on another forum and they hit the nail on the head...."you know its safe, you can explain why it's safe but another part of your mind is constantly repeating nope this is dangerous nope this is dangerous".....its just the consequences of an unforeseen problem are too large.

Now I know it works I may change the refrigerant out for a safe alternative, but that is very expensive and considering the size and cumbersome nature of this build I may just sink those funds into an evolution of this design....OH GOD SOMEBODY STOP ME.

I'm thinking I may use a low pressure refrigerant, one with a boiling point around ambient, that would mean the liquid could be left in the chamber so no need for a freezer to recover it.

I'm also thinking of doing away with the ac unit's evaporator as it makes the chamber size so large, I'd just have the suction and capillary lines directly entering the chamber.

This would allow for a much smaller chamber....I'd have to make my own custom compressor circuit and include an oil separator in there for lubricant return, additionally I've also thought of a means to collect any lubricant that gets passed the separator into the chamber.

I could even put the compressor assembly on a flexible line so it could be house outside like a split ac system.

But I need to get some skills to do this and some advise from a HVAC specialist...?

Additional thought.....could even use my current build with a low pressure refrigerant and so dispose of the butane and the freezer?....that could be a nice stepping stone and test along to the route to the final goal.

Last edited by technogiant; 10-20-2013 at 08:33 AM.
United Kingdom  Offline
    Register to Reply to This Post
Old 10-21-2013, 01:09 AM   #62
Orthello77
Overclocker
Senior Member
 
Posts: 410
Last Seen: 12-19-2015
iTrader: 0 / 0%
I do believe its a world first , one i don't think many people without your chemical knowledge would dare make ! major kudos to you

low pressure refrigerant would definitely be simpler, i'm a bit unhappy about the size of my cooling solution, wish it was smaller so i know where you are coming from. Less things to potentially go wrong too. What happens in a power cut with that freezer currently ?

On another note , all parts arrived for my new pump build today , i pulled the trigger on a 4770k and mobo also ... sick of this asus board (HT bug still present with 3770k) .. bought an asrock one this time. So should have some 4770k delidding happening soon.

Any more clocks on those GPUs from the lower temps ?
New Zealand  Offline
    Register to Reply to This Post
Old 10-21-2013, 02:12 AM   #63
technogiant
Overclocker
technogiant's Avatar
Senior Member
 
Posts: 335
Last Seen: 09-15-2014
iTrader: 0 / 0%
Thanks Orthello, yes I'm seriously considering using a low pressure refrigerant...it would be much safer, reduce the set up size and greatly simplify and quicken the usage procedure.

Trouble is as always finding something suitable and available, refrigerants are so darn expensive and mostly not available for public purchase unless you have some kind of HVAC certification.....these are the major reasons I went with butane...It's cheap and readily available to joe public and I knew at least in theory it is possible to negate the flammability issues....I'm thinking of R-123....but will just have to see if I can get it.

Good news on your stuff arriving.....looking forward to seeing some more activity on your thread....As your doing a mobo replacement have you given the TIM anymore thought?.....I understand GC extreme is often used with LN2 benching and is one of the best for low temp usage.....also consider Indigo Extreme, that would remain solid in your temp range.

The only concern I have about Indigo ex is one of thermal shock.....I mean you know how our systems react when cpu stress testing.....the cooling is so powerful you go from +30c to -30c in seconds when the test stops...not sure if differences in thermal expansion rates between the Indigo ex and IHS would cause it to detach.

Haven't done anymore testing on my system yet...doing a lot of thinking and trying to rationalize why I'm getting such a large temp delta on the cpu.

I mean hitting +50c when my liquid temp is -30c seems insane to me....especially when you consider my setup....seriously....if I was just using cold liquid without the phase change aspect I would have thought that would provide sufficient cooling...I mean it's not like I'm dealing with the relatively small area provided by a waterblock...the entire 9cmx9cm heatsink is in fluid contact on both it's upper and lower surfaces....the liquid will penetrate into the socket around the pins and is also cooling the back side of the mobo.....now add to that the fact that it is phase changing removing 220watts of heat for every 1ml per sec of evaporation......I just don't understand it.....I thought I was going to get a maximum delta of about 10c....not an 80c delta

I'm considering it is perhaps my home made tim, but I know it works I've tested it, I may have put a little bit too much on there, but I don't think I did such a bad job to cause such a delta as I'm seeing here.

The only other thing I can think is that the HS needs to have that special nucleation coating......but then again LN2 pots which use the same passive boil off cooling don't use that to my knowledge....really got me baffled.

I'm starting to think its actually to do with internal cpu limitations....I mean once your cooling gets to a certain power then if the bottle neck become internal within the cpu you can do no more.

Now I'm sure someone could come and say but you get better temps with a triple cascade so that argument doesn't hold......but tbh I'n not sure what actual temps these uber cooling systems produce at the silicon level.

Most just quote their cooling head temps not the load cpu silicon temp..and most results I've seen only use less intensive stress tests like Pi ....not IBT on max as we have been doing.

I'd be really interested to see what actual silicon temps these uber systems can hold at 5.5GHz running IBT on max setting? anyone?
United Kingdom  Offline
    Register to Reply to This Post
Old 10-21-2013, 02:58 AM   #64
Orthello77
Overclocker
Senior Member
 
Posts: 410
Last Seen: 12-19-2015
iTrader: 0 / 0%
Suitable and Available are always the limitations here in lil ol NZ lol .... i know what you mean.

The temp delta seems high, although i have a spreadsheet of IBT testing somewhere with my 2700k at around 5.3 - 5.4 ghz on Maximum , so it could give you an idea how it compares to waterblock cooling as i have recorded liquid temps also. I'll see if i can find it and PM it to you , might giving you an idea how it compares. From memory i think you are getting 10-15c higher core temps than i was at the time if you are using IBT max, albiet you are 100 mhz - 200mhz faster, and i'm usure of your voltages vs mine.

My gut feeling is the tim or the block is not seated right, or contact is not a good as it should be. Evapouration should be more efficient than cooling via waterblock. Would be interesting to see if you can evacuate it - if the tim has held up. Did you try good old MX4 ? I wish i could get the size of HS you have there watercooled , infact i'm working on it , hopefully have something to show soon.

I think the arguement for a Sandy hitting the brick wall is pretty common, the imc's just don't budge past a point regardless of extra cooling, or there could be significant temperatures walls. Hwbot avg on 2700k liquid nitrogen 5592 mhz. Avg on cascade 5602 mhz. So you are actually bugger all in % terms (less than 2%) off the avg for a cascade and or liguid nitrogen. It could just be the silicon lottery and your at the brick wall. Interestingly the avg for liquid nitrogen is less than cascade so goes to show the sandys certainly reach a limit quite early on in terms of extreme cooling.

Ok just found that spreadsheet , highest IBT maximum testing with HT off (as it produced higher temps and Gflops ~ 145) was at 5.3 ghz. 1.512vcore, core temps max 34-48c with -23.5c liquid.

Last edited by Orthello77; 10-21-2013 at 03:15 AM.
New Zealand  Offline
    Register to Reply to This Post
Old 10-21-2013, 10:22 AM   #65
technogiant
Overclocker
technogiant's Avatar
Senior Member
 
Posts: 335
Last Seen: 09-15-2014
iTrader: 0 / 0%
Quote:
Originally Posted by Orthello77 View Post
Ok just found that spreadsheet , highest IBT maximum testing with HT off (as it produced higher temps and Gflops ~ 145) was at 5.3 ghz. 1.512vcore, core temps max 34-48c with -23.5c liquid.
Thanks Orthello....that is eerily close....at 5.4Ghz 1.51vcore (nominal vcore...although my system does raise vcore considerably under load) liquid temp I guestimate at -30c I got min/max 43c/55c

That's too close to be coincidental.....I'm thinking there are internal cpu temp limitations...what I mean by that is beyond a given cpu heat production actual heat transfer within the cpu is becoming the limiting factor so no matter how powerful your cooling system you are going to be stuck with a certain cpu silicon temp regardless of whether there is a 50,80 or 150 deg C delta between the silicon and your cooling system...the cpu internally just can't transfer anymore heat.

That has helped to clarify it.....so until someone with a cascade system pops up and says I can run IBT max on 2700k at 5.4ghz on 1.51vcore at a silicon temp of something like 5 degree C then I'm fairly set in my mind that these relatively high temp deltas are a failing of internal cpu heat transfer rather than our cooling systems.....and I think chasing improvement may be as fruitful as going on a unicorn hunt....lol

Ps...regarding the TIM....I cant use standard tim based on silicone grease...they dissolve in the fluid.....I used a home brew of syrup and 9 micron alumina powder....weird I know but it tested out well and I had to use something that chemically would not dissolve in the fluid and mechanically would not wash away...the syrup is superviscous at low temps so should stay put...the only concern I had with it is I may have put too big a pea on there.....but will have to wait till I crack open the vault to check on that and I'll probably give indigo extreme a go.

Last edited by technogiant; 10-21-2013 at 10:31 AM.
United Kingdom  Offline
    Register to Reply to This Post
Old 10-22-2013, 03:28 AM   #66
Orthello77
Overclocker
Senior Member
 
Posts: 410
Last Seen: 12-19-2015
iTrader: 0 / 0%
That's a really interesting idea you bring up about internal cpu temp limitations. I guess with the new waterblock i'm working on we'll test that theory as it will have orders of magnitude more metal / liquid ratio - even latent heat storage in the block will be a lot higher.

Its almost incredible how fast say IBT maximum changes cpu temps , 20c+ in split seconds, the water blocks surface wouldn't have change temp by more than a degree in the same time i'm sure. So your idea of only so much heat can dissipate / sec out of the package might be right on the button. I realise heat transfer is latent and dependent on the heat deltas - the higher the delta the more heat transferred, possibly this explains it , but its an interesting theory.

One thing i've been thinking about is why are gpu temps miles lower than cpu temps , eg my gpus are always below 0c if liquid temps are lower than -20c. Cpu can be 15-20c north of that in a gaming situation with the same liquid flowing through, even less wattage. Given GPUs have much larger metal waterblocks / heatsinks .. but theres most likely more to it than that. They have very even distribution of power as they are highly parrallel, cpus on the other hand may have half their re estate taken up with cache etc .. leaving even less area to effectively spread heat from. GPUs have much larger asics so thats definitely helping too , i guess the watts per mm2 on the GPUs are way down compared to cpus. I think battling that delta down with efficiency may not be realistic beyond a certain point. Thats probably why liquid nitrogen does so well , you could have an 80c delta and still be miles below 0c.
New Zealand  Offline
    Register to Reply to This Post
Old 10-22-2013, 08:23 AM   #67
technogiant
Overclocker
technogiant's Avatar
Senior Member
 
Posts: 335
Last Seen: 09-15-2014
iTrader: 0 / 0%
Yes I had also though that of the gpus, the do have a much larger surface area especially when you consider that much of the cpu surface is taken up by integrated graphics as well.

The internal limitation theory could well be the explanation.

If that was the case I don't even know or think that LN2 cooling would make any difference......all they ever go for is max oc....they are not interested in the silicon temp at the max stable overclock on full load with IBT max settings.....may be their silicon temps would be the same?

But it is hard to imagine where the limiting bottle neck would be in the cpu....silicon itself has a thermal conductivity of 149w/m/k and certainly with soldered IHS designs like the 2700k...so perhaps this is rubbish and I need to drastically tweak my build....lol

But as I've said before...really not much more I can do...hells teeth I've drowned it in liquid gas already.

Additional Comment:

I've been doing some calculations on thermal conductivities and temperature deltas and think I've come up with the answer.....

I think it is due to limitations in the cpu.......not the silicon itself which has a high thermal conductivity of 149 w/m/c but the solder layer......the lower temp melting point solders which they must use, from the info I've found have conductivities of between 20 - 50 w/m/c

Sandybridge total die area is 216mm^2......only about half of that is dedicated to cpu so 108mm^2

Now I'm not sure of the power usage of a 2700k at 5.5ghz and 1.51 vcore core temp says about 140watts but most people say 200-250watts for a clocked cpu so lets be conservative and say 200watts....and lets go down the middle with the solder and say it has a conductivity of 35w/m/c and a thickness of 1mm.

Q(heat flux watts) = K (thermal conductivity w/m/c) x A(contact area M^2) x (T2-T1 temp delta) / T(thickness M)

200=35 x 0.000108x(T2-T1) / 0.001

So resolving this for delta T you get a temp delta of 52.9 deg C just across the solder interface.

Like I thought.....it's the high heat density interfacing with the solder causing such a large delta...not much can be done about that...just try and reduce the deltas at the other interface but you are saddled with a +50c delta you can do nothing about.

Last edited by technogiant; 10-22-2013 at 08:23 AM. Reason: Automerged Doublepost
United Kingdom  Offline
    Register to Reply to This Post
Old 10-23-2013, 04:00 AM   #68
Orthello77
Overclocker
Senior Member
 
Posts: 410
Last Seen: 12-19-2015
iTrader: 0 / 0%
Really interesting stuff TG. I gotta love your formulas !!

I think a mm of thickness for tim is probably a bit much though , i mean the tim is been squeezed out until the ihs or die is sitting on the water block or heatsink. I would think its more like a fraction of a mm.

Still it goes to show there is latency and delta between each interface. So trying to minimize the delta can only be done by changing tim and waterblock / heatsink or the medium taking heat off the heatsink. as long as it can remove all the heat given with a the lowest delta possible. Thats where a larger water block could still help me, i could gain 5c + there as the waterblock has to heat up past the fluid temp for heat transfer to occur. A large block will transfer more heat per degree c of delta.

Check this out : Similar to Indigo Extreme , but better i think , lower melting point 59c ?, possibly easier to apply and burn in ?
http://www.coollaboratory.com/pdf/sa...d_englisch.pdf
Just read that it requires burn in at around 80c . Possibly too risky for my liking after my bad experiences with indigo extreme.

This review is really interesting , looks like GC extreme is the best of the non liquid metal pastes, most likely what i will go for for the haswell.
http://www.tomshardware.com/reviews/...rk,3616-9.html

Last edited by Orthello77; 10-23-2013 at 04:26 AM.
New Zealand  Offline
    Register to Reply to This Post
Old 10-23-2013, 06:28 AM   #69
technogiant
Overclocker
technogiant's Avatar
Senior Member
 
Posts: 335
Last Seen: 09-15-2014
iTrader: 0 / 0%
Hi Orthello, not sure if we got our wires crossed there....I wasn't refering to the TIM but the solder layer between the IHS and die on the 2700k.

I'm not sure how thick that would be....think 1mm may be a ball park figure....there are a few guesstimates I've plugged into the equation for lack of actual figures but I think it gives an idea of the approximate size of the fixed temp delta that you can do nothing about.

Its because of the small surface of the cpu die....only about 1cm^2 of the 2700k is actual cpu....the compounding factor is that the solder it interfaces with has a lowish thermal conductivity, not sure of the spec but low melting point solders are 20 - 50 w/m/c ....they may use a fluxless solder that may be as high as 80 w/m/c so that delta is quite a rough guesstimate but I think that quite a considerable heat delta is present because of this and there's nothing can be done unless I attempted a delid on a soldered cpu which I'm not going to try.

Hey thanks for the link to that coolab product....I like the sound of that, it's a better idea as the pad means there alloy is already there in place rather than having to rely on it flowing as you do with the indigo extreme....also I can use for the graphics card...excellent.

Couldn't see any thermal conductivity figures for it though?

Last edited by technogiant; 10-23-2013 at 07:25 AM.
United Kingdom  Offline
    Register to Reply to This Post
Old 10-29-2013, 04:19 AM   #70
Orthello77
Overclocker
Senior Member
 
Posts: 410
Last Seen: 12-19-2015
iTrader: 0 / 0%
Ah my bad re the 1mm tim .. i need to not read your posts so late at night lol ..

I havn't come across thermal figures for it yet, but from the testing at toms on GPUs i think its a no brainer for GPUs atleast. I'm going to give gelid ultra a shot on the 4770k and see how that goes. I won't try the liquid pro / ultra .. i'd rather have consistent performance and not know what i'm missing in a sense.

Hows the project coming along ?
New Zealand  Offline
    Register to Reply to This Post
Old 10-29-2013, 04:42 AM   #71
technogiant
Overclocker
technogiant's Avatar
Senior Member
 
Posts: 335
Last Seen: 09-15-2014
iTrader: 0 / 0%
No probs...there is a lot of leeway in that calculation anyway.....but I think when increasing the power consumption greatly as we are when overclocking to this extent then we are hitting internal heat transfer problems on the cpu, which explains why the cpu is not hitting the same low temps as the gpu which have a lower heat density.

Sadly I think as its a combination of the extreme heat density and the interfacing material then its a problem that will be present in Haswell just as in IvyB.

Your going to really need that interfacing material to be as thermally conductive as possible.

I've knocked my project on the head now....the reallity of using 18 liters of liquid butane and 135 liters of butane vapor was really just too scary....had an incident where I thought a leak had developed sucking air in which struck the fear of god into me....as it happens there was no leak of consequence but it's not worth the risk continuing.

I was considering changing to a different refrigerant...hfc 227ea...but the expense was extortionate $3500 for 30kg shipped from China....just don't have that sort of money to spend...also considering that there may be unforseen chemical incompatibilities....like it dissolving the sealant and perhaps ruining the refrigerant then that is not a financial risk I can take.

So I'm leaving it as a successful proof of concept built...it was too large and ungainly to use as a 24/7 system so just doesn't warrant much further investment.

In the long run I may pursue this idea further using a low pressure refrigerant with the pc essentially built into the evaporator of the cooling system....but no immediate plans to do that.
United Kingdom  Offline
    Register to Reply to This Post
Sponsored Links:
Please Register to Post a Reply

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT -6. The time now is 02:37 PM.

Copyright ©2000 - 2016, Jelsoft Enterprises Ltd.
Powered by vBulletin
Copyright ©2000 - 2016, EXTREME Overclocking