Biggest Copper Heatsink?

[XyBeRWaReZ]

Does anyone out there know who makes the biggest copper heatsink and the model number?

I'm looking for a basic rectangle shape all copper heatsink without any type of copper tubing in the design. A lot of the new stuff coming out has copper tubes in the heatsinks in one way or another.

So basically the requirements are a flat copper bottom with copper fins on top. Hopefully the fins will be pretty high and not low profile.

[bobo5195]

they make them with tubes in because it is better like that.

[XyBeRWaReZ]

Originally Posted by bobo5195

they make them with tubes in because it is better like that.

Well I plan to use it in a water chiller setup for my water cooling setup. So if it's more efficient, then which one would be the best?

[bobo5195]

see other thread.

Waterchillers are not that efficant and you have to be very very careful that you don't get condensation.

Gernally better (by a long long way) to just mount the pelt on the cpu and have done with it.

[XyBeRWaReZ]

Originally Posted by bobo5195

see other thread.

Waterchillers are not that efficant and you have to be very very careful that you don't get condensation.

Gernally better (by a long long way) to just mount the pelt on the cpu and have done with it.

I've read a lot of forums and have actually seen pictures of people and thier water chillers. There's even a company that has actually gotten the idea to work and is selling the product right now. Please check out

www.coolitsystems.com

and look at the new Freezone CPU coolers which implement 6 peltiers to chill the liquid. Here is a site that reviewed that product:

http://www.extrememhz.com/freezone-p1.shtml

I don't mind if people have a lot of theories and doubts, but if we don't try, then how are we supposed to break out of the barrier? So my plan is to stick on a pelt on the CPU, and to test out my own design for a water chiller, since it has been proven to work, but hopefully a design I feel is slightly more efficient, I might get better results.
Also condensation won't worry me as I've read up and found instructions on how to insulate the board and the CPU properly as to not have any problems.
Being that I do live in Hawaii, I don't think that it would hurt me any more as temperatures around here average over 30C! At night it doesn't get too much cooler either, so I couldn't say overclocking only at night works much better. Maybe if I lived somewhere like Alaska, then it might not make a huge difference. Since I don't, I have to do something, otherwise my overclocking dream, will only remain a dream.
Also isn't joining a forum, where collective minds all come together with their experiences and knowledge and sharing them what it's for? I'm joined up in quite a few different forums, not only just for overclocking, but other things I have interests in, and I'm always surprised when I see people telling each other to go look for it. Why should I join a forum if I should go look for it myself? Doesn't that defeat the purpose of why I, you, and everyone else who joins the forums?

[Brians256]

You misunderstand us. We aren't saying it won't work. We are just saying that it is less efficient than putting the pelt directly onto the CPU.

In order to "break out of the barrier", you need to make a setup where the secondary losses from the tubing are much lower than the losses in a direct contact scenario. Very difficult. Especially when you have to contend with pump heat.

[XyBeRWaReZ]

Originally Posted by Brians256

You misunderstand us. We aren't saying it won't work. We are just saying that it is less efficient than putting the pelt directly onto the CPU.

In order to "break out of the barrier", you need to make a setup where the secondary losses from the tubing are much lower than the losses in a direct contact scenario. Very difficult. Especially when you have to contend with pump heat.

Sorry, I only posted that part about not everyone not saying it won't work is because I have read a lot of forums already and actually have been following this subject for over a year now, and I see that a lot.

As for pump heat, does it really genereate that much more? I'll be using an Aqua Via 1300. I thought that newer pumps nowadays with an inline setup, and not submerged, generates a substantially less amount of heat. Please correct me if I'm wrong about that.

[bobkoure]

OK, I looked at that freezone kit.
I see no indication that it's made to be run below ambient - AFAICT the TECs and "heat exchanger" are just a way of getting the performance of a really big radiator into a small package - at the expense of, well, being more expensive, plus using a bunch more power (which does matter if you're running 24x7).

[XyBeRWaReZ]

Originally Posted by bobkoure

OK, I looked at that freezone kit.
I see no indication that it's made to be run below ambient - AFAICT the TECs and "heat exchanger" are just a way of getting the performance of a really big radiator into a small package - at the expense of, well, being more expensive, plus using a bunch more power (which does matter if you're running 24x7).

Sorry, I just read the reviews on the product and they say that the temperature never hits over 30C even under full overclocked load.

[bobkoure]

Did they spec room temp, fan db and how they are getting their temps?

I'm sitting at a desktop box right now which is moderately overclocked, w/c ed set up for low noise (low powered pump, swiftech 5002, montego heater core, very undervolted nexxos fan).
My diode reported temps stay within 8C of room temps with normal usage (including photoshop stuff which can push the CPU pretty hard but only for minutes at a time) and within 12C when multi-hour testing with CPU Burn. Not that diode temps mean anything, BTW, except that, on the same motherboard and CPU, higher numbers probably mean hotter
At one point, I tried turning the fan up - my temps drop a few C, but I don't get any additional overclock, so, back to quiet.

So...using TECs, and the extra complication and extra power consumption to stay "under 30" just isn't that interesting - at least not to me. Feel free to buy one, if you like.
The major advantage seems to be that because you can get the fins of your "heat exchanger" (sure looks like a rad to me and it's not phase change) hot in comparison to the air being passed through it, which might allow good heat transfer even with low to moderate airflow, so quiet fans maybe and TECs are quiet.
Of course, if the TECs are running off your PC PS, the fan for that is going to stay pretty ramped up.
And there's the whole issue of consuming more power than you need to get the job done - but who am I to talk about that: I'm still using big CRT monitors, and they're energy hogs.

PS: Do they mention if the CPU temps can ever go below ambient? Without proper setup, this can be bad.

[maxSaleen]

Let me get some things sorted out here:

XyBerwhatever....

Do you want an air or watercooled heatsink?
Do you plan on building a chiller with pelts or with a compressor?

Don't knock liquid chillers guys... they can get some damn low temps. For examples look over on extremesystems.org.

Dealing with the condesnation, I'll conceed, is a major PITA. I gave up on a project becuase the amount of inulation I had to use wouldn't fit inside of a respectable case.

Remember, when using a compressor (think AC system or fridge) the amount of heat energy transported is NOT a direct result of the power draw of the compressor. The type of gas used plays a huge role in trasnsporting heat. This allows compressors to be substantially more efficient than their TEC brethern.

[bobkoure]

I'm not knocking liquid chillers - the question is whether the expense, complexity and energy usage are worth is for 8C over ambient (assuming, for the sake of discussion, that that's accurate).

Originally Posted by XyBeRWaReZ

I'm looking for a basic rectangle shape all copper heatsink without any type of copper tubing in the design. A lot of the new stuff coming out has copper tubes in the heatsinks in one way or another.

Why no tubing? Do you know why the tubing is there? Or what the tubing is? Is there some way this is going to be a problem clamped to the back side of a TEC? (which is what you seem to be asking about). Were you planning on building your own freezone-type rig, but with air cooling "big copper heatsink" on the TECs? IMHO the major challenge you are going to run into is controlling the TECs, although if you had a mess of 'em you could just switch individual ones on and off - applies to a dual loop chiller as well.

[XyBeRWaReZ]

Okay, let me get some things straight.

1. I'm not knocking on water cooled guys, because I have water cooling equipment myself.
2. The kit I listed above, well check out the reviews. They said the system runs at less thn 15db, which is incredibly respectible for a unit that can cool without as much setup as a watercooled unit. I know, remember I have it too.
3. I'm interested in this project because I want to see if it really can work, through phsyical trial and error and not just by thinking about the theories.
4. If it does work out and I can build a pretty good unit, even though I will understand that this will be in no way efficient as pelts are damn inefficient, I can learn a whole boatload and share with the community and either prove or disprove everyone's theories.
5. It doesn't hurt anyone except for my pocketbook when I shell out the money to learn from my own curiousity.

In the end, no one but me would lose anything. But in actuallity I wouldn't be losing, I'd be gaining a learning experience. It's not like they have schools that you can go to learn just about this.
So with that said, I will continue with my endeavour whether it be stupid or not, but I'd say it's more stupid to not know at all.

[Ls7corvete]

http://www.swiftnets.com/products/mcx462plusT.asp

HHHhhhmmmm, trial and error huh?

I know this can work, I also know how and by how much. The question is do YOU. Research your setup and get a good calculated estimate of your final result. I did this with my last experimental setup and I was happy to be within 1-2c when the final product came around.

[XyBeRWaReZ]

Well that was what I'm trying to figure out. With some help of others who have experience with pelts, maybe they can share some info on what they achieved as I can apply it to my setup.
I had that heatsink before, but without the TEC unit. It was a b@stard to get on! Good though. I was thinking about picking one up again and trying it with that.

[Ls7corvete]

Originally Posted by XyBeRWaReZ

Well that was what I'm trying to figure out. With some help of others who have experience with pelts, maybe they can share some info on what they achieved as I can apply it to my setup.
I had that heatsink before, but without the TEC unit. It was a b@stard to get on! Good though. I was thinking about picking one up again and trying it with that.

Just run the numbers, heat load->cooling loop efficiency->TEC efficiency->HSF efficiency

Find the result and what is optimal for your application, if it all checks out then build it.

[XyBeRWaReZ]

Cool thanks for the calculation method.

I've got my CPU wattage ouptout. It's rated at 67W full load not overclocked. But I have it overclocked and I'm currently on air to 1950mhz. So I'm not sure how much more heat and wattage it's making. Also I haven't unlocked my chip yet, which happens to be an AMD XP 1600+ AGOIA stepping. I will be pumping more voltage in this unit but not just yet as I'll be upgrading the mobo, ram, and video card.
The upgrades will be an Abit NSF-7 v2.0 mobo, 2GB OCZ PC3200 ram, and an eVGA 6800GS video card.
I still have a bit more numbers to run though.

[bobkoure]

One way to figure your wattage if you've an idea of your PS efficiency, is to us a kill-a-watt power meter on your AC (they're about $35).
1) measure with processor un-overclocked and un-overvolted (and either idling or full out - seems to me that AMD wattage specs were for full out, but I could easily be wrong).
2) overclock the processor
4) measure in the same state as above (full out or idle)
5) take the difference (you did use more when overclocked, right?), apply the PS efficiency, then add that to your base 65W to get a notion of how much power you're using

[XyBeRWaReZ]

Originally Posted by bobkoure

One way to figure your wattage if you've an idea of your PS efficiency, is to us a kill-a-watt power meter on your AC (they're about $35).
1) measure with processor un-overclocked and un-overvolted (and either idling or full out - seems to me that AMD wattage specs were for full out, but I could easily be wrong).
2) overclock the processor
4) measure in the same state as above (full out or idle)
5) take the difference (you did use more when overclocked, right?), apply the PS efficiency, then add that to your base 65W to get a notion of how much power you're using

I didn't up the voltage on my CPU yet, just the FSB and I'm able to pull 1950. But I do know for a fact that my CPU temp does go up from stock speed though.

[Brians256]

Has anyone ever been able to put a shunt resistor "after" the mobo power MOSFETs (and the LC filter that smooths it out) that feed the CPU? That'd be a better way to estimate CPU wattage. Sounds like a heck of a dangerous mod though, and it might be difficult due to routing. Might have to desolder throughhole devices just to get to the high current output of the DC-DC converter and it still might be too noisy to get accurate current measurements without expensive equipment.

OK, done with random thought.

[bigben2k]

Yeah, someone took a shot at it (can't remember if they succeeded). It was either Incoherent or Groth.

[bobkoure]

Originally Posted by XyBeRWaReZ

...
2. The kit I listed above, well check out the reviews. They said the system runs at less thn 15db, which is incredibly respectible...

It's not so much "incredibly respectable" as not believable. Even an undervolted nexxos fan pushing air through a radiator is going to be more than 15 dB - and that's about as quiet as fans get. Then there's a pump - pumps can be quiet, but not 0dB - and together this and the fan are going to be 15dB?
Something like this makes me doubt the rest of the measurements.

[XyBeRWaReZ]

Originally Posted by bobkoure

It's not so much "incredibly respectable" as not believable. Even an undervolted nexxos fan pushing air through a radiator is going to be more than 15 dB - and that's about as quiet as fans get. Then there's a pump - pumps can be quiet, but not 0dB - and together this and the fan are going to be 15dB?
Something like this makes me doubt the rest of the measurements.

Sorry, I'm just quoting what I read off of a review, so it's not from personal experience. I don't know if I'd ever pay for that thing, maybe if the price was good enough, but I highly doubt I'll get the chance to test it out for myself.

[Shoebox9]

bolkour is correct, there is no way a fan (92mm?) blowing air through a small rad (was it even shrouded?) like that with enough air speed to think about cooling 80w of CPU heat + 200?w from the TECs, is going to be even wildly close to 15db. The rad alone will produce heaps of air turbulance & cause all sorts of loud air whoosh. Then there's the pump. I wish I could buy a pump that only gave off 15db without submerging it first.

15db is VERY soft, most quiet rooms have more background noise than this.

I am very impressed by the TEC control module however. I'd like to see high power software controlled versions of these available (like mCubed's T-balancer, only with the ability to handle the power load of serious TECs).

[modster]

so what kind of advantage do you see in waterchillers? if you dont have temperature control like that unit, you are going to have to worry about condensation (assuming you are using some pretty powerful TEC). That just add another level of complication to the direct contact method. Second, why would you want to cool the water and the cpu when you can just cool the cpu? anyway, i am not trying to stop you from doing your project so good luck