Water Cooling Heresy

The alternative to direct air cooling

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Water Cooling Heresy

Post by ces » Fri Mar 12, 2010 12:11 pm

1. MOVING THE HEAT TO AMBIENT AIR

Whether you are direct air cooling or water cooling, don't you have to run the same thermal load into the ambient air whether through the air based fins or the water based fins to dispose of the heat into that air?

So then don't you need the same amount of air current to do so, at least on a sustainable basis?

And doesn't the water pump just add additional noise?

2. BENCHMARKED PERFORMANCE

Here look at the frostytech rankings
http://www.frostytech.com/articleview.c ... 501&page=5

The highest ranking water cooler is the CoolIT Domino ALC. It is out performed by a fair number of top line aircoolers, and it is loud! It is about 17db louder than the top ranked air cooler.

This is with a 150 watt thermal load.

I don't know how good or bad the CoolIT Domino ALC system is. It does perform well in the Frostytech rankings, but it is loud. And there are direct air systems that out perform it. And these direct air systems seem to be able to handle the heat load generated by any of todays CPUs, even overclocked ones.

3. COST

At NCIX:
the Noctua NH-D14 sells for $89.99 Canadian
the Sunbeam Tuniq Tower is 58.08 Canadian
the Zalman CNPS10X Extreme is $85.99 Canadian

4. AGAIN

Don't you have to move the same amount of air whether you are water cooling or using direct air?

Why would it be quieter moving this air through a water cooling radiator?

5. SAFETY

If your pump breaks what happens to your computer?

If one of the three fans on a Noctua NH-D14 breaks, what happens to your computer?

6. CAPACITY

Xbits has some data on the Noctua and the IFX 14
http://www.xbitlabs.com/articles/cooler ... d14_7.html

Their test platform is a challenging one:
Core i7-920, 4.01 GHz 1.36875

Their delta at full load using Linpack 64 is about 46C (70-24) using two 2000rpm fans. Delta at idle was about 10C (34-24). Probably you can get the same performance with less noise using 3 Apache PWM fans.

Three Apache PWM fans loping along at 500rpms under light to moderate load, probably generate almost no noise. They will kick in only under high stress conditions, and even then won't be that noisy. And light to moderate load probably includes gaming.

A Prolimatech with 2 Apaches will perform almost as well. As will a stock single fan Tuniq Tower 120 Extreme

Here is the performance data on stock single fan Tuniq Tower 120 Extreme performance at high and low speeds (using the single stock fan that comes with the heat sink):
http://www.frostytech.com/articleview.c ... 480&page=5

Note that at low speed it handles 150 watts just fine and is quieter than a standard Noctua Tower cooler.

7. WHAT AM I NOT GETTING? WHAT AM I MISSING?

I don't own a watter cooling setup. I admit I may be missing something. But what?

No one has yet articulated to me what I get in return for the greater complexity, cost, maintenance headaches and likelihood of failure?

And when a water cooling systems fails, it is likely to be a more catastrophic failure than losing one fan of two or three on a high end air cooler. Nothing good can come out of water leaking on a motherboard.

Big Pimp Daddy
Posts: 175
Joined: Mon May 30, 2005 11:26 am
Location: Sunny Swansea

Post by Big Pimp Daddy » Fri Mar 12, 2010 4:34 pm

OK, I don't use WC (but am very tempted to make the switch) so this may seem hypocritical, and is probably a less impassioned response than you would get from a true WCer.

1. Yes, the laws of thermodynamics don't change, and yes the pump makes noise, though I believe there are some good quiet pumps, and they can be undervolted too.

The benefit of watercooling is that you can move the heatsink (radiator) to wherever you like, either within or outside the confines of your case. You can also have a much larger cooling surface area than would be possible near the CPU. For example, a 480 radiator (one that accepts 4 120mm fans) positioned elsewhere in the case will have a hugely greater surface area than any conventional CPU heatsink, hence less forced airflow (less fans + lower speed = less noise) is needed to achieve the same level of cooling. If you are willing to get a big enough radiator it is possible to cool the highest power system passively, no fans means the only noise will be coming from the pump, which can be softmounted and encased in the bowels of the case to conceal its noise.

As a side benefit, it takes a lot longer for water to heat up than air, meaning that short periods of high processor activity will not raise system temperatures greatly, hence no need to increase fanspeeds for a while. In my aircooled system sporadic high loads will cause the CPU fan (which is temperature controlled) to speed up and down as the load changes. This is a lot more noticeable and intrusive than a constant speed.

2. I think comparing something like the CoolIT Domino ALC to a "true" watercooling setup is a mistake. The CoolIT Domino ALC is a ready-made Fisher-Price style cooler, personally I do not see the point in systems like this for the very reasons you have outlined. A "true" watercooling setup comprises a pump, reservoir, radiator(s) and waterblock(s). If you so choose, you can cool your CPU, GPU, VRAM, chipset, mosfets, system RAM, and even hard drives in the same loop by adding waterblocks to them. This removes the need for multiple fans in separate locations, some of which may not be conducive to large low speed fans.

The ability to move the heat to a centralised location or locations where you can use less large fans is a boon.

3. I concur. You have not factored in the greater cost of a proper watercooling setup, so the cost differential is possibly more than you imagine.

However, a good watercooling setup will be able to cool any system in the foreseeable future, unless thermal loads really get carried away, but that does not seem to be the way the trend is going. With most waterblocks, you can simply buy a new mounting kit whenever a new CPU socket is released and transfer the whole cooling system. So total cost of ownership might work out better/similar, but initial outlay with WC is definitely higher.

4. See above

5. Same as would happen if your fan (I only have one) breaks, the temperature rises and the safety features in your motherboard shut down the system automatically. You can also get a flow rate sensor which will shut down the system as soon as the flow stops, meaning less chance of thermal damage to components than with air.

6. I don't see a question here, just a comparison of 3 air coolers?

The maximum cooling capacity of WC is far greater than with air, hence why serious overclockers all use water (or phase change).

7. See above.

The "water + electricity = bad times" does worry me somewhat, but the use of distilled water which has a very low electrical conductivity can minimise this risk. Also it is essential to leak test the system before installing it into a computer, as the system is fairly hardy, no leaks should occur over time, save for large physical shocks to the computer case, which should generally be avoided in any scenario.


I hope that has cleared a few things up, I'm not trying to convert you to the ways of water, just to point out that a lot of people use it (there is even a whole section of the SPCR forums dedicated to it), and they have good reason to do so. Not all of these reasons are to do with silence, but some are.

1398342003
Posts: 324
Joined: Wed May 07, 2003 10:35 pm
Location: Surrey, B,C

Post by 1398342003 » Fri Mar 12, 2010 7:27 pm

Note
Since you made two posts with pretty much the same content I'm going to quote my message and your reply.
1398342003 wrote:According to this computer power scale, SPCR's test CPU draws about 152w (stated in the SPCR test bed article to be 160w, so I'll use that) at 100% load when overclocked and overvolted. It's also stated that the temps are from the hottest CPU core, not the sensor labeled 'CPU'.

At about 19dBA the Noctua NH-D14 has a delta of about 44c. That's about 0.275 degrees celcius per watt.

My i5 750 runs at 4300 MHz at 1.4v, giving me 179 watts. It's watercooled, and I don't know how loud it actually is, as the fans and the pump are undervolted. My loop also cools the GPU, a stock 9800 GTX+. Even with the GPU, my hottest core has a delta of 30c, making the c/w rating of my system less than 0.18 as the GPU adds an unknown amount when idle.

When I played games on my old air cooled system I'd have to turn up the fan speed to keep it stable, especially during the summer. Watercooling can keep my system quiet and stable. At the time, air couldn't, and I don't want to risk it.

Also, as far as cost; a complete watercooling setup with a radiator, pump, CPU block, GPU full coverage block, tubing, and fluid shouldn't cost more than 400 Canadian before taxes.

Pump - a DDC is about $70
CPU - Apogee XT is $80 or EK-SUPREME LT Nickel is about $55
Radiator - PA120.3 triple 120mm fan $105
GPU - Swiftech KOMODO-HD5800 $105
Tube - 10' is about $10 - $20
Fluid additive is about $10
Total - $390
ces wrote:
1398342003 wrote:
At about 19dBA the Noctua NH-D14 has a delta of about 44c. That's about 0.275 degrees celcius per watt.

My i5 750 runs at 4300 MHz at 1.4v, giving me 179 watts. It's watercooled, and I don't know how loud it actually is, as the fans and the pump are undervolted. My loop also cools the GPU, a stock 9800 GTX+. Even with the GPU, my hottest core has a delta of 30c, making the c/w rating of my system less than 0.18 as the GPU adds an unknown amount when idle.
OK. But...

1. MOVING THE HEAT TO AMBIENT AIR

Whether you are direct air cooling or water cooling, don't you have to run the same amount of air through the air based fins or the water based fins to dispose of the thermal load into the ambient air?

So then don't you need the same amount of air current to do so, at least on a sustainable basis?

And doesn't the water pump just add additional noise?

2. BENCHMARKED PERFORMANCE

Here look at the frostytech rankings
http://www.frostytech.com/articleview.c ... 501&page=5

The highest ranking water cooler is the CoolIT Domino ALC. It is out performed by a fair number of top line aircoolers, and it is loud! It is about 17db louder than the top ranked air cooler.

This is with a 150 watt thermal load.

I don't know how good or bad the CoolIT Domino ALC system is. It does perform well in the Frostytech rankings, but it is loud. And there are direct air systems that out perform it. And these direct air systems seem to be able to handle the heat load generated by any of todays CPUs, even overclocked ones.

3. COST

At NCIX :
the Noctua NH-D14 sells for $89.99 Canadian
the Sunbeam Tuniq Tower is 58.08 Canadian
the Zalman CNPS10X Extreme is $85.99 Canadian

4. AGAIN

Don't you have to move the same amount of air whether you are water cooling or using direct air?

Why would it be quieter moving this air through a water cooling radiator?

5. SAFETY

If your pump breaks what happens to your computer?

If one of the three fans on a Noctua NH-D14 breaks, what happens to your computer?

6. CAPACITY

Xbits has some data on the Noctua and the IFX 14
http://www.xbitlabs.com/articles/cooler ... d14_7.html

Looks like their test platform probably generates an even greater thermal load than yours.

Their delta at full load using Linpack 64 is about 46C (70-24) using two 2000rpm fans. Idle was about 10C (34-24). Probably you can get the same performance with less noise using 3 Apache PWM fans.

Three Apache PWM fans loping along at 500rpms under light load, probably generate almost no noise. They will kick in only under high stress conditions, and even then won't be that noisy.

A Prolimatech with 2 Apaches will perform almost as well. As will a stock single fan Tuniq Tower 120 Extreme

Here is the performance data on stock single fan Tuniq Tower 120 Extreme performance at high and low speeds:
http://www.frostytech.com/articleview.c ... 480&page=5

Note that at low speed it handles 150 watts just fine and is quieter than a standard Noctua Tower cooler.

7. PLEASE CLARIFY

"My hottest core has a delta of 30c" What kind of load is that?

What is your delta with Linpack 64? How noisy are your three fans pushing air through your radiator at that thermal load? They can't be quiet.

8. WHAT AM I NOT GETTING? WHAT AM I MISSING?

I don't own a watter cooling setup. I admit I may be missing something. But no one has yet articulated to me what I get in return for the greater complexity, cost, maintenance headaches and likelihood of failure?

And when a water cooling systems fails, it is likely to be a more catastrophic failure than losing one fan of two or three on a high end air cooler. Nothing good can come out of water leaking on a motherboard.
----

1. Yes, the heat must be transfered to the air, but a radiator has more surface area than a heatsink. Ex. I have two heatsinks, both are 120mm square, made of aluminium with an equal number of fins that are equally thick, however one is 1" thick, the other is 4" thick. The 4" one can dissipate the same amount of heat with less CFM from the fan.

I have a PA120.3, it is optimized for use with low speed fans, and it uses three of them. Since the fans are in parallel each additional fan adds to the total air flowing through the radiator. Heatsinks with multiple fans arrange them in series, so each fan adds to the pressure.

Check out the graphs on this page.

Replace:
'pumps' with 'fans'
'system curve' with 'airflow resistance'
'head' with 'overall pressure'
'flow rate' with 'CFM'

The serial graph is an example of the heatsink, when you add three fans the max pressure triples, but the max CFM doesn't go up, only the point where the curves meet is further along the CFM line.

A radiator behaves differently though. Since each fan has it's own bit of radiator to cool, both the 'fan' and 'airflow resistance' curves stretch to three times their length.

Effectively, this means that three times more air goes through the radiator, and it goes over more hot material.

2. The CoolIT Domino ALC has three problems:
a. Thin 120mm radiator
b. Unknown pump, likely weak
c. Unknown waterblock, likely inefficient.

It's hard to defend this system, as it does seem to be really bad. The radiator is a thin single 120mm model, the pump seems unimpressive, the tubing is thin AND corrugated (which hampers flow), and the waterblock is probably poorly designed. But I can't find any actual inside pics of the waterblock, or any specs on the pump, so I may be wrong.

EDIT: Most watercooling kits are like ThermalTake coolers, flashy but loud and hot.

3. I know that air cooling is less expensive than WCing, I wanted to point out that $500 Cad could WC the CPU and GPU, not just the GPU. It'll do a better job too.

4. See 1. and 7.

5. and 8. This is a pretty big risk, so I'll address it.

A. Pump stops, CPU block heats up, computer starts to thermal throttle, then auto shuts down due to over heating. Modern CPUs (5 years and newer) have max temperatures that will shutdown the computer if they are reached.

B. I once had a fluid leak because I didn't tighten a fitting properly, and I knew that I hadn't, but left it anyways. Using a Feser One additive kept my system from failing for at least a month before I noticed the leak. (it was a small drip) eventually my sound card's copper PCI contacts corroded. There was a small pool of liquid on the back of the card, and it never worked again. (though for some reason the manufacturer didn't put in 5.1 digital, only 5.1 analogue so it needed to be replaced anyways) The rest of the system was fine though.

Feser One coolant can be sprayedon a running computer without causing a short. My leak lasted a while, so there was some kind of corrosion over a long period.

There is a bigger risk than air cooling, but most pumps for WCing are pretty reliable. My pump has been running for 2-3 years without any trouble.

7. I'll answer this first, then 6. My hottest core, core0 has delta of 30c with Prime95 at 100% load across all 4 cores. The other cores are 2c - 4c cooler.

I hadn't even heard of Linpack 64, so I'll have to try it out later.

I don't know what the actual dBa of my system is. As a comparison I have a 640GB version of this drive completely enclosed in a 1/4" aluminum box, sitting on foam, which competes with my main system as the most aggravating part of my computer as far as volume.

6. Okay, first x-bit labs, their delta at full load with 2000RPM fans is 46c, which is hotter than my 30c, and it's unacceptably loud. I never put any of my fans above 1000 RPM. Ever. I can't live with them at that level.

At 800RPM the achieve a 55c delta, with (according to the Antec power thingy) a 255w processor. That gives a c/w rating of 0.216. Watercooling could have the fans spin even more slowly, as 255w in my system would give about 46c.

Also note that the CPU's picture showed a copper colour, indicating lapping was done, and the thermal goop wasn't specified. These are important differences that must be accounted for in any testbed comparison. (Speaking of which, I use arctic silver 5 on my non-lapped CPU. Don't know what Mike uses, though his CPU isn't lapped.)

Lapped CPUs run cooler than non-lapped ones, further invalidating the comparison.

Also note that FrostyTech does not use a reference fan like SPCR. Every cooler is tested as the manufacturer intended, so the results are useful as a comparison of the package. The results are not comparable as far as which is the best -heatsink-. The moment the fans change everything else changes too.
Last edited by 1398342003 on Mon Mar 15, 2010 11:54 pm, edited 1 time in total.

~El~Jefe~
Friend of SPCR
Posts: 2887
Joined: Mon Feb 28, 2005 4:21 pm
Location: New York City zzzz
Contact:

Post by ~El~Jefe~ » Fri Mar 12, 2010 9:07 pm

There is nothing that beats watercooling for cooling.

Do you drive a corvette and say... hey, lets just take off the water system and put a big fan on the motor!

No.

Air can get to a lower temperature than water as air is always new and always the same temperature as it is outside. Says 70 degrees out today, then it can cool to 70 degrees F. 30? then it can cool below freezing. Always new. Water? Heats up. Now, say you have 500 watts of power going. The fan would have to be like 10 inches wide and blaring away to get it to cool down that that much heat. Water always is a higher temperature, but never wavers even after you crank up the wattage. IT is temperate and stable, also easily silent if you use low flow submersible pumps like Eheim 300's.

it's just a different planet. there is no harm done to a CPU or GPU running at 45 degrees C. Ever. It is always happy. It can go lower, but on load this is where they hang out on a WC system on lower flow.

I have said this before but sometimes it has to be said again. Temps are meaningless, it is all about heat capacity and total heat removal ability. Water is just far far far superior to any air setup. And, it can be made at less volume than the equivalent air setup for whatever wattage you are running.

it just costs a lot initially and is a pain to do

Also, if your radiator is OUTSIDE the case, you are blowing air (or in my case I go fanless) onto the radiator and the case has no heat being stored in it. When you have a fan blowing on a GPU, that heat is roasting the rest of the computer. It goes into heating up the hardrives, the psu, and the motherboard parts. WC takes this heat completely away from the system. Your psu's and hd's will be operating at super low temperatures and the fan on the psu might not even spin up past low idle.

Also, WC systems don't fail. They just don't. I havent ever heard of one failing aside from random forum people. In reality, in all the people I know shoping at MIcrocenter, none ever failed and leaked. Why? The parts on quality products are designed for some serious shit a computer will never see. Like medical and industrial equipment type shit. Lil computer parts generating 200 watts total with a few gallons per minute flow... the parts just last forever. You should change stuff around if you have the time.

My reserator has cooled: 754, 939, and now my 757 intel setup. It will cool also my i7 930 setup. Can an air system do all that with only 30 dollars spent ever 3-4 years? My old block fit on all three. i7 will need a new block as there are no adapters. You just normaly buy a new 7 dollar adapter and use the same setup. I even use the same GPU cooler for 3 cards! 9800 pro, x1900, and 5770. I used a specific cooler for the 3870 because I was feeling fancy back then.

Wibla
Friend of SPCR
Posts: 779
Joined: Sun Jun 03, 2007 12:03 am
Location: Norway

Post by Wibla » Sat Mar 13, 2010 5:56 am

Actually, a cooler like the Thermalright Ultra120 Extreme will do the same, you need to buy new brackets for it, but thats not really a problem. It will fit the i7 no problem, just needs a $7 bracket... :D

The U120E was put on market in may 2007 and is still showing very good performance, even on the i7, so good air coolers also last a while.

1398342003
Posts: 324
Joined: Wed May 07, 2003 10:35 pm
Location: Surrey, B,C

Post by 1398342003 » Sat Mar 13, 2010 10:29 am

I have said this before but sometimes it has to be said again. Temps are meaningless, it is all about heat capacity and total heat removal ability.
Temps are not meaningless, the simply need to be taken in context. Knowing temps + watts shows heat removal. We were talking about temps in relation to watts, or the thermal ability of various setups.

Also, we were talking about delta temps, so add the ambient air temp. My hottest core at load is 50c, ambient 20c, so delta 30c. SPCR's test system gets 70c with a 20c room temp.

The TRU120E is one of the few that gets adaptors, some heatsinks get a new version that must be paid for at full price, most are obsolete when the socket dissappears.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Post by ces » Sat Mar 13, 2010 11:26 am

It's going to take me a while to absorb all this info. Maybe someone else would be willing to attempt to distill it into a simple summary.

So far it does seem like there are two major advantages to water cooling:

1. It's easier, or at least more common, to make use of cooler ambient air because typically the radiator is located outside of the case.

2. Radiators can have greater fin area, meaning you can move larger volumes of slower air to accomplish the same amount of cooling. Slower air means slower fans. Even if you have more fans, slower is quieter.

xxxxxxxxxxxxxxxxxxxxxxxxxx

It seems like a key point of reference would be the size of the fin arrays for some larger high end direct air heat sinks:

* Noctua NH-D14 ============> 12,020 sq. cm.
* Scythe Mugen 2 ============> 10,663
* Thermalright IFX-14 =========> 10,323
* Cooler Master V10 ==========> 10,659

* Prolimatech Megahalens =======> 7,008
* Thermalright Ultra-120 eXtreme ==> 6,789
* ThermoLab BARAM ==========> 7,503
* Zalman CNPS10X Flex ========> 8,200
* Zalman CNPS10X Quiet =======> 6,817
* Cooler Master V8 ===========> 8,594

* Zalman CNPS9900 LED ========> 5,402
* Noctua NH-U12P SE1366 =======> 4,971
* Thermalright AXP-140 =========> 4,787
* Thermaltake BigTyp 14Pro ======> 5,080


As compared to the size of the fin arrays of a water cooler.

xxxxxxxxxxxxxxxxxxxxxxxxxx

There is probably a level of thermal load below which water cooling is probably not the most effective use of money to improve the performance of a computer. And there is a level of thermal load above which it is the only practical route. And some grey area in between.

I'm not certain what that would be. Most probably the key variables are the CPU temp, the GPU temp and the db level. And that might vary from person to person. In fact of course it will vary from person to person.

Assuming you don't want an early failure of any component. What is a comfortable temp for the CPU and for the GPU? And is that measured at idle, at moderate use levels, at 100% utilization. At the highest utilization that you are likely to engage in for more than a hour at a time... or what?

It's not possible to compare levels of noise from one person's environment to another. So it is not practical to talk about db levels. But temperature is fairly objective.

What do you think?
Last edited by ces on Sat Mar 13, 2010 11:42 am, edited 1 time in total.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Post by ces » Sat Mar 13, 2010 11:42 am

You still have to get the thermal load to the fins. Then you need to get the heat load from the fin array to the ambient air. Some of the direct air coolers are substantially better at than than others.

The Prolimatech Megahalens are almost half the size of the Noctua NH-D14, yet perform almost as well.

I would expect that the water coolers are pretty much equally capable of delivering their thermal load to their fin arrays, at least at the same flow rate. If they move X cubic centimeters of water per hour, that will determine how much load thermal they deliver.

Two critical areas are likely the seating on the CPU chip (or as applicable the GPUs) and the design of the fin array.

The fin array or radiator design is probably a key determinant of how loud or quiet the system will be. If there are substantial differences in effectiveness between the direct air coolers there must be similar differences between water cooler radiators.

If you have a lot or surface area, loosely spaced fins (so the fins themselves don't generate noise), and quiet fans, you would expect a water cooling system to be able to do more cooling per unit of noise... because it can support a larger and more free flowing (for the air) fin array.

But even then looking at the comparison of the Prolimatech Megahalens to the Noctua NH-D14, there is lots of room to do it poorly or to do it well. And both of these companies have superb engineering prowess. (look at the size of the NH-U12P and how well it performs against much larger competitors)

If there can be that much difference between two high end engineering driven companies. there can be even greater differences between not so great competitors or between companies that have different design objectives.

Are there any clear leaders in the water cooling market in terms of quiet and efficient heat sinks? I would guess that efficiency should not be judge by size, but by the amount of cooling it can extract from a given stream of air.

1398342003
Posts: 324
Joined: Wed May 07, 2003 10:35 pm
Location: Surrey, B,C

Post by 1398342003 » Sat Mar 13, 2010 12:28 pm

The basic WC points are:
1. Radiating area can be almost unlimited
2. The Radiator itself can be placed outside of the case
3. Fans work in parallel, lowering speed and noise
4. Replaces loud GPU fans

Those are the main advantages of watercooling.

Watercooling systems work somewhat similarly to a huge heatpipe, as they provide a quick exit for heat. But the main difference is that a radiator has lots of thin, elliptical water paths with very small fins on them. Ex.

|>|<|>|<|
|>|<|>|<|
|>|<|>|<|

And so on. This means that most of the fin is quickly heated by the water channel, unlike heatsinks, where the heatpipes are fewer and round. Round items have the greatest volume for a given surface area, ellipticals have more.

Radiators have the fins soldered on to the pipes, providing a better joint. Heatsinks are usually press fit.

----

Anything below 85c with an i5 or i7, or 50c with older processors would be acceptable to prevent failure. OCed CPUs will fail if the temps get too high, which can drive that down.

----

It is difficult to compare dBa ratings, but if we both have a metric to compare it to those ratings can have some meaning. Mike does sound recordings of the items he tests. When my system is as loud as a Nexus 120mm fan at 7 volts, you know that it's X dBa.

Some CPU thermal sensors are poorly calibrated, and remounting a WB or HSF can change the results by 5c, so I may have the best my block can do, or I may be able to go lower.

----

Essentially, designing waterblocks well is a very difficult process. Fins or pins? Center inlet or offset? Here is a very good starter for info on some of the older waterblocks. The basic design has stayed the same for most companies, with minor variations.

European blocks are designed to run well with high pressure, low flow pumps, as they tend to prefer 3/8" tubing over 1/2". North American blocks tend to benefit more from low pressure, high flow pumps, although that has been changing in the last three years to medium pressure, high flow blocks.

Even on the radiator side there are many differences. Black Ice Labs produces thin radiators with high fin density, Thermochill produces thick rads with low density fins. This guy has pics of the PA120.3 and the 480GTX side by side so you can see the fins on them. More fins means more dissipating area in less space, but slower fans have a harder time pushing air through them.

PhilW
Posts: 27
Joined: Sat Mar 01, 2008 10:06 pm
Location: Singapore

Post by PhilW » Sat Mar 13, 2010 4:34 pm

I started with stock HSF, followed by third party coolers and watercooling.

Its actually quite simple.

Watercooling

Pros
- Better cooling performance (water conducts heat way better than air)
- Lower noise
- Unlimited upgradablility

Cons
- High cost
- More complicated to set up
- Higher maintenance

Phil

~El~Jefe~
Friend of SPCR
Posts: 2887
Joined: Mon Feb 28, 2005 4:21 pm
Location: New York City zzzz
Contact:

Post by ~El~Jefe~ » Sun Mar 14, 2010 6:39 am

1398342003 wrote:
I have said this before but sometimes it has to be said again. Temps are meaningless, it is all about heat capacity and total heat removal ability.
Temps are not meaningless, the simply need to be taken in context. Knowing temps + watts shows heat removal. We were talking about temps in relation to watts, or the thermal ability of various setups.

Also, we were talking about delta temps, so add the ambient air temp. My hottest core at load is 50c, ambient 20c, so delta 30c. SPCR's test system gets 70c with a 20c room temp.

The TRU120E is one of the few that gets adaptors, some heatsinks get a new version that must be paid for at full price, most are obsolete when the socket dissappears.
temps are meaningless.

You will have 45C on load on a gpu with a 150 watt system.

You will have 48C on load on a gpu with a 350 watt system.

Try saying the same thing with air! WC systems operate with a higher temperature yet never budge. You could go 600 watts of draw on a WC system and never bust like low 50's for a very quiet system. The more heat, the better the rate of dissipation. On air, the more heat, the more it builds up and starts to cook.

I run a 1 scythe fan at 7 volts, no fan in the psu and I have a slightly overclocked e8400 (overclock on the fly with it, software), and a 5770 that I sometimes overclock but normally do not. I run no fans on the Reserator. My temps on load for gpu are low 40's. The cpu is around 41.

1398342003
Posts: 324
Joined: Wed May 07, 2003 10:35 pm
Location: Surrey, B,C

Post by 1398342003 » Sun Mar 14, 2010 2:14 pm

More heat --> more dissipation applies to both air and water cooling. Every watt used must go out to the air. Watercooling simply has an intermediate heat absorber, the water, and greater dissipating area.

You state that temperatures are not useful, than what is a useful metric for watercooling performance? How is my system better or worse than yours?

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Post by ces » Mon Mar 15, 2010 9:50 am

See for reference

Corsair H50 Hydro Series Pump noisy for a Silent PC
viewtopic.php?p=503095#503095

spookmineer
Patron of SPCR
Posts: 749
Joined: Sat Nov 11, 2006 6:02 pm

Post by spookmineer » Mon Mar 15, 2010 4:13 pm

ces wrote:See for reference

Corsair H50 Hydro Series Pump noisy for a Silent PC
viewtopic.php?p=503095#503095
That is to watercooling what a boxed cooler is to aircooling.

1398342003
Posts: 324
Joined: Wed May 07, 2003 10:35 pm
Location: Surrey, B,C

Post by 1398342003 » Mon Mar 15, 2010 11:54 pm

All I see there is someone who didn't do enough research and bought something that would not do a good job. Any forum with a serious watercooling board will tell you that, with very few exceptions, you should never buy a watercooling kit. Swiftech kits are better than most, but there radiators aren't the best.

Frostytech's audio testing method is also suspect. They reviewed the Nirvana 120 but so didSPCR. They rank the dBa at 45 for high power, SPCR rates it at 31dBa. The Scythe Zipang has 33dBa (FT) vs 23dBa (SPCR). This massive disparity in sound levels puts most of what they test in doubt. I don't really trust them, since they differ from a trusted source.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Post by ces » Tue Mar 16, 2010 6:17 am

1398342003 wrote: Frostytech's audio testing method is also suspect. They reviewed the Nirvana 120 but so didSPCR. They rank the dBa at 45 for high power, SPCR rates it at 31dBa. The Scythe Zipang has 33dBa (FT) vs 23dBa (SPCR). This massive disparity in sound levels puts most of what they test in doubt. I don't really trust them, since they differ from a trusted source.
You can never compare db testing results from one tester to another. There are too many different ways to measure db.

Temperatures are almost as bad.

That is why Frostytech and SPCR standardize on a testing platform so at least you can compare results between and among the tests that they do on that platform.

Don't forget the same heatsink will perform differently with different types of chips. Especially when you are talking about AMD to Intel or 1156 to 1366 or i5-750 to i5-650.

And from what I gather from random commentary on the internet (we all know how reliable that is) it seems like it isn't difficult for different mountings on the same platform to vary 2 to 4C. And in the case of a bad mounting to vary even more.

I do agree with you that I have more trust in Mike's diligence than that of any other source of benchmarking. He seems to be more careful than most and I think he is more trustworthy. Also it doesn't hurt that he tends to measure the attributes that I am most interested in.

But even he has made mistakes. Though he is the only benchmarker I have ever seen who then admits the error and then goes back and corrects it.

Frostytech has a much larger data base of benchmarks (though even they don't benchmark all the interesting heatsinks). They are good for scanning the horizon and generating first impressions. But Mike is the person I trust before making a buying decision.

But I trust no single source. And I was disappointed when Mike ranked the Pico-PSU as recommended and the Winmate as merely acceptable. He never explained why, and by any measure those ranking should be reversed.

It seems like he downranked the winmate because it is sold by an enthusiast compared to the Pico-PSU which is sold by a corporation that has been around a while. That is the only reason I could surmise that would support such rankings. If that was the reason, he should have said so.

But that is probably my only quibble with SPRC.

1398342003
Posts: 324
Joined: Wed May 07, 2003 10:35 pm
Location: Surrey, B,C

Post by 1398342003 » Tue Mar 16, 2010 5:55 pm

The main problem with Frostytech's temps is the lack of a universal reference fan. If he runs the system with whatever was in the box than you know the out-of-the-box performance. OOTB ability doesn't really matter compared to the ability with the Nexus 120.
And from what I gather from random commentary on the internet (we all know how reliable that is) it seems like it isn't difficult for different mountings on the same platform to vary 2 to 4C. And in the case of a bad mounting to vary even more.
Very true, though I saw that it was up to 5°C.
But I trust no single source. And I was disappointed when Mike ranked the Pico-PSU as recommended and the Winmate as merely acceptable. He never explained why, and by any measure those ranking should be reversed.

It seems like he downranked the winmate because it is sold by an enthusiast compared to the Pico-PSU which is sold by a corporation that has been around a while. That is the only reason I could surmise that would support such rankings. If that was the reason, he should have said so.
You should probably ask him in the forum page for the review. It seemed to me that it simply didn't get put on the recommended page, as the review's 5th page says that it's SPCR Recommended.

ascl
Posts: 279
Joined: Tue May 05, 2009 1:15 am
Location: Sydney, Australia

Post by ascl » Wed Mar 17, 2010 3:11 pm

Heatsink (and waterblock) tests are generally pretty terrible. There are a BUNCH of factors:
-ambient temp
-airflow (ie fan used, case (or testbench) used)
-heat source (ie cpu)
-mounting
-thermal paste used
-limited temp sensors (both numbers and accuracy)
-mods performed (ie lapped CPU, washer mods etc)

Ideally, all should be kept the same, and multiple mounts performed, and the outliers discarded. The closest to this I have seen in reviews is skinee labs, although they only review waterblocks (and other WC gear). SPCR is also very good, although I cannot find info on how many mounts are performed (its possible its there and I just missed it). I have trouble believing any db tests other than SPCR.

If heatsinks are tested with whatever fan they come with, you are testing something quite different (although it is arguably a valid test) than purely heatsink efficiency.

So, bottom line, take any test results with a grain of salt. They will, at best, tell you how certain heatsinks (fans etc) compare with other heatsinks in the test environment. This may give you some broad indication of how they perform for you, but it will only be a broad indication, as there are too many variables.

Now, given that, to answer your original questions.... my experience, given my equipment:

1. All (passive) coolers move heat to ambient air. Where water can win out is you have far more flexability in the placement of the heat dissipation device (heatsink, radiator), which, in turn, allows you to have a much larger surface area to move the heat. A larger surface area requires less air movement. Yes pumps add noise, similar to adding a fan. However, since pumps don't cause the 'rushing of air' noise, I find them easier to dampen.

2. As per above. Ignore most reviews. In MY experience, with my equipment... I gained a 15c drop in idle temps and a 30c drop in load temps on my CPU, and even higher on my GPU, all for significantly less noise.

3. No need for debate here, air cooling is cheaper, and probably better value

4. No. Greater surface area requires less air movement to dissipate the same amount of heat. Also, the freedom to place the radiator wherever you like, means you can put the noise source outside, or in a cupboard or another room or whatever. Obviously no good for portable set ups, but its a possibility.

5. Pump breaks, CPU gets hot and shuts down. At least, assuming you have appropriate bios settings enabled. The GPU has no such bios setting, so that can be more of a risk, however, various software (speedfan) allows you to set actions based on sensor readings... ie shutdown if GPU hits <pick a number>. Of a bigger risk is leaks, however, a properly designed system, that is properly leak tested, has a very small chance of leaking. My system got moved internationally by post, and was still leak proof after all that shaking around (obviously I leak tested it before firing up the system!).

6. Water cooling, because of the (potential for) greater surface area, will allow much greater cooling capacity. However, if you are not overclocking, air coolers will most likely suffice.

7. Cooling performance for less noise. Thats the bottom line for me. I have a highly overclocked system which I rarely hear (depending on my ambient noise). Especially noticeable is getting rid of the graphics card noise! Its not for everyone, and its definitely not cheap, but it does work very well.

Hope this helps.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Post by ces » Wed Mar 17, 2010 3:38 pm

ascl wrote:Hope this helps.
In my personal opinion, the most thorough, balanced and complete response to date, though some of the others are close.

Migi06
Posts: 12
Joined: Mon Jul 06, 2009 3:44 am
Location: Finland

Post by Migi06 » Fri Mar 26, 2010 2:20 am

1. Frostytech is shit.
2. T.R.U.E is shit with low speed fans.
3. I have passive watercooling and cpu is [email protected]

So if you need more better tests read this:
http://www.hardwareluxx.de/community/f1 ... 53427.html
Translate:
http://translate.google.fi/translate?hl ... 53427.html
Watercool Heatkiller 3.0 CU vs. Prolimatech Megahalems Rev. B
Image

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Post by ces » Fri Mar 26, 2010 3:04 am

Migi06 wrote:1. Frostytech is shit.
2. T.R.U.E is shit with low speed fans.
3. I have passive watercooling and cpu is [email protected]
1. Frostytech has its flaws. But in the world of the blind, it has one semi-functioning eye.

2. I have been wondering about that. It would make sense if that were so. But I did see a review on the Venomous X, which is darn close in configuration where the Venomous X appeared to perform very well at low fan speeds. It made me wonder if maybe it was quiet too.

3. What do you mean " passive watercooling" can you through up some pictures of it?

b_rubenstein
Posts: 273
Joined: Tue Aug 04, 2009 7:03 am
Location: Brooklyn, NY

Post by b_rubenstein » Fri Mar 26, 2010 3:19 am

Passive water cooling probably means no pump. This is old technology know as thermosyphon:

Thermosiphon (alt. thermosyphon) refers to a method of passive heat exchange based on natural convection which circulates liquid without the necessity of a mechanical pump. This circulation can either be open-loop, as when liquid in a holding tank is passed in one direction via a heated transfer tube mounted at the bottom of the tank to a distribution point - even one mounted above the originating tank - or it can be a vertical closed-loop circuit with return to the original vessel. Its intended purpose is to simplify the pumping of liquid and/or heat transfer, by avoiding the cost and complexity of a conventional liquid pump. http://en.wikipedia.org/wiki/Thermosiphon

ascl
Posts: 279
Joined: Tue May 05, 2009 1:15 am
Location: Sydney, Australia

Post by ascl » Fri Mar 26, 2010 3:28 am

I would have guessed passive meant no fans.

Migi06
Posts: 12
Joined: Mon Jul 06, 2009 3:44 am
Location: Finland

Post by Migi06 » Fri Mar 26, 2010 4:22 am

ascl wrote:I would have guessed passive meant no fans.
Yes.. The radiator is Mora 2 (almost got three 3*120mm radiator cooling power and that cost me about 85€):
http://www.realredraider.com/vbulletin/ ... php?t=5197
I could take some pictures later when I get Fractal Design case. Right now its bolted to my desk :D
ces wrote: 2. I have been wondering about that. It would make sense if that were so. But I did see a review on the Venomous X, which is darn close in configuration where the Venomous X appeared to perform very well at low fan speeds. It made me wonder if maybe it was quiet too.
Thermalright Rev.C and Venomous X should be alright. I did use that first revision TRUE (washer mod, lapped and push&pull config) before I went to watercooling.

Thermalright Ultra-120 eXtreme Rev.C
http://www.hardwareluxx.de/community/f1 ... 64210.html

Thermalright Venomous X
http://www.hardwareluxx.de/community/f1 ... 95450.html

~El~Jefe~
Friend of SPCR
Posts: 2887
Joined: Mon Feb 28, 2005 4:21 pm
Location: New York City zzzz
Contact:

Post by ~El~Jefe~ » Fri Mar 26, 2010 1:31 pm

1398342003 wrote:More heat --> more dissipation applies to both air and water cooling. Every watt used must go out to the air. Watercooling simply has an intermediate heat absorber, the water, and greater dissipating area.

You state that temperatures are not useful, than what is a useful metric for watercooling performance? How is my system better or worse than yours?
heat capacity and rate of absorbtion. the actual chip is a tiny amount of heat. It just is concentrated. Getting a large pool of heat away from the chip and doing it quickly is why water is in cars for cooling instead of a heat sink and fan.

Monkey Puzzle
Posts: 37
Joined: Tue Apr 13, 2010 5:14 am
Location: UK

Re: Water Cooling Heresy

Post by Monkey Puzzle » Wed Apr 14, 2010 7:42 am

ces wrote:1. MOVING THE HEAT TO AMBIENT AIR

Whether you are direct air cooling or water cooling, don't you have to run the same thermal load into the ambient air whether through the air based fins or the water based fins to dispose of the heat into that air?

So then don't you need the same amount of air current to do so, at least on a sustainable basis?

And doesn't the water pump just add additional noise?

2. BENCHMARKED PERFORMANCE

Here look at the frostytech rankings
http://www.frostytech.com/articleview.c ... 501&page=5

The highest ranking water cooler is the CoolIT Domino ALC. It is out performed by a fair number of top line aircoolers, and it is loud! It is about 17db louder than the top ranked air cooler.

This is with a 150 watt thermal load.

I don't know how good or bad the CoolIT Domino ALC system is. It does perform well in the Frostytech rankings, but it is loud. And there are direct air systems that out perform it. And these direct air systems seem to be able to handle the heat load generated by any of todays CPUs, even overclocked ones.

3. COST

At NCIX:
the Noctua NH-D14 sells for $89.99 Canadian
the Sunbeam Tuniq Tower is 58.08 Canadian
the Zalman CNPS10X Extreme is $85.99 Canadian

4. AGAIN

Don't you have to move the same amount of air whether you are water cooling or using direct air?

Why would it be quieter moving this air through a water cooling radiator?

5. SAFETY

If your pump breaks what happens to your computer?

If one of the three fans on a Noctua NH-D14 breaks, what happens to your computer?

6. CAPACITY

Xbits has some data on the Noctua and the IFX 14
http://www.xbitlabs.com/articles/cooler ... d14_7.html

Their test platform is a challenging one:
Core i7-920, 4.01 GHz 1.36875

Their delta at full load using Linpack 64 is about 46C (70-24) using two 2000rpm fans. Delta at idle was about 10C (34-24). Probably you can get the same performance with less noise using 3 Apache PWM fans.

Three Apache PWM fans loping along at 500rpms under light to moderate load, probably generate almost no noise. They will kick in only under high stress conditions, and even then won't be that noisy. And light to moderate load probably includes gaming.

A Prolimatech with 2 Apaches will perform almost as well. As will a stock single fan Tuniq Tower 120 Extreme

Here is the performance data on stock single fan Tuniq Tower 120 Extreme performance at high and low speeds (using the single stock fan that comes with the heat sink):
http://www.frostytech.com/articleview.c ... 480&page=5

Note that at low speed it handles 150 watts just fine and is quieter than a standard Noctua Tower cooler.

7. WHAT AM I NOT GETTING? WHAT AM I MISSING?

I don't own a watter cooling setup. I admit I may be missing something. But what?

No one has yet articulated to me what I get in return for the greater complexity, cost, maintenance headaches and likelihood of failure?

And when a water cooling systems fails, it is likely to be a more catastrophic failure than losing one fan of two or three on a high end air cooler. Nothing good can come out of water leaking on a motherboard.
1. Moving the same amount of air may be necessary (assuming the thermal material is the same, which it isn't, with copper used in radiator fins being almost twice as good at transferring heat compared to aluminium), but one of the major beneits of watercooling is the ability to transfer the heat to water, a moveable medium, so that the area for cooling can be much larger. Moving the same amount of air through a larger radiator over a far bigger area makes much less noise than having to force air through densely packed fins (as they must be on an air-cooled cpu heatsink that must fin on the mobo).

2. The domino isn't great. If you want really decent watercooling then you should really do a bit of reading and get custom parts (good waterblock, good pump, good radiator). The beauty of watercooling is if you wanted to you can use something silly like a MORA2 and run it passive or oveclock the cpu.

3. Cost - yep, watercooling sucks in terms of expense. But then at SPCR people are quite anal about noise and pay £50 a hit for hdd enclosures. The importance of better cooling and silenece depends on the user really.

4. See 1.

5. With modern CPUs, if the pump breaks the CPU throttles and shuts down long before any significant damage is done to the processor. I had an old e4300 in a watercooling loop and the pump broke down. It throttled and shut down. I ran the watercooling quite happily pumpless with the CPU underclocked to half it's normal speed with the fans turned up to 11. :)

6. The fins on a TRUE are too close (1mm) for it to cool particularly well with a low-static pressure fan. If silence is what you are after you would do much better with something like a scythe ninja or orochi, which have far wider-spaced fins.

7. Watercooling is a lot of faff, and it's expensive. Depends whether you enjoy that faff really. It's not usually dangerous though, as once things have been leak-tested when set-up before turning the pc power on they very very rarely will spontaneously then leak.[/quote]

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Post by MikeC » Wed Apr 14, 2010 9:53 am

Big Pimp Daddy wrote:As a side benefit, it takes a lot longer for water to heat up than air, meaning that short periods of high processor activity will not raise system temperatures greatly, hence no need to increase fanspeeds for a while. In my aircooled system sporadic high loads will cause the CPU fan (which is temperature controlled) to speed up and down as the load changes. This is a lot more noticeable and intrusive than a constant speed.
This is simply a matter of increasing the hysteresis of the feedback loop... or making the fan run at a steady speed just high enough to keep the CPU in this side of throttling at full load -- but then it will be way cool enough at all other times and probably still pretty quiet if you have a big heatsink.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Re: Water Cooling Heresy

Post by ces » Wed Apr 14, 2010 10:16 am

Monkey Puzzle wrote: 1. Moving the same amount of air may be necessary (assuming the thermal material is the same, which it isn't, with copper used in radiator fins being almost twice as good at transferring heat compared to aluminium), but one of the major beneits of watercooling is the ability to transfer the heat to water, a moveable medium, so that the area for cooling can be much larger. Moving the same amount of air through a larger radiator over a far bigger area makes much less noise than having to force air through densely packed fins (as they must be on an air-cooled cpu heatsink that must fin on the mobo).
Thanks for your very complete and thoughtful response.

Copper fin arrays son't seem to help air cooled coolers all that much. I understand that copper is good at conducting heat. The other part of the equation is that air has to accept the heat.

The larger cooling area make sense. I just wonder if the copper provides that much of an advantage. It certainly doesn't appear to do much for air cooled cpu coolers.

Also, my understanding is that, contrary to most people's natural assumptions, those heat tubes are highly efficient at transmitting heat - much more so than watter - and that there is little difference in temps between one end and the other.

ascl
Posts: 279
Joined: Tue May 05, 2009 1:15 am
Location: Sydney, Australia

Re: Water Cooling Heresy

Post by ascl » Wed Apr 14, 2010 12:36 pm

Monkey Puzzle wrote: 5. With modern CPUs, if the pump breaks the CPU throttles and shuts down long before any significant damage is done to the processor. I had an old e4300 in a watercooling loop and the pump broke down. It throttled and shut down. I ran the watercooling quite happily pumpless with the CPU underclocked to half it's normal speed with the fans turned up to 11. :)
There is an awesome video of one tech site doing a test back in pentium days, where they have quake running (software rendering), and pull the heatsink off some chips. All the intel chips throttle (the game moves to a slide show), and then picks up once the heatsink is replaced. The AMD chip starts smoking! (before I get torched by AMD fans, yes I know they should react properly these days).


My 2c:
Water cooling is not worth it for the average SPCR. It will be easier, cheaper and require less knowledge to build a quiet air cooled system... providing you are willing to make some performance sacrifices (and these days its not even a big sacrifice). My reason for WCing is to have a highly overclocked CPU and a decent gaming GPU running quietly.

Regarding copper vs aluminium. It makes a difference even with air coolers (look at TRUE copper vs TRUE), however the difference is not huge, and the extra weight can be an issue. Heat pipe air coolers are very good, but no matter how good the heat pipe is (and I'd believe it that they are more efficient than water in tubes), you are still limited by the surface area of fins you can (realistically) have. At the end of the day, you need surface area, regardless of how efficient (or not) your transport is.


EDIT: I found it! Back when Tom's was doing real stories...
http://www.youtube.com/watch?v=06MYYB9bl70

Vicotnik
*Lifetime Patron*
Posts: 1831
Joined: Thu Feb 13, 2003 6:53 am
Location: Sweden

Re: Water Cooling Heresy

Post by Vicotnik » Wed Apr 14, 2010 3:06 pm

ascl wrote:There is an awesome video of one tech site doing a test back in pentium days, where they have quake running (software rendering), and pull the heatsink off some chips. All the intel chips throttle (the game moves to a slide show), and then picks up once the heatsink is replaced. The AMD chip starts smoking! (before I get torched by AMD fans, yes I know they should react properly these days).
It was Tom's Hardware and in the P4 days. Back then the P4 had built in thermal protection - if it got to hot it throttled down and avoided a hard crash. AMD CPUs also had thermal protection but relied on the motherboard to halt the system to avoid damage. Some older motherboards lacked support for that feature and then the CPU would not shut down.
The video was seen by many as Intel propaganda and further fueled the "AMD=hot" myth. I think this video (as well as other things) has contributed to Tom's low standing among enthusiasts today.

This was really a non-issue already back when the video was released.

Post Reply