Quiet Liquid Cooled Gaming PC Build Guide

Want to talk about one of the articles in SPCR? Here's the forum for you.
Olle P
Posts: 711
Joined: Tue Nov 04, 2008 6:03 am
Location: Sweden

Re: Quiet Liquid Cooled Gaming PC Build Guide

Post by Olle P » Wed Jan 21, 2015 6:50 am

Interesting article! Incredible that it could be that quiet with a triple rad.

I would have opted for a considerably bigger radiator outside the case, the wall mounted type normally used for heating rooms.
No fans needed for it, so it's quieter than the triple rad. With enough water (10+ litres in the system) it's futile to wait for the temperature to peak since it will take many hours.(1)

There are a few advantages with water cooling. Each being more or less relevant depending on circumstances:
  1. Cooling (transfer of heat to the surrounding air) can be done where there's room for it, not necessarily at the hot spot.
  2. Case size doesn't limit the size of the radiator, if it's placed outside the case.
  3. Thermal inertia. As noted on page 9 of the article the temperature change is slow, which prevents the thermal fatigue associated with heatpipe coolers. (Your CPU/GPU can survive for decades rather than years.)(2)
The disadvantages of water cooling are cost and weight, plus the requirement of some maintenance every now and then.

(1) With 25 litres of water, starting at +20C, and NO radiator the tested rig could easily be running for one and a half hours at full test load without going anywhere near "hot" (the water only reaching +40C).
(2) Thermal fatigue of course isn't a problem for a system running 24/7 at near constant load, occupied with bitcoin mining or distributed computing and such.

Darxus
Posts: 32
Joined: Sun Dec 25, 2005 4:07 pm
Contact:

Re: Quiet Liquid Cooled Gaming PC Build Guide

Post by Darxus » Wed Jan 21, 2015 10:36 am

"Since the radiator is where most of the heat in the system will end up, the main task of any fans not directly on the radiator should to ensure as much cooler air to the radiator as possible."

I have no doubt that you would've liked the opportunity to keep fiddling with this thing for months, but I'm wondering if it would have worked better without the non-radiator fans. I find it difficult to believe they're doing anything useful, while adding at least a tiny bit of noise. Seems like it would be better to tape up any gaps in maybe the top half of the case to make sure any of the vented hot air doesn't get sucked right back in.

I've noticed, with my system, when the power supply fan hasn't kicked on, air gets sucked in at the power supply, right after venting by the case fan right under it :(

Thanks for doing this, it's very well timed for me.

Darxus
Posts: 32
Joined: Sun Dec 25, 2005 4:07 pm
Contact:

Re: Quiet Liquid Cooled Gaming PC Build Guide

Post by Darxus » Wed Jan 21, 2015 11:04 am

Olle P wrote:I would have opted for a considerably bigger radiator outside the case, the wall mounted type normally used for heating rooms.
I've been trying to figure out options for large passive radiators. Do you have any info on using radiators like that?

First I was looking at Alphacool Cape Cora, then I tried finding other options. I wondered about building something to mount on the side of a case out of the radiators you're talking about, which turn out to be called hydronic heating elements. I also noticed that refrigerator evaporators look quite good.

And there's this thing: http://www.globalindustrial.com/p/hvac/ ... ng-element
"All types have heating elements with 1/2" nominal (5/8 O.D.) copper tube connections"

I've also wondered about getting some aluminum sheet, and (3/8"?) copper tube, and building my own.

Galvanic corrosion is a thing that comes up a lot in water cooling. Apparently bad things happen when you have multiple different metals in a water loop. Looks like it's only a problem if they're electrically connected though, so the problem could be solved by making sure any dissimilar metals are not grounded?

This water loop, using a 10 gallon aquarium basically as a passive radiator, seems worth mentioning: http://wwww.homstardware.com/nate/HomsLoopArticle.html

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

Re: Quiet Liquid Cooled Gaming PC Build Guide

Post by MikeC » Wed Jan 21, 2015 11:24 am

Darxus wrote:"Since the radiator is where most of the heat in the system will end up, the main task of any fans not directly on the radiator should to ensure as much cooler air to the radiator as possible."

I have no doubt that you would've liked the opportunity to keep fiddling with this thing for months,
No, not at all! :shock:
Darxus wrote:but I'm wondering if it would have worked better without the non-radiator fans. I find it difficult to believe they're doing anything useful, while adding at least a tiny bit of noise. Seems like it would be better to tape up any gaps in maybe the top half of the case to make sure any of the vented hot air doesn't get sucked right back in.
Perhaps I should have included all the data on the iterative sessions? nah, most readers would have been snoozing after half a dozen. To address your wondering... yes, the two back and front intake fans did need to be on to keep non-CPU/GPU temps under control. They did have an impact on the main cooling as well, but much less than the top fans, as you would expect.
I've noticed, with my system, when the power supply fan hasn't kicked on, air gets sucked in at the power supply, right after venting by the case fan right under it :(
With the top fans spinning so slowly, their "reach" is much reduced, compared to higher RPM. Yes, even w/o any intake fans, cooler air would naturally be pulled into the case and up to the radiator/fans, but the volume/rate would be very modest. With the 400W+ thermal load, I just didn't feel it safe to rely only on the radiator fans. OTOH, for real-world use -- ie, gaming load, not furmark+p95 -- it might be OK to pull or turn off the other fans. Obviously this is where DIYers who decide to follow this guide will experiment for themselves.
Thanks for doing this, it's very well timed for me.
You're welcome. So what are you planning?

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

Re: Quiet Liquid Cooled Gaming PC Build Guide

Post by MikeC » Wed Jan 21, 2015 11:33 am

Darxus wrote:I've been trying to figure out options for large passive radiators. Do you have any info on using radiators like that?
Here's one from a fellow Vancouverite who used to hang out on SPCR forums. Ron was in touch recently to say he's finally dismantled his copper pipes. http://www.overclockers.com/pc-water-co ... -radiator/

For my guide, I avoided anything that might look DIY -- hence, anything with an exposed radiator on the outside was not considered (which explains the case search). That was a common enough option for some enthusiasts back in 2000, but it is really not where most current gamers want to go.

Darxus
Posts: 32
Joined: Sun Dec 25, 2005 4:07 pm
Contact:

Re: Quiet Liquid Cooled Gaming PC Build Guide

Post by Darxus » Wed Jan 21, 2015 12:00 pm

MikeC wrote: No, not at all! :shock:
Haha, okay.
MikeC wrote: Perhaps I should have included all the data on the iterative sessions? nah, most readers would have been snoozing after half a dozen. To address your wondering... yes, the two back and front intake fans did need to be on to keep non-CPU/GPU temps under control. They did have an impact on the main cooling as well, but much less than the top fans, as you would expect.
Nice, I really appreciate that information.
MikeC wrote: So what are you planning?
Something around a Radeon R9 290, and probably an AMD FX-9590. Aimed at 4k HTPC, and light gaming (not max detail 4k, I don't do much of this). I'm tempted to put a water block on everything (video, CPU, motherboard, RAM, drives), and try cooling it with only a fanless radiator. So the only moving parts would be the pump, and probably one big slow scratch drive. I figure without a radiator size limit, it should be doable. But it would be nice to be able to keep it small enough to bolt to the case.

So, of course, I'm among those who would love to hear your opinions of pump volumes while wet.

I've wondered if a vertically oriented passive radiator would work better with a sleeve / chimney wrapped around it, in hopes of increasing convection.

I'm really frustrated that nobody makes a Radeon blower wide enough that it can run a fan slow enough to be quiet.
Take one of these, venting externally: https://www.youtube.com/watch?v=QQhqOKKAq7o
Make it (fan and heat sink) 4 or 6 slots wide, so the fan can run much slower, and it seems like that would be much more tolerable, and make me not need water cooling. But... I guess I don't really mind the excuse to play with water cooling finally.


A decade ago, based largely on the info I found here, I built a system in an Antec P180, with an Arctic Cooling video cooler, and a massive fanless Scythe Ninja CPU Cooler ducted to the only (case) fan. Made that duct out of a Lucky Charms box. Worked beautifully. I'm still surprised that doesn't seem to be a common thing.
http://www.chaosreigns.com/gallery/P180 ... T.jpg.html
http://www.chaosreigns.com/gallery/P180 ... t.jpg.html
viewtopic.php?t=27923
So, I've been appreciating your work for a long time :)

Darxus
Posts: 32
Joined: Sun Dec 25, 2005 4:07 pm
Contact:

Re: Quiet Liquid Cooled Gaming PC Build Guide

Post by Darxus » Wed Jan 21, 2015 12:08 pm

"As long as the motherboard remains frozen in a form factor never meant for video cards that exceed the power and thermal profile of CPUs, liquid cooling will continue to attract."

How would you like it to be changed?

I saw a system that used a ribbon riser to lay the video card down flat next to the motherboard: https://www.reddit.com/r/buildapc/comme ... tercooled/
It's a pretty extremely custom build, but if you want space for ventilation, there are options.

CA_Steve
Moderator
Posts: 7650
Joined: Thu Oct 06, 2005 4:36 am
Location: St. Louis, MO

Re: Quiet Liquid Cooled Gaming PC Build Guide

Post by CA_Steve » Wed Jan 21, 2015 1:38 pm

Darxus wrote:Something around a Radeon R9 290, and probably an AMD FX-9590. Aimed at 4k HTPC, and light gaming (not max detail 4k, I don't do much of this). I'm tempted to put a water block on everything (video, CPU, motherboard, RAM, drives), and try cooling it with only a fanless radiator. So the only moving parts would be the pump, and probably one big slow scratch drive. I figure without a radiator size limit, it should be doable. But it would be nice to be able to keep it small enough to bolt to the case.
The FX-9590 has a 225W TDP....and is outperformed by a 65W i7. An 84 or 88W i5 is in the same price class as the FX and performs about the same - better for games, worse for multithreaded transcoding. Trying not to bash the CPU, it just is contrary to my nature (lower wattage-> cooler PC -> quieter PC).

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

Re: Quiet Liquid Cooled Gaming PC Build Guide

Post by MikeC » Wed Jan 21, 2015 3:57 pm

Darxus wrote:"As long as the motherboard remains frozen in a form factor never meant for video cards that exceed the power and thermal profile of CPUs, liquid cooling will continue to attract."

How would you like it to be changed?

1. GPU should be parallel to CPU and be allowed as much room, around and above -- like 2011/1366 socket mounting holes
2. If CPU coolers can be used, that'd be just about perfect. Some standard VRM & RAM positions on the card would be good; this could be cooled by a secondary cooler.

Off the cuff thoughts --

Imagine a modified mATX board with only 2 slots, with an additional slot at the bottom edge of the board that would access a completely separate graphics board on the same plane as the motherboard. The graphics board could be in 2 sizes -- standard one maybe 4x9.5"; big one at maybe 6.5x9.5". It would screw on to the case the same way as the motherboard. Total length of the motherboard + standard graphics board would be about the size of standard ATX board; the bigger board would obviously be taller. But now you can have two tower coolers for both CPU and GPU, with all the airflow going in the same direction.

Very few applications call for all the ATX mobo slots to be used; so much is integrated anyway, which is why mITX is a viable solution for both mainstream and specialized apps these days.

Haven't thought about how you'd get multiple graphics boards in play here...

A less extreme change:
Make more room between CPU and first PCIe x16 slot with the GPU on the same side as the CPU -- ie graphics board PCB faces away from the CPU. Then airflow around/to the CPU and GPU can be shared.

Cistron
Posts: 618
Joined: Fri Mar 14, 2008 5:18 am
Location: London, UK

Re: Quiet Liquid Cooled Gaming PC Build Guide

Post by Cistron » Wed Jan 21, 2015 4:54 pm

MikeC wrote:
Darxus wrote:"As long as the motherboard remains frozen in a form factor never meant for video cards that exceed the power and thermal profile of CPUs, liquid cooling will continue to attract."

How would you like it to be changed?

1. GPU should be parallel to CPU and be allowed as much room, around and above -- like 2011/1366 socket mounting holes
2. If CPU coolers can be used, that'd be just about perfect. Some standard VRM & RAM positions on the card would be good; this could be cooled by a secondary cooler.

Off the cuff thoughts --

Imagine a modified mATX board with only 2 slots, with an additional slot at the bottom edge of the board that would access a completely separate graphics board on the same plane as the motherboard. The graphics board could be in 2 sizes -- standard one maybe 4x9.5"; big one at maybe 6.5x9.5". It would screw on to the case the same way as the motherboard. Total length of the motherboard + standard graphics board would be about the size of standard ATX board; the bigger board would obviously be taller. But now you can have two tower coolers for both CPU and GPU, with all the airflow going in the same direction.

Very few applications call for all the ATX mobo slots to be used; so much is integrated anyway, which is why mITX is a viable solution for both mainstream and specialized apps these days.

Haven't thought about how you'd get multiple graphics boards in play here...

A less extreme change:
Make more room between CPU and first PCIe x16 slot with the GPU on the same side as the CPU -- ie graphics board PCB faces away from the CPU. Then airflow around/to the CPU and GPU can be shared.
Amen! Mike's first approach would also mean that one could mount an exhaust fan close to the GPU.

Olle P
Posts: 711
Joined: Tue Nov 04, 2008 6:03 am
Location: Sweden

Re: Quiet Liquid Cooled Gaming PC Build Guide

Post by Olle P » Thu Jan 22, 2015 2:20 am

MikeC wrote:... anything with an exposed radiator on the outside was not considered ... it is really not where most current gamers want to go.
Most current gamers don't opt for a <20 dB(A) rig in the first place.
For a more "professional" look a pair of Alphacool Cape Cora 642 should do the trick.
Last edited by Olle P on Thu Jan 22, 2015 2:29 am, edited 1 time in total.

Olle P
Posts: 711
Joined: Tue Nov 04, 2008 6:03 am
Location: Sweden

Re: Quiet Liquid Cooled Gaming PC Build Guide

Post by Olle P » Thu Jan 22, 2015 2:29 am

Darxus wrote:I've been trying to figure out options for large passive radiators. Do you have any info on using radiators like that?
I got the idea from a Norwegian guy who posted on an overclockers forum several years ago.
He used a smaller radiator of the classic "accordion look" type. Had it laying on the floor under his desk keeping his feet warm.
He claimed that he used it to heat the room, adjusting the room temperature by varying the level of overclock on the CPU.
The water temperature was stable no matter the amount of overclocking...
Darxus wrote:Galvanic corrosion is a thing that comes up a lot in water cooling. ... Looks like it's only a problem if they're electrically connected though, so the problem could be solved by making sure any dissimilar metals are not grounded?
First there are plenty of additives to the water that prevents corrosion (and algae growth).
Second I agree that differentiating metal should not have any other electrical connection to the rest of the loop but the water. A wall mounted radiator not grounded and connected to the computer through plastic hose should pose no galvanic problems.

Darxus
Posts: 32
Joined: Sun Dec 25, 2005 4:07 pm
Contact:

Re: Quiet Liquid Cooled Gaming PC Build Guide

Post by Darxus » Thu Jan 22, 2015 6:39 am

MikeC: I noticed that when I search for pump decoupling, there are a lot of web and image hits, which suggests to me that there is actually a fair amount of interest in quiet pumps. So if you *wanted* to test lots of pumps wet, with your nice tools, and build an authoritative list, I believe there would be interest. I also feel like such high quality measurements could encourage manufacturers to do better.
Olle P wrote:Most current gamers don't opt for a <20 dB(A) rig in the first place.
Maybe that's because they don't know how?

Also, I suspect I'm not the only one curious how you'd decouple the MCP655-PWM pump you used if it was not in the drive bay reservoir you used. Something like this:
http://www.frozencpu.com/products/16242 ... alled.html
It seems like the drive bay reservoir you used was easier to suspend, but you've done this a lot more, so I'm curious what you would do with this type. Or, if that's hard, maybe the pump alone with the reservoir detached, maybe like this: http://www.frozencpu.com/products/12198 ... A-MBK.html
I noticed a number of people are laying them on a sponge.

I did a search on why not to run a DC motor unloaded:
"One of the drawbacks/precautions about series-wound DC motors is that if they are unloaded, the only thing limiting their speed is the windage and friction losses. Some can literally tear themselves apart if run unloaded." - https://www.pc-control.co.uk/dc-motors.htm
Olle P wrote:First there are plenty of additives to the water that prevents corrosion (and algae growth).
I just looked them up. This page seems like a good start, although it's 11 years old: http://www.overclockers.com/pc-water-co ... y-part-ii/
"There are numerous corrosion inhibitors on the market, most of which have been designed for automotive cooling systems (Water-Wetter, Hy-Per Lube Super Coolant, Purple Ice, Zerex Racing Super Coolant, etc). However, a couple products are being specifically marketed for PC water-cooling systems (Swiftech HydrX and InnovaProtect)."

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

Re: Quiet Liquid Cooled Gaming PC Build Guide

Post by MikeC » Thu Jan 22, 2015 10:12 am

Darxus wrote:MikeC: I noticed that when I search for pump decoupling, there are a lot of web and image hits, which suggests to me that there is actually a fair amount of interest in quiet pumps. So if you *wanted* to test lots of pumps wet, with your nice tools, and build an authoritative list, I believe there would be interest. I also feel like such high quality measurements could encourage manufacturers to do better.
Yes, I have the acoustic measurement facility & tools but what of all the flow, resistance, etc measurements done by the likes of Martin? If pumps make different sounds under different loads, then they'd need to be tested in a variety of standardized conditions -- like we do currently with fans. And how many pumps are we talking about? iirc, Martin had a list of around 20 he'd tested, and many just seemed like minor variants. It seems like there can't be THAT many. Maybe if ALL major pumps could be collected up, this might be doable for a 1-2 part roundup. Lots of research to get the testing procedure right tho.
Darxus wrote:Also, I suspect I'm not the only one curious how you'd decouple the MCP655-PWM pump you used if it was not in the drive bay reservoir you used. Something like this: http://www.frozencpu.com/products/16242 ... alled.html It seems like the drive bay reservoir you used was easier to suspend, but you've done this a lot more, so I'm curious what you would do with this type. Or, if that's hard, maybe the pump alone with the reservoir detached, maybe like this: http://www.frozencpu.com/products/12198 ... A-MBK.html I noticed a number of people are laying them on a sponge.
It's easy to suspend just about anything in elastic cord, just takes a little imagination. With the pump attached to the cylinder, I'd consider the techniques used to shock mount microphones. Pump alone is probably easier... too many way to list. Putting it on soft foam probably gets >80% of the way, and might be no less precarious than elastic cord (though if you use enough runs of the cord and keep it under modest tension, it'll pretty much last indefinitely.)
I did a search on why not to run a DC motor unloaded:
"One of the drawbacks/precautions about series-wound DC motors is that if they are unloaded, the only thing limiting their speed is the windage and friction losses. Some can literally tear themselves apart if run unloaded." - https://www.pc-control.co.uk/dc-motors.htm
Hmmmm....

flinx
Posts: 111
Joined: Wed May 21, 2008 5:19 am
Location: Norway

Re: Quiet Liquid Cooled Gaming PC Build Guide

Post by flinx » Thu Jan 22, 2015 4:07 pm

By the way, if you're adding a 290X to the setup, it'll be downgraded to a regular 290 so as to match the existing 290 card. Just something to keep in mind.

Olle P
Posts: 711
Joined: Tue Nov 04, 2008 6:03 am
Location: Sweden

Re: Quiet Liquid Cooled Gaming PC Build Guide

Post by Olle P » Fri Jan 23, 2015 4:16 am

Darxus wrote:
Olle P wrote:Most current gamers don't opt for a <20 dB(A) rig in the first place.
Maybe that's because they don't know how?
The vast majority don't care, thinking 30dB(A) is just about dead silent and 40dB(A) is considered very quiet.
Darxus wrote:I did a search on why not to run a DC motor unloaded:
"One of the drawbacks/precautions about series-wound DC motors is that if they are unloaded, the only thing limiting their speed is the windage and friction losses. ...
That's plain wrong. The main limiting factor is the induced voltage.

As the rotor spins the motor acts like a generator creating a voltage opposite to the one driving the motor.
This voltage increases with speed. The current through the motor is defined by feeding voltage minus the induced voltage, divided by the resistance in the windings. The current will stabilize when it is just sufficient to create a magnetic field strong enough to have the motor spinning at constant speed.
Only if you apply to high voltage will a motor fall apart or overheat.

Darxus
Posts: 32
Joined: Sun Dec 25, 2005 4:07 pm
Contact:

Re: Quiet Liquid Cooled Gaming PC Build Guide

Post by Darxus » Fri Jan 23, 2015 9:56 am

Video discussing radiator fans, push / pull / push + pull.
https://www.youtube.com/watch?v=UyC3lZ5WFMk
Says for performance, push vs. pull doesn't matter, and push + pull doesn't help as long as you have fans optimized for pressure. Of course, I'd love to see it tested.

There are some big in-case radiators available.
http://www.alphacool.com/product_info.p ... a-480.html
(One more fan, and ~2x the thickness of the one used for this article.)

W. G. Bergamot
Posts: 11
Joined: Sun Jun 01, 2014 4:30 pm

Re: Quiet Liquid Cooled Gaming PC Build Guide

Post by W. G. Bergamot » Sat Jan 24, 2015 5:51 pm

I've very excited by SPCR's foray into liquid cooling. It's an area I've avoided due to alleged noise concerns and a very real aversion to SLI (costs, microstuttering). There hasn't been noise testing for liquid cooling done anywhere to SPCR standards; there's no way I'm dropping the kinda coin liquid cooling requires without it. The testing you've begun may finally make this jump possible, and those in the liquid cooling business would be wise to support you in your testing, because I suspect there may be more of us waiting in the wings.

In liquid cooling, it seems there is a category of radiator that begs to be tested: "slim-profile" or "low-profile" radiators that are ~30mm thick and may (it's not immediately clear) also feature fewer Fins Per Inch (FPI). The SPCR mantra of reducing impedance for quieter computing can be tested.
Example 1
Example 2
Example 3
Example 4
Example 5 (34ish thickness)
etc.

The radiator you reviewed is on the thin side as far as radiators go @ 38mm thick. A cursory inspection of common radiators reveals 45mm, 60mm, and even 80mm thicknesses. Could a radiator that is ~8mm thinner be a tradeoff that works, as it gives you less surface area with less impedance?

-edited for coherence-
Last edited by W. G. Bergamot on Sat Jan 24, 2015 8:43 pm, edited 1 time in total.

Darxus
Posts: 32
Joined: Sun Dec 25, 2005 4:07 pm
Contact:

Re: Quiet Liquid Cooled Gaming PC Build Guide

Post by Darxus » Sat Jan 24, 2015 7:42 pm

I'd also be very interested in more data on radiator thickness vs. cooling capability for a fixed noise level.

This is a neat start: http://martinsliquidlab.org/2012/06/08/ ... h-testing/

Particular the part where he realized if he sandwiched an additional radiator on top of an existing one, cooling *decreased*, apparently due to constricted airflow.

So many questions. What if we had even lower Fins Per Inch radiators? What if people made double thickness quiet fans that were further optimized for pressure?

But "What thickness of radiator is best for use with quiet fans?" would be great info to have.

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

Re: Quiet Liquid Cooled Gaming PC Build Guide

Post by MikeC » Tue Jan 27, 2015 12:41 pm

POSTSCRIPTS ADDED - 15th & 27th of Jan: http://www.silentpcreview.com/article1437-page11.html

This is a followup in two parts to the question of pump noise first raised by the screeching noise exhibited by the EK-SBAY DDC 3.2 PWM in dry testing. The first part explores the cause of that noise; the second part presents acoustic measurements for the pump in wet operation, and compares it to the Swiftech MCP655 PWM Pump used in the build guide.

To say I have egg on my face is not enough; I'm also eating crow and humble pie. My sincere apologies to EK, VisionTek and/or Laing for any negative effects of my original report on the acoustics of this pump.

quest_for_silence
Posts: 5275
Joined: Wed Jun 13, 2007 10:12 am
Location: ITALY

Re: Quiet Liquid Cooled Gaming PC Build Guide

Post by quest_for_silence » Tue Jan 27, 2015 3:32 pm

MikeC wrote:POSTSCRIPTS ADDED - 15th & 27th of Jan: http://www.silentpcreview.com/article1437-page11.html
Thanks for being such a fair, honest, thoughtful reviewer, Mike. :mrgreen:

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

Re: Quiet Liquid Cooled Gaming PC Build Guide

Post by MikeC » Wed Jan 28, 2015 9:02 am

quest_for_silence wrote:Thanks for being such a fair, honest, thoughtful reviewer, Mike. :mrgreen:
Thanks Luca, I try. Let's hope others are as forgiving.

Verskon
Posts: 1
Joined: Sat Feb 21, 2015 4:24 am

Re: Quiet Liquid Cooled Gaming PC Build Guide

Post by Verskon » Sat Feb 21, 2015 7:03 am

Thank you for the well written review.
I too cringed when you wrote about dry running pumps. Sort of like running a car engine once you've drained all the oil to get the last of the 'used' oil' out ><'.

I've been running my first watercooled system almost 24/7 for 7 years now. It's gone through a number of upgrades since its original AMD64 6000+ cpu and 8800gts, to a AMD A8-3850 and AMD 5850HD and now a I5-3570K @ 4.3GHZ and Radeon 7970GHZ. Rad > Pump > Cpu > GPU > Rad. The CPU and GPU blocks were upgraded 2 years ago from a D-TEK FuZion CPU and Swiftech MCW60 VGA blocks to the Koolance CPU-380i and MCW82 VGA blocks. The 360mm x 12mm rad is the same, as is the D5 Var pump, Antec P182SE, 3/8 ID 5/8 OD Tygon tubing, Copper T- joint and 1/2'' worm clamps. I drain, fiddle with the plumbing and refill with distilled water + some copper sulfide drops system every time I upgrade and every year or two in between. System water level is topped off every 6 months with a T-Fill tube just in-front of the pump intake. Fans are replaced every 3-4 years, or as they start to make noise. I have a simple 4x knob variable speed fan controller rad fans and a hdd fan are hooked up to.

Now, for why I registered and gave all that above. Your idea to suspend the Pump+Rad was wonderful, as well as taking a scientific approach to running down each source of noise. Long ago, I calculated that my setup could dissipate 500-800 watts of energy easily with powerful fans. Instead, I went with a powerful, but quiet, system like many of us who invest in watercooling. I had gotten a P182SE case and cut the top out, dremeled the internal partitions away and re-positioned the HD cages for more room. I stand-off mounted the radiator on top of the case with 3x pusher 120mm fans inside the case and a 19mm (3/4'') gap between to allow for better better flow. Packing and window sealing tape were used to block gaps and force air through the bottom half of the case at higher static pressure, for more passive flow over HDD's, GPU and PSU. The 7970GHZ has copper heatsinks stuck to the memory, capacitors and power components with a 140mm PWM fan tilted toward it for more stable operation at max utilization.

This worked adequately till recently when the system hum had gotten more noticeable and annoying. After reading your article, I tore the system down and systematically plugged in each component that made noise to identify the hums and rattles. I cut up some old tubing to act as 'feet' for the non-mounted 140mm GPU cooling fan. I removed the 140mm PWM fan from the Thermaltake 750XT PSU, replaced the rad fans with Hyperborea case fans and suspended the D5 pump. The pump had been sitting on a 25mm thick silicone vibration isolation pad for those 7 years, but, by suspending the pump, the worst of the annoying 'hum' was eliminated, much like you found in your setup. I then stress tested the system for an hour with Prime95 and Furmark for max heat and power consumption. The fans were set for 'just audible air noise' and pump @ setting 3. I was registering 400+ watts of power consumption for the PSU (600+ system with triple 26'' monitors and peripherals) and reached 80*c CPU Core temps and 52*c GPU (95% max temp reached after 10 minutes), all in line with previous stress tests. The D5 has a rated max of 50*c water temp, however, because the pump is immediately after the rad, it always receives cooler water, in this case about 40*c.

These changes have made the system much more tolerable and saved me several hundred dollars where I was about to purchase a number of 'silent' components (passive rad tower and setup, passive psu). My next project for this system will be to tackle the last aspect that bugs me yet. Even when Idle, my system has a moderate amount of air noise and light electronic hum. Using an iPhone decibel meter, my house reads 38 db with no noise noticeable and 42 db @ 1m with the computer system on. An air cooled system can easily achieve 'whisper' quiet or inaudible operation when internet browsing or idling. I may de-tune my CPU and GPU as I'm not needing full performance for gaming atm and attempt similar <20db results (with my reference db readings so high, I may end with a dedicated sound level meter too). The other possibility is a PWM controlled pump & fan setup. This should allow very quiet or silent operation when idling and automatic ramp up as the system needs more cooling. If gaming, more system noise isn't an issue. I've seen pwm multipliers or you can make your own.

There may be more base level noise with a watercooled system, but, not having it shriek when gaming and having the cool factor of a custom water system are just two of the many overwhelming benefits! It's unfortunate that many people are put off by the perceived 'dangers', 'fiddly bits' and spergy forum zealots.

scrawl
Posts: 1
Joined: Sun Jan 31, 2016 8:51 am

Re: Quiet Liquid Cooled Gaming PC Build Guide

Post by scrawl » Sun Jan 31, 2016 9:30 am

That was a good article, but the end result could have been better! You called the project a success, but I see so much untapped potential, would love if SPCR could give watercooling another go in the future.

Rads/fans:

The main reason to watercool is scalability... you used a triple rad, this is what most people will consider the bare minimum for a GPU + CPU, not enough for a very quiet system. Double the amount of radiator space, and you can run your fans at completely inaudible speeds. Use as many fans as you can fit (push + pull), more fans = slower speeds = less noise.
The only location that is practical is the top panel unless you can find a case with ~40cm height front intake, which would probably make it at least 60cm tall, and hot air would blow out from the front. On any watercooling radiator, using the fans to exhaust makes most sense; no point blowing the heat back into the case!
It doesn't matter if the radiator fans are intake or exhaust, the difference in cooling will be negligible. You could even argue that intake can perform better since you're cooling the radiator with cool intake air, not with warm case air. Generally, watercooling gives you so much temperature headroom that you don't need to care about 1° or 2° improvements. Personally I use both, 1 intake radiator and 1 exhaust radiator. Just make sure that overall case airflow is balanced, erring on the side of more intake so that you have positive pressure keeping dust out of your system.

The Coolstream PE 360 Triple radiator you used isn't optimized for slow speed fans, get a thinner radiator with lower FPI (fins per inch). I use 30mm Alphacool ST rads.

Pump:

The reason your pump is audible is because it's mounted in that EK reservoir. Even though you decoupled the reservoir with the elastic cord, the pump vibrating against the reservoir body still produces airborne noise. You need to get a bay reservoir that is optimized for silence, or you get a separate pump top instead of a pump/res combo. I have the Monsoon S2 D5 pump/reservoir combo mounted in optical bay, and my pump is completely inaudible at 800 rpm over my be quiet PSU idling. Things to look for in a quiet pump/reservoir combo would be a soft material, a cover/housing for the back of the pump, and anti vibration screws. The Monsoon one even features "sound absorption chambers", though I can't say how much of a difference they make.

quest_for_silence
Posts: 5275
Joined: Wed Jun 13, 2007 10:12 am
Location: ITALY

Re: Quiet Liquid Cooled Gaming PC Build Guide

Post by quest_for_silence » Mon Feb 01, 2016 1:55 pm

scrawl wrote:The Coolstream PE 360 Triple radiator you used isn't optimized for slow speed fans, get a thinner radiator with lower FPI (fins per inch). I use 30mm Alphacool ST rads.

Actually the PE is a 21mm core with an average 18FPI, while the ST is a 18mm core with an average 16FPI, not so different I'd say: as a matter of fact the PE should have a slightly better dissipation at low fan spinning than the ST.

Olle P
Posts: 711
Joined: Tue Nov 04, 2008 6:03 am
Location: Sweden

Re: Quiet Liquid Cooled Gaming PC Build Guide

Post by Olle P » Mon Nov 20, 2017 6:22 am

tim851 wrote:... some of the things they found out quickly were
a) flow rate doesn't matter
b) hose diameter doesn't matter
c) loop order doesn't matter
d) multiple independent loops don't make sense
...
Let's add "pump speed doesn't matter" to this list, and I've heard them all...

Let's investigate them a bit deeper:
* Hose diameter and pump speed will influence the flow rate, which is key.
* The flow rate does influence the difference in temperature between "cold" and "warm" water. Assuming there's only one convector in the system the temperature difference is given by
dT=0.86*P/F where dT is the temperature difference in degrees kelvin, P is the amount of heat transferred in watts, and F is the water flow in liters per hour (lph).
- While F>P all of the "findings" are true. The water temperature will stay essentially the same throughout the loop, so any changes to the flow will have a minimal impact.
- While F<<P you have a situation where loop order really can matter and multiple loops may make sense!
Assume for example that you want to use the water to cool a 150W GPU and a 100W CPU, for a combined power of 250W. Then if the flow rate is 8.6 lph (caused for example by running a weak pump slowly in combination with using narrow tubes) the water temperature will go up by 15 degrees while passing the GPU block and by 10 degrees while passing the CPU block. It's obvious that there will be a significant difference in CPU and GPU temperatures, depending on the order of the cooling blocks. It's also quite possible that the pump will overheat if located in the "warm" section of the loop. Any increase in water flow is beneficial, so pump speed and hose diameter does matter!

When a (very) low flow rate is considered a must to keep things quiet it thus become sensible to use multiple loops to prevent feeding pre-heated water to a cooling block.

Post Reply