Would suggest re-reading the original post, the OP made no mention of:
Increasing performance
Needing a Faster GPU
Lowing GPU temps
Installing aftermarket air cooling
None of the answers you provided will do anything to address his stated goal .... reducing noise... The AIO is louder than the factory solution.
Now for the OP ... two things wanna say up front:
A. What was the best 3 years, 2 years, 6 months ago is not likely to be the best today.
B. Best varies by necessity for each individual's goals ... maximizing cooling performance is a goal ... reducing noise is a goal ... aesthetics is a goal.... functionality is a goal ... no product is *best* in all these areas.
I would recommend moving away from / ignoring a couple of common mindsets:
1. All components should come from the same manufacturer... among experienced WC'ers, this pretty much never happens.
2. What was the best and brightest in one year, will not likely hold that position a couple of years down the line. This applies greater to GPUs than the WC components themselves. For example, looking at GPU reviews ... up thru the nvidia 600 series, Asus was go to manufacturer that usually had the best performance ... since then, performance wise, it's been MSI winning the title in TPU reviews with smaller guys trading place behind them. EVGA is oft considered to be a quality manufacturer but they have the highest number by far of failed designs. The VRMs on their 570 often fried, the heat sink on the 970 ... 1/3 of it "missed" the GPU The 10xx series had excessive failures, even fires, because they cheaped out and didn't include thermal pads.
3. Watercooling isn't always about improving performance, for me it's about noise reduction ... seems this is your major concern. So water cooling a card is never pointless. No card needs to be worthy of WCing. A 3080 on a box w/ a 1080p monitor can only be described as supreme overkill. But there's no reason any card should not be made quieter id it'sd annoying the user. If my svreen goes dark and i can hear that the sytem is on, I'm gonna want to throw it out the window. But performance should always be a consideration, if not the primary consideration
The go to source for WC components back when we doing a lot of WCing was martinsLiquidlab ... but martin "retired" several years ago (partial site still up) ... you will notice there that AlphaCool has the highest performance ratinmg there, but those are old numbers. A great sources for tests and reviews is overclock.net ... not necessarily published there but oft quoted in the forums. But even these reviews are lacking because lotta things they don't go into detail on. Another thing about forum opinions; if you ask in forums about pickup trucks, you will find that there are Ford guys, Dodge Guys and Chevy guys. Stacks of data and reports that go against their preconceived notions handed down from father to son will not ever change anyone's mind. Anecdotal evidence from the family's teeny sample size, will always outweigh published data sets invorelving thousands of trucks. The same is true w/ PCs ... when a person gets involved in building PCs, the advice they get in that 1st year of forum activity often sticks forever. The most common example is the refence to "needing high SP fans" ... In the 90s where rads were thick and had 30 fpi ... yeah, we used high SP fans to overcome the thickness and tight fin spacing. With today's slim and low to medium fin spacings, it is no longer a "real" thing"... the higher rpm of high SP fans just adds noise and you are limited at low speeds. A 2000 rpm fan will dottom out at about 500 rpm.
Things to consider:
A. Radiators
a) Some rads have 2 ports, others have 7 which provides much greater routing flexibility, especially when using rigid acrylic tubing
b) Some rads have screw protectors, some don't ... good safety feature, especially for 1st timers
c) Some rads are not cleaned very well after assembly
d) Are you going to use a chemical cleaning kit or manufactured coolants ?
e) Thickness matters .... not only does increased thickness mean some extra performance, but it also dictates the fan type and rpm you want to use.
f) The 1st law of water cooling is no dissimilar metals ... go all copper or all aluminum, later is as good thermally but if budget does not allow you can buy safe, all alumium cooling systems. Don't recommend it, but galvanic corrosion is not an issue as laong as all metals are same.
Did you run the calcs to determine radiator size ?
A few peeps have asked me to provide this info in an easily accessible / findable spot so the don't have to remember what thread it is in or take snippets from multiple threads. If ya find it useful, great .... if ya like something better, by all means use what ya comfy with. First off, I did...
www.overclock.net
Given the above, the best performance, best aesthetics, best functionality all mean choosing a different radiator
B. Pump
a) In the past, when SLI'd GPUs was a no brainer cause twin x70s was 40% faster than single x80, we only used the Swiftech 35X2 dual pump system and piped the GPUs in parallel.
GPU manufacturers actually lost money with dial cards as there is more profit in selling the top end cards and as such have intentionally reduced the viability of this technolgy
and this option is no longer attractive. As such single pumps make more sense ... we use the Swiftech D5 / DDC variable speed variant.
C. Waterblock
a) This is mostly determined by the card you have.... EK generally has the widest number of cards supported.
b) Use "full cover" water blocks only. Avoid AIOs
D. Tubing
a) Depends on the fittings you use. We prefer rigid acrylic tubing so I don't want to make a recommendation here.
E. Fittings
a) We have always used Bitspower but that's mainly a function of array of choices, availability and wide product stack
b) For flex tubing, we recommend compression type over cheaper barb type.
F. Coolants
a) We prefer manufactured coolant where you get a concentrate and add distilled water.
G. Reservoir
a) We prefer separate reservoir over pump / res combos. This is the only place I generally use flexible tubing to provide vibration isolation between the pump and res.
b) Recommend leaving a space at top of res to allow visual indicatorof coolant level and allow for thermal expansion
c) use a "fill tube" such that return coolant to res us dicharged bekow the water line to preveny splashing which will entrain air in the coolant
H. Fans
a) Remember, your ONLY stated goal here is to reduce sound.
b) 140mm fans become audible at about 850 rpm
c) Fans are one of the best examples where the "presumed" best is based upon recommendations from a bygone era. For example, if you take the Noctua fans off a Noctua cooler, and replace them with Phanteks, your CPU temps drops by 6C. We always looked at martin's site to get fan perfomance data, but since he retired, it's harder to find ... one good site to check for noise issues is silentpcreview
140 mm fans from Noctua, Phanteks, and Xigmatek battle in our latest fan roundup with one emerging as a clear winner in both thermal performance and acoustics. May 21, 2013 by Lawrence Lee Last month we staged an epic shoot-out among some popular 120 mm fans. Some great sounding fans emerged but...
silentpcreview.com
Scroll down to the table at bottom of that last link and see the Performance / Noise ratio table and not the following quote:
"The Phanteks PH-F140HP/TS is the clear winner in every respect. It edged out the new Noctuas every step of the way, delivering the best overall results of any fan we’ve tested thus far. To top it off, it had cleanest, smoothest sound of all the new fans in this roundup. If we had to start from scratch, this might be our new reference model." These are SP (standard pressure) fans, the HP/TS designation refers to frame type as when mounting a 140mm fan on a 120mm cooler and it lines up with cooler screw holes.
d) Another long held mindset is PWM fans rule. Improvements over the last decade have rendered such "truisms" obsolete. PWM's weakness was low speed hum but that has since been eliminated, of course they remain very expensive compared to DCV fans. DCV fans weakness was that if called to run at low speed froma dead stop, they didn't have anough torque to overecome intertua and get the fan moving. Todays DCV fans are designed to 'start" at a higher voltage and then ramp it down once intertia is overcome. A PWM control feature for DCV fans is available on mny fan hubs and game Mobos.
e) Follow manufacturer's recommendations and ***ALWAYS*** install rad fans as intakes. The mindset that cooling your components with air preheated by with hotter internal air rejects all the laws of thermodynamics. Every cubic inch of air entering the cost MUST exit the case. And, even if it were true, if there is a component of concern with respect to the interior case air being warmer, why doesn't that component get a block ? You put a block on ya GPU /CPU because it's getting hotter than you want it to .... and / or ya want to reduce noise. Trust that the manufacturer of the components knows more abouyt how they best perform than the home builder who builds a system every 3-4 years..
View attachment 188915
I. TIM
a) The best TIM for CPUs is generally not the best for GPUs. With a CPU, you have one small surface ... with GFX Card ...you can sometimes need to apply to GPU, Memory Chips and VRMs ... sometimes also requiring application on the side of the backplates. See attaced PDF
Again, other than fans .... what manufacturer you want depends on what aspects are important to you ... A 1st timer might be best served w/ Alphacool rads with screw protectors ... Someone concerned with aesthetics may be drawn to the finish on Hardware Labs. Some one concerned with aesthetic arrangement or tubing might lean toward Alphacool's 7 ports.
Control System for our test box:
CPU MoBo Header ==> 1st Pump Motor
CPU_OPT MoBo Header ==> 1st Pump Motor
CHA_1 Header==> PWM / DCV Fan Hub 1 ==> (6) DCV fans on 420 rad
CHA_2 Header==> PWM / DCV Fan Hub 2 ==> (4) DCV fans on 280 rad
CHA_3 Header==> PWM / DCV Fan Hub 3 ==> (6) DCV case fans
Channel 0 for the fan motors sees a range from 2200 - 4400 rpm
Channel 1 - 3 have a manual range from 325 - 650 rpm
Channel 1 - 3 have see a typical range from 375 - 825 rpm when stress testing
Channel 1 - 3 have see a typical range from 375 - 625 rpm when gaming
Channel 1 - 3 shut off when curve calls for less than 375 rpm
Fan speeds are dampened by ramping up to speed over 10 seconds
Rad Fans are dampened by ramping down over 90 seconds to remove latent coolant heat
Case Fans are dampened down by ramping down over 90 seconds to remove latent coolant heat
The system is completely inaudible, GPU temps are 42C at 650 rpm ... also have since removed the push fans from the origonal push pull arrangement and droppe 3 case fans. Most times, if not gaming or stress testing, all fans turn off.
) The housing thou seems to be the same so I should be able to replace the pump later if it fails, correct?
