Powerful but silence first, for i5-9600KF & poor airflow

[NewbieOne]

TL;DR — maximum silence within $150 budget for i5-9600KF, for a person with very good hearing and sensitivity issues. Performance comes after silence. Cost savings within budget are not important unless it's paying double for +1 degree (but paying double for -1 dbA is OK, even preferable).

 

Longer version:

 

The CPU is a 9600KF mounted on a Strix Z370-H with a three-fan 1070ti plus an X-fi (PCIE) and two M.2's, nothing else but fans.

 

The case is a Phanteks Enthoo Pro with:

 

Front: 1x230mm intake (some sort of Cooler Master, good stuff but can't pinpoint the exact model)

Bottom: 2x120mm intake (Gelid Phantom fans)

Rear: 1x140 mm exhaust (factory Phanteks)

Top: 1x230mm exhaust (same as above)

 

Filters and mesh everywhere except rear (just honeycomb grille).

 

Current contenders:

 

• Dark Rock Pro 4. Because it's generally supposed to be quieter than D15 while still a greater performer. But after looking at behaviour charts, I suspect that's because of settings, while D15 would still be colder at the same decibels, quieter at the same temps.
• Dark Rock 4 non-pro. Because this isn't a 9900K, and the non-pro tends to be quieter by 1–2 dBA.
• Le Grand Macho RT. Because chances are its fan would never have to spin to audible levels. I suspect D15 would still be colder at the same decibels, quieter at the same temps.
• Ninja 5. Because it often leads the charts on silence. I suspect D15 would still be colder at the same decibels, quieter at the same temps much of the time, though its floor noise level may be worse.
• Fuma 2. Because while not reaching the Ninja's level, it's still very quiet and high-performing.
• Archon, or True Spirit Power + additional fan.
• Arctic Liquid Freezer II 360 (unless 280 is quieter). Worried about the pump's audibility, though.

 

Already owned heatsinks, for which better fans could be bought:

 

Thermalright Ultra 120 (the older revision with only four heatpipes)

Gelid Phantom

 

But, the effect of even the best two fans on either of these two coolers would probably not be as good as a complete cooler available for the same price.

 

Thanks!

 

 

[DefineOG]

Noctua D-15

If you want to get something a little cheaper then get Fuma 2 and noctua fans.

(the fuma + noctua fans are cheaper in Australia, not sure about the U.S or wherever)

[Derrk]

any consideration for full open loop? with lots of radiators, you can slow the fans down and it'll be quieter than any air-cooled configuration. Plus you'd be able to take the fans off the gpu, and replace with a block

 

Edit: would be over budget but:

6 minutes ago, NewbieOne said:

Cost savings within budget are not important unless it's paying double for +1 degree (but paying double for -1 dbA is OK, even preferable).

You'd be able to drop several decibels

[narrdarr]

ninja 5 for 3 reasons

1. the cooler can handle alot without the fans even turning on

2. the fans are silent

3. when the fans do kick they cant run at some of the lowest rpm that i've seen and still push air throw the fin stack.

 

[freeagent]

Just curious, the people recommending coolers.. have you used them before? 

[NewbieOne]

9 hours ago, DefineOG said:

Noctua D-15

If you want to get something a little cheaper then get Fuma 2 and noctua fans.

(the fuma + noctua fans are cheaper in Australia, not sure about the U.S or wherever)

Thanks. Will the Fuma heatsink + Noctua fans be better than the Ninja heatsink + Noctua fans?

 

  

9 hours ago, Derrk said:

any consideration for full open loop? with lots of radiators, you can slow the fans down and it'll be quieter than any air-cooled configuration. Plus you'd be able to take the fans off the gpu, and replace with a block

 

Edit: would be over budget but:

You'd be able to drop several decibels

Would freaking love to, and that's the reason I picked the Enthoo Pro, largely. Unfortunately, I have a bunch of bills to pay plus some opportunities to invest in my freelance business (and those are endless sinkholes for money), so I wouldn't be able to justify spending the money on a CLC rather than e.g. a new professional certificate or website or whatever.

 

  

6 hours ago, narrdarr said:

ninja 5 for 3 reasons

1. the cooler can handle alot without the fans even turning on

2. the fans are silent

3. when the fans do kick they cant run at some of the lowest rpm that i've seen and still push air throw the fin stack.

 

Thanks. Would the Ninja show improvement with a fan swap? I'm asking because while there are fans that test better than those shipped with the Ninja (Kaze Flex?), I've also seen some tests showing that a lot of heatsinks tend to perform better with their 'native' fans, presumably due to having been designed around them.

 

For the record, the Ninja's long heatsink would probably span just the length of my top exhaust 230 mm fan, in addition to the Ninja's exhaust fan getting close to the case's rear exhaust. Maybe this would be helping some?

 

 

Edited August 25, 2020 by NewbieOne
Appending quotes with replies gradually

[pureexe]

40 minutes ago, freeagent said:

Just curious, the people recommending coolers.. have you used them before? 

I use Noctua NH-D15 Chromax.black with Ryzen 9 3900x
For me, I feel that this one isn't significantly improvment from Stock cooler that ship with 3900x in term of tempurature. but it better in term of noise. D-15 is very quite for me.

 

Spoiler

[NewbieOne]

8 minutes ago, pureexe said:

I use Noctua NH-D15 Chromax.black with Ryzen 9 3900x
For me, I feel that this one isn't significantly improvment from Stock cooler that ship with 3900x in term of tempurature. but it better in term of noise. D-15 is very quite for me.

 

  Reveal hidden contents

Let's say you push that baby to the maximum realistic, non-synthetic gaming load. How loud will it be? Will you struggle to hear whispered dialogues or make out the various tiny sounds of nature in the game's ambient?

 

And in office applications — let's say you have a bunch of instances of text editors, calc sheets, browser windows (30 tabs or so), some two or three applications decide to update all at the same time, and a virus scan expresses its desire to join that exciting party. Will there be fans ramping up or will the system take that in silence?

[pureexe]

10 minutes ago, NewbieOne said:

Let's say you push that baby to the maximum realistic, non-synthetic gaming load. How loud will it be? Will you struggle to hear whispered dialogues or make out the various tiny sounds of nature in the game's ambient?

 

And in office applications — let's say you have a bunch of instances of text editors, calc sheets, browser windows (30 tabs or so), some two or three applications decide to update all at the same time, and a virus scan expresses its desire to join that exciting party. Will there be fans ramping up or will the system take that in silence?

This is my daily life. working with a lot of chrome tab, power point and mobile gaming in the same time. Temp is about 50c with with 1300rpm fan speed. it still hear the fan noise but it lower compare with stock fan. and i still hear the cricket sound in the background (i live in tropical area)

Spoiler

 

[freeagent]

Le Grand Macho RT is as strong as D15, but quieter. True Spirit 140 Power is as strong a Le Grand Macho RT but cheaper. I haven’t used a Scythe cooler yet. I have owned a D14 though.

[NewbieOne]

Oh, didn't explain this: When I wrote 'poor airflow', I didn't mean the fans or fan paths but the fact that the case sits in the very corner of the room, under the desk. There's no other realistic placement for a case this size. Rear exhaust sits about 10 cm from the wall, and top exhaust some 15 cm from the desk floor.

 

I've been thinking about switching to a bare frame / test bench — though that may have to wait until the next platform upgrade, so I could also reduce the mobo form factor to a minimum. That would sit on top of a wardrobe some two metres away from my ear.

[NewbieOne]

23 minutes ago, pureexe said:

This is my daily life. working with a lot of chrome tab, power point and mobile gaming in the same time. Temp is about 50c with with 1300rpm fan speed. it still hear the fan noise but it lower compare with stock fan. and i still hear the cricket sound in the background (i live in tropical area)

  Reveal hidden contents

 

I tend to have complete silence during the day and a barely noticeable overall PC sound at night with either the Thermalright Ultra + some fan, or even the Gelid Phantom, during my normal work (Trados, a bunch of Word/Excel/PP/Adobe/whatever instances, some 30 browser tabs, Steam, etc.), but file copying tends to ram the fans up, and so does fast browsing in Facebook (video thumbnails, etc.). I probably wouldn't mind, except my current BeQuiet Silent Wings 3 hi-speed tend to be obnoxious about speeding up, and the Gelid Phantom fans are nasty in terms of sound quality on some rpm levels.

 

I'm not really a complete freak about demanding complete silence even during the most taxing system operations — as long as the hum is quiet and pleasant I can definitely live with it. High-pitch, rattle, wobble, whine, vibration and other such 'pathological' sounds are what I have a problem with. Barely audible humming during updates/virus scans isn't the end of the world, though I hate the sound of my PC waking up on its own as I try to fall asleep (but can't switch off due to needing the tabs where they are plus a bunch of unsaved low-priority work, generally notes writing down marketing ideas as they come to me, etc.).

 

  

11 minutes ago, freeagent said:

Le Grand Macho RT is as strong as D15, but quieter. True Spirit 140 Power is as strong a Le Grand Macho RT but cheaper. I haven’t used a Scythe cooler yet. I have owned a D14 though.

 

Yeah, I've seen the TechPowerUp test and one or two others concurring. What I worry about is that a bunch of other tests seem to disagree and place LGMRT worse than D15, including a couple of decibels louder in some high OC/high load situations.

 

The above is also the problem with the Ninja in some tests and perhaps also True (though there seem to be fewer tests for True).

 

Both with LGMRT and Ninja 5 there seem to be several tests giving them glowing results just a little better than D15's, and perhaps a somewhat larger number of tests placing them on more mundane levels.

 

By the way — I obviously realize that my Ultra 120 is a poor cousin, but if True Spirit Power 14 is so spectacular with a similar construction (though larger in size and more pipes), wonder if my Ultra 120 couldn't be turned into a decent cooler by a pair of powerful but civilized fans. What would you think?

 

[pureexe]

D-15 not that silent. I still notice that my PC still there.

Maybe you just build normal pc and put it into another room. then, use KVM switch (Keyboard Mouse Video switch) to route cable back to your room. 

If it PC for working or causal gaming. This method is fine.

[freeagent]

Yes you could still use your ultra 120 extreme if you have a mount for it and some good fans. I still have mine but no mounts, well not all the pieces anyways. Been meaning to get a new mount from amazon but right now my coolers are fine. With that dense fin pack you need strong fans to get the best of it. Silence wouldn’t be your friend in this instance.

[NewbieOne]

27 minutes ago, pureexe said:

D-15 not that silent. I still notice that my PC still there.

Maybe you just build normal pc and put it into another room. then, use KVM switch (Keyboard Mouse Video switch) to route cable back to your room. 

If it PC for working or causal gaming. This method is fine.

Yeah, been suggested more than once. A bundle of cables including a huge, classless, in-your-face external power switch (red, of course, how else) could be quite awesome to have. I could use a bare metal frame and put it on the fridge in the kitchen.

 

12 minutes ago, freeagent said:

Yes you could still use your ultra 120 extreme if you have a mount for it and some good fans. I still have mine but no mounts, well not all the pieces anyways. Been meaning to get a new mount from amazon but right now my coolers are fine. With that dense fin pack you need strong fans to get the best of it. Silence wouldn’t be your friend in this instance.

Yeah, I was lucky to find an adapter years ago. Possibly still needed an adapter for 775, then 115x… Can't even remember. This thing is so old it could be my child and make me a grandfather. I would love to keep it for countercultural value, and I even already feel like a cheater for buying that Gelid Phantom — but the whole thing was like three bucks more expensive than a single Noctua fan, and rated 200W (yeah, my foot, 200W… perhaps at more decibels than my room fan that looks like it's been taken from a WW2 fighter plane). What I don't have — and am not going to buy (ten bucks apiece? no, thanks) — is fan clips, which are supposed to 12 cm. Instead, I'll buy 14 cm fans and fix them with zip ties. Right now I use a silicone (har har) meat-roasting (har har) cord.

 

I suppose the situation would be actually calling for proper pressure fans, as opposed to the modern CPU cooling equilibrium in which airflow fans perform better even through a heatsink? Where I live, used NF-F12's cost 10 bucks apiece, sometimes shipment included. And then there's Arctic P14, NIB, for three bucks less. Could perhaps also buy two 12 cm rubber washers to put underneath for vibration absorption.

[NewbieOne]

Re: Le Grand Macho RT, I've just found Tweaktown's review, with test outcomes similar to TechPowerUp's, albeit on an Intel i7-6700K, which is probably going to behave closer than an AMD CPU to a 9600K. This covers both stock and OC under load, as well as both PWM and 100% 12V. Sigh.

 

However, this test (video; look at 8:13) shows RT with worse results than D15 and Dark Rock Pro 4. This test is newer, so perhaps newer CPUs act differently from older CPUs despite similar wattage? TIM problems with soldered vs not? (9600K is soldered, of course.)

 

Do we have any Macho owners on board? Can you guys hear it in heavy gaming sessions or when system maintenance coincides with application updates/downloads/whatever?

 

Edit: I forgot there were supposed to be some new Noctuas in late 2020, perhaps 2021. Might be worth waiting for, if going air. For example these — beefier D15, new total-passive, more. A 14 cm version of U12A, along with new 14 cm version of the NF-A12 fan, is also supposed to be coming out eventually, which would likely be somewhat better than the 12 cm version, and since the 14 cm version would still fit in my case, the 12 cm one would make less sense.

 

Speaking of NH-U12A, how do you guys think this thing would turn out with 2 of those NF-A12x25 fans? (At half the cost of an U12A.) It has seven pipes, with 3 or 4 being 8 mm and the rest 6 mm, and the base is copper. In my experience the cooler doesn't perform as well as its looks tend to suggest, though fans may be the cause. Or perhaps my i3-8350K (@ 4.6 GHz and up) was a particularly hot unit.

 

An alternative would be to buy two of those 14 cm fans from the Macho (TY-140-something?), for little more than a single NF-A12x25. Or just P14, which I'll probably end up having to test anyway, being too interesting to pass up on. Three would be a possibility — especially with cheap Arctics — but from what I've seen a third fan adds more noise than performance (like 1 degree at the cost of 1–2 decibels, roughly speaking, based on three-fan D15 experiments from a tech site). And of course NF-A15 are a possibility and would fit with the included clips rather than a ghetto method.

 

… And eventually after the new Noctuas coming out I could just put such fans on the case.

 

 

 

[freeagent]

I have Le Grand Macho RT, outside of True Spirit 140 Power it’s the best air cooler I’ve used yet. The only reason I can say that is because I still have my x58 setup that I ran D14 on. That and my old screen caps told me what I needed to know. Everyone can say whatever they want about reviews, I can only go by my own experience.

 

As for Ultra 120 Extreme, I still have one from 2007 that I bought used on overclockers.com. My D14 was their review sample.

[NewbieOne]

Can't recall when I got the ultra, but I think I probably jumped at a super low opportunity, considering I had already had a fan, and the cooler's legend at the time had been a huge factor — I had to have it, even several years after the hype. The mount adapter I purchased separately cost more than the cooler itself — meanwhile I spent some months on an Intel box cooler with my i5-6600, so getting the Ultra back in my case was an emotional moment. So yeah, well, I'm quite attached.

 

However, the Gelid Phantom may be a better heatsink. I guess I'll have to test each for a week or so in succession.

 

But, I've also always wanted to go water. AIO is no custom loop, no skill required, but it would kinda still have a symbolic meaning.

[freeagent]

Oh yeah that old Thermalright cooler is only good for around 240w or so. Still good, but you can do better these days.

[narrdarr]

7 hours ago, NewbieOne said:

Thanks. Will the Fuma heatsink + Noctua fans be better than the Ninja heatsink + Noctua fans?

 

  

Would freaking love to, and that's the reason I picked the Enthoo Pro, largely. Unfortunately, I have a bunch of bills to pay plus some opportunities to invest in my freelance business (and those are endless sinkholes for money), so I wouldn't be able to justify spending the money on a CLC rather than e.g. a new professional certificate or website or whatever.

 

  

Thanks. Would the Ninja show improvement with a fan swap? I'm asking because while there are fans that test better than those shipped with the Ninja (Kaze Flex?), I've also seen some tests showing that a lot of heatsinks tend to perform better with their 'native' fans, presumably due to having been designed around them.

 

For the record, the Ninja's long heatsink would probably span just the length of my top exhaust 230 mm fan, in addition to the Ninja's exhaust fan getting close to the case's rear exhaust. Maybe this would be helping some?

 

 

 

You can switch fans if you want too. But that would defeat the purpose. The biggest pro to the fans that come with the ninja 5 is the fact that they can have real low rpm and still have enough static pressure to push air through the fins. 

 

I noticed the post for you daily use case. Still think this the best best cooler and fan combination(stock fans). The le grand macho is would be my next option as both the n 5 and lgm can both handle alot as a passively cooler. Some where around 125w passively. The fans on the lgm are awesome along with all the other cooler mentioned but none of the fans from any of the cooler have as low of a starting rpm or capped rpm as the kaze flex fans

 

Not saying that the other coolers/fans aren't great. they all are awesome. 

[NewbieOne]

With a single quiet 12 cm fan on an i3-8350K @ 4.6 GHz Prime 95 was hitting 100 (EDIT: on the Ultra), though I have no idea how many Watts that was outputting. Clear indication the heatsink's own passive potential was not enough and fans became very relevant at a much earlier point. (The 200W-rated Gelid Phantom also seemed to struggle and IIRC wasn't able to stabilize even on max RPM — there was a very small but slow but constant gain, meaning the temps would eventually hit unsafe levels.) On a 65W processor without OC (i5-6600) and under heavy loads I never heard it unless I wanted sub-75 temps, and then it took some work to pick out the CPU cooler's individual sound. Back with my old C2D E8600 that heatsink and that single silent fan had been enough to OC from 3.3 to >4 GHz. Can't even recall how much above 4, but temperatures weren't really the problem there as much as stability.

 

I still keep looking at things, and I'm becoming more and more convinced that NH-D15 might after all be the safest of all options, as in it will always manage and won't crap out even if the wattage should start nearing on 400.

 

However, I'm kinda loathe to give up on the situational advantage Ninja 5 or Le Grand Macho RT could offer with their potentially lower noise levels under medium or medium-high loads.

 

Things would be easier if Noctua already had the 14 cm version of NF-A12x25 out along with the associated heatsinks. And their prototypes do show beefed-up heatsinks. Judging by what was said in 2018 to 2019, much of the stuff should already be out by now, except it isn't. Hope Noctua are better than George Martin about this. If there's hope for a 2020/early 2021 release, Ultra or Phantom with P14s could be a viable temp solution.

 

Alternatively, I could just go AIO with the Freezer 360, but I worry about the pump and the thing's ability to stay really quiet under low loads (no worries about the high loads it was made for). What seems good to me about water is the longer time to reach equiibrium, like 5 minutes instead of 1.5, which should eliminate most of the spikes from applications starting, closing or updating or file copying or brief ganging up of several updates at once. But I do want the thing to be dead silent under low (Windows + Office) loads at night with low ambient.

 

Oh, and my work is single-core-intense. Basically, it's computer-assisted-translation (CAT) software — fundamentally an XML-based word processor making use of XML-based translation memories getting read and written all the time and assisted by term glossaries operating on the same principle. The application is old, 32 bit, single-core in almost all aspects, compiled in a high-level programming language (possibly even Visual Basic) and badly underoptimized, while the files are sometimes huge and the search parameters quite intense. There is sometimes input lag to be felt as you translate a 50-page document with tons of footnotes and other stuff, let alone a ridiculously large term database such as the European Union behemoths. I had bought that i3-8350K for this reason, actually (total single-core focus + hopes for OC, with 4 cores generally good enough for my games). Avoiding spikes while having a generous overclock would make some of my projects much less of a royal PITA (those where you have a micro wait in between every sentence due to the constant memory/term base parsing).

 

[NewbieOne]

3 minutes ago, narrdarr said:

 

You can switch fans if you want too. But that would defeat the purpose. The biggest pro to the fans that come with the ninja 5 is the fact that they can have real low rpm and still have enough static pressure to push air through the fins. 

 

I noticed the post for you daily use case. Still think this the best best cooler and fan combination(stock fans). The le grand macho is would be my next option as both the n 5 and lgm can both handle alot as a passively cooler. Some where around 125w passively. The fans on the lgm are awesome along with all the other cooler mentioned but none of the fans from any of the cooler have as low of a starting rpm or capped rpm as the kaze flex fans

 

Not saying that the other coolers/fans aren't great. they all are awesome. 

Yeah, they're all awesome, and the choice paradox is kicking in. TBH the choice overload is why I'm looking at the $150 range — hoping for less compromise to analyse, with the ability to just pay for not having to worry about stuff and buy an overkill cooler in so far as doing so won't be counterproductive for noise under lower loads. Wasted expense on extra headroom is no concern (if anything, I'm saving money by cutting deliberation time and being able to get back to work faster). The only concern is that a totally beefed up cooler made with way heavier OCs and loads in mind could actually be noisier in my typical use than a more modest product, netting me with additional decibels and not just some wasted money.

 

Heck, I might as well increase the budget further (more than I paid for the CPU) to avoid any more time loss as jobs are coming in like there's no tomorrow and I keep watching tests on YouTube, unable to focus on work — I'm basically losing work time splitting hairs over a CPU cooler. I just don't want to go into the high hundreds building a custom loop — for more reasons than just the money, but if there existed, let's say, a dead silent AIO (like passive 350W TDP) priced at $500, I'd be jumping at it probably.

 

 

 

[narrdarr]

Well give it to you straight top air coolers for you are going to be 

Ninja 5

Le grand macho

D 15

Silver arrow ib extreme

 

 Honorable mention

assassin 3

 

Since the silver arrow hasnt been mentioned. It been going to head to head and has been trading blows with the d15 for years.

 

 

[freeagent]

LGMRT, SA IBE are both rated at 320w, TS140P is rated for 360w. Probably won’t mean much to you since many are like well TDP is meaningless. D15 is nice, but it blocks ram, so to me it’s not that nice. Unless you just use one set of ram then it shouldn’t make a difference either way.
 

When you start looking at top of the line coolers, you can’t really go wrong.

[NewbieOne]

I can sacrifice the first RAM slot, but I'd rather keep the second one for Dual Channel. And indeed some coolers collide even with the second one when using a 14 cm fan.

 

I do know all of those coolers are great and you can't normally go wrong with any of them if you just want a good cooler, but I end up obsessing about two things:

 

1. The extra quietness you could possibly miss out on, like getting 29 dbA where D15 would get 30.

2. Gimping yourself for high gaming loads when you choose too light-weight of a cooler.

 

(And perhaps 3. Sacrificing too much performance by skimping on CPU rpms where case fans have to be something cheaper and running at higher RPMs anyway. I won't be decking the whole case out with like 10 $30–40 Noctua fans.)

 

Obviously, just getting a D15 would still produce a way better than decent result, as would any other cooler mentioned here, including the 360 AIO, but the optimizing obsession is still there.

 

Edit: Is Ninja 4 noticeably weaker than 5?

 

Edit2: Let's take a look at these two TechPowerUp reviews: LGMRT (tested with a 6700K) and Ninja 5[/url] (tested with an 8700K). From what I see LGMRT is 1 dbA louder at max RPM than both Ninja and True (38 vs 39). At 50% RPM Ninja and LGM tie at 35 dBA, while True wins with 33 dBA. And at 25% True produces 31, LGM 32 and Ninja 33. From what I know at this level a 1–2 dBA difference can mean being twice louder to human hearing. Now, dBA at %% rpm isn't telling the whole story without the full curve, so how indicative is this? And next question: while D15 ends up being a bit louder on paper (which could mean much louder in real life), it still has enough headroom with the temps to perhaps set it to a more silence-focused curve. But, I suppose setting the fans below 25% could fail to provide enough dissipation unless D15 is good at near-passive heatsink performance. Hence, Ninja, Macho and True could all be better performers than D15 for my purposes.

 

On the basis of these two reviews, among Macho, Ninja and True, which one would you guys suggest?

 

Theoretically, True is the highest rated (360) of the three, though on the basis of temps from tests I doubt it can hold more than Macho (not sure about Ninja).

 

Edit 3: Another idea: I can get those 800-capped Flexes for $10 apiece and put them on my Ultra or Phantom.

 

Edit 4: Anyone familiar with Thermalright Archon IB-E X2? I like the looks of it, but it costs $100 for what looks like Ultra 140 with +1 fan -10% TDP and +50% price. Though I almost fell in love at first sight. Those fans are available for like $17 apiece. Could mount them on my own Ultra.

 

Edit 5: Mugen Max, possibly based on Mugen 4 — 14 cm, single fan — GlideStream rather than Kaze. Any thoughts?