Corsair H115i unimpressive vs powerful air cooler

[Kirby Wurm]

I recently upgraded to an 8700K with ASRock Z370 Extreme 4 motherboard and 16GB 3200MHz CL16 DDR. Using my venerable Thermalright Archon SB-E X2 I was able to get a stable 4.8Ghz OC using 1.25v for which the temps didn't exceed 79c albeit with a fair amount of noise. As such I opted to buy a H115i to try and reduce noise and possibly attempt a 5GHz OC.

I can't help but wonder whether I've done something wrong during setup as right now it seems to perform worse than the Archon. In the balanced profile it very quickly ramps up to an audible 1500RPM (no louder than the HSF but similar) but the temps are worse, up to 86c. I wasn't expecting miracles but surely this is high? Have I applied too much thermal paste (Noctua NT H1 in an X pattern)? Any advice would be much appreciated.

 

[Bluntwrapped]

The Corsair backplate bracket can be quite loose depending on which board you have. The instructions for my H80i v2 don’t mention it, but in the box you’ll probably find some rubber washers. These are intended to fit over the treaded legs of the bracket to give a snug fit against the motherboard when the block is screwed down tight. Without these there may be a bit of movement and probably not enough mounting pressure.

That said, I had this issue, stripped the system down, rebuilt and while it did help a fair bit.. I’m still finding the H80i v2 to be a lot of crap and not at all impressive.

Too much thermal paste won’t ever cause a performance issue as the mounting pressure of the cooler will force the excess out.

 

[doyll]

Yeah, there is likely something not quite right in your mounting. But Archon SB-E are very good coolers, I think as good as H115. Reason many reviews show CLCs with lower temps is because they test in cases and don't optimize airflow so air coolers are ingesting warmer air then CLCs are .. and every degree warmer the air is translated into about a degree hotter CPU is. I only use a rice size dob of TIM in middle of CPU. Only area of IHS that transfers heat is area over CPU chip, maybe a couple mm more. The rest of IHS size is to spread the mounting pressure out to edges of CPU PCB so it does not deform / flex when cooler is mounted. IHS is poorly named, becauset it is not really an 'integrated heat spreader' at all. It is a mounting load spreader. It is too thin to radiate heat horizontally more than a couple mm.

I suggest returning the H115. If you want to push higher overclocks look at AIOs that are not CLC with higher flow pump and 280 or 360 copper radiator like be quiet! Silent Loop, Fractal Design Kelvin and Alphacool Eisbaer & Eisbaer LT. But even then you will only gain a couple degrees lower temps at same noise levels. With AIO fans at full speed you would see 3-4c, maybe 5c, but it will be 4-5dB more which is about 1.5 times louder to our ears. Even better are Swiftech H240 X2 or H320 X2, but I would wait because they are just about to release their new X3 series. Swiftech pump is rated 550 L/h while all the other AIO above are rated 62 L/h, and more flow does make a difference. Swiftech also has better built quality.

But all of these AIO and CLCs, even custom loops require more maintenance and with more moving parts have a higher chance of problems than air cooling.

 

[Ross Thomson]

most reviews are done on test benches and no one runs their AIO fans full speed. mine run around 600rpm until I start rendering then they speed up to about 1000rpm.

 

[doyll]

most reviews are done on test benches and no one runs their AIO fans full speed. mine run around 600rpm until I start rendering then they speed up to about 1000rpm.

Maybe most of the reviews you look at are done on test benches, but at least have of all testing is done in case built systems and maybe 15% are monitoring air temp going into cooler / radiator in real time like they monitor CPU temp. Air coolers even on bench builds are generally receiving warmer air than CLC radiators simply because air cooler iss mounted on motherboard with heat radiating off other components warming the air coming to it while a radiator is usually hanging off to one side of motherboard. We all know warmer on and close to a functioning test bench than elsewhere in the room and checking room ambient temp somewhere else in the room before and after test run does not give us the 2-5c warmer it is right above the motherboard during a stress test run.

 

[Kirby Wurm]

Thanks for the suggestions and advice. Looking around at various user results mine do seem to fall in line with what many experience. I reset to stock and ran Prime95 and got high 20's/30c at idle and around 70c at load. When gaming it doesn't ramp up too high and is very usable with the quiet profile even at 4.8GHz. I've decided to do a delid on the CPU just to lower temps and hopefully voltage. I'll also check the mounting when I take the CPU out but it seems solid enough. There don't seem to be any rubber washers in the box but the backplate looks snug enough.

 

[Ross Thomson]

Maybe most of the reviews you look at are done on test benches, but at least have of all testing is done in case built systems and maybe 15% are monitoring air temp going into cooler / radiator in real time like they monitor CPU temp. Air coolers even on bench builds are generally receiving warmer air than CLC radiators simply because air cooler iss mounted on motherboard with heat radiating off other components warming the air coming to it while a radiator is usually hanging off to one side of motherboard. We all know warmer on and close to a functioning test bench than elsewhere in the room and checking room ambient temp somewhere else in the room before and after test run does not give us the 2-5c warmer it is right above the motherboard during a stress test run.

There always has to be an excuse for why air coolers are not as good as an AIO. Its quite pathetic really.

 

[buttkinz]

iv stopped people getting AIO's, theyre all pretty useless compared to a noctua NH-D15 air cooler, by the time you get a AIO thats as good as the noctua you might as well get a custom loop >< the do look good with RGB bling but YHMMV

 

[doyll]

There always has to be an excuse for why air coolers are not as good as an AIO. Its quite pathetic really.

What is pathetic is people saying reality is untruth because they don't understand the truth of how things work having no understanding of how to do accurate testing but instead believing all the shills and Carny barkers telling them CLCs is the way to go. CLCs are poorly build, over-priced, not near as dependable, don't cool any better, and when they dies (99.999% pump failure) have no cooling until replaced. to put down others who do know better instead of learning the truth. Just look at how many X-CLC people there now are like buttkinz who say basically the same things I've been saying for 7 years here .. that air cooling is definitely better when noise, cost and life expectancy are all taken into account and that much of the review testing is garbage. Look at how many review testing sites have changed from case testing to open bench and monitor air temp going us into cooler/radiator usually on open bench versus how many did 7 years ago when I was preaching the same things I am now.

 

[Ross Thomson]

do it in a case and you are limited by the fans in the case which you would have to test as bought against the AIO drawing cool air in from the front. on a test bench the fan is drawing hot air in from the mobo against the AIO being away drawing cool air in. sorry but that's the point of the AIO. it doesn't need to be next to the mobo or require a fan pushing air onto the fans to supply cold air.

I didn't have a desktop 7 years ago and I don't care about results from 7 years ago either, 7 years ago is irrelevant. See I have nothing against air coolers and would/have recommended various ones for people that don't want water cooling. Not everyone has a case that can fit a massive heatsink or what the visuals of one. No good recommending a noctua NH-ND15 if it don't fit the case or clear the ram.

The fact is you don't know better and people do actually see that. Air is not better than water and vice versa as each persons requirement is different. A 911 is not a better car than my Octavia as it doesn't perform the same. while I agree they are expensive compared to an air cooler (although mine was only £20 more than a NH-ND15) they are not poorly built. If they were they would be failing left right and centre which they are not.

 

[doyll]

Or do it in a case that has airflow tuned / optimized for the system in it.

Or test the AIO/CLC with radiator as exhaust which is a more logical way to do it considering everything in the case is air cooled except what the CLC is on. But even that is not optimizing case airflow for system in it. The point of AIO is to cool one component. There is no 'point' about it not using the same temperature air as an air cooler would. In fact the 'point' of testing them with same temperature air flowing through them is so we know the actual cooling ability of one versus the other and not the falsely inflated numbers resulting from different air temp going into them. But more to the 'point' why would it be good to use a CLC as intake heating the air going into a system with everything being air cooled except what CLC is cooling? That is totally illogical.

Not having a desktop 7 years makes no difference, nor does your not caring about results back then. True not all cases support all coolers. But the lower cost, more dependablity and lower noise of air coolers makes them the logical choice in most cases (no pun). If someone wants to use a CLC or hopefully an AIO that is not a CLC that is their choice. All I'm trying to do is educate the consumer as best I can before they make their choice. I don't think I said to get D15, but D15 will clear almost all RAM and many cases.

Actually from what you are posting here I do know a lot more about coolers than you do .. and people do actually see that. No idea how cars got into this discussion. A good air cooler is generally a better choice than a CLC, but there are rare exceptions. Air coolers being better than CLCs is one thing. What a user chooses to buy is another. Much of the advertising, review testing, and other marketing is geared to draw the buyer into CLCs, and the reality is they are not the better choice. The cost more, generally make more noise (pump noise is always there), don't last as long (and when they die there is no cooling until replaced .. not cheap) versus an air cooler that only thing that can go bad is the fan, which usually give plenty of warming by making noise, system will still work at low load or with whatever fan user has held on with rubber-bands until suitible repalcement is in hand .. and new fan is a frction the cost of new CLC.

£20 more than NH-D15 which at £84.95 is most expensive air cooler in OcUK's store (with PhononicHEX 2.0 being same price and not near as good a cooler). Genesis and D15 SE-AM4 are close and Dark Rock Pro 4 is £79.99. All the others are sub £75. There are now several CLCs in this same price range, but again not as good as the air coolers for all the already listed reasons.

Sorry, I got personal in this post, but your saying I don't know what Ii was talking about when it is so obvious you are the one who doesn't kinda left me no choise.

 

[Ross Thomson]

Or do it in a case that has airflow tuned / optimized for the system in it.

but that's not how you buy something or what people have. you are introducing factors that people don't do. you'll be buying more fans therefore more adding more cost plus you can't test in all cases.

£20 more than NH-D15 which at £84.95 is most expensive air cooler in OcUK's store (with PhononicHEX 2.0 being same price and not near as good a cooler). Genesis and D15 SE-AM4 are close and Dark Rock Pro 4 is £79.99. All the others are sub £75. There are now several CLCs in this same price range, but again not as good as the air coolers for all the already listed reasons.

only £5.05 more than the MH-D15 then which doesn't even fit in my case or cools are well

Sorry, I got personal in this post, but your saying I don't know what Ii was talking about when it is so obvious you are the one who doesn't kinda left me no choise.

I didn't read it and you clearly don't know what you are talking about given your comments.[/QUOTE]

 

[AiiR]

I've used liquid coolers since the original Corsair H50, only ever had an issue with a 2nd hand H55 which was my own fault for putting too much stress on the tubing. Currently running an NZXT X62 and with the fans at 400-800RPM the temperature on a delidded 5.2GHz 7700k never exceeds 80c under full load, and is silent.

The issue you have is probably either to do with mounting/thermal paste application or the pump speed in the bios. For best results it should be full speed but might be a bit noisy so go as high as possible until it becomes too loud. You should be able to achieve good temps with the fans at very low speeds.

 

[doyll]

but that's not how you buy something or what people have. you are introducing factors that people don't do. you'll be buying more fans therefore more adding more cost plus you can't test in all cases.

only £5.05 more than the MH-D15 then which doesn't even fit in my case or cools are well

I didn't read it and you clearly don't know what you are talking about given your comments.

People I know buy a case, motherboard, CPU, GPU, power supply, needed fans, then assemble and tune things to get performance, cooling and noise levels to their liking.

First you said

although mine was only £20 more than a NH-ND15

Now you say

only £5.05 more than the MH-D15

So which is the truth?

You post in reply to my posts but without even reading them? Kinda hard to even know what I said without having read it.

 

[Ross Thomson]

First you said

Now you say

So which is the truth?

You post in reply to my posts but without even reading them? Kinda hard to even know what I said without having read it.

I thought the Nh-D15 was cheaper, the cost of my AIO didn't change.

I didn't read your personal attack(s) which I clearly stated in my post.

 

[doyll]

I thought the Nh-D15 was cheaper, the cost of my AIO didn't change.

I didn't read your personal attack(s) which I clearly stated in my post.

I didn't attack, only defended myself against your false statements by pointing out how your posts show off little you actually know.

 

[Bluntwrapped]

Having had my rig under custom water, aesthetics were my main focus when I decided to buy the H80iv2. Some great air coolers out there but the giant block looks pretty hellish when the rest of the build is kept to a minimum.

I’m half tempted to try and replace the radiator with a black ice rad, just to see how much potential has been wasted in crap design.

 

[Kirby Wurm]

Well this went a little off topic!

I've used liquid coolers since the original Corsair H50, only ever had an issue with a 2nd hand H55 which was my own fault for putting too much stress on the tubing. Currently running an NZXT X62 and with the fans at 400-800RPM the temperature on a delidded 5.2GHz 7700k never exceeds 80c under full load, and is silent.

The issue you have is probably either to do with mounting/thermal paste application or the pump speed in the bios. For best results it should be full speed but might be a bit noisy so go as high as possible until it becomes too loud. You should be able to achieve good temps with the fans at very low speeds.

I looked at the Kraken but didn't deem the extra 25% cost worth it for a small improvement in thermals, and the Corsair fits the aesthetic of my system better. Looks an excellent piece of kit.

The pump at the faster speed (only two settings, quiet and performance) is noticeable in volume but not too intrusive. When gaming the quiet profile is okay at 4.8GHz and I imagine would be even quieter at stock, it just isn't any better than the Archon.

I'll be refitting it once my CPU is delided tomorrow and will be careful with the mounting (there may have been a bubble or two last time using the X pattern for thermal paste).

 

[doyll]

Having had my rig under custom water, aesthetics were my main focus when I decided to buy the H80iv2. Some great air coolers out there but the giant block looks pretty hellish when the rest of the build is kept to a minimum.

I’m half tempted to try and replace the radiator with a black ice rad, just to see how much potential has been wasted in crap design.

Might help, but H80 pump is only has about 40-60 L/h flowrate compared to AIO's that are nto CLC having 72-550 L/h. By comparison the D5 pump (probably the most popular custom loop pump) has 1500 L/h.

 

[Bluntwrapped]

Might help, but H80 pump is only has about 40-60 L/h flowrate compared to AIO's that are nto CLC having 72-550 L/h. By comparison the D5 pump (probably the most popular custom loop pump) has 1500 L/h.

Fairly low powered then. I suppose the pump compares more to the mini Alphacool DC-LT more than any other pump. The black ice rads have significant restriction but I’m thinking if the H100i uses the same pump with a 240mm rad, it should be powerful enough.

That said, the Chinese eBay has ID cooling cpu blocks/pump separate for DIY custom AIO’s. I may be tempted to buy one and build my own AIO just to see how it compares to the Corsair.