Disappointing Corsair H50 Performance

[pantherx12]

It should out perform it, I've a Domino ALC which I've modded so I can put fans on how ever I like.

And it out does a Xig 1283 and True 120ex by a fare ammount.


10c of max load temps and idle is pretty much ambient.

 

[phanbuey]

Turn the fan round so it blows through the rad instead.

That will knock a big ol chunk off those temps.

Other then that have you tried cleaning up and reaplying thermal paste? using as little as possible so it just about covers the CPU.

Also, touch the main block area, is it vibrating ?

Not on the h50 it won't... push fans only work better when RPM is above 1800... with fans that have less static pressure, pull is much better.

You need more CFM and/or more rad. Unfortunately a 120MM rad is just not enough for a quad, and actually has never really been. Best thin you can do is to use a shroud... and as has been said before, use "fresh" air from outside the system (ever seen those reviews that compare air coolers on a test bench vs a case and the test bench is always 5C or so better? - same thing applies).

Pull = better when static pressure of the fan is not enough to compress air through the rad.

Push = better when you have a 38MM fan at a high rpm/CFM (75+cfm) and a high static pressure...

 

[PaulieG]

The biggest problem with this cooler is the size of the rad. I'ts just too damn small and inefficient to cool the liquid enough for i7 chips. I had an old Swiftech setup that was similar to this in one of my s775 rigs, and honestly I had better temps with a TRUE. That's when I moved to custom water cooling for awhile.

 

[phanbuey]

The biggest problem with this cooler is the size of the rad. I'ts just too damn small and inefficient to cool the liquid enough for i7 chips. I had an old Swiftech setup that was similar to this in one of my 5775 rigs, and honestly I had better temps with a TRUE. That's when I moved to custom water cooling for awhile.

yeah me too... except for me it was the swiftech kit and a Q6600 at 3.6Ghz thing would load at 80C, thats when I went custom.

@ OP - alot of people think that watercooling's benefits lie in the "water's ability to cool better"...

The real benefit of watercooling is that it allows you to increase the surface area with which you can cool. It allows you to use 3 120mm fans to cool the cpu, for example... try to fit an air cooler with the same surface area on that lol.

Cooling, in this case, is mainly a function of surface area*airflow.

Unfortunately, a 120mm radiator has the same surface area as a tower cooler (if not even less) and will cool actually a bit worse since you also have the heat of the pump in the loop.

Unless you're going with a doube or a triple rad, or using watercooling only for the videocards, a single rad isn't worth it. Thats why all these cooling solutions with a single 120mm radiator get ripped apart in our forums.

Here are some ways to improve the performance of what you have now...

1. A shroud - bascially take an old fan, cut the 'fan' part out, and use it to put space between the stock fan and the rad. This has shown a temp performance increase of ~15% by skinee labs by increasing the amount of air that flows through the 'dead spot' in the rad.

2. A more powerful fan - I highly recommend something like the arctic cooling turbine xtreme or a 38MM yate loon - stay away from scythe slipstreams as they get slaughtered by rads... their cfm drops off pretty drastically. Noctua's are good as long as you guy the newest one and not the ones that have a ton of space between the blades.

3. Give the rad the coldest air you can... its hard with the H50, but see if you can put the rad somewhere where it isnt using already warmed air to try to cool your proc.

 

[TheShad0W]

^Cheers for that, Will try the shroud method because I have no shortage of spare 120mm fans, I'd already ordered the scythe slipstreams when I posted this :S, but hopefully they'll do better than the fans I have on it at the minute, and the PC already draws in cold air from the back.


Another little side-thought - how much do you think adding a basic air filter to the intake would reduce the overall flow rate through the radiator? It's the only intake fan on my system without any kind of filter at all (NZXT puts them into all the "intake" fan slots as standard...) and I don't want to take this setup apart to find it full of dust in 6 months time.

 

[phanbuey]

Another little side-thought - how much do you think adding a basic air filter to the intake would reduce the overall flow rate through the radiator? It's the only intake fan on my system without any kind of filter at all (NZXT puts them into all the "intake" fan slots as standard...) and I don't want to take this setup apart to find it full of dust in 6 months time.

Dust filters will increase your temps pretty drastically.

I just take a can of air to the case every 2 months or so (mine is acryllic now, so I can see when it starts to need it...)

 

[douglatins]

Temps look good, though i dont see ambient temps

 

[Wile E]

Not on the h50 it won't... push fans only work better when RPM is above 1800... with fans that have less static pressure, pull is much better.

You need more CFM and/or more rad. Unfortunately a 120MM rad is just not enough for a quad, and actually has never really been. Best thin you can do is to use a shroud... and as has been said before, use "fresh" air from outside the system (ever seen those reviews that compare air coolers on a test bench vs a case and the test bench is always 5C or so better? - same thing applies).

Pull = better when static pressure of the fan is not enough to compress air through the rad.

Push = better when you have a 38MM fan at a high rpm/CFM (75+cfm) and a high static pressure...

Not true. Depends on the fin density of the rad. My swiftech rads all do better with push fans, regardless of rpm. Even my old original-style BIX 360 did better with 1200rpm fans in push. Never tested with a lower speed fan to comment. Push is almost always better than pull.

 

[phanbuey]

Not true. Depends on the fin density of the rad. My swiftech rads all do better with push fans, regardless of rpm. Even my old original-style BIX 360 did better with 1200rpm fans in push. Never tested with a lower speed fan to comment. Push is almost always better than pull.

Skinnee labs tested with an XSPC rad which is not very dense at all... also did shroud and push - pull testing. All showed that the ~1600 RPM it made no difference and that below that Pull was better... Push was better above that.

their testing methodology was solid, so im inclined to believe them.

http://martin.skinneelabs.com/Radiator-Fan-Orientation-And-Shroud-Testing-Review.html

same one posted by Martin himself ...

 

[Kantastic]

A E5200 at any insane clocks cannot even match the heat given out by any Quad core CPU especially an i5

My i7 at 4.0 1.28v tops out at 70C.

 

[Wile E]

Skinee labs tested with an MCR 320 rad which is not very dense at all... also did shroud and push - pull testing. All showed that the ~1600 RPM it made no difference and that below that Pull was better... Push was better above that.

I cant find the link to the article atm bc im at work, but its the same one that they did the shroud testing - google it.

their testing methodology was solid, so im inclined to believe them.

http://martin.skinneelabs.com/Radiator-Fan-Orientation-And-Shroud-Testing-Review.html

same one posted by Martin himself ...

My methodology is just as solid. Absolutely nothing changed between setups except for fan orientation. Even ambients were the same. No blocks were removed or shifted in any way. Tubing was not moved or altered, nothing.

This is with YL fans and no shrouds, so that may make a difference as well. Push has so far always done better for me for anything 1200rpm and above.

 

[phanbuey]

My methodology is just as solid. Absolutely nothing changed between setups except for fan orientation. Even ambients were the same. No blocks were removed or shifted in any way. Tubing was not moved or altered, nothing.

This is with YL fans and no shrouds, so that may make a difference as well. Push has so far always done better for me for anything 1200rpm and above.

I mean those are some completely contradicting results. Especially with the MCR series as their FPI is relatively low. They tested with and without shrouds, and either way, a shroud didnt change whole lot.

I would love to see a your results as well (if you can post them). I personally have never noticed a massive difference either way - for me pull is easier to clean dust out of the rad, makes less fan noise, so i use that with shrouds...

Ive only ever seen one real, methodical review done to answer the question. And that was it. It basically said that push is better when the fan can overcome the restrictiveness of the rad - which makes sense.

Where that line is in terms of RPM is completely up to the FPI of the rad... i agree with that.

In any case, I would be very interested to see your numbers... ill flip the fans on my triple too - to see if that will make a difference.

 

[Wile E]

I mean those are some completely contradicting results. Especially with the MCR series as their FPI is relatively low. They tested with and without shrouds, and either way, a shroud didnt change whole lot.

I would love to see a your results as well (if you can post them). I personally have never noticed a massive difference either way - for me pull is easier to clean dust out of the rad, makes less fan noise, so i use that with shrouds...

Ive only ever seen one real, methodical review done to answer the question. And that was it. It basically said that push is better when the fan can overcome the restrictiveness of the rad - which makes sense.

Where that line is in terms of RPM is completely up to the FPI of the rad... i agree with that.

In any case, I would be very interested to see your numbers... ill flip the fans on my triple too - to see if that will make a difference.

Ahh man, I posted them a long time ago. I wouldn't even know where to begin looking for the results anymore. After a few formats, they aren't on my hard drive anymore.

But you also aren't considering setup differences between Martin's setup and the setups mentioned here. That test was done using a completely different rad, a very specific flow amount, and different fans from the setups we are commenting on in this thread. Not exactly apples to apples.

At any rate, it wasn't entirely significant. On the order of 2-3C under load. Don't think that's gonna matter much for his H50 anyway.

 

[TheShad0W]

Temps look good, though i dont see ambient temps

On mine? It's 24-26 in here usually.

 

[PP Mguire]

These temps seem about right. Its a tad lower than my true with 1 fan at 4ghz.

 

[PP Mguire]

On second observation

 

[TheShad0W]

^Try 15 minutes of IBT on a 3072MB problem size, see how it does then

If I only run a couple of passes on IBT, or run something else like prime95 on blend my core temps don't pass 80.

 

[PP Mguire]

No need to waste my time. My temps never go above that anyways when playing a game or anything else.

 

[3volvedcombat]

Is there any mods you can do to the H50 that will better the temps overall. I cant lap the base, which i would almost do and i might go buy some nice thermal paste. Is there anything better the Arctic silver 5. Piont me to it cause i would love to see a 1-3c drop because changing thermal paste did the job ;D.

I know theres better paste then AS5 but for a water cooling of some sort application the best thermal paste would be?

 

[phanbuey]

put another fan on it for push/pull...

 

[3volvedcombat]

put another fan on it for push/pull...

I have a antec 3 speed fan and the corsair fan in a P&P setup, now should i add a spacer or what, I dont have room for a spacer anyways because my rad is almost hitting the block its self. T_T

 

[PP Mguire]

Higher RPM fans work to.

 

[Tom20]

Ouch!

 

[Nick89]

the H50 is garbage compared to an actual water cooling system with a decent radiator. The single 120mm Rad is easily overwhelmed by the overclock on your i5.

Is there any mods you can do to the H50 that will better the temps overall. I cant lap the base, which i would almost do and i might go buy some nice thermal paste. Is there anything better the Arctic silver 5. Piont me to it cause i would love to see a 1-3c drop because changing thermal paste did the job ;D.

I know theres better paste then AS5 but for a water cooling of some sort application the best thermal paste would be?

AS5 is the best thermal paste. The only other thermal paste that gets the same performance is: Shin-Etsu MicroSi G751

80-way Thermal Interface Material Performance Test: http://benchmarkreviews.com/index.p...k=view&id=150&Itemid=62&limit=1&limitstart=12

 

[Kantastic]

the H50 is garbage compared to an actual water cooling system with a decent radiator. The single 120mm Rad is easily overwhelmed by the overclock on your i5.

An actual water cooling system with a decent radiator is garbage compared to liquid nitrogen. The decent rad is easily overwhelmed by the 2.0v on your overclock.

See what I did there?