I expected more out all in one coolers

Chestnutphantom · 23 Mar 2017 at 10:09

So I upgraded yesterday from a hyper 212 to a Deepcool Captain 240 EX Red and Black, I was expecting a much greater temperature difference as I only seen about a 10 degree difference in temps and that was before I let it sit there for a while for it to level out. So I am starting to think of delidding my E3 1230V5 Xeon as I am thinking the bottleneck is the thermal paste on the die. Any suggestions on the matter, I do have to further testing as I didn't get much time to test yesterday, but these were my first impressions.

danny77 · 23 Mar 2017 at 10:34

I'd say a 10 degree drop in temps good going and i wouldn't have expected to see as much of a drop .

ubersonic · 23 Mar 2017 at 10:35

danny77 said:
I'd say a 10 degree drop in temps good going and i wouldn't have expected to see as much as a drop .

And I would agree.

Glanza · 23 Mar 2017 at 12:26

10 degree is a good drop without going into a custom loop.

ttaskmaster · 23 Mar 2017 at 13:42

Depends what temperatures you're currently getting. 10ºC is a good drop, though.
You won't ever be able to get below ambient temperature anyway, without chucking a load of LN2 at it... so a quick, easy and cheap (potentially free) solution may be to just open the windows for a bit and drop your room temp.

Bluntwrapped · 23 Mar 2017 at 13:53

I definitely wouldn't delid it. By the sound of things your new AIO is doing the job nicely.

I remember reading somewhere delidding x99 only makes sense if you run the chip bare die. At that point you're almost guaranteed to damage pins in the socket. With the price of x99 boards it's just not worth the risk.

I delidded my 3770k a few years back and tried bare die before remounting the IHS. The results were really awful and I ended up damaging the board in the process

Brizzles · 23 Mar 2017 at 14:19

What are you current temps and what were they on air ?

DONOHUE07 · 23 Mar 2017 at 14:27

What were you expecting out of a matter of interest?

Legend · 23 Mar 2017 at 16:24

As mentioned, I'm not sure what you were expecting here. It takes 5 minutes of research to see that top end air coolers are on an equal footing with AIOs. Just because it has liquid in it doesn't make it a magical super powered cooling solution. Yes you will see a bump over a Hyper 212, but 10 deg is more than even I would expect tbh, so you're doing well to get that.

Danny75 · 23 Mar 2017 at 16:38

It can depend on the max temps. Just as an example, say you were getting 70C with Hyper 212 and 60C with 240mm AIO. But then you raise voltages, and the Hyper 212 is less able to deal with more heat and goes up to 80C whereas the 240mm AIO might better cope with extra heat and only go up to 65C, and make a 15C temp difference. In other words the temperature difference can become less or more pronounced, depending on the heat you put through it.

Chestnutphantom · 23 Mar 2017 at 16:56

Thanks for the realistic point view, I dunno I just expect temps to be I dunno about 20 degrees lower best case just be cause it's a bigger investment. By the way is MX 4 still one of the better thermal pastes, or is it worth applying a different thermal paste?

ttaskmaster · 23 Mar 2017 at 17:27

Chestnutphantom said:
Thanks for the realistic point view, I dunno I just expect temps to be I dunno about 20 degrees lower best case just be cause it's a bigger investment.

They can only ever go as low as the temperature of the air that passes through the rad, really and even then it's usually a few ºC above.
I really liked my AIO and it did me proud right up to when I had to get a custom loop.... because the GPU waterblock was only £40... would have been stupid not to, at that point!

Chestnutphantom said:
By the way is MX 4 still one of the better thermal pastes, or is it worth applying a different thermal paste?

I think it is, yes.
IC Diamond is really rated by some, along with Gelid and the rather awesome Thermal Grizzly!

Rroff · 23 Mar 2017 at 17:31

The advantage AIOs have is that they can often be run on quieter fans while keeping temperatures down - a big tower heatsink usually needs faster fans to get the same performance. It isn't a straightforward story though as I've found things like pump whine can be bad sometimes on AIOs :|

ttaskmaster said:
and the rather awesome Thermal Grizzly

The only one that I've noticed much difference from plain old MX4 or AS5 is the Thermal Grizzly and even then it often isn't earth shatteringly different. When people see bigger differences between thermal materials it is usually due to how specific the application method has to be to get best results - some are more tolerant than others.

ttaskmaster · 23 Mar 2017 at 17:46

Rroff said:
It isn't a straightforward story though as I've found things like pump whine can be bad sometimes on AIOs :|

Ja.
Mine wasn't bad, but you noticed it.
I find they're better in smaller cases, as there's less resonance and vibration.

Rroff said:
The only one that I've noticed much difference from plain old MX4 or AS5 is the Thermal Grizzly and even then it often isn't earth shatteringly different.

Grizzly really likes being thoroughly warmed up pre-application, though,. It's really thick stuff and not easy to spread.
I think it mostly comes down to personal preferences.
Not a fan of Arctic Silver myself, though.

Rroff · 23 Mar 2017 at 17:49

ttaskmaster said:
I find they're better in smaller cases, as there's less resonance and vibration.

An interesting point - my recent experiences are with the 540 Air case which is kind of cavernous and pump noise was pretty terrible.

ttaskmaster · 23 Mar 2017 at 18:03

Rroff said:
An interesting point - my recent experiences are with the 540 Air case which is kind of cavernous and pump noise was pretty terrible.

All this stuff comes with the caveat that your own experiences may differ, depending on your exact, particular combination of kit...

Just last night (and into the early hours of this morning) I was helping a friend swap his fans around. Enthoo Primo with XSPC and EKWB 360 rads, one high FPI, one very low - Running Vardar 4 120ERs on both, the fans were noisy as heck... swapped them to another (I can't recall, but likely XSPC) rad in his 350D-housed HTPC and they were almost silent as the grave. None of this is cheap or crap kit, but it's just the particular combination.

Chestnutphantom · 24 Mar 2017 at 09:06

So I tightened the mounting plate for the all in and temperatures are much better as I can run my CPU at 1.4 volts or over and not go into 90 degrees celsius, but instead around low eighties and high seventies as the fans ramp up and down when they hit 80.

chef-borjan · 24 Mar 2017 at 11:33

I was heavily considering getting an AIO, as I have just upgraded from a 4690K to a 4790K but still have a evo 212. Important to note that I have an ITX case.

Always better to see what you are playing with first, so I installed the CPU, OC'ed to 4.6ghz at 1.28V and fired up some 3D mark benchmarks to see how it handled things.

Pleasantly surprised. Need to do more extended testing, but during the CPU test at max load it was only about the upper 60s in temp! I was sure it was going to be much much worse from everything I have read on the net.

A couple of games of overwatch later (where it's really nice not to see 4 cores maxed out but instead the load balanced across 8.. usuage maybe around 50% each?), the temps just about reached 60.

Part of me feels that I need to check that the sensor data isn't bugged. Haha.

But if this is real I am very happy. Sure my fan profiles are quite aggressive but I play with headphones. No need to splash even more cash on an AIO!