TickTockBoom :
1) According to many reviews, there are lots of problems with the CPU AIO you suggested!! Are you sure the Corsair 115i is bad because it seems to be popular over the internet with many great reviews?
Who is the you and which AIO are you talking about ? Much of PC component selection is based upon "impressing the friends" rather than the careful performance analysis. That fact that folks are spending 50% mire for CLCs that can't cool as well and are 12 times as loud as cheaper air coolers is kinda amazing. If you're referring to the Swiftech, please document / include links to these many reviews.
The Swiftch units are assembled from pre-existing custom water loop components that have garnered a level of success and proven performance for years. The laws of physics and thermodynamics are not "bendable" and not subject to debate
-Copper moves heat better than aluminum
-The H110i has been measured at 0.11 gpm... the Swiftcej over 10 times as much
-Mixing copper block and aluminum rad results in the creation of a galvanic corrosion cell
https://martinsliquidlab.wordpress.com/2012/01/24/corrosion-explored/
-Corrosion inhibitors have a shelf life, that's why we change the coolant in our car radiators
-The results show that the Swiftech 2 x 120mm unit outperforms and is quieter than 3 x 120mm units
-Water cooling can not cool a GFX card VRM if they don't put a water block on them
2) why will exhausting fans on a CPU AIO make the CPU hotter. I thought that if air was being taken in, the heat from the radiator would be carried into the case causing a hotter CPU, motherboard and GPU!!
You thought wrong. Before we get to the thermodynamics, I think the better question is why does Corsair, the folks who make the cooler you suggested direct you to install them as intakes:
http://www.corsair.com/~/media/corsair/download-files/manuals/49-000175_rev_ab_h100i_qsg_web.pdf
Page 3, 1st paragraph:
For the best cooling performance, we recommend mounting the fans as an air-intake into your case
Cooling performance is based upon what i called "Delta T", the difference between the air passing thru the radiator and the coolant temp inside it. Lets say you just boiled an egg, which would cool it faster .... a bowl of 20C water or a bowl of 40C water ?
Each cooling system is cable of managing a specific Delta T ... lets assume 15C
a) If you use outside the case (ambient) air of 23C, the system will then be able to deliver a coolant temp of 38C
b) If you use inside the case (preheated) air of 28C, the system will then be able to deliver a coolant temp of 43C
No the CPU will not get hotter from the interior case air because the CPU is not exposed to it. It's temperature will be determined by the cooling capability of the copper block, the rate at which the water moves thru the copper block and the temperature of the coolant. 23C ambient air will give you 43C coolant, 28C interior air 38C coolant.
You have a CPU cooler on your shopping list ... Is there a HD cooler ? A MoBo cooler ? There reason there is not is because you have recognized that, unlike the CPU, the performance of your system will not be impacted by 2-5C of increased case temps on any of those devices.
3) I am aware of the custom loop cooling options, but I do not feel it is necessary and it would require more maintenance and if there is a leakage, usually I would not have time to put it right and I need a pc for uni work most days. DO you really think a GPU AIO is that bad since I have heard the three I have listed are great and enough to get temps <40ºC
Hightechlegion reported the Swiftech AIO as the easiest AIO thay had ever installed. If you looked at the video, connecting the prefilled water block into the system is no more difficult than connecting a Molex Cable.
What evidence do you have to support the maintenance conclusion ?
a) Seen anything like this ?
http://www.tomshardware.com/forum/281843-29-corsair-exploded
b) If the system develops a leak, you can not fix a CLC, can't tighten a connection or do anything. An OLC is easily fixed by tightening the connection.
c) In both cases we have a liquid coolant running thru a system with metal parts. Any contact between metals and coolant has the potential for corrosion based upon the galvanic potential.
https://martinsliquidlab.wordpress.com/2012/01/24/corrosion-explored/
c.1) The CLC with aluminum radiator and copper block has a galvanic potential of 0.40 ... the Swiftech has a galvanic potential of 0.05. The rate of corrosion which will occur in the CLC is there for 8 times what it is in the Swiftech.
c.2) Corrosion inhibitors have a useful life of about 12-18 months. You can not augment what was originally supplied with the CLC. Once the originally supplied inhibitors loses its effectiveness, corrosion will occur. Asetek and the folks they supply to, are betting that you build anew system before the metal corrodes enough for the system to fail.... in the meantime, cooling effectiveness is lost as crud builds up which cuts back flow and serves as an insulating leter on the block reducing its effectiveness.
So what are we left with ... a choice between a system which has a large corrosion potential that has no means to augment the corrosion inhibitors when the lose their effectiveness after 18 months.... or a system with just 12% of the corrosion potential which does allow the addition of corrosion inhibitors (3 minutes of effort) if you choose to.
d) The data shows that hybrid CLCs don't accomplish anything. Are any air cooled 10xx series AIB cards throttling ? Have there been any reported instances of GPU temps / VRM temps inhibiting the performance of any AIB 10xx cards (other than the EVGA SC / FTW which had VRMs hitting 114C before the fix ?
As far as "reviews go", it takes some effort to distinguish reputable sites from fluff sites that just reprint press releases and run a few benchmarks. Also remember that web sites are supported by advertising ... and many sites have a policy about not getting too harsh on any brand or product. The fact remains that when tested w/o limiting factor which could skew the results, air cooling remains the superior alternative to CLC type water cooling. It's performance to noise ration is mich higher and it's performance per dollar ratio is higher.