If you've not yet read part one, two, three, or four, then please do take some time to have a look - they're packed full of info and explain all the jargon; we've covered the most important questions to ask before part picking, what to look for in a GPU, all the need-to-know when it comes to CPUs, and stripped back the excess from motherboards to make their specs a little less confusing too.
In today's post, we'll be tying up the loose ends by covering the rest of your system's components, from the memory (RAM) and storage options, through to your power supply, the case itself, and all the cooling to keep your hardware running at its best. These are possibly some of the easiest components to pick for a build, since their spec lists are fairly short and mostly self-explanatory. We'll go through a few of the key points for each, explain why you might prefer some options over others, and put you at ease with some of the finnicky bits.
So without further ado, let's get into the post, kicking things off with Memory...
RAM (Random Access Memory), otherwise called memory, is a form of extremely fast volatile storage and is necessary for just about every task your PC performs. Whenever you install programmes on your computer, all of the relevant files are stored on some kind of hard drive where they can be accessed and utilised by your system. However, when you launch an application and use the functions within it, not every single file will be required; there will be a number of common "core" files which are used in a multitude of situations, various files used only as and when they are requested, and sometimes certain operations will actually create additional temporary data designed to be deleted after use. This is where RAM comes into play. These files are loaded into your system's RAM, which is essentially copy-pasting that data for quick access by other hardware such as the CPU.
So, rather than retaining files and data indefinitely, RAM only retains data temporarily, allowing other components and software to quickly access and utilise that data for whichever particular task you may be performing, whether that's gaming, photo-editing, browsing online or checking emails. This data is then "forgotten" when you close whichever applications you were using, or when you shut down the PC.
As such, how much RAM you need will often come down to the types of programmes you use, and how much multitasking you do. For the vast majority of users, 16GB of DDR4 RAM is usually plenty enough to deal with large games, many editing softwares and plenty of multitasking (think several tabs open in your browser, and maybe 4-5 other applications running simultaneously). But if you regularly use multiple editing softwares concurrently, with particularly large files like 8K raw video footage, as well as keeping tens of tabs open in your browser (particularly Chrome), and keep a heavily modded videogame running in the background at all times, then you may well be a candidate for 32GB, or even 64GB.
RAM works most efficiently in pairs which is why you'll often see kits of two and four modules. It also comes in a variety of speeds and timings that further affect the speed and performance of the memory. These are generally of less concern for the average user though, as for the most part, the overall performance gain in a system isn't hugely affected by these specs.
Volatile Storage - Simply refers to storage that does not retain stored data once it loses power.
Speed/Timings - The speed of RAM is measured in MHz and simply denotes how quickly the module can perform a given task, meaning higher numbers equate to a faster memory module. Meanwhile, timings refer to the latency between the RAM chips as it performs said tasks, with lower numbers indicating less latency and thus higher responsiveness. But with how fast RAM is by design, these differences will be largely imperceptible in overall system performance by the average user. However, there is one instance where it is worth paying attention to RAM speed; AMD's Zen-based Ryzen CPUs have an "ideal" RAM speed (see part 3) which allows the processor to perform at its best. Although going higher than this speed will not yield any noticeable results, going lower can affect how fast the processor can perform certain functions, artificially limiting its full capabilities.
DDR* (Double Data Rate) - All RAM will be prefaced, or list in its specs, its DDR rating. At the time of writing, the most commonly available are DDR3 and DDR4, with the latter being the default option for the majority of enthusiast builders (on account of its greater speeds).
XMP (eXtreme Memory Profile) - More often than not, when you purchase almost any RAM module, that module will run at a base frequency, or "profile," which limits the memory to speeds below those that are advertised on its product page. When you install these modules for the first time, they will operate at this lower speed by default. This is because technically speaking, the faster speed advertised is actually a verified overclock for those modules. As such, it is important to always enable the XMP setting in your system's BIOS, in order to actually take advantage of that higher frequency.
Where this is located will vary between motherboards, as each manufacturer has their own UEFI design for their BIOS. But essentially, you just need to find the tab or menu containing the options for your memory. For example, this is where you would find the XMP option in Gigabyte motherboards.
As you can see, the XMP setting is disabled by default, and for this particular system, that means the memory is only running at 2133MHz. To change this, we would simply click the drop down and select the first available profile. The listed 2133MHz would then change to reflect this, and would instead list whatever the maximum speed of the memory purchased is. All that remains to do after, is save and exit from the BIOS.
Though not explicitly a RAM spec per-se, and certainly not something high up in most people's priorities, the physical height of a RAM module is something that may need to be considered depending on which CPU cooler you are intending to use. Some larger coolers, and especially those that utilise multiple fans too, will overhang the DIMM slots on your motherboard and so taller, bulkier RAM designs, such as the T-Force Nighthawk modules, won't be able to fit under the cooler.
Generally speaking, this shouldn't be an issue, but it's definitely worth looking into if you're particularly interested in some of those wackier designs and non-AIO CPU cooling options. If you aren't entirely sure, check whether the RAM kit lists the modules as being "low-profile." If it does, then these modules should fit under just about any cooler that will overhang your DIMM slots. And of course, when in doubt, you can always check user forums and Reddit - usually if you aren't sure if something will fit, someone else has had the same doubts and asked the question online. Nine times out of ten, someone will have confirmed whether there's any issues with clearance.
Depending on where you look, you'll often find heated debates about whether it's better to have two or four sticks of RAM installed in your system, and to complicate matters even further, whether those modules should be dual rank or single rank. Some will insist that two is the ideal configuration, whilst others will preach that four is best. Well, actually, you don't need to worry - it's fairly irrelevant. Many people have started testing and benchmarking with dual and quad channel set-ups to see whether there's any significant difference in overall system performance and the findings are fairly unanimous: there's little to no difference whatsoever. The biggest hit to performance you'll ever see when it comes to RAM is when you use only a single stick, instead of a pair.
As such, personally I'll always recommend opting for dual channel kits as they tend to be cheaper than their quad channel kit counterparts. If you're interested to learn a little more, then you can check out the video below from Hardware Canucks, but fair warning, it is very in-depth. It does, however, perfectly demonstrate how minor the differences are across the board, so even just taking a quick look at their benchmarks will tell you all you need to know. And it's also worth bearing in mind that the gaming benchmarks are very exaggerated, on account of being very well-optimised esports titles, running at 1080p, meaning higher resolutions and different game genres will likely have even more minimal differences than those in these tests.
When it comes to storage for your system, there's only really three things you need to know; how fast is it, how much can it store, and how much does it cost? This makes it extremely easy to choose for the most part, because in the simplest sense, you'll just be choosing between M.2 Storage (either SATA or NVMe), 2.5" SATA SSDs and 3.5" SATA HDDs.
There are also some use cases for each type of storage and it is fairly common to see a combination of these three in many desktop solutions, which is usually done to maximise the speed, capacity and cost of the system, so we'll cover these too.
SATA - Serial Advanced Technology Attachment. Simply refers to a type of interface used by a number of devices. In this case, hard drives.
3.5" HDDs (Hard Disk Drives) - Also known as Mechanical Drives, these are the oldest and slowest of the three storage types. This is due to how they work; data is written and read by moving mechanical parts, consisting of an arm for reading and writing data, and spinning magnetic platters where said data is stored and accessed. Of course, by its very nature, this means that the speed of a hard disk drive is limited to how fast that arm and those platters can move, with higher RPMs (Revolutions Per Minute) meaning more responsive drives. The most common RPMs for modern HDDs are 5400 and 7200, with read/write speeds increasing fairly linearly - the latter usually offers around 20% faster speeds than the former - and without much of an increase to cost.
Although the slowest storage type, they are however the most affordable, meaning it's very easy to buy several TBs of storage, giving you more room than you could ever want for storing your data. This makes them great for archiving, which is certainly of benefit to those that work with a lot of editing programmes, and who need somewhere to keep finished projects around as back-ups. Though they do have the highest rate of failure on account of those moving parts (still a fairly low percentage chance, with single figures in the first 3 years, and only about a 10% chance thereafter), they are however often still salvageable, meaning they're a great option for those with particularly important files that they can't afford to lose.
However, when it comes to your operating system, games and applications, you'll really notice those slower speeds, and start to feel like you spend all day waiting for everything to load...
2.5" SSDs (Solid State Drives) - SSDs were historically a more expensive option but for good reason. Although still connected to a system via a SATA cable, they're extremely fast compared to HDDs, averaging around 3-4x faster read/write speeds. They are quieter and generally more reliable, since they use flash memory meaning there's no mechanical parts involved, and are significantly smaller making them easier to integrate into new or existing systems (2.5" HDDs do exist, but they are a rarer and often marginally more expensive breed, usually used in laptops). Over the last few years, these improvements have continued to widen the performance gap between the two and with prices falling all the while, it's now almost as affordable to go with a mid-range SSD as it is with a high-end HDD up to around the 1TB mark, making it fairly senseless to opt for an HDD in a system using 1TB of storage, or less. And for most people, especially gamers who don't mind occasionally uninstalling and reinstalling the odd few games that require massive amounts of storage, 1TB will be more than enough.
But if you aren't convinced, allow me to help put things into perspective. My current set-up includes two 1TB SSDs because I am an incredibly lazy gamer and hate managing storage; even with 60 games installed concurrently, several of which are larger titles like Fallout 4, XCOM 2 and The Witcher 3 (plus all of their respective DLCs and over 300 installed mods combined), I still have 650GB of free storage remaining - that's two-thirds of one of those 1TB drives! So if you're someone who actually uninstalls games they don't play anymore and don't mind a little storage management say, once every couple of years, then you'll likely struggle to ever hit that 1TB mark.
As fast as SATA SSDs are though, and certainly when it comes to load times in games, they still aren't the fastest option available, with our third type of storage being the ideal for your operating system.
M.2 SATA & NVMe SSDs - M.2 drives are very similar to SATA SSDs in that they use flash memory to store data. The key difference however, is that they are even smaller still, and do not require a SATA cable to be connected, instead mounting directly into an M.2 interface on the motherboard (check our beginner's guide to motherboards for more on this interface and compatibility). So what exactly is an M.2 SATA SSD if it isn't using SATA? Well, in this case, SATA is used in the name to refer to the fact that these particular M.2 SSDs use the same read/write speeds as their equivalent 2.5" SATA SSDs. This means that the main benefit of this form factor is that you'll be able to optimise the room you have available in a system, since you won't encroach on space below the power supply shroud with added cables, or at the very least, won't be occupying one of your 2.5" SSD mounts. You'll also technically get a slightly more responsive system, since mounting the drive on the motherboard gives more direct access to neighbouring components than the 2.5" drives connected via a cable, and as such, a minor reduction in latency - though you'd be hard pressed to really notice the difference.
However, if you wanted to gain the full benefits of the M.2 interface, then this is where M.2 NVMe SSDs come into play. NVMe M.2 options not only benefit from those points we just made, but also see a substantial increase to their read/write speeds over the standard SATA SSD counterparts, since they can take advantage of PCIe (also covered in our motherboard post) rather than SATA. This results in a form of lightning-fast storage unmatched by anything else on the consumer market today.
This makes M.2 NVMe storage an ideal option for installing an operating system on, since you'll see an impressive boost to system boot times. Of course, this isn't just limited to your operating system either; you could also use this added speed to your advantage and use M.2 NVMe storage for some of your most frequently used applications, such as a favourite game or particular editing software.
Currently, not all games will benefit from the increased speeds of an M.2 NVMe SSD, or only marginally so, since many are designed with the likes of HDDs in mind, due to the fact that most systems (including up to the last generation of consoles) would be using them. However, with the new, current generation consoles now operating on all SSD storage, and the likes of Unreal Engine 5 making its way into the hands of game developers, this could change going forward.
Understandably, you'd probably think that all of this would mean M.2 storage options are extremely expensive, but you'd be surprised - buying an equivalent mid-range M.2 SATA option to a mid-range 2.5" SSD (that is to say, an M.2 option with the same read/write speeds as the 2.5" SSD) doesn't see much of a price difference, and in some cases, can even be slightly cheaper. But even once you move into higher-end M.2 NVMe options, which feature those even faster read/write speeds - take, for example, Samsung's 870 Evo (2.5" SSD) and 970 Evo Plus (M.2 NVMe) - there's only around a 17.5% increase (£20-25 at that 1TB mark), for which you'll be getting around 4-6x faster performance than a 2.5" SSD. Not too shabby at all.
To summarise then, in an ideal world, a desktop system would benefit most from a hybrid of these storage types: M.2 NVMe for system OS and favourite applications, a 2.5" SATA SSD for most frequently used programmes and games, and one or more high capacity, 7200RPM 3.5" HDDs for storing small or infrequently used files.
Power Supply Units (PSUs) are possibly the most underrated and yet most essential part of any system. Granted, they certainly aren't nearly as exciting as a GPU or RGB, but nonetheless, they are certainly quite important. What's more, when it comes to new builders, they usually fall into two camps: those that don't worry about them at all, and those that do nothing but worry for fear of blowing up their lovely new hardware. And with both usually comes either a total underestimation or overestimation of what rating and wattage they actually need for their components.
If you want to be smart about choosing your power supply, then you should be aiming for something in the middle - by no means should you skip over the details and just hope everything will be fine, but equally, it really isn't stressful at all when you know what you're looking for. And what is that exactly? Well, there's four key things to pay attention to: Wattage, 80 Plus Rating, Modularity and Form Factor.
Wattage - This is fairly straightforward and is pretty much exactly what you'd expect - the maximum consistent power that the PSU can supply to your system. How many watts you need in a system will depend on the components you are using (surprising I know), which you can generally estimate by taking the total power draw of all of your components and then adding about 30% to that number. This is because it's difficult to know exactly how much power your components could draw at a given moment (and in particular, their peaks); it's better to have some leeway than end up on the wrong side of the cusp. It also means that if you upgrade something like your graphics card later in your PC's lifetime, then you should still have plenty of breathing room for any extra power it might need compared to your previous card.
80 Plus Rating - When you browse around for PSUs, one thing you'll see a lot of is the assortment of coloured stickers ranging from white to titanium, vaunting a very bold "80 Plus." This is a certification indicating the overall efficiency of the power supply, which guarantees a minimum 80% energy efficiency at 20%, 50% and 100% loads. It also acts as a good indicator for the overall quality of the power supply, with many manufacturers adding in a few low-cost extras to make the higher tiers more worthwhile, such as custom sleeved cables or additional mounting brackets for broader case compatibility. This can vary from manufacturer to manufacturer, so don't take it as a given. The only thing it completely guarantees is the efficiencies, which is far more important than the extras; you can check the details below if you're interested to know exact percentages for each tier.
Non-modular power supplies come with all the cables you might possibly need hardwired into the unit. This is great if you intend to use all or the majority of the attached cables, as these PSUs are generally more inexpensive than the other two options. However, most average users won't need all of these cables, and so the downside becomes the cable management you'll have to tangle with (excuse the pun) when it comes to building your PC. In the best scenario, this is a minor inconvenience and could result in a slightly messier looking interior (if you can actually see inside your case to begin with), and worst-case scenario, the cables restrict the airflow design of your build (more on this later) and become a real hotbed for dust.
Semi-modular, as the name would suggest, are power supplies offering a middle ground between non and fully modular designs. Featuring only the required power cables hardwired in (those being the power for your motherboard and CPU), any remaining cables that might be needed can be added where necessary, such as those used for a graphics card. This makes them marginally more expensive than non-modular, but still relatively cheap when compared to fully modular. They're also perfect for gamers and builders planning to start with a smaller budget, since you can design a system using integrated graphics and use a semi-modular power supply - this will give your system an affordable and functional starting point, with room to add-in additional components at a later date, without having to worry about the superfluous cable management that would come with the non-modular options.
Fully modular, once again, in the name of the design, are power supplies offering complete freedom with cabling, allowing you to connect only precisely what you need in your system. This allows for optimal cable management and helps keep your system free of clutter, air chokes and dust nests. You'll also be able to add or remove cables whenever necessary, which is especially useful if you ever need to RMA your power supply - just unplug the cables from the PSU and you're done; no faffing around undoing all of your meticulous cable management. It also allows you to make use of custom sleeved cables, so for those who enjoy the aesthetic side of PC building (and let's be honest, that's a lot of people), it's possible to really spice up that interior with different coloured cabling. Of course, this freedom and flexibility comes with a price, but statistically, fully modular power supplies are quickly becoming the go-to choice for many PC builders, both veteran and new.
Form Factor - Much like with cases and motherboards, power supplies also come in a few form factors to allow for compatibility with differently sized systems. The most commonly used PSU size is the standard ATX12V, which will fit the vast majority of mainstream desktop systems, from full tower cases right down to the smallest midi-towers, and even a handful of larger ITX cases too. Occasionally, you may come across some smaller or quirkier designed cases which will often list "SFF" or "SFX" compatibility only. SFF/SFX simply stands for Small Form Factor, and indicates that a smaller PSU design will be required in order for it to fit in this particular case. There are a few variations under the umbrella SFF/SFX designation though, with some cases only requiring the narrower and shorter SFX12V designs, and others needing the longer, lower profile designs (TFX12V). In these instances though, the exact maximum measurements are usually listed in case specs, and you can compare these to the listed dimensions of PSUs to find the right fit.
For some fair few years, the PC building scene saw a lot of debate around rails in PSUs. This discussion revolved around two different PSU designs for power delivery, one using something known as a single rail, and another using multiple rails. It was indeed an important debate to be had, with the key concerns surrounding the idea of having an additional fail-safe in power supplies. Without going into too much detail, the essence of the debate was that multiple rail designs offered an extra layer of safety by limiting how much current could be sent along any of its connections and therefore avoiding the risk of overheating cables and essentially melting parts of your system (known as OCP - Over Current Protection). On the contrary, and the main dilemma within the debate, was the way in which new hardware was developing at the time, with the likes of CPUs and graphics cards becoming more and more power hungry; this would mean that multi-rail designs would not have the capacity on an individual rail to deliver the required power, since multi-rail designs, as a result of limiting current for the OCP feature, could not share voltage across each rail. That could have meant plugging in a new graphics card only to find that the PSU couldn't keep it sufficiently powered.
The consensus, in the end, was that the likelihood of too much current being delivered by a single connection would be incredibly slim, and in properly utilised systems (that is to say, normal everyday PCs, workstations and Gaming systems, rather than the likes of Crypto-currency Mining Rigs), should only ever occur in the event of a short-circuit in a system. This itself is incredibly unlikely to happen, and in the rare event that it does occur, there are already short-circuit fail-safes in place that would kick in to prevent damage to the system. And as PC hardware has improved over the years, a lot of these concerns are of even less significance now.
As such, these days the debate is reserved for more niche system designs and the vast majority of PSUs actually use single rail designs anyway. So in 99.9% of cases, this is what most people will end up using, and I can almost guarantee it's what someone reading this will use too.
For those that may still benefit from that additional layer of protection however, multi-rail PSUs are still being manufactured and are usually seen in much higher wattage PSUs, as well as those used in the likes of servers and workstations, where even that miniscule chance of failure must be avoided at all possible costs. Even then, many of these PSUs will actually have a hybrid functionality where through either a physical mechanical switch or software, the power supply can be alternated between single and multi-rail, according to the needs of the user.
If you'd like a more technical breakdown, look no further than the video below, where Jon Gerow (aka the PSU man) talks about all things power supplies.
As a final note, if you're ever unsure about what sort of wattage you'll need for a new system, then make use of the wonderful tool provided by OuterVision. It's a very clever calculator that allows you to fill out a few details about the rest of your system, then tells you roughly what it expects that system to use power-wise. It'll even recommend a few power supplies that might fit the bill too.
Last but by no means least, we have cases and cooling solutions. Possibly the simplest part of any build, there are far fewer specs to consider compared to your other components. They also offer the most freedom when it comes to the overall aesthetics of your build - after all, it's the case you'll be looking at more than anything else and your system fans are generally where you'll get the RGB flair from.
When it comes to cases, you'll mainly be thinking about form factor (and maximum hardware measurements), airflow, and if it's important to you, aesthetics too. As for cooling, you'll want to consider the airflow design of your case and thus how many system fans you need, what size those fans will be, and whether you'll be opting for an aftermarket CPU cooler or an AIO solution. To an extent, all of these factors will also be affected by whether you want to have a system that operates as quietly as possible under load, or one that stays as cool as possible, regardless of noise, which we'll explore more momentarily.
Form Factor - When it comes to cases, form factor simply refers to the overall size of the case. For the most part, cases will be categorised under three main form factors: full tower, midi-tower and ITX. We discussed some considerations in our first post (find it here), but to summarise, you'll need to decide the size of your case based on three things - the space you physically have available for the PC, your motherboard form factor, and how cool you want your components to stay (with larger cases offering more airflow than smaller ones).
Hardware Support - As for hardware support, your primary concern will be the motherboard sizes supported by your case, since the smaller the case, the less it will support. Full towers will allow you to mount almost every form factor without issue (see our motherboard post for more on "E-ATX"), whilst midi-towers will generally support the main three motherboard form factors: ATX, Micro-ATX and ITX. Meanwhile, ITX cases will generally only support ITX motherboards.
You should also check any "maximum" measurements listed in a case spec too, such as the maximum supported GPU length or CPU cooler height, as well as which power supply form factors are supported, especially so when looking at smaller cases - ITX cases mostly make use of SFF PSUs, but as mentioned earlier, there are some larger ITX cases around, like the Fractal Define Nano S, which actually support the larger ATX power supplies.
Another reason to pay attention to the maximum supported GPU length and CPU cooler height, is to verify whether there will be any changes to these measurements based on the installation of an AIO radiator and fans. How many fans or radiators, and which sizes are supported in each possible mounting location (front, top, rear) will sometimes encroach on your available space, particularly when it comes to GPUs (as you can see in the above image). Knowing these measurements, and how they could change, will help you evaluate which cards and coolers will comfortably fit in any given case, as well as helping you determine whether to opt for front or top-mounted AIOs. These fine details are quite important too - miscalculating by even 1mm can result in incompatibilities, and I can personally vouch for this from experience... but we'll talk more on AIOs in a moment.
Where you eventually put your PC is a fairly obvious factor to take into account, particularly when you don't have a great deal of room to work with. But where possible, it is worth considering not just where it could go, but where it would be best situated from an airflow perspective, whether that's beside or under your desk, or up and next to your monitor.
The placement of a PC can affect a few things like heat and dust build-up, both of which can be problematic in the long run. We generally recommend having your PC up off of the floor, at least on a shelf or other solid flat surface, as this will help keep the underside unobstructed - your PSU is usually mounted in a way that allows it to draw cool air from the bottom side of your case, so having a PC directly on top of a carpet, especially those with taller piles, can really choke this point of intake, as well as picking up a lot of dust which is then pushed into the rest of your system.
Equally, placing a PC with the back and topsides of the case particularly close to other surfaces, or in small enclosed spaces such as cupboards and cubby-holes, can also restrict the effectiveness of your fans, which will have to work harder to move hot air away from your PC, whilst also struggling to freely pull in cooler air. This can result in more noise as the fan speeds ramp up to compensate, as well as preventing your hardware from operating at cooler temperatures, which can hamper their performance in turn.
When it comes to maximising your component cooling, you'll need to decide on a few things: how many fans you want for moving air through your case, the type of CPU cooler you want (or need), and how much noise you feel is an acceptable level to work or game with.
Ideally, when you design your system, you should aim to keep air moving in the same general direction - you could have a case with 100 fans inside it, but if they're all trying to push and pull air in different directions, that system will likely be no cooler than a system with four well-placed case fans. Essentially, however many fans you choose to have, you need to think about where air is coming into the case and where you want that air to be exhausted, as well as balancing the force of that air (air pressure), which will be affected by the number and speed of fans moving it in any given direction - in simple terms, if you have two fans pulling air in, then aim to balance this by having two fans pushing air out.
Since hot air rises, most people consider a great system airflow design to be one that has cool air pulled through the front of the case and expelled from either the back and/or top of the case. Technically speaking, this isn't going to make too much difference with proper fan set-ups, since hot air is easily moved by any fan spinning at around 1000RPM, so theoretically you can direct it wherever you want. That said, we would still recommend going with this option, as it's a tried and tested design that will often yield the best results for the vast majority of systems.
In terms of standard aftermarket CPU coolers, first and foremost, you need to make sure the cooler is compatible with your CPU socket type (read more in our CPU and Motherboard posts). After checking socket support, you'll mostly be considering the size of the heatsink, its fans, and how fast those fans spin, as well as ensuring it won't obstruct any other components (as per our earlier discussion on RAM height).
The key points to remember here are that, as a rule of thumb, the larger the heat sink, the more fans mounted to it, and the faster those fans can spin, the cooler it will be able to keep your CPU. Of course, with more fans and higher speeds comes more noise, which if you're looking for a cool and quiet system, isn't ideal. In this case, an alternative to one or more high-speed fans is slower but larger fans - bigger fans can move more air at the same speeds as a smaller fan, so for some people it might be worthwhile using 140mm fans at 1000RPM, rather than 120mm fans spinning at 1600RPM, which is also true for case fans, as discussed at the start of this section.
This also leads us to the topic of All-in-one Coolers (AIOs).
If you've ever wanted to keep a PC as quiet as possible, a great way to achieve this is to go with water cooling. There are two ways to do this with custom PCs: custom loop, where you design the entire loop yourself (sourcing all the individual pipes or tubing, radiators, fans, water blocks, pumps and reservoir), or AIOs, which offer a complete, closed loop design, premanufactured and tested, which simply requires you to mount it into your case. The latter is, of course, ideal for beginners, since the process is really no more difficult than mounting a standard aftermarket CPU cooler.
AIOs come in a few common sizes, which are 120/140mm, 240/280mm, and 360/420mm, with 140mm and 420mm AIOs very rarely making an appearance. In each of these cases, and dependent on the manufacturer, they will be paired with the equivalent sized fans and sometimes (particularly for the 120/140mm designs) will offer twice as many fans to allow for "push-pull" configurations, where fans are mounted on both sides of the radiator. Very similarly to standard aftermarket CPU coolers, you'll need to check socket compatibility before anything else, and the rule of thumb for cooling effectiveness is the same; the larger the radiators (essentially the heatsink) and the larger the fans, the cooler they can keep a CPU. The key difference here though, is that the fans will generally be able to operate at much lower speeds than those mounted on standard aftermarket coolers (on account of using larger "heatsinks" and water cooling), thus making them quieter on the whole.
An AIO works by having a pump (situated in the CPU water block) push water or an equivalent coolant around a set of tubing, passing it through a metal (often aluminium) radiator which can have up to 6 fans mounted, depending on its size. As water passes over the CPU, heat from the processor is transferred to the water. The warm water is then carried to the radiator, where it is cooled as it passes through the series of pipes (which are kept cool by the fans), before returning back to the CPU where the process begins again. This is important to know, as mounting the pump and radiator in certain orientations can cause problems for the longevity of an AIO's individual parts.
When mounting an AIO in your case, it is important to be mindful that the pump is always positioned in the lowest part of the loop - in other words, wherever possible, you should always have the radiator higher than the CPU block, and never below it. Due to the way AIOs are manufactured, it is impossible to have no air whatsoever within the loop, no matter how hard a manufacturer may try. As such, when an AIO is installed, the air will naturally want to move to the highest point in the loop. Pumps in an AIO are not designed to move air and rely on the water block being completely filled with water - ergo, if they are not, i.e. the pump is situated at the highest point in the loop, this is where the dreaded "pump burnout" can occur, rendering the AIO useless. Therefore, it is president for the longevity and functionality of an AIO, to mount it in a suitable orientation, such as those pictured below. It is also worth bearing in mind that the tubing of an AIO should also be free of any harsh angles which could causes kinks in the piping, hindering or completely blocking the flow of water around the loop - a more obvious issue that can be easily avoided.
Phew! You wouldn't believe it, but that really is everything! If you've been reading since part one, you've officially been clued in on every part and every spec that you could possibly want to know about! Hopefully now you feel much more confident when it comes to picking custom PC parts.
If this series has been of help in anyway, please do share it with your friends and let us know down in the comments below - we love to hear your feedback and are more than happy to answer any questions you may have!
14 Jul 2021
We caught up with PNY's David Flood to discuss how the 4th industrial revolution GPU refreshes and other major tech developments are powering incredible and exciting transformation in the media and entertainment industry.