Oh yeah, I totally forgot about cramming big components into tight cases. Once in the past I decided I don't want to deal with that. Built my latest desktop in a Define R5.

Trying to build a high performance system in a mini-ITX case will shorten your lifespan for sure. ;-)

Anyway, how do you troubleshoot a build without any spare parts? (I assume you don't have any spare PSUs, motherboards, CPUs, etc. gathering dust?)

.... sup? .... I've been building my own for a little while now (and for everyone else in the family; office workstations, entertainment builds and most importantly, gamers); it really isn't hard to do anymore, almost everything these days are plug and play.

... but if you really don't want to just dive in yet, I would recommend IBUYPOWER.com for a pre-build. I've bought two machines from them and was impressed with everything from the ordering to the delivery; of course this was years ago, hopefully they haven't changed any. The first one lasted 6 yrs until the CPU fried (about average for a heavy online gamer); the second one is still gaming after 7 yrs. the only thing that has been replaced was the GPU (video card) which is expected after 5 or 6 yrs of heavy gaming ....

...happy hunting

I don't have many spare parts, though usually there's at least a spare PSU lying around. I might have other components but they tend to be too old to try on a new system ('cause sockets and slots change). Or they're things like disks which are usually easy to figure out anyway.

Really I haven't needed to do much troubleshooting. But when I do, common sense can get you quite far. Like, there are various ways to test whether you're getting any juice at all from a PSU. Usually it's obvious though, and the more pressing question is whether you're getting enough juice at load (reboots during heavy loading would make me suspect the PSU first).

If the system powers on, well most mobos these days have a whole bunch of diagnostic leds. Some exist solely to indicate the presence of a particular error (or the good working state of a particular component), and there's often a small 8-segment display that gives specific codes giving you insight into the boot sequence and any potential errors. Usually I buy more than one stick of RAM, so I can take them out and test each individually if I suspect a faulty stick. The other thing to check is just to make sure everything is seated and all the power & data & etc. connectors are there. Such common sense really should get you far.

So if you've triple checked that everything is seated, the thing is trying to boot/post (getting juice), and you swapped out ram sticks, that really leaves only three critical components: cpu, mobo, and gpu. These diagnostic leds really help here; if they indicate that everything is normal, then the most likely culprit is the GPU (system boots ok, but monitor gets no signal). If the leds are telling you something is wrong, it's more likely to be between the CPU and mobo (or their interaction with RAM). But you can still yank out the GPU to make sure it's not interfering (you shouldn't need one to boot, though some boards are retarded). If that doesn't help, and you've triple checked CPU seating, and diagnostic lights aren't giving an answer, assume it's the mobo because it's *very* rare for CPUs to arrive DoA (although the same applies to the other parts too and this is why I just don't generally don't need to troubleshoot my own new builds). In fact, based on two anecdotes, I hold it more likely that a store-assembled PCs would arrive DoA (or just in a bad but fixable state) because they fucked up, or transit fucked up.. ;-)

Really swapping parts is not necessarily the best troubleshooting tactic; it is usually the last thing I need to do, and by the time I'm doing it (which is practically never), I already have a very strong suspicion that it really is the faulty part, and I have been able to rule out most other parts with high confidence. At that point, I might as well buy the new part and get prepared to return the suspect faulty part.

In fact the kinkiest of problems seem to be related to new platforms, like Ryzen. I know many people had systems that wouldn't get as far as posting. This isn't a hardware fault as such, and in that scenario the best troubleshooting comes from forums where other builders are dealing with the same issues and someone is likely to have a similar build with similar issues. Even in these cases the mobos nudged people in the right direction with their diagnostics, and people quickly found a RAM configuration that worked.

Apparently there were some really crazy issues with specific cooling solutions. I'm still not sure what the issues were caused by; some postulate that it was too much (or too little?) pressure on the CPU. Others speculated that the rubbery material that goes behind the CPU is ever-so-slightly conductive and it's in contact with some pins on the bottom of the CPU socket. Either way, browsing these forums should quickly reveal if you're stuck with such a problematic cooler.

Again, all these issues were related to a completely new platform which wasn't perfectly compatible with all the (coolers and) RAM sticks out there. By now you likely wouldn't run into these woes. So I guess one could tell a newbie to avoid completely new platforms. But experienced builders were bitten by those issues too and they had to lurk the forums anyway; so it doesn't really matter. Except that (I suppose) a newbie could find it too frustrating.

Sometimes the force just isn't with me. Some years ago, I had an Athlon64 X2 4200+ system and one day I couldn't cold boot or restart it anymore without first unplugging the power cable and waiting 5 minutes. Then one day I upgraded to a Core2, new motherboard, new RAM, new PSU (not a cheap one - a Seasonic, new GPU. And guess what? Same boot problem. To this day I still have no idea what was wrong with it; and no amount of reading online helped.

I just spent a couple of days troubleshooting random BSODs on that Ryzen system. At the end I got a new GPU. Why? Because it only happened in games and the same thing happened on the old computer (which had only the GPU in common). Turns out that that wasn't the problem at all! It was the XMP profile.
On my i7, when the memory frequency was set too high it just stopped at POST/Bios - no vague BSODs or anything.

Thanks for the long reply!

When I say consoles aren't what they used to be, I mean they just aren't the simple "plug and play" boxes they were a couple of gaming generations ago. Buying physical media is becoming silly when the games usually aren't complete anyway and DLC is coming just months (or sooner) down the road. Things like massive updates and mandatory game installations just make me ask why I'm doing this on a console when PCs offer so much more.

I suppose my biggest concern is how often I'll have to be troubleshooting games or trying to solve performance issues. I hate that stuff. What I like about consoles is just turning on the game and playing - the end. I don't mind changing a couple of things in the menus if necessary, but I don't want half of my PC gaming experience being spent outside the games. I think this is probably an irrational fear from someone who hasn't gamed much on a PC, so it's good to get input from experienced PC gamers.

That's interesting about the longevity. It would be great if I could get 4 to 5 years out of a build, with a GPU upgrade midway through. Being someone who is content with the visuals provided by consoles (blasphemy, I know), I don't really care so much about ultra high settings as long as the experience is smooth.

Yea, this is the thing with Ryzen. But after reading some forums, I would have almost definitely suspected the memory (settings) before the GPU as the leading cause of BSODs, even if it happened only in games. In fact it looks like the vast majority of troubleshooting people had to do was ultimately related to finding the right memory settings. And as far as GPUs go, I haven't actually had one cause BSODs yet. The symptoms of a bad GPU have always been quite different. BSODs [as far as I know, I haven't actually used windows in 10+ years] are caused by bad memory access from the OS side (kernel or drivers) while devices [as far as I know] effectively bypass page protections. So you'd get BSODs for either a software (most likely driver) bug or corruption that leads to bad memory access -- such corruption taking place in the CPU, memory, or the interface between them.

I don't know what kind of testing you did apart from games, but there are other workloads (various cpu & memory stress tests, big multithreaded compiles) that *likely* would have exposed the issue. In fact games rarely push a system to the limit.

So who ended up paying for the damaged components? I presume you just took it back to their store and told them "fix this shit or give my money back".

I'm surprised Systemastore didn't then test what they had assembled? I think I mostly used a small PC store called "DataDog" because it was pretty close to where I lived, and I was happy with their operation. They were a small shop with only a couple of employees, and I know for sure they always did test the systems they assembled, before informing their client that the system was ready to be collected.

I think I bought several desktop PCs from that store, also to others who asked what kind of PC to get and from where.

The very first PC I ever bought wasn't as peachy though (different store, I don't recall the name anymore). The PC itself worked fine, but the delivery to me just got more and more delayed. The mistake I made was to pay for the PC up front as that was a bit cheaper than paying on delivery, and that apparently put me to the end of the priority list.

When I finally started to demand my money back and threatening them with legal action (as if I would have really known how to do that; call the police?), they instantly contacted me that my PC is ready to be collected and I get twice as big hard drive free of charge.

I reseated the CPU and managed to bend the heatsink clip back into shape well enough (it still felt looser than I like but the PC's been running great for years now). So nothing was permanently damaged, and I didn't even contact the store because I felt that would've been more trouble than it's worth. Though I guess with luck they could've given a refund or something..

Yes, drivers, and devices like graphics cards use drivers. ;-)
Also, the DLLs mentioned in the BSODs were often DirectX related. I like waiting with a Memtest86 to the very last moment because it takes so long. And in this case, the result might've been that the memory was fine (because it isn't faulty).

In that case, it would be a software bug that is in all likelihood reproducible, reported, and soon to be fixed with a new driver version. I would hesitate to buy a new GPU over it. I tend to run memtest86 first (leave it running overnight after doing the build). It's true that it doesn't flag all memory related errors, but if and when it does, it saves you some headache down the road.

If you're getting corruption and bsods, it doesn't really matter whether it's faulty memory or just bad memory settings. From the system perspective it's all the same -- things get corrupted or memory accesses fail. In this case, the problem might be linked to the CPU's memory controller so loading the CPU and memory simultaneously may be better than memtest86. That's why the second step I take (after installing an OS) is various system stress tests, benchmarks, and big compiles that batter both the CPU and memory heavily. I tend to leave games and anything that stress the GPU for the last, for when I have high confidence that the rest of the build is rock solid. That is, if I test games at all. I don't play much to be honest.

Having to troubleshoot games on PC is a rare thing these days. That is, if you play new games or stick to GOG for old games. Trying to get old disc-based games to work in a recent Windows is where the challenges lie.

Getting 4 to 5 years out of a system is easy nowadays. A GPU upgrade might even not be necessary, depending on the kind of games you like, the graphics settings and which card you pick as your first.
My GTX770 is still doing fine and I have no plans to upgrade. I only have 2 games that push it to it's limits and those are The Witcher 3 and Dragon Age: Inquisition. I don't like the AAA games of the last couple of years.