Night 3: Motherboard setup

After taking a couple nights off to enjoy the outside world, Sunday night seemed like the time to get back in. For one thing, it enabled me to avoid giving my full attention to the Chicago Bears, which was in hindsight a pretty good call. I put the game on in the background while I got back to work with the motherboard. This was pretty unremarkable (or so I thought, we’ll discuss why that wasn’t accurate in a future post). I popped off the stock heatsinks covering the VRMs, and the cover of the IO back plate (complete with LED cable that was frustrating to detach from the board) and set to cutting thermal pads. The preinstalled pads were pretty mediocre, one of them was too thin and not really straight, so again this tells me that all this stuff isn’t exactly rocket surgery, right? This water block installation was a snap compared to the GPU, so after slapping some thermal paste on the CPU and peeling off the clearly colored warning sticker from the inside of the water block I popped it on the board with only a modicum of fuss. Humming right along, I installed both SSDs, and put the RAM in. This is, of course, way easier to do before mounting the board in the case. At this point I was cruising and the Bears were really losing. A smarter reader than I might assume it was time for a bench test before I put everything in1 the case, but I figured I’d built plenty of PCs before so let’s just move this along. I mounted the motherboard in the case, again pretty easy with the preinstalled standoffs, and decided to call it a night.

Night 4: Component installation, cable routing, loop test

I’m actually writing this on night 5 after a pretty rough go, but I’m saving that for a subsequent post. I just want to start this section off by saying that cable management absolutely is not fun. I know some people enjoy it, and they make truly lovely creations, but I just don’t like it. I don’t like the inevitable loss of some blood in the case as I stupidly cut my hand on some jagged bullshit. I don’t like the cables that seem intentionally made to be unwieldy. It’s just not fun.

Also, and this is what really got me on night 4, the amount of times you need to tear down part of the build to solve some minor problem can start to wear on you. This is definitely where I’m currently at, and getting there was … interesting.

With the motherboard in, it seemed time to wire up what I could from the case. This case has two sets of USB 3.1 Gen 1 ports (those are USB-A ports that speak USB 3.1). My motherboard only had one header for 3.1, but I knew this ahead of time and picked up an adapter for the internal cable to go to a USB 2 header instead. Don’t really care that deeply about the front ports anyway. This thing is going on a wall and they’ll (almost) never get used. Now, the cables for these ports are universally fucking terrible. They are split, they are thick, and they’re a nightmare to maneuver. Hilariously, Lian-Li gave me one set with flat cables (much easier to deal with) and another with round and incredibly inflexible cables. The USB-C cable was also just awful to work with. Super thick, inflexible, annoying to route within the back of the case. Universal suckiness bus, amirite??

There was also the dance of the front panel headers (power LED, HDD LED, power/reset buttons) and this situation somehow has not improved at all in the three decades I’ve spent inside computers! How is this header not just standardized? Why isn’t this a block I slap on the board by default? If people have a weirdo board where these aren’t all together sell them a breakout cable. Don’t make normal humans suffer through this. My motherboard even came with a block for this. It’s just silly. Anyway, that wasn’t actually difficult, just weirdly annoying and anachronistic. The HD-Audio cable (another set of ports I may not use even once) was similarly a simple installation, and I was pretty set up. I figured I would set aside wiring of on-board RGB stuff for the very end (this was a good call!) and decided it was time for the radiators and case fans. Again, maybe2 a good time for a bench test, but I knew better.

I set about installing my four non-radiator case fans, and mounting the pump. I thought I was being very clever here by offsetting the pump mount instead of mounting it through the fans. Turns out I would have to undo this arrangement later, but the idea at least was clever. Once I had the four fans and pump installed, I set about building out my two radiators with their attendant triplet of fans (bottom radiator pulls, top radiator pushes). Here I made a really egregious mistake. I installed the radiators before finishing the cabling! Don’t do this. Cable the radiators last. If you need to buy a fan hub and shove it in the back of the case, that might not be the worst thing?

Why was this radiator debacle bad? Well, only afterwards did I decide to actually wire up the already-installed case fans. And guess what? The four-way splitter I intended to use for them was about 2-3cm too short for the furthest-away fan in the case. Fuck. I had to pull out the top radiator (which was going to end up being necessary anyway), and I pulled out the three case fans that are mounted in the side of the case (a rather interesting and unique feature of the O11D), swapping out the gap-blocking plate from the top of the case to the bottom. This also, of course, meant removing and re-mounting the pump. Cleverness defeated! But I got everything installed (again) and got the case fans wired to the motherboard.

I want to talk about that for a second, for two interesting reasons. Fan splitters are an easy to get thing, but there’s not great literature on whether to use then, when you need a powered fan hub, etc. I did some quick math off the specs of both my fans and the motherboard to make sure I was actually in the clear here, turns out I was. Each case fan claims 1.5W draw @ 12V from the board, and the fan headers on the board were all rated for 1A, so this gives me lots of headroom (each fan wanted 0.125A @ 12V to get 1.5W) on my worst, four-way split. The EK fans had a slightly higher draw (~2.2W) but even with 3 of them joined together that was only 0.55A total! My old case came with a built-in, powered fan hub. I’m skeptical of the value of these unless you’re using a ton of high-RPM (and thus high power draw) fans for some reason, or you want to put eight fans on a single PWM header. For 120mm fans, at least, it seems you can chain 3-4 of them together.

At any rate, I again felt like I was cruising along. I did some work with cable management, and decided it would save me a lot of pain if I just pulled the two 3.5” HDD bays (it did, I should’ve done that from the start) to wire things behind them. This is also where I discovered that the top radiator was going to need to stay out, because the 4 and 8 pin ATX cables were never going to install properly with the radiator in the way during installation. They do fit, snugly, next to that radiator, but it’s definitely a tight fit and the radiator absolutely must go in last. So this was what I ended up doing, I just got here in a roundabout way. I had all the fans on the motherboard with their various splitters, and was only frustrated once because the fan header locations on the Crosshair VIII Hero variants are almost great. Except … the header for the CHA_FAN 1 (chassis fan 1) sits right next to the rear IO area and right where a CPU cooler (or water block) would go. It’s miserable to reach. I do not know why it is there. It should go somewhere else

This had already been a lot of work, but I was really itching to install the liquid lines at this point. I had all the radiators in, I had the video card in, I had all the stuff plugged into the board. There was no need3 to bench test, progress could be had! Had it was, and over maybe half an hour or 45 minutes I cut tubing, and installed ports on all the places, tubed them up, and put my drain valve into the pump! The tubing installation was frankly unremarkable and simple, but I quickly learned a couple of things. First off, with soft tubing on a barbed fitting, you’re not getting that tubing back off without either an inordinate amount of effort, or scissors and a razor. I recommend the latter. Second of all, it’s critical to strike a balance between a very tight run and having too much tubing. An overly long tube is going to kink and wreck your flow, and removing it after the fact is a bit annoying. Still, this process felt remarkably simple, and the EK fittings felt very snug and safe.

So I went ahead and watered the bad boy up! This is a pretty simple process of laying down paper towel (just in case you have a leak), and then filling the pump. I plugged the pump into a PSU and used the motherboard cable jumper to get ‘er going, adding water as the loop gradually filled. Spotting no immediate issues, I left the loop going for the night4. Tomorrow was going to be a big day!

footnotes

  1. It was absolutely time for a bench test. ↩

  2. Still a pretty solid time for that bench test! ↩

  3. There was so much need. So much. ↩

  4. It’s important to note that there was absolutely no power going into the computer itself. The pump has a molex connector and you can wire that to a power supply outside the case and use a jumper on the PSU’s 24 pin MB cable to induce it to provide power as desired. ↩