Need a different case

[Grogan]

This case has some issues for me. I didn't think they were serious, but I noticed some things that are dire.

Firstly, the board doesn't really go in the case quite right. It fits up and mounts properly, but the rear i/o ports are too far forward (inside the case) to get the i/o shield on. I don't care about that and there's nothing that can be done about it. Even if I could move the tray where the motherboard mounts (which I can't) the cards wouldn't seat.

Secondly, the case is a bit too small for the hardware and cooler etc. The fans don't really work out. Let me explain. First of all, the heatsink and fan assy is too big for the motherboard with all 4 ram slots filled with jacketed Corsair sticks. I have to mount the fan on the other side (and can't put two on the cooler because of the RAM). That's not a problem that's going to be solved by a case, but at least I could have better fan placement. As it is, the rear case fan and the CPU fan are kind of working against each other. I tried just taking the rear case fan off, but it's needed. (front fans aren't really strong enough on their own). I would mount a fan at the top if I had room. I'd rather not buy another cooler. (though I think I need to remount this one with more/better paste... my temps are going through the roof at -j24 and probably I'm being throttled because the cores never go one degree over 100C). At first I thought it was the coretemp sensor giving bogus values but one of the sensors on the Nuvoton NCT6687D motherboard chip corroborates it.

Now... the show stopper.

I have an ESD problem! If I walk up and touch the top of the case while the machine is running, the screen blacks out and it stops responding to input in every way (not even ACPI power button shutdown). If I discharge myself first (dissipate on metal table frame) it doesn't happen. The case somehow isn't grounded right. I fucking went over everything last night and I can't see anything I could change to fix that. I filed off some paint from the case and the metal frame on the table and attached a ground wire, but that didn't help. Walk up with these shoes, touch top of case (only the top cover, sides and front and front switch/port area don't seem to do it) and blammo. Possibly the bottom too (that would make sense) but that's inaccessible. So it's something I can avoid happening until I get the guts in a different case.

How I discovered it was, I was measuring the holes for a top fan with a tape measure. As soon as I touched it, my screen blacked out. After the hard boot, I thought "fucking XFCE power manager" and I uninstalled it. I went back to my measuring and blammo again. At first I thought I had a short, but that didn't make sense because it's a metal case top anyway and it would be shorting without me touching it. Then I realized it was ESD, for if I dissipated and didn't move my feet, it didn't happen. Most of the time I can feel a minor static discharge, but it still happens even if I don't feel a little tick. Have to discharge.

I DO have the metal standoff lugs that came with the case on every screw hole where the board mounts.

I'm wondering if I defeated some of the grounding by removing that drive bay. It was attached to the bottom of the case by 4 screws, and one screw that attached it to the motherboard mounting tray

Anyway, I'm not too happy with this case. It should be fine for ATX motherboards, but this one, not so much. I would have lived with everything but that grounding problem, and I don't think I'm going to solve that.

Ever have anything like that happen before? You've built a fuck of a lot more systems than I have. I don't think I've ever seen that (that I know of, on other people's computers) certainly never on any of mine.

P.S. I suppose it's remotely possible that the PSU has a faulty ground, though that would be odd, for the ground is a passive thing. It would have to be not connected internally or something.

[Grogan]

I've got a big black, tall, ugly Antec case I might try to use. The computer that used to be on the other desk (also X58 board and nehalem corei7). I could take the guts out and use that... ugly, but practical if the motherboard mounting works out (should... it's a full tower case). My good old system has a nice antec case, but it's got non-standard fans and I wouldn't want to have to replace them again on a 15 year old case. The fan at the top of the case is a huge low RPM high volume fan. The last time I ordered just the front case fans from Antec they were the right size, but had a different speed control switch that didn't fit. Mine had rheostat knobs that poked through the front. I had to turn it up to high and have it inside the housing. It was tricky and I wouldn't want to take it off again. Also, I don't want to disturb that system. Let it live its life. I'd buy a new case before I did that. The big ugly antec is an older, less fancy model and the fans are more easily replaceable.

[Zema Bus]

Sorry to hear that. I haven't had the degree of static discharge the way you describe, and being very dry here during the fall and winter months there is a ton of static electricity to contend with. I have often had monitors briefly turn off just as a result of moving around near them. Did you put some rubber feet under the case? Since these cases don't provide them I put some on the ones I have mainly to keep the cases from sliding when on smooth surfaces but it should help with static discharge too. I got them on Amazon but they should be easy to find locally.

Fans interfering with the nearest memory slot is very very common on boards when using tower coolers, in fact I've run into that more often than not. I have the same board and cooler with four memory slots filled and fans on both ends and didn't have any clearance issues in this case but it will vary depending on the particular memory you have and the thickness of the fan since some have taller heat spreaders. But when I do encounter that I just attach the fan slightly higher so that it clears the top of the memory module. I'm not sure how your fans are fighting each other - are they all blowing in the same direction? They should all be blowing out towards the back of the case. I've used a lot of these coolers as well as similar coolers by other brands and never had them fighting each other as long as they were blowing in the same direction (air moves into the blade side of the fans, and there should also be an arrow indicating air flow).

These processors are known to run hot under load and hitting 100 F isn't uncommon. Are you using the single mediocre fan that came with the cooler? I wouldn't use that especially with these processors. I'm using a 3000 rpm Noctua fan (this one) along with another fan on the other side (both blowing in the same direction). Even better would be two of these noctua fans one on each side. Having good fans is one thing you don't want to skimp on with these processors, and good thermal paste properly applied helps too. Here's an article by MSI about these processors and cooling, and a similar question on Tom's Hardware. There are lots of other discussions about them. So get some better fans

[Grogan]

It has feet on it, I don't know if they are rubber. It's on a wooden table top though (just the frame is metal... it's a folding office table.

Yes, that does seem like a shitty CPU fan. It's a big housing, too.Though no matter what you get, it's a shitty clip mounting system on the radiator that would have to work out. If I had a slimmer one I could mount it properly on the RAM side, but then... would it clip on.

The fans are fighting each other, because they are about the same height in the case, and the case fan is blowing out, and the CPU fan is blowing in. It would not do to turn the CPU fan around because it would just suck air from the sides and not cool the radiator. (it needs to blow on it, it can't really draw air through it).

I thought about turning the case fan around, but then I'd have front fans blowing air in, and rear fan blowing in.

I'm aware that it's normal for these processors to spike at 100C, but I don't want to be compiling with all cores 90 to 100C. It likely won't damage it (throttling... and I know it's happening) but that's no good, very bad.

P.S. I know I didn't put enough goop, too. It's always been my practice to put very little, I'd rather have metal on metal than degraded shitpaste in between. I put WAY less than all the pictures I see on the Internet. I guess that's not such good practice anymore, especially not with 18 year old AS5.

[Grogan]

The cooler is working, just too poorly for my uses (I think I can remount it without taking the board out of the case, as long as I can get at the fan clips with my hands). I didn't even realize I had a problem because it idles with the cores between about 37C and 42C and it comes down instantly when the compiling stops. I'd only done builds that took a minute and a half so I didn't notice it WHILE I was compiling. But yesterday I did.

It's not a problem while gaming, as there are lots of cores to juggle and it's not sustained usage like compiling.

I'm definitely going to put this stuff in a new case, soon. I might actually try laying a piece of rubber at the top for now anyway. The case might actually work better if those fan mounting gratings at the top are covered.

I'm not sure how dangerous that ESD issue is, but we can't have it.

[Grogan]

Fuck... so even water cooling would only keep the temps around 88C for these CPUs while compiling ("heavy content creation workloads"), from what I just read. I guess I'll just compile with -j8 or something so cores can juggle more (still have to fix the cooling though).

[Grogan]

Sonofabitch... does static ever travel. I got some rubber carpet underlay and tested it with my multitester. It is non conductive, even with the probes a millimeter apart, and pressed on either side of it.

It didn't work. OK, not thick enough, so I put two, and three pieces (just for testing). Nope, if I walk across the carpet, onto my vinyl floor mat, touch the top of the computer, and blammo. If I discharge the static and don't move my feet, I can press on the top with both hands as long as I like and nothing happens. (same as with no rubber on top)

I guess it would have to be a different type of rubber. There are many non-conductive materials that DO build static charges obviously.

[Grogan]

While it didn't help with the static touch (maybe I'm really some sort of Marvel villain and I don't know it yet lol), covering the top like that did help with air flow in the case. I just did a mesa build (ninja -j24) and while I wouldn't say the temps were good, cores were only spiking up to 100C, mostly in the 80's and 90's. I'm not polling with any kind of sensor app, I'm just typing "sensors", then up arrow autocomplete "sensors" every couple of seconds, so I'm not seeing it in real time, but that's better behaviour than yesterday.

But there's no fucking way static should be affecting any components inside a case. It should be going to ground. So the transplantation surgery is going to have to happen. I don't know, maybe this case might be OK for a smaller board.

It may be a matter of the static affecting some frequency of something while it goes to ground. There's a lot of juice going through and poor grounding might not be enough. I don't know, this defies my understanding of electricity. (e.g. then, why didn't my ground wire help)

[Grogan]

Code: Select all

coretemp-isa-0000
Adapter: ISA adapter
Package id 0:  +32.0°C  (high = +80.0°C, crit = +100.0°C)
Core 0:        +27.0°C  (high = +80.0°C, crit = +100.0°C)
Core 4:        +30.0°C  (high = +80.0°C, crit = +100.0°C)
Core 8:        +26.0°C  (high = +80.0°C, crit = +100.0°C)
Core 12:       +32.0°C  (high = +80.0°C, crit = +100.0°C)
Core 16:       +24.0°C  (high = +80.0°C, crit = +100.0°C)
Core 20:       +27.0°C  (high = +80.0°C, crit = +100.0°C)
Core 24:       +29.0°C  (high = +80.0°C, crit = +100.0°C)
Core 28:       +26.0°C  (high = +80.0°C, crit = +100.0°C)
Core 32:       +28.0°C  (high = +80.0°C, crit = +100.0°C)
Core 33:       +28.0°C  (high = +80.0°C, crit = +100.0°C)
Core 34:       +28.0°C  (high = +80.0°C, crit = +100.0°C)
Core 35:       +28.0°C  (high = +80.0°C, crit = +100.0°C)
Core 36:       +27.0°C  (high = +80.0°C, crit = +100.0°C)
Core 37:       +27.0°C  (high = +80.0°C, crit = +100.0°C)
Core 38:       +27.0°C  (high = +80.0°C, crit = +100.0°C)
Core 39:       +27.0°C  (high = +80.0°C, crit = +100.0°C)

Idle temps are great... that's about 10 minutes after compiling.

[Grogan]

Another thing about this. I have aggressive motherboard settings, "XMP Profile 1" which affects CPU clock and memory speed. (I want my memory running at 3200 MHz). It reports a 5 GHz processor. Now... LINUX, is VERY aggressive (at least with only "performance" governors etc.). It seems to run all your cores at max turbo boost when needed, regardless of TDP lol. So there's that to consider.

See this? This wouldn't be happening on Windows

Code: Select all

Idle

[grogan@nicetry ~]$ cat /proc/cpuinfo | grep MHz
cpu MHz		: 800.000
cpu MHz		: 800.000
cpu MHz		: 800.000
cpu MHz		: 800.000
cpu MHz		: 800.000
cpu MHz		: 800.000
cpu MHz		: 800.000
cpu MHz		: 800.000
cpu MHz		: 800.000
cpu MHz		: 800.000
cpu MHz		: 800.000
cpu MHz		: 1590.761
cpu MHz		: 1202.648
cpu MHz		: 800.000
cpu MHz		: 800.000
cpu MHz		: 800.000
cpu MHz		: 800.000
cpu MHz		: 800.000
cpu MHz		: 800.000
cpu MHz		: 800.000
cpu MHz		: 800.000
cpu MHz		: 800.000
cpu MHz		: 800.000
cpu MHz		: 800.000

---------------------------------------

Full sustained load while compiling for a few minutes, it stays pretty much the same with the penis cores and effiminate ones.

[grogan@nicetry ~]$ cat /proc/cpuinfo | grep MHz
cpu MHz		: 5000.049
cpu MHz		: 5000.050
cpu MHz		: 5000.100
cpu MHz		: 5000.000
cpu MHz		: 4912.411
cpu MHz		: 4912.348
cpu MHz		: 5000.000
cpu MHz		: 5000.000
cpu MHz		: 5000.000
cpu MHz		: 4999.950
cpu MHz		: 5000.000
cpu MHz		: 5000.000
cpu MHz		: 5000.000
cpu MHz		: 5000.050
cpu MHz		: 5000.000
cpu MHz		: 5000.000
cpu MHz		: 3799.991
cpu MHz		: 3800.067
cpu MHz		: 3799.952
cpu MHz		: 3800.012
cpu MHz		: 3799.995
cpu MHz		: 3800.074
cpu MHz		: 3799.988
cpu MHz		: 3799.942

[Grogan]

Yeah, the temps aren't as bad now with the top of the case covered. That would have been just taking in air and blowing it everywhere and not doing much cooling. I'm doing a sustained compile (compiling LLVM with LLVM... need to ditch the distro packages for that... I build mine differently) and it's only one or two cores showing 100C sometimes when I type sensors. All in the 80's and 90's most of the time. Also, before I wasn't really compiling long enough for the fan speeds to get turned up. They are running at higher RPMs now.

P.S. 26 minutes to build the whole thing (including my lib32-llvm which also builds lib32 Clang and installs the libraries so it can be used to compile 32 bit code) in one shot (scripted). That would have taken me probably 180 minutes or so on the old box. I'd have started it and gone to bed.

[Grogan]

As I often say, I think I've been had. This isn't bad grounding, it's bad "faraday caging". The only thing that makes sense here is that I'm generating EM interference when I'm all charged up like that and touch the case and its disrupting a signal (or sending a spurious one). It's less likely to happen (but still does) if I only touch it with one hand too. I tested it 3 times with the carpet underlay rubber on top (3 layers of it the final time). I did not feel anything, any of those times like I have when touching the metal and I can't imagine much static charge is going through it. I think the static discharge is incidental.

Up at the top there, there is almost nothing, it's just very thin metal with a zillion holes in it for mounting two 6 inch fans. No EM shielding, quite the contrary.

Either way, the case has to go.

[Zema Bus]

I watched this video today, and thought it might be relevant to the temp issue:



If you switch to your 15 year old case make sure you have enough clearance with the tower cooler to be able to put the side panel on - older cases were narrower. I found that out when was going to put a newer motherboard in my old InWin case from 2001, I couldn't put the side panel on with a tower cooler installed. Yours not being as old as that might be wide enough but it's one thing to check for. Also front panel connectors may not all be compatible, I ended up having to get a set of modern from panel connectors for the case to replace some of the leads.

[Grogan]

I wasn't thinking about voltages, but I knew that the turbo timings curve was aggressive (I also mentioned that Linux frequency governor is aggressive with that too).

It won't harm the CPU though, and if it gets throttled, it doesn't matter. You could set less aggressive settings, but then you're wasting CPU time all the time. (I don't agree that "you're not getting what you're paying for")

If there was anything wrong with the XMP Profile 1 settings I loaded, my computer would not be as stable as it is.

Some of my cooling problems were corrected by covering the top of the case, and when I get the new paste and remount with a more generous application we'll see about that. In the Advanced OC settings it will show the real settings.

As for defaults, my motherboard defaults were very anal. 3 GHz base clock reported, and 2133 MHz memory clock. It wasn't until I enabled the XMP profile that things were ramped up.

[Grogan]

Hahah... I had the CPU cooler bolted on backwards, that's why there wasn't enough clearance for the fan and the RAM modules. My eyes don't judge distance as well as they used to and it looked symmetrical.

But I noticed it today when remounting it (got the Arctic MX4 tube... and put a bit more paste than I normally would this time)

So now the air will at least be blowing the right way with the fan mounted on the right side.

Also, I noticed in the Advanced OC settings in the bios, there's a CPU power setting for different cooler types. Normal air cooler, tower air cooler, and liquid cooler. Guess what mine was set for? That's right, liquid cooling. So I put it to air tower cooling, for whatever that will do (that's probably wattage/voltage, not frequency because I saw the same core clock behaviour while doing a test compile)

Well, now to compile something big and watch the behaviour.

So this leaves only that ESD/EM interference problem, but a serious problem it is. I tested the PSU ground today (with my multitester, that PSU tester doesn't test ground I don't think) while I had it on the bench and I'm getting continuity on screw holes and rivets and stuff. The tiny trickle of current to test continuity isn't enough to overcome dissipation further away on case metal.

I still haven't decided what I'm going to do about a case, but I'll order something next week. As long as I don't trigger the EM fault I don't really have any problems now.

[Grogan]

Code: Select all

coretemp-isa-0000
Adapter: ISA adapter
Package id 0:  +95.0°C  (high = +80.0°C, crit = +100.0°C)
Core 0:        +89.0°C  (high = +80.0°C, crit = +100.0°C)
Core 4:        +83.0°C  (high = +80.0°C, crit = +100.0°C)
Core 8:        +93.0°C  (high = +80.0°C, crit = +100.0°C)
Core 12:       +91.0°C  (high = +80.0°C, crit = +100.0°C)
Core 16:       +94.0°C  (high = +80.0°C, crit = +100.0°C)
Core 20:       +95.0°C  (high = +80.0°C, crit = +100.0°C)
Core 24:       +95.0°C  (high = +80.0°C, crit = +100.0°C)
Core 28:       +96.0°C  (high = +80.0°C, crit = +100.0°C)
Core 32:       +77.0°C  (high = +80.0°C, crit = +100.0°C)
Core 33:       +77.0°C  (high = +80.0°C, crit = +100.0°C)
Core 34:       +77.0°C  (high = +80.0°C, crit = +100.0°C)
Core 35:       +77.0°C  (high = +80.0°C, crit = +100.0°C)
Core 36:       +73.0°C  (high = +80.0°C, crit = +100.0°C)
Core 37:       +73.0°C  (high = +80.0°C, crit = +100.0°C)
Core 38:       +73.0°C  (high = +80.0°C, crit = +100.0°C)
Core 39:       +73.0°C  (high = +80.0°C, crit = +100.0°C)

Code: Select all

PECI 0.0:        +98.0°C

This is as bad as it's getting now, with my P cores all blazing at 5 GHz and Ecores around 3800 during -j24 compiles.

A more useful sensor to watch is the Intel PECI sensor, which averages all the core temps over a 256ms interval. Mine got up as high as 98.0 in that particular sampling, during this test. That's likely an anomalous reading though, as the cores weren't reading that high. it's usually slighly lower than the coretemps.

That one will be on a motherboard sensor chip though, as it will likely be used for fan control (Intel recommends using the PECI sensor). It's not the same as the coretemp "package" temperature, as that's on a hotter part. They read similarly though.

P.S. My PCH chip sensor is doing better now too with the fans not working against each other

Code: Select all

PCH CHIP:        +68.0°C

That was running about 78C while compiling, and that one takes a while to come down after too. That was OK though, they are meant to run hot.

[Zema Bus]

That's much better. After decades of being used to much cooler running processors it's kind of hard to get used to these higher temps being normal lol! I've heard the latest AMD processors, the 7000 series, are the same way, and when users complained following their initial introduction AMD told them those temps are normal.

[Grogan]

It does seem alarming, but "compiling" (with all cores) is not actually normal use of a consumer CPU. You'd never hit that gaming or under normal desktop workloads or even multimedia encoding/processing. These boards have those aggressive settings for gamers.

I rewatched that video last night to confirm, but that "CPU Cooler" CPU power setting changed exactly that. My board had the exact same settings with the aggressive defaults in XMP Profile 1 (the board does more than just clock my RAM to 3200 vs. 2133 with the optimized defaults)

He was spot on, my upper and lower power limits were 4096W ("virtually unlimited") so the CPU would never downclock, relying on the CPU's own thermal throttling at 100C (and if it were to get to 105C it would halt and I checked, I am NOT overriding thermal protection though I do have that setting). I'm still a bit too aggressive with 288W limits and 512A max current limit (but setting a limit doesn't matter if it doesn't hit that... it's the wattage limits that count). I'm happy with the current settings (not throttling anymore, some 95C cores at worst), but to get anything else I'll have to tweak it manually. If I choose the lowest "CPU cooler" setting, "boxed cooler" it sets them a bit too low. On that setting it drops the max current limit down to 255A (should be 360A for these CPUs). I don't think I'd want to hit a lower amperage limit, that would need more voltage.

(That sounds nuts, eh? 512A (but makes sense in context of 1.3'sh V). With higher voltage, that would FRY you to a crisp. DC current, holy shit. What you can do with electricity by transforming, rectifying etc. We have 400A service, which means our whole house wouldn't draw that. That's actually double what most people have)

[Grogan]

So I ordered this case yesterday:

Corsair 5000D Airflow Tempered Glass Mid-Tower ATX PC Case - Black
https://www.amazon.ca/gp/product/B08M49WW51/

It's a little larger than the 4000D, but it's still a Mid Tower case! Most cases are, Full Tower cases aren't common anymore. This one will be more appropriate for the hardware though.

I really wanted the Corsair 7000D (their full tower case) but that's $400 and will fit nicely up their asses. I got the 5000D for $204 CDN, which is 11% off according to scamazon.

Anyway, I ordered the case and I'll figure out what all fans I'm going to need to buy when I see it. I'll have one front fan and one rear fan included and I can take the rear fan from the Fractal case (not part of it) and mount it turned around on the front for now. Or at the top. We'll see.

According to Amazon I should have it Sunday.

[Zema Bus]

Nice! Similar design as the older ones I have. Too bad they can't be had with 5.25" drive bays anymore. The biggest thing I wouldn't like about that is the glass side panel, but that's hard to avoid these days. If I get a case with a glass side panel in the future I'll probably consider taking the panel to a shop to replace the glass with acrylic. And power supply covers, another thing that's becoming impossible to avoid.

[Grogan]

I would have preferred steel, or if it has to be transparent, plexiglass, but apparently I can buy an aluminum side panel for this case if I break the glass lol

I gave up on the 5.25" drive bays

I know I'm not going to like that PSU enclosure, but I already bought an unsuitable case because of my stubbornness to not accept "the new way". (though we couldn't have anticipated that EM issue)

[Grogan]

The new case showed up a while ago. It is indeed big, probably more so than my Antec Nine Hundred Two case (which I thought was a Full Tower case, but it's actually a Mid... I just realized that now)

However, this is a weird case. That's why I didn't buy it in the first place (including not having optical drive bays in my decision). I really, really dislike the way the power supply goes in and I haven't even started yet. Now because I didn't buy a side-connecting Corsair PSU, I'm going to have a harder time poking those fat PSU cables through openings and shit. What a silly, unnecessary pain in the ass that design is. I wouldn't mind if the top of the PSU enclosure came off, but it doesn't, it's two rivets at the top not screws (like it looked in the picture). I have a long PSU and it's going to be hard to get in there from the side. The hard drive bay is going to have to be moved (I think it can be moved forward by the look of it). I'll take it out permanently if it gets in my way!

I don't like how the side panels go on. (I'll have that tempered glass broken in no time fumbling with it).

I'm also not impressed at being chintzed on fans. For an expensive, enthusiast's case, that's poor. My Antec Nine Hundred Two case came with all the fans. (Nice fans)

Oh well, it will be a good case for the hardware once I'm done cursing.

[Grogan]

Well, one thing I like about this case is that there are fan mounting rails that will take either 120mm or 140mm fans, and they can be positioned anywhere on the rails.

I'm still in the procrastination stages, but I went to the computer store and bought a 140mm fan for the top and I'm installing that now (and figuring out where to tuck, and ultimately connect the wires). I actually have a little board with fan headers behind the other side panel of the case that looks pretty convenient. It looks like it connects to the motherboard with two fan headers, then I have 4 fan ports on that little board. However, it's just pins, no plastic notch to guide it, so they will have to be connected the right way. (i.e. the fan connectors are 3 pin, the headers are 4 pin)

For now, I'm going to have two 120mm front intake fans, one 120mm outtake rear fan, and a 140 mm fan on top blowing upwards. (Corsair SP140, but an older red LED model that I can't find online anymore... probably sitting on the shelf a few years. I won't be connecting any LEDs, I really don't like that shit). The case is really meant to have a big liquid cooler block on top, but I'm not getting one of those.

P.S. Oh... I guess I will be having the red LED on that 140mm fan, it's built in

[Zema Bus]

Next thing you know you'll be installing magnetic LED light strips lol!

[Grogan]

Well, here goes... the time of procrastination is over. I'm really dreading this and I have butterflies in my stomach. The difference is, now I'm disrupting my good computer whereas before it was just parts in boxes. Also, this is the most complicated computer case I've ever seen. I have octopi of wires and connectors tied up at the back with velcro that I have to make sense of (I won't be using all of it)

It is a popular case though, and people who know what they are doing put these systems together in minutes in youtube videos. (They speed up the video while they are doing up screws and stuff... it looks funny). Unfortunately they all buy fancy cable extensions (so where they are tucking cables won't work out for me).

Know what else you can do in these cases is vertically mounting the video card. I don't have the accessory for it, but I would imagine it's cabling with a connector that connects to the PCI-E slot on the motherboard. Imagine not having a big, fugly, double slot video card on your motherboard. HOWEVER, those video cards are designed to be mounted the way we do, with the fans on the bottom. I would imagine vertically mounting would defeat some of the graphics card's own cooling (you'd need an array of case fans... you can put like a dozen or so in this case)

I'm not a "fanboy" though. I say it like that, because back in the 90's, the very first time I heard the term "fanboy" was in reference to AMD fanboys. At the time I thought it was because they were always bragging about their fans (which they need to cool their ridiculous hardware... like multiple voodoo2 accelerator boards and all the shit those folks had.

[Grogan]

Graphics cards are difficult to remove from this board once snapped in. It's not they fit tightly in the slot or anything, no, it's because the slots have a locking latch more like a RAM stick now and it's difficult to snap open with a screw driver without raking it across the motherboard. There isn't much room to work with the board's heatsinks etc.

I've been hearing that a lot of Gigabyte Nvidia 3090 and 4090 (and similar) cards have been failing because that locking tongue on the PCB (on the edge that goes in the slot) have been cracking, and it just so happens there are circuit pathways around there. It is thought that it's either happening because of sag (nothing supporting card in case, poor fastening of rear tang by manufacturer) OR people busting it taking it in and out of the motherboard.

I was a little rough with mine while I was having trouble getting at it (I had to move stuff to even get eyes on it) and thought I could wiggle it out lol, I hope I didn't damage anything (the card looks OK, I'm worried about the PCI-E slot)

[Zema Bus]

I've seen those video cards mounted that way, just so that the fan side of the card faces out when viewed through a windowed side panel. That always looks weird to me, and seems like it would be hard to work around when trying to get your hand in there to access sata ports etc. I think they use riser cards or some specialized variant of a riser card. I don't know how they secure them without bracket screws, some of those big high end 3 slot cards are heavy and even when mounted in the normal way require using an anti-sag mechanism.

[Zema Bus]

Yeah you have to hold it in while pulling out on the card. About 10 years ago I ended up breaking off the latch on a motherboard while trying to get a card out. It didn't otherwise do any damage and the slot remained functional.

[Grogan]

10 years ago, they were just a little lever you had to press and hold on a little while taking the card out. Now, on this board, it's a snap that takes quite a bit of force with a screwdriver to unsnap. (once unsnapped you don't have to hold it, but if you're working with it laying down, it'll easily snap back down in the slot. It goes in easy, but seems to be hard to unsnap

[Grogan]

I realized I actually can use both M.2 slots on my motherboard, I thought the top one couldn't be used because of the CPU cooler, but it's down under enough and I can still add/remove drives there with the cooler mounted. I was using the adapter card for my second NVME (games). The third one (that would be directly behind the graphics card) is E key, for mini wifi cards and stuff. That's useless anyway, where it is.

What I did was, edited one OS (my Bollux) for nvme0 and left the other (Arch) nvme1 in fstab and the grub.cfg on Arch. This way, if it switches, or if it doesn't, I'll be able to edit one OS from the other, whatever happens. I hate booting with USB to fix shit.

P.S. My rationale is don't mount anything, anywhere you have to take other shit off to get at, if you can help it.