Log Experience with itx so far

[spiral]

So...

Coming from a Caselabs M8 i wanted something smaller since my watercooling days are over. I went for the LianLi TU-150, since it can fit a beefy cooler.
So initally i put my trusty Dark Rock 4Pro in to the case, but it seemed to restrcit the airflow for the rest of the components. Switched over to a Arctic Freezer 34 that was in my sons pc, and it seems to work a bit better, alteast the noise is down a bit. All in all i was not expecting this level of noise , so is to be expected ?
Specs are:
Ryzen 3600
Gigabyte x570 (my b550 order got pushed to august)
Asus Strix 2070s
16gb trident neo
960evo/860evo

Memory controller temp pic

Edit: Forgot to add that i use noctua for intake/exhaust. NMo room for fans under the gpu.

 

[scatterforce]

"Drive Temperature 2" on Samsung 960 Evo

Hi, the "Drive Temperature 2" reading on my 960 Evo is at about 62 C idle, 85+ C just from running a (relatively short) crystal disk mark benchmark, and can hit up to 90 C just from the GPU being pretty hot. Needless to say, those temperatures are pretty toasty and might be a cause for...

www.hwinfo.com

.

With a quick google-fu, it looks like over 70C (your peak temp recorded) is not so great. It won't damage anything, but it could throttle the drive. If you like your case, try using a heat sink. The Gigabyte board has a fan over the NVME drive and it allows room for a short heat sink.

The bottom line is the TU-150 has a single intake fan that is restricted. Have you tried looking at slim 120's for the bottom of the case?

 

[spiral]

I can't fit slim fans in the bottom, since there is only 12mm clearing. With some testing back and forth it seems that it goes to that temp in the first couple of mins, so i guess it's before the fan kicks in or smth. I get's to the same temp while playing a game on the other drive, so for now i'll just ignore it. raised the case on some improvised feet, some 5cm or so and gpu temp is way down, and cpu temp is better (more like 64c in realistic loads) I guess i can put some fans on the outside under, but that's a bit too much hassle i just opt for a atx at that point.

 

[scatterforce]

I think you can control that fan over the chipset and NVME in the bios or gigabyte fan control software. Set a minimum RPM and that might fix the issue.

 

[spiral]

After bulding a atx build for my son in a H500 it seems i'm a noise snob. I guess i'm just too used to zero noise from wc, that i can't stand any sort of fan noise. Right now the =11 mini looks appealing as i want a small(ish) case and no noise at all. That said i'm not too impressed with the tu-150, there is space for dual 120mm intakes if they wanted to, and the compartment on top seems like a waste of space compared to the aircooling potential.

 

[smitty2k1]

If you want quiet SFF, IMO you can't beat the NCASE M1. Big heatsinks for both the CPU and GPU and SFX PSU means you can get the nearly silent Corsair SF600 gold/platinum.

 

[spiral]

Picked the pc apart and put it back together today, and now it's silent all of sudden. I have no idea what the issue was, but it seems to be working now. The only thing i can think if is that everytthing was not 100% connected on the fan splitter for the case fans. So i guess i can start with a small oc now. Getting 64c all core @4.2 so there is some headroom. I put some paste on the thermalpad for the m.2 and it seems to max out at 59c now.
Seems going itx and amd at the same time was a bit overwhelming, but i'll get there in the end.

Considering getting extension cables for the psu, i guess there is no downside compared to getting cable set from Corsair ? It seems the replacement sets also are a bit short on the 8/24 pin.

 

[spiral]

So after having a couple of normal summer days here (25c), i'm done with this case. Playing games i can do about 30 mins before the system restarts, due to the gpu overheating. Currently using it with the sideåanel removed, put this is not a temporary sulotion. The question that comes to mind is why this case gets so good reviews by some youtubers, it's not like i have high end hardware. The gpu does sit around 90-100% load while gamiong given the 3440*1440 resolution, but the issue is there is no room for the hot air from the gpu to go fast enough. Is sff not meant for heavy workloads, or is this just a mediocre case ?

 

[catawalks]

That's not normal at all. Your computer should not be shutting down due to thermals unless something is very wrong. These devices have lots of protections and logic to throttle them hard before thermal shutdown happens. What are the temps looking like for the GPU during those 30 minutes of game play?

 

[spiral]

Using Afterburner the osd only says 70 for gpu and 65 for cpu before it restarts. The backplate on the gpu however is quite hot to touch. Tctl/Tdie is 76 max.

 

[thelaughingman]

Using Afterburner the osd only says 70 for gpu and 65 for cpu before it restarts. The backplate on the gpu however is quite hot to touch. Tctl/Tdie is 76 max.

Your CPU and GPU temps look perfectly fine. I'm gaming at close to 80C on my GPU and it never shuts down. I run benchmark full loading my CPU at 75C and it never shuts down. Your restart is probably because of other issues, not temps

 

[spiral]

you should think so, but removing the side panel there are no issues. So it's pretty obvious it's a temperature issue.

 

[thelaughingman]

then Afterburner is giving you some bs # cause throttling/shutdown temps for the CPU/GPU is 95C/100C not 65/70. maybe try HwInfo64 and do a full log while you're gaming with side panel on?

 

[tinyitx]

then Afterburner is giving you some bs # cause throttling/shutdown temps for the CPU/GPU is 95C/100C not 65/70. maybe try HwInfo64 and do a full log while you're gaming with side panel on?

Good suggestion. If the difference comes between having the side panel on and off, then it is very likely a thermal issue. But, 70C (GPU) and 65C (CPU) are not shutdown temp (not even throttle temp).

Since Hwininfo64 logs so many parameters, I suggest just logging the thermal readings in this initial investigation.

 

[spiral]

Yea, i'm using hwinfo but i don't get any logs when the pc crashes. I know the temps looks fine, and it should throttle rather then restart. Chipset/ram is 52-53c. The m.2 gets quite toasty peaking at 72c before the fan spins up, but that's not for long before it goes down to low 60's.
Windows eventviewer only reports the kernel error saying windows didn't shut down properly.

 

[tinyitx]

Yea, i'm using hwinfo but i don't get any logs when the pc crashes. I know the temps looks fine, and it should throttle rather then restart. Chipset/ram is 52-53c. The m.2 gets quite toasty peaking at 72c before the fan spins up, but that's not for long before it goes down to low 60's.
Windows eventviewer only reports the kernel error saying windows didn't shut down properly.

So, you have Hwinfo in a continuous logging mode and there is no log file after the PC crashes? This is weird.
Maybe you just do some light load work (like browsing) rather than gaming and make sure a log file is being saved first.

 

[catawalks]

Yea, i'm using hwinfo but i don't get any logs when the pc crashes. I know the temps looks fine, and it should throttle rather then restart. Chipset/ram is 52-53c. The m.2 gets quite toasty peaking at 72c before the fan spins up, but that's not for long before it goes down to low 60's.
Windows eventviewer only reports the kernel error saying windows didn't shut down properly.

Anecdotally speaking, I've had to RMA a STRIX GTX980 after manhandling it for a few days as a test mule. I ended up finding the cooler had come loose, but the damage had already been done. I try to stay away from those cards now due to the entire heat sink that runs the length of the card was only held on by the 4 litte screws around the GPU die. Picking the card up by the tail end would put a tremendous amount of leverage on those little screws and it's just not to my liking.

Do you have any other card you could use to swap and test to see if the shutdowns continue?

 

[spiral]

Got the log to work now. There is nothing strange in it, cpu at 63c and gpu at 73c when the system rebooted. vrm/ram and the other stuff low 50's. The only thing that doesn't have a sensor is the psu. It only took 12mins for a restart playing Division 2.

 

[catawalks]

Got the log to work now. There is nothing strange in it, cpu at 63c and gpu at 73c when the system rebooted. vrm/ram and the other stuff low 50's. The only thing that doesn't have a sensor is the psu. It only took 12mins for a restart playing Division 2.

Interesting. Something definitely sounds wonky. Side panel off consistently works, and side panel on consistently crashes, but measurable temps are not showing issues. What PSU are you running, does it have OTP, and is the fan spinning while the system is under load?

Honestly I would start swapping parts and testing different pieces until I could isolate the faulty one. The temps you mention are not an issue, like @thelaughingman said, serious throttle temps are way higher and shutdown temps are even higher.

 

[thelaughingman]

yeah, swapping out parts is probably the next step to take now.