[SOLVED] My AIO looks misaligned..help?

Page 2 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Solution
I noticed the AIO pump looks misaligned..is this serious or a risk at all?
the block is definitely misaligned.
whether this is due to incorrect installation or cheap manufacturing, it's impossible to tell by the video or image.

what make & model is the cooler?
i've never seen this logo that i remember.
if it's just some generic cheaply manufacturer product, i would return the entire system for something better.
if they've skimped out on the cooler, then they probably have with other hardware also.

whether there is any serious risk would really only be any possibility for leakage from a low quality product.
if temperatures are within good ranges, 25°min - 65°max(C), then it is doing it's job as it should for now.
Yes they make those but the way he done it is use a 6CFM fan then stuck heatsinks all over the top of the backplate of the card it's probably running hotter since no airflow.

The fans on top of the backplate are exhausting the heat faster from the memory region that is suffocated by the backplate as it actually does not have any active cooling going on in the thin region.

I have reduced the temperatures so I guess its working!
 
even with a heavy philips screwdriver you can't loosen it?
if so, that could mean the threads are damaged.
usually those thumbscrews come off pretty easy.

If I am being honest here I am afraid of unscrewing it and not being able to screw it back in properly. I have never fiddled with the pc and I have found these little issues so i dont want to actually cause a problem and void my warrenty!

The tech support did say all I have to do is unscrew the top and bottom plates and then slot the top plate in place. Thing is I dont know if the pump would pop out or if if I have to put it back in a certain way etc.
 
keep a copy of this response they sent readily available.
if you follow their instructions and any sort of damage occurred it would be on them.
though this could still lead to some added difficulty if/when ever trying to RMA the system again.

This is the response from tech support;

I have had a look at the images you have provided THE AIO does seem slightly of center however if there is no performance issue or heat issues then this should not be a problem as it receiving sufficient cooling, however we can look to straighten it out for you if needed.

Please let me know if you would like to send the system back to us so that we can look to get these issues resolved for you however as stated they will not effect performance or cause any issues so this is entirely up to yourself.
 
My apologies AIO is misaligned because the top plate isnt all the way in, my intention with this thread was whether it is still safe to use or not etc!

No it is actually at the hands of the company who built my pc, they first said maybe it was because my phone knocked it lol...and then possibly during transit it may have moved. But then again the plate as well as the screw is so tight I cant imagine transit being the reason.

Its a 3090 ROG Strix.

Initially playing warzone I was getting 82-86c in the memory junction. After using the heatsinks and fan the temps are reduced by 2-4c. However I also undervolted and now I am in the 64-68 region.
Ah ok, well if the temps are fine then yes I would say it was safe to use, it does as you say look flat as well it's just slightly wonky. It is quite easy to install these AIO's though, you could take the water block off and position the backplate correctly and reinstall it. If you did I wouldn't bother reapplying the thermal paste, but you'd just need to be careful when you take it off not to get it everywhere.

Thanks, I have a Palit 3080, my memory temps are 81C. That doesn't sound too bad, some 3080's apparently have been running at 105C on the memory. When you say undervolt, do you mean you undervolted the GPU memory directly or the GPU itself?
 
Ah ok, well if the temps are fine then yes I would say it was safe to use, it does as you say look flat as well it's just slightly wonky. It is quite easy to install these AIO's though, you could take the water block off and position the backplate correctly and reinstall it. If you did I wouldn't bother reapplying the thermal paste, but you'd just need to be careful when you take it off not to get it everywhere.

Thanks, I have a Palit 3080, my memory temps are 81C. That doesn't sound too bad, some 3080's apparently have been running at 105C on the memory. When you say undervolt, do you mean you undervolted the GPU memory directly or the GPU itself?

On the basis that for the plates to be fully screwed in, it would mean the pump is flat however it is at an angle but still..I have a few friends who open up and clean their pcs on a weekly basis etc so I can ask them if worse comes to worst.

Just to inform you, Micron memory chips are completely safe and function between 0-95c. From 95-105 it will throttle and anything after 110 becomes permenantly damaged;
View: https://www.youtube.com/watch?v=ph4_Sr8YNXI


I undervolted the GPU core clock; 775mV x 1700MHz
I have not touched the memory clock of the GPU, that is the default stock setting. *does this count as an overclock at all if I kept the memory clock the same but changed the core clock?
 
Last edited:
I have a Palit 3080, my memory temps are 81C
even though it may not be dangerous i would consider that unacceptable.
i'd be working to lower that for sure.
i've never had a card's memory junction over 78°C with over 20+ years of building and over 10+ years of overclocking GPUs and i'm in Florida where it's 90-100°F 6-8 months out of the year.
currently have a game running in the background as i'm browsing here and my 3080 Ti is still only hitting 65°C.

do you just have terrible airflow throughout the case or do you keep the fans running at like 500RPMs?
 
even though it may not be dangerous i would consider that unacceptable.
i'd be working to lower that for sure.
i've never had a card's memory junction over 78°C with over 20+ years of building and over 10+ years of overclocking GPUs and i'm in Florida where it's 90-100°F 6-8 months out of the year.
currently have a game running in the background as i'm browsing here and my 3080 Ti is still only hitting 65°C.

do you just have terrible airflow throughout the case or do you keep the fans running at like 500RPMs?

View: https://www.youtube.com/watch?v=ToTWaZtGOj8


This video explains the issue for the 80s & 90s

In my case I have 3 120mm corsair fans at the front, and the two AIO cooler fans exhausting from the top. I have no fan for the back though. Should I put there as an exhaust?

View: https://imgur.com/a/tPyZsiE
 
Micron memory chips are completely safe and function between 0-95c.
and most CPUs are safe up to high 90s but why allow them to when you can keep them much cooler?

my main reason for keeping all temperatures as low as possible to is to keep the ambient temp around the system as low as possible and doing it while still keeping noise relatively low has become my main focus.
i'd rather not be sitting next to a heater on my desk after sweating my butt off at work all afternoon.
pulling in as much cool air as possible also keeps the exhausted air much lower.

but if someone else is happy with their own setup that's all that really matters in the end.

*and i don't care what some idiot YouTuber states or achieves.
i've got two newer ASUS cards, RX 6700 XT & RTX 3080 Ti, and neither reaches those max temps ever.
if they are mentioning Nvidia specifically then they are referring to FE cards, not 3rd party versions.
 
Last edited:
and most CPUs are safe up to high 90s but why allow them to when you can keep them much cooler?

my main reason for keeping all temperatures as low as possible to is to keep the ambient temp around the system as low as possible and doing it while still keeping noise relatively low has become my main focus.
i'd rather not be sitting next to a heater on my desk after sweating my butt off at work all afternoon.
pulling in as much cool air as possible also keeps the exhausted air much lower.

but if someone else is happy with their own setup that's all that really matters in the end.

That is why I did the undervolting plus the heatsinks and fans!

Anything above the 70s and im uncomfortable.
 
  • Like
Reactions: JohnBonhamsGhost
even though it may not be dangerous i would consider that unacceptable.
i'd be working to lower that for sure.
i've never had a card's memory junction over 78°C with over 20+ years of building and over 10+ years of overclocking GPUs and i'm in Florida where it's 90-100°F 6-8 months out of the year.
currently have a game running in the background as i'm browsing here and my 3080 Ti is still only hitting 65°C.

do you just have terrible airflow throughout the case or do you keep the fans running at like 500RPMs?
I only checked because I read articles about 3080's having extremely hot GDDR6X like 100C common on a number of cards. I've never had a card with memory that hot either but this seems specific to GDDR6X.
My GPU temp is quite modest, 65C in Cyberpunk with max graphics including RT.

No I don't have poor airflow, I have a Fractal Meshify C, cables are all nice and tidy with a 140mm fan right in front of the GPU. I agree with you I don't like it that hot, I've since downclocked the memory and it's at 76C. The 81C I quoted was at stock when running Cyberpunk.

Not 100% sure on the fan speed, can't remember which fan was plugged in where. I believe 728 and 933 RPM for the front ones at the moment.
 
I only checked because I read articles about 3080's having extremely hot GDDR6X like 100C common on a number of cards. I've never had a card with memory that hot either but this seems specific to GDDR6X.
My GPU temp is quite modest, 65C in Cyberpunk with max graphics including RT.

No I don't have poor airflow, I have a Fractal Meshify C, cables are all nice and tidy with a 140mm fan right in front of the GPU. I agree with you I don't like it that hot, I've since downclocked the memory and it's at 76C. The 81C I quoted was at stock when running Cyberpunk.

Not 100% sure on the fan speed, can't remember which fan was plugged in where. I believe 728 and 933 RPM for the front ones at the moment.

Sorry these temps you are referring to are memory junction temps or the GPU temp?

Also to answer your question earlier:

I undervolted the GPU core clock; 775mV x 1700MHz
I have not touched the memory clock of the GPU, that is the default stock setting. *does this count as an overclock at all if I kept the memory clock the same but changed the core clock?
 
Sorry these temps you are referring to are memory junction temps or the GPU temp?

Also to answer your question earlier:

I undervolted the GPU core clock; 775mV x 1700MHz
I have not touched the memory clock of the GPU, that is the default stock setting. *does this count as an overclock at all if I kept the memory clock the same but changed the core clock?
I was referring to the memory temps. My GPU doesn't get that hot, I have my GPU running at stock clocks, but I pulled back the memory clock speeds from 9500Mhz to 8500Mhz, that lowered temps from 81C to 76C.
 
I was referring to the memory temps. My GPU doesn't get that hot, I have my GPU running at stock clocks, but I pulled back the memory clock speeds from 9500Mhz to 8500Mhz, that lowered temps from 81C to 76C.

I believe playing with lower frames also reduces the temperatures too.

Sorry to mention it again but, is there an overclock if the memory clock remains stock but I just reduced the core clock?
 
I believe playing with lower frames also reduces the temperatures too.

Sorry to mention it again but, is there an overclock if the memory clock remains stock but I just reduced the core clock?

That would be underclocking not overclocking. I mean I wouldn't mention it if you ever had to RMA the card but reducing clocks and voltages has no chance of damaging anything. If you were increasing the voltage and clock speed of either the memory or the GPU itself that would be considered an overclock, that has the potential to cause abnormal wear.
 
That would be underclocking not overclocking. I mean I wouldn't mention it if you ever had to RMA the card but reducing clocks and voltages has no chance of damaging anything. If you were increasing the voltage and clock speed of either the memory or the GPU itself that would be considered an overclock, that has the potential to cause abnormal wear.

Oh ok thank you.

Thank you all, you guys have been really helpful here!

If somebody can help in regards to improving GPU temps with fans etc that would be amazing!

I ordered a Noctua 120mm PWM fan to install on the back of the case as I dont have a fan for the back.
 
What would you suggest by looking at the photo?..
If somebody can help in regards to improving GPU temps with fans etc that would be amazing
just what i mentioned earlier;
a PCIe fan bracket for 2x 80mm intakes
or a bottom case 120/140mm intake if the case provides adequate airflow directly into the GPU.

how are all of the fans setup in the case currently
and which case is it?
 
depending on the make & model of fans that may not be very much cool air coming in.

i let my Silent Wings intakes max @ ~1050RPM but they have higher than average airflow & air pressure ratings so the three front plus one bottom provide an awful lot of cool air intake.
The fans I have are:

There's also a Noctua 80mm fan in the hard drive bay.
 
fans I have are:
the 140mm GP-14 are rated with pretty low airflow & air pressure.
may as well let them run @ full speed(1000RPM).

the 120mm GP-12 offer slightly better specs.
i would also max those @ full speed(1200RPM).

if you ever get the urge to replace them, some
be quiet! Silent Wings 3 High-Speed PWM
or Noctua NF-A14(140mm) or A12(120mm) would do you much better.