• 22.08.2025, 15:30
  • Registrieren
  • Anmelden

  • Sie sind nicht angemeldet.

 

Leakshield and kryographics next rtx 2080 ti: pressure drops when GPU heats

Flop Ysh

Junior Member

Leakshield and kryographics next rtx 2080 ti: pressure drops when GPU heats

Freitag, 13. August 2021, 15:47

Hello
I have a strange behaviour. When the GPU is heating due to load (standard game), the pressure drops then the LS has to run frequently. See the attached screenshot. That phenomenon is not happening when the CPU us heating the liquid. Water reaches 48-50°C. (since my fan header of the D5 Next doesn't work I can't adjust the fan&pump rpms to the water temp yet)

Moreover, water bubbles can be seen on the GPU WB (mounted vertically) at the top left, while the loop has been drained several times (see photo). It sounds like the waterblock has not been perfectly sealed and when the temperature rises, the oring and/or the mechanical constrains makes the WB less sealed, and only the 'very' low pressure applied by the LS can identify it.
if the operating pressure is set at around 250mbar, that issue doesn't occur. And it seems that, when I hadn't the LS yet, the water was spreading a bit beyond where the PMMA contacts, up to the oring, at the exact same spot.

Has anyone a Leakshield AND a kryographics next WB? and if yes, experienced that kind of pressure drop? Or am I experiencing again a fourth issue with AQ products?
»Flop Ysh« hat folgende Dateien angehängt:

Remayz

Senior Member

Freitag, 13. August 2021, 16:04

you water is WAY too hot. that's one thing, but it would create high pressure..
What's weird in your photos, besides a dangling strip :p is the little bubbles we can see on the oring sealing surface towards the bottom left. worth checking if all the screws are correctly tightened around the acrylic face of the waterblock.

Flop Ysh

Junior Member

Freitag, 13. August 2021, 16:23

Thank you for your reply
I know that's quite high but, hey, that's summer ^^ and I can't adjust the fans to the water temp yet. And I've planned to add more intakes too.


good catch about the ones at the bottom left of the pic! everything is pointing to a faulty WB So again a quality issue with AquaComputer products...

that's just the third timeI've installed the loop in July. I've got:-a D5 Next pump with the fan header not working
-a kryographics connection terminal too tight for the VISION to fit
-now a kryographics next WB not tighten enough.


Wunderbar

Plus the fact I've paid a replacement screen for nothing! And they don't want to make me a discount on a Quadro to actually make than fan control on water temp as planed even after all those issues :!:

Remayz

Senior Member

Freitag, 13. August 2021, 16:37

you'd have to check the screw tightness :) that's what it looks like to me but if it's just that, it's solved in a couple of seconds.
Usually you can see the contact surface between the acrylic and the squished oring, a dark black surface with no bubbles in between. No need to tighten hard. As long as the acrylic is touching the copper block, the oring will be pressed enough

RobinXe

Junior Member

Freitag, 13. August 2021, 16:53

Hopefully this is not applicable, but under 1/2 bar vacuum the boiling point of water is down to about 80 degrees C. If water is heated to this temperature locally then it would 'boil', which would introduce vapour to the system. This would increase the pressure/decrease the vacuum of the system.

Flop Ysh

Junior Member

Freitag, 13. August 2021, 18:22

yes indeed

at 550 mbar absolute which is around 450mbar relative displayed by LS, pure water boils at 84.1 °C

ingame, when pressure drops, the hotspot temp indicated by the GPU itself says 75°C, quite steady

there is a bit of a margin but actually that was an interesting thought Thx

Dieser Beitrag wurde bereits 1 mal editiert, zuletzt von »Flop Ysh« (13. August 2021, 18:26)

Remayz

Senior Member

Freitag, 13. August 2021, 19:10

if it came to boiling you'd have more pressing problems hehe
75 is air cooling temperatures, so there's work to do there :p the water temp is usually not much below the GPU water block temperature, so that's a big delta you have here. Maybe not enough flow to start with. i wouldn't go far below 50% pwm on any kind of loop just because of that.
These waterblocks are pretty good so i would expect to see 10-15°C difference between water and GPU temp under load.
That big bubble, plus your 25° temp differential = maybe low flow.

Flop Ysh

Junior Member

Freitag, 13. August 2021, 19:26

that's the hot spot from HWinfo , which were around 90°C with air cooling and can reach +100°C.
https://www.reddit.com/r/nvidia/comments…ks_like_in_the/ https://www.techpowerup.com/forums/threa…eadings.281908/


The gpu 'standard' temp is at 55 so very close to the water one so that's ok

Ähnliche Themen