[SOLVED] Increase Minimum Temp Limit

[Benjimangos]

Hello, I have an Gigabyte RTX 2080 Gaming OC and when it reaches its "Minimum" temperature limit, the fans max out which is quite loud and unnecessary.

I have tried a custom fan curve and increasing the temp limit on Afterburner, Firestorm and GPU Tweak 2 separately to no avail. I have tried a different PCI-E slot, reconnecting the fan headers and I also used DDU to do a clean install of graphics drivers and also tried switching to an older graphics driver all with still the same issue.

On GPU-Z in the advanced tab under temperature limit it looks like this,
Current: 83.0 c
Minimum: 65.0 c
Default: 83.0 c
Maximum: 88.0 c
Adjustment Range: -22% to +6%

When I adjust the temperature limit in any of the overclocking software it only adjusts the "Current" temperature limit value. The card also has close to zero dust on it.

Any input is greatly appreciated.

 

[Phaaze88]

Minimum: 65.0 c

What this means is that you can't set the gpu's temperature limit any LOWER than 65C via software. It's built into the vbios.
Same deal with how it's not possible to set Maximum over 88C.
The fans shouldn't kick into full gear until 83C, but they're probably going bonkers at lower temperatures for a different reason.


How comfortable are you with flashing bios?
https://www.gigabyte.com/us/Graphics-Card/GV-N2080GAMING-OC-8GC/support#support-dl-bios
Gigabyte has 3 different bios updates here. 2 of them contain updates to the fan curves, but they are only compatible with certain bioses, which you'll see listed at #5.
To check which bios your gpu is currently on: Gpu-Z Advanced tab, dropdown box to Nvidia Bios, find message and you'll see something like this:
GV-N108TGAMING OC-11GD/F1/0543 (what's highlighted in red is the bios version)

Instructions on how to flash are further down the page in the hyperlinked 'How to reflash bios?'
A number of their 20 series cards had problems with rogue fan curves and they made patches with vbios updates. Your issue may be related to that.
If not that, then crazy high gpu hotspot temperature comes to mind. If you need an example of what a high hotspot is - around 100C, with a thermal limit of 110C.

 

[Benjimangos]

I installed F13 through the Aorus engine just a bit ago. I also tried from that website just now and it says This BIOS version does not match. This unfortunately has not alleviated the issue. The BIOS version in GPU-Z says 90.04.23.40.6F and next to "message" it looks like this: GV-N2080GAMING OC-8GC/F13/0807

I ran a benchmark again but this time looking at GPU hotspot temperature using HWiNFO64 and it reached 107.2 C. When it gets around 106.8 C the fans max out. Could this be a thermal compound issue? It looks like it has thermal pads right now.

 

[Phaaze88]

Hotspot is the hottest sensor off the gpu die. The gpu core reading we're all used to seeing is more of an average.

The high hotspot can mean a couple of things:
A)Uneven paste spread, or paste has possibly dried up. Obviously, you need to repaste the die here, but there's a need to be a bit more selective in which pastes to use. Thicker and stickier - higher viscosity - pastes are ideal.
NT-H2, Prolimatech PK-3, Gelid GC Extreme, and Kingpin KPx are a few of those. The highly praised TG Kryonaut and NT-H1, for example, are not so good for these applications.

B)Poor cooler cold plate to die contact due to significant gpu sag. There's anti-sag brackets for those.

 

[Benjimangos]

Its definitely a paste issue then. My gpu is vertical. Thanks a bunch for your help!