I noticed my RAM junction temperatures on the stock RTX 3080 FTW3 Ultra were relatively poor - regularly reaching 95°C. I originally wanted the 3080 FTW3 Hybrid, so I purchased the hybrid cooler as soon as I saw it go in stock.
 
After install, I see my GPU temperatures have predictably dropped - but my RAM junction temperature has increased. It can now exceed 100°C under load. What are my options at this point? EVGA doesn't seem to supply thermal pads, and the stock thermal pads seem to be nonstandard sizes.
 
Any ideas on how I can get the hybrid temperatures down? If not, how can I reinstall the stock cooler since I can't obtain thermal pads with the correct dimensions?

I have the exact same issue. I've just converted my RTX 3080 FTW3 Ultra to Hybrid a few hours ago, everything looks ok except for the Memory Junction Temperature that's being read by HWINFO.
 
 

Wow just saw this. Looks like I am not buying the hybrid kit unless EVGA has another version that helps lower VRM temp.

RMV07
I have the exact same issue. I've just converted my RTX 3080 FTW3 Ultra to Hybrid a few hours ago, everything looks ok except for the Memory Junction Temperature that's being read by HWINFO.
 
 

Same issue as you with 3090 Hybrid. Someone on reddit mentioned about getting Gelid Ultimate thermal pads to put on that dropped their junction to 88C.
 
Thread here: https://old.reddit.com/r/...g_thermal_pads_on_the/

rustyjn
Ok, well darn. I just ordered the hybrid kit an hour ago for my 3080 FTW ultra. The main reason was to get lower Vram temps. Was that a mistake?

I can tell you my results, if nothing else. This is ETH mining tests.
 
3080 FTW3 Ultra Gaming
+1135 memory clock (absolute clock 10386)
731mv core voltage
Runs around 220w
Around 98MH/s.
 
When on air: memory junction at 90c at 60% fan. Quiet. Efficient. Don't have to increase case fan speeds.
 
Hybrid: memory junction at 96c at with VRM fan at 75%. Hybrid rad in a push/pull config, blowing out the case. Doesn't really change anything for VRAM temps if I increase rad fan speeds beyond 50% or so. That's after I added some 3mm thermal backs to the back to try and let some of the heat get dissipated to the backplate. (Haven't changed pads in the card itself or added addition heatsinks/fans outside the card yet.)
 
Overall, it's been a ****ty experience. Louder when gaming. Louder when mining. Cooling was perfect on air. I'll probably go back to it rather than try to heckin' paste on heatsinks to the back of my card.
 
It's louder.

EVGATech_AnthonyS
I have actually sent you an email about this.

I’d like to know what it said because I ordered one of these for the exact reason of lowering vram temps and sure don’t want to go through it all if it not going to improve them but actually worsen them

northo00

rustyjn
Ok, thanks for that info. Probably gonna try and cancel my order. Sounds like a waste of time and actually going in the other direction. If not using this cooler what the next best step I should take to get the vram temps down you think?

I plan on custom looping for the first time except there's been reports of cards dying on EK water blocks. Not sure of other brands.

Other than that probably not much other than placing a fan and heatsinks on the backplate and run it as normal?

Yep, I think I’m just going to add some
Thermal pad to the backplate and see what that does and then maybe try some heatsinks if need and stop at that point regardless. Thanks for the help

While it can be alarming to see the memory junction hit 90+ degrees Celsius, it is specced to run at 125 degrees Celsius before encountering problems. You'll typically be hard pressed to find better memory junction temps, as other brands are having the same temps. Hope this was able to clear some information up.