Question Ambient Temps and Overclock Stability -- like day and night?

[LordTimothyDexter]

Hello! I am not new to overclocking, but I don't know everything. I have my suspicions regarding the answer to my own question, but I was hoping you guys could help with maintaining my sanity =)

I recently got a stable overclock on my 10900k to 5.1ghz all core. It passed 2 hours of RealBench (using 16GB ram, not 32 because it won't allow), P95, and was working great. All of a sudden, I am getting L0 Cache error in HWinfo64 when running RealBench today. Here's the info for my stable overclock (at the time it passed 2 hours RB)

10900k
Asus Z490-E gaming Mobo
2x16gb G.Skill Trident RGB 3600 CL18 Ram with XMP II enabled
Arctic LF 420 AIO cooler in push pull
Offset mode
SVID: Best Case
LLC = 5
vCore offset = +.035
V/F offset at 51x = +0.030
BCLK aware adaptive = disbable
BCLK Spread = disabled
MCE = Enabled
All other settings in BIOS are Auto

Voltage under load in RealBench (typically the same in P95 small FFT w/AVX disabled) = 1.332v
Temp under load in this scenario: Average 78/79C with a max of 82C
Ambient temps were about 23C when stress testing was passed, because I did it all at night.

Today, during the day when its warmer, I am running RB again, and within 10 minutes I get a L0 Cache WHEA error in HWinfo64.

The only difference in today is the ambient temperature. It was 23-ish when stable, today it is 28.6C in here. This affected my temperature in RealBench testing as well, since today its averaging about 82/83C with a max of 85/86C during the test.

Could the 5 degree difference in ambient temperature take my overclock from stable to unstable? My guess is that, when it was 23C, the settings I had were right on the edge of stability. Did the change in ambient temps tip the scale over to the side of instability since it's now averaging about 4-5C hotter?

 

[LordTimothyDexter]

After doing more research, I'm seeing that this can indeed affect stability. 80C is probably that area where going above means you'll see a greater need for more voltage to be stable compared to what you needed at cooler temperatures. And my ambient change moved temps from High 70s to low 80s. That 1.332v under load in RB is not enough to get it done during the day when ambient temps are high. But once night comes and its around 22C, my overclock is stable. So this begs the new question from me: if this is true, should I bother upping the voltage or should I just not run stress tests when its hotter than 27C (80 farenheit) in the room? I suppose I should determine how much additional voltage is actually needed to be stable during the warmer days when running Realbench.