Question Very weird POST issue

[plee82]

I have been having a very weird POST issue.

My specs (all components 1 month old):
PSU: ToughPower GF1 750W Gold
Mobo: B550 Aorus Pro (latest bios)
CPU: 3700x Ryzen 7
GPU: RTX 3070 Asus Dual OC
RAM: G.Skill RipJaws V series 16GB 3600 CL16

If I play a GPU intensive game and then power off the PC normally through windows and try to power on the PC after 5 secs, POST fails with CPU and RAM debug lights flashing back and forth. The CPU led will clear and then move to DRAM, it stays there for 2 secs and then moves back to the CPU led and back to RAM, and keeps repeating. If instead of Powering On the PC right away after shutoff I wait 2-3 minutes, no issues. Now, If I don't do anything that's intensive, such as browsing, youtube, etc, and power off the PC and power on right away, no POST issues. None of these issues happen if I just restart Windows, no POST issues at all regardless of what games I was playing. My temps when gaming for GPU are 71C and CPU 70C. I can game all day without issues, hell I can run cyberpunk and then run cinebench r20 in the background and I have no issues with the PSU. I would not have found this issue if it was not for an email I forgot to write after a gaming session. I shut off my PC and then remember I had to write an email and turn it on right away and noticed the POST issue. I have been testing to see what triggers it and it seems to be consistent with having a GPU intensive game running for a while before shut off. If I play a GPU intensive game and then quit the game and just browse for half an hour and then shut down the PC and turn it back on, no issues.

 

[RealBeast]

This will be pretty hard to hunt down IMO. How about you make it a moot point by just having a blank screen saver at 5 minutes and leave it on more, shutting it off when you won't use it again for the day?

I have several machines that get restarted a maximum of once a month and the monitors turn off at 15 minutes and I've never had any issues.

 

[plee82]

This will be pretty hard to hunt down IMO. How about you make it a moot point by just having a blank screen saver at 5 minutes and leave it on more, shutting it off when you won't use it again for the day?

I have several machines that get restarted a maximum of once a month and the monitors turn off at 15 minutes and I've never had any issues.

Yup, I am at lost. I initially thought it was the RAM, so I disabled XMP but the issue still happened. Could it be heat? The GPU is blowing directly into the cmos battery, but I never heard issues related to that.

 

[RealBeast]

Yup, I am at lost. I initially thought it was the RAM, so I disabled XMP but the issue still happened. Could it be heat? The GPU is blowing directly into the cmos battery, but I never heard issues related to that.

Yeah, I've not seen that in the 30+ years that I've been building lots of machines, which is why I think that it will be very difficult to figure out -- and I have a dozen machines and boxes of parts. Theoretically, it could be that shutdown immediately after the game turns off the GPU fans and they don't have time to cool it down so require a bit of time to drop back below their thermal limit. But who knows.

 

[plee82]

Yeah, I've not seen that in the 30+ years that I've been building lots of machines, which is why I think that it will be very difficult to figure out -- and I have a dozen machines and boxes of parts. Theoretically, it could be that shutdown immediately after the game turns off the GPU fans and they don't have time to cool it down so require a bit of time to drop back below their thermal limit. But who knows.

Yeah at this point I will just live with it. I even switched off the PSU, and pressed the power switch on the case to discharge the capacitors to see if it works right away but NO, I have to wait regardless after a heavy gaming session lol.

 

[plee82]

Yeah, I've not seen that in the 30+ years that I've been building lots of machines, which is why I think that it will be very difficult to figure out -- and I have a dozen machines and boxes of parts. Theoretically, it could be that shutdown immediately after the game turns off the GPU fans and they don't have time to cool it down so require a bit of time to drop back below their thermal limit. But who knows.

Just an update. Contacted Thermaltake and explained my issue. They told me to RMA the PSU. Very interesting, surprised they decided this just by the description of my issue.

 

[RealBeast]

Just an update. Contacted Thermaltake and explained my issue. They told me to RMA the PSU. Very interesting, surprised they decided this just by the description of my issue.

Nice! That's the kind of service that makes me recommend stuff.

 

[plee82]

Nice! That's the kind of service that makes me recommend stuff.

Welp, happens with a new PSU as well lol. This is going to remain as a mystery lol.

 

[plee82]

Nice! That's the kind of service that makes me recommend stuff.

I know this is very old now but I just wanted to give my final update. I received a new Thermaltake GF1 750W PSU 3 months ago and experienced the same issue. I just gave up and kept using my PC for a few months. Two weeks ago, I decided to buy a Super Flower Leadex III 750 Gold PSU that was on sale. No more issues. I can shutdown my PC and turn it on right away, and no POST issues. I am not sure what the issue with Thermaltake ToughPower GF1 PSU's are but the Super Flower does not have this issue.

 

[DSzymborski]

It's quite surprising, actually. Thermaltake doesn't sell a lot of quality PSUs, but the Toughpowers do tend to be quality. There have been instances with 30 series GPUs cause problems in otherwise fine GPUs; these are spiky cards, power-wise, and older SeaSonic Focus GPUs and some FSP-made ones using active clamp topology have had reported issues. But this is a modern CWT platform, so it's a bit surprising.

I'm glad the Super Flower works though; I haven't had any problems with my Leadex III and a 3080. It's basically an EVGA G4 750, but Super Flower no longer manufactures EVGA's good PSUs, to EVGA's detriment. Might be able to dig in with a load tester and oscilloscope, but otherwise it's hard to track down what the issue with the Thermaltake is.

 

[plee82]

The Toughpower had zero issues except for this very weird edge case. Most people would not even notice this. My theory is that the PSU is not delivering enough power during the edge case because I noticed that the LED on my Aorus mobo does not light up when the issue gets triggered.