Question New Ryzen 9 7950x Died After Thermal Paste Upgrade (HELP)

[THEPASHKA]

I have upgraded my pc to a R9 7950x with Gigabyte X670E Aorus Master and Corsair 32gb (2x8) DDR5

During initial build/upgrade (2 weeks or so ago). I found that I have ran out of thermal paste. There was a bit left, but I didn't feel like it was enough, despite that I decided to apply what's left and buy some new thermal paste later.


PC worked fine, temps ware a normal for that CPU at 100% it was at 95c while boosting to 5.35 GHZ (Running with AIO), had some issues with AMD's EXPO mode, to get RAM running at 5600mhz. After I have enabled it, It booted showing only 24gb instead of 32gb. After I did a reboot (It was long, around 5 - 15 min) and re-enabled EXPO, and it was working correctly at 5600mhz. (I did hear a low pitch noise from RAM sticks under load and sometimes in idle)

So a few days ago finally ordered well described Noctua NT-H2 paste.

It arrived yesterday, today I woke up, I booted up the PC for around 10 min and turned it off again. Waited for about 30 min for it to cool down (just in case). Removed the AIO cooler, cleared old thermal paste, applied new one, assembled everything back, turned it on, and immediately noticed that I didn't plug power & data cables for AIO Pump. Quickly powered off the system (it was running like that for less than 5 seconds).
Made sure everything is plugged in. Turned it on again, fans started spinning, everything looked normal, but there wasn't any post codes (I do not remember, if there was any during first "aborted" boot)
Onboard numeric display did not show any POST codes, and Q-LED was switching from DRAM to CPU for 1 second, every 10 - 20 secs. After a few minutes in this loop, it goes back to DRAM and stops switching.


I was careful when applying thermal paste,
I'm sure I did not over tightened the AIO,
I have tried using every single DRAM slot with both ram sticks, (except 1st one, AIO pipes covers it),
I have q-flashed the BIOS to latest version,
I have made sure all the connectors on RAM and ram slots are clear,
I have made sure all other connectors are connected properly,
I have tried using clear CMOS button & jumper,
I have tried unplugging the GPU,
I have tried booting completely without any ram sticks, (Same Q-LED indication as booting with ram).

I have pulled the board out of the case, and tried booting outside (Not sure what I was expecting, but yea..)
I have even gone to disassembling the board's backplate, to gain access to the CMOS battery, but the backplate have tones of thermal compound between it and the actual board, and I thought I might lose warranty if I started pulling that apart despite none of the screws having any "Void if removed" stickers.

Tried booting without any coolers attached, so I could feel the CPU, but it wasn't even warming up (I have touched it on the sides, and it was only running for less than 30 sec)

USB devices do light up, if trying to run the board completely without CPU, it doesn't even attempt to start.

I still think, it might be dead RAM. I did have problems with it, and I did hear low pitch noise when it was working, that noise was really noticeable when under load / stress testing. But I didn't think much of it, since DDR5 has a voltage controller thing on the actual stick that might have been emitting that noise

So, what should I do, other than going through RMA?
Or what have I done wrong, what could have possibly killed it?


P.s. This is way far from my first time upgrading components, replacing thermal paste etc... But never had anything similar.

 

[CountMike]

I have upgraded my pc to a R9 7950x with Gigabyte X670E Aorus Master and Corsair 32gb (2x8) DDR5

During initial build/upgrade (2 weeks or so ago). I found that I have ran out of thermal paste. There was a bit left, but I didn't feel like it was enough, despite that I decided to apply what's left and buy some new thermal paste later.


PC worked fine, temps ware a normal for that CPU at 100% it was at 95c while boosting to 5.35 GHZ (Running with AIO), had some issues with AMD's EXPO mode, to get RAM running at 5600mhz. After I have enabled it, It booted showing only 24gb instead of 32gb. After I did a reboot (It was long, around 5 - 15 min) and re-enabled EXPO, and it was working correctly at 5600mhz. (I did hear a low pitch noise from RAM sticks under load and sometimes in idle)

So a few days ago finally ordered well described Noctua NT-H2 paste.

It arrived yesterday, today I woke up, I booted up the PC for around 10 min and turned it off again. Waited for about 30 min for it to cool down (just in case). Removed the AIO cooler, cleared old thermal paste, applied new one, assembled everything back, turned it on, and immediately noticed that I didn't plug power & data cables for AIO Pump. Quickly powered off the system (it was running like that for less than 5 seconds).
Made sure everything is plugged in. Turned it on again, fans started spinning, everything looked normal, but there wasn't any post codes (I do not remember, if there was any during first "aborted" boot)
Onboard numeric display did not show any POST codes, and Q-LED was switching from DRAM to CPU for 1 second, every 10 - 20 secs. After a few minutes in this loop, it goes back to DRAM and stops switching.


I was careful when applying thermal paste,
I'm sure I did not over tightened the AIO,
I have tried using every single DRAM slot with both ram sticks, (except 1st one, AIO pipes covers it),
I have q-flashed the BIOS to latest version,
I have made sure all the connectors on RAM and ram slots are clear,
I have made sure all other connectors are connected properly,
I have tried using clear CMOS button & jumper,
I have tried unplugging the GPU,
I have tried booting completely without any ram sticks, (Same Q-LED indication as booting with ram).

I have pulled the board out of the case, and tried booting outside (Not sure what I was expecting, but yea..)
I have even gone to disassembling the board's backplate, to gain access to the CMOS battery, but the backplate have tones of thermal compound between it and the actual board, and I thought I might lose warranty if I started pulling that apart despite none of the screws having any "Void if removed" stickers.

Tried booting without any coolers attached, so I could feel the CPU, but it wasn't even warming up (I have touched it on the sides, and it was only running for less than 30 sec)

USB devices do light up, if trying to run the board completely without CPU, it doesn't even attempt to start.

I still think, it might be dead RAM. I did have problems with it, and I did hear low pitch noise when it was working, that noise was really noticeable when under load / stress testing. But I didn't think much of it, since DDR5 has a voltage controller thing on the actual stick that might have been emitting that noise

So, what should I do, other than going through RMA?
Or what have I done wrong, what could have possibly killed it?


P.s. This is way far from my first time upgrading components, replacing thermal paste etc... But never had anything similar.

Does not mean that CPU is dead but maybe you should RMA it. Until you are 100% sure it's 100% working all troubleshooting will not be valid,
That type of socket is very sensitive just like with Intel, should be checked with magnifying glass Even one pin in wrong place bent and it won't work, You also have to have at lest some fan connected to CPU_FAN or BIOS would refuse to start it,

 

[THEPASHKA]

Does not mean that CPU is dead but maybe you should RMA it. Until you are 100% sure it's 100% working all troubleshooting will not be valid,
That type of socket is very sensitive just like with Intel, should be checked with magnifying glass Even one pin in wrong place bent and it won't work, You also have to have at lest some fan connected to CPU_FAN or BIOS would refuse to start it,

Thanks for the response!
Surprisingly, I did inspect pins with my phone's macro lens, and I did have AIO fans hooked up to CPU_FAN header. The main problem, the POST sceen should be showing something, 00 is for dead CPU etc... but it's just turned off showing nothing

 

[Noedell]

Try removing the CMOS battery from the motherboard and put it back in. I had the same problem once after trying EVERYTHING. Clearing the CMOS didn't work. This is a know issue on Gigabyte motherboards.

Please let me know if this works. Why I am specifically asking this is because I also thought this was ridiculous, but it actually worked like a charm.

 

[guilhermeol]

Remove ram, power without it, put ram back again and try it.

 

[THEPASHKA]

Try removing the CMOS battery from the motherboard and put it back in. I had the same problem once after trying EVERYTHING. Clearing the CMOS didn't work. This is a know issue on Gigabyte motherboards.

Please let me know if this works. Why I am specifically asking this is because I also thought this was ridiculous, but it actually worked like a charm.

Thank you for the suggestion. I thought of that, my old PC (also with gigabyte board) it didn't boot, don't really remember why exactly, but removing battery did help

But wouldn't it void my warranty? It sounds stupid Ik, but. On my board the CMOS battery is located under a front shield / heatsink, that covers 2 PCIE controller chips, and that shield is screwed from the back of the board... And yea, the back also has a massive aluminium plate, that covers it, between that plate and the board there are tones of thermal pads. When I've tried pulling that apart (gently) it felt like it being stuck to the board due to them thermal pads.

 

[THEPASHKA]

Try removing the CMOS battery from the motherboard and put it back in. I had the same problem once after trying EVERYTHING. Clearing the CMOS didn't work. This is a know issue on Gigabyte motherboards.

Please let me know if this works. Why I am specifically asking this is because I also thought this was ridiculous, but it actually worked like a charm.

Did manage to remove the CMOS battery, waited over 30 min and still nothing.
However, I have noticed that CPU does get warm a little bit, I am able to hold a finger on it, so possibly the RAM is dead