2nd set of ram died, any ideas?

Skidder · 23 Sep 2023 at 17:50

Hi, I have
TeamGroup Delta RGB 32GB (2X16GB) DDR5

PC5-44800C30 6000MHz Dual Channel Kit

in a
Asus ROG Strix B650-A Gaming WIFI (Socket AM5)

DDR5 ATX Motherboard

Everything was fine until a couple of months ago, first one stick of ram failed and then the second. These were RMAd and replaced.

Again everything fine for a while and then one stick of ram failed and then the other.

Won't boot with either stick in any slot, have reset CMOs etc.

Went through exactly the same process with OcUK for RMA and the symptoms are identical this time.

Is this just dumb bad lack? Is this something potentially wrong with mobo or PSU causing an issue (there doesn't seem to be any issue elsehwere in the system)

I will contact OcUK for an RMA but any tips would be great. thanks

Tetras · 23 Sep 2023 at 18:39

RAM from the same bad batch maybe?

Skidder · 24 Sep 2023 at 10:26

Tetras said:
RAM from the same bad batch maybe?

yeah could be, hoping that's the case. i do feel like switching to another manufacturer. might RMA and sell this lot

wookiee87 · 24 Sep 2023 at 12:26

Sounds like it could be your cpu to blame, very very odd for 2 sets of ram to fail in the same way, the memory is controlled via the cpu's IMC so if thats not working right no amout of kits you use will work right at expo speeds, default speed of 4800mhz may be a diffrent story, which cpu are you running?

Finners · 24 Sep 2023 at 12:32

Does that memory have temp sensors on them? sometimes visible in software like HWinfo64

Skidder · 24 Sep 2023 at 13:37

wookiee87 said:
Sounds like it could be your cpu to blame, very very odd for 2 sets of ram to fail in the same way, the memory is controlled via the cpu's IMC so if thats not working right no amout of kits you use will work right at expo speeds, default speed of 4800mhz may be a diffrent story, which cpu are you running?

7900x 3d

Skidder · 24 Sep 2023 at 13:38

Finners said:
Does that memory have temp sensors on them? sometimes visible in software like HWinfo64

not sure, i can't post at all now

Skidder · 24 Sep 2023 at 13:39

wookiee87 said:
Sounds like it could be your cpu to blame, very very odd for 2 sets of ram to fail in the same way, the memory is controlled via the cpu's IMC so if thats not working right no amout of kits you use will work right at expo speeds, default speed of 4800mhz may be a diffrent story, which cpu are you running?

would it work fine for a period and then stop though if the CPU. both sets have been fine until they weren't.

wookiee87 · 24 Sep 2023 at 13:41

Skidder said:
7900x 3d

official ram support is 5200mhz for x3d, if you get a good chip 5600mhz ddr5 may be possible, i would look at a clear cmos and see if you can boot, if so set 5200mhz and see if it works with your ram or replacment sets.

Skidder · 24 Sep 2023 at 14:08

wookiee87 said:
official ram support is 5200mhz for x3d, if you get a good chip 5600mhz ddr5 may be possible, i would look at a clear cmos and see if you can boot, if so set 5200mhz and see if it works with your ram or replacment sets.

i've cleared cmos and tried to restart but won't post. i'll give that another shot with both sticks separately. then may be get a bigger set of slower ram

wookiee87 · 24 Sep 2023 at 14:11

Skidder said:
i've cleared cmos and tried to restart but won't post. i'll give that another shot with both sticks separately. then may be get a bigger set of slower ram

if you still cant boot after a cmos clear, then take the cpu out and reinstall, it should fool the board in thinking a new cpu is installed, then you can hopefully boot and set expo at 5200mhz and see if it works.

Skidder · 24 Sep 2023 at 15:27

wookiee87 said:
if you still cant boot after a cmos clear, then take the cpu out and reinstall, it should fool the board in thinking a new cpu is installed, then you can hopefully boot and set expo at 5200mhz and see if it works

do i need to power it up with the cpu out for it to realise that it has been taken out?

wookiee87 · 24 Sep 2023 at 15:29

Skidder said:
do i need to power it up with the cpu out for it to realise that it has been taken out?

shouldn't have to as long as the cpu is out of the socket, place it back and the board should show at post screen with "new cpu installed enter setup" message
when i cleaned my loop and took my 7950x out to clean and repaste, i got new cpu installed even though i haddent changed anything, luckly all my settings were saved when i got to the bios, exited and back into windows.

when you have the cpu out, just give it a quick look over and check theres nothing bulged where the gold pads are on the reverse of the cpu, give the socket a look over too and check for any dips, dont want to worry you but with all the issues surronding x3d before the bios updates, its best to check.

do you know what kind of SOC voltages you were running before the problems started?

Skidder · 24 Sep 2023 at 15:40

wookiee87 said:
shouldn't have to as long as the cpu is out of the socket, place it back and the board should show at post screen with "new cpu installed enter setup" message
when i cleaned my loop and took my 7950x out to clean and repaste, i got new cpu installed even though i haddent changed anything, luckly all my settings were saved when i got to the bios, exited and back into windows.

when you have the cpu out, just give it a quick look over and check theres nothing bulged where the gold pads are on the reverse of the cpu, give the socket a look over too and check for any dips, dont want to worry you but with all the issues surronding x3d before the bios updates, its best to check.

do you know what kind of SOC voltages you were running before the problems started?

thank you will try that

re voltages i think just on auto, i haven't got around to much fiddling yet

wookiee87 · 24 Sep 2023 at 15:45

Skidder said:
thank you will try that

re voltages i think just on auto, i haven't got around to much fiddling yet

the soc on auto is probally way too high, x3d via bios updates locked down SOC to 1.3v to stop the burn out problems, may explain why the ram isnt working, the IMC on the cpu may be getting damaged or already has been damaged, if you can get it to boot set expo as normal but under ddr5 speed change from 6000mhz to 5200mhz and lower down in the voltage area find CPU SOC voltgae change from auto to manual and in the box just below enter 1.3v max, sometimes you can use lower but its not 100%, best is 1.2 - 1.25v ideally but upto 1.3v is the very max for x3d

Blackjack Davy · 24 Sep 2023 at 19:03

Is the voltage high and more to the point how hot do they get? There no sensors on RAM sticks that I know of so most people don't give it a second thought and samsung b-die is particularly sensitive to temperature in my experience if they get hot to the touch chances are they're going fail sooner or later had two sticks of said die on me one after the other and it wasn't even the voltage it was the heat from an FE card blowing all over them

wookiee87 · 24 Sep 2023 at 19:30

Blackjack Davy said:
Is the voltage high and more to the point how hot do they get? There no sensors on RAM sticks that I know of so most people don't give it a second thought and samsung b-die is particularly sensitive to temperature in my experience if they get hot to the touch chances are they're going fail sooner or later had two sticks of said die on me one after the other and it wasn't even the voltage it was the heat from an FE card blowing all over them

hwinfo is a good tool if your dimms have temp readouts, my dominators do and i have seen peak temps of 63.8c for both sticks, i run 1.4v and oc to 6200mhz, tbh if i see temps in the 70's i will look to active cooling which should bring down temps a fair bit but for now everything is ok, a bit warm at 64c, but working fine with no issues.

The Asgard · 27 Sep 2023 at 08:10

I wouldn’t trust auto on ram voltages. You best bet is to find out what type of chips you have, use Hwinfo timing. Then have a look for the standard voltages people are using.

I picked up 2*32GB Hynix A-Die yesterday and just playing with the timings but I have 6000 32,36,36,30 with tight sub timings.

SOC 1.1
Ram Voltage 1.35
VDDQ 1.25
VPP 1.8
VDDIO 1.25

These voltages are safe if your system is stable. YMMV. You can work from there to potentially bring the ram voltage down, but depends on your chips. SOC is low so you may have to bring that up to 1.2

With regard to heat just run Cinebench R24 and feel how hot they get. Mine just get warm.

Skidder · 3 Oct 2023 at 10:04

Have just ordered some replacement ram. I will do the RMA and sell that set but need to get up and running after being away.

The ram I have ordered is AMD Expo compatible. If my bios and drivers are up to date should I be alright using Expo or still cautious?

I will check the processor before I rebuild.

Cheers

wookiee87 · 3 Oct 2023 at 10:23

Skidder said:
Have just ordered some replacement ram. I will do the RMA and sell that set but need to get up and running after being away.

The ram I have ordered is AMD Expo compatible. If my bios and drivers are up to date should I be alright using Expo or still cautious?

I will check the processor before I rebuild.

Cheers

if all your drivers are upto date and your bios too then you'll be fine to run expo, in the bios set expo I as that will adjust the primary timings of the ram, expo II will do the primary and will adjust some secondary timings, but to play it safe use expo I, which ram have you ordered btw?