help Help needed with persisting DRAM error, components keep dying (MB, RAM, CPU) PSU replaced, 4 MB, 2 CPU and RAM

[maxmaster]

48 minutes ago, Lobster84 said:

I have uploaded some pictures of the bios and voltage settings. Maybe this could be of some help? Like I said, I have no idea on how to tweak this to the right settings.

U have 17h hours to buy one of the best memory sticks there is a korting atm 

 

https://www.alternate.de/Team-Group/DIMM-32-GB-DDR5-7600-(2x-16-GB)-Dual-Kit-Arbeitsspeicher/html/product/1878742

 

If u have hynix a-die then u dont have to, but i know that there are Microns and ... it is hard to make them stable, also default xmp might not work. With samsung max is 6200mhz at normal voltage usually pcmi doesnt allow to go higher than 1.43v. So the best is: check what ram u have, if it is samsung than it will work at 6000, if micron than buy teamgroup delta 7600. It is very good ram. 

 

Ok so try to set the vddp 1.010v or 1.005v, the vddg iod can be like that also can be 1.050. But or in range of 10 or much higher. Try that

Vddp 1.005

Vddg ccd 950

Vddg iod 955

Enable expo

Enable low latency

Set vdd 1.35

 

[Lobster84]

25 minutes ago, maxmaster said:

U have 17h hours to buy one of the best memory sticks there is a korting atm 

 

https://www.alternate.de/Team-Group/DIMM-32-GB-DDR5-7600-(2x-16-GB)-Dual-Kit-Arbeitsspeicher/html/product/1878742

 

If u have hynix a-die then u dont have to, but i know that there are Microns and ... it is hard to make them stable, also default xmp might not work. With samsung max is 6200mhz at normal voltage usually pcmi doesnt allow to go higher than 1.43v. So the best is: check what ram u have, if it is samsung than it will work at 6000, if micron than buy teamgroup delta 7600. It is very good ram. 

 

Ok so try to set the vddp 1.010v or 1.005v, the vddg iod can be like that also can be 1.050. But or in range of 10 or much higher. Try that

Vddp 1.005

Vddg ccd 950

Vddg iod 955

Enable expo

Enable low latency

Set vdd 1.35

 

So, if I understand correctly, these settings are for a stable run of the RAM with the Ryzen CPU?
If so, I can check on that but first I will need to read a little into tweaking these settings.

The ram you linked is a nice one. I know Team Group, I had them before and they worked wonderful. Also had Crucial but those were not to my satisfactory.
Though, my budget is 0 at the moment. Even a great discount like the one linked is to much. I lost my job two months ago and have a smaller income now.
When I get a job again, I will safe up for some RAM and GPU.

I saved your messages, so I can get them alongside me when I decide to go tweak these settings.
I do know that the shop that I bought from, would probably be less nice if I damage something because I did something wrong.

 

As of now, the next step is to monitor my system. It is working now but I haven't really done much on it except using the browser.

I will do some testing, Memtest86 too, and see what it shows.
I will also run some software and a game or benchmark to see if the GPU will cause an issue after load.

 

I keep things updated in here.

Thank you so far!

[Lobster84]

59 minutes ago, maxmaster said:

But if the computer post "is working", doesnt mean it is stable. It is expo there for. Just try to turn it. Vddp and vddg are on the tweaking tab, the penultimate possition. Well u need to learn a little, otherwhise it might be hard.

 

There is a lot of corsair vengence stocks 5600 36 36 36 , it is samsung kit, 100% it wil run 6000 cl 36 36 36, but actually all the ddr5 5600mhz sticks are able to run 60x. So u might try to run xmp, enable expo, enable low latency, set multiplyer 60x, and set the voltage VDD to little higher than default xmp. Like if it is 1.3 on auto set 1.355

By the way, how do I find out if it is a Samsung or Micron Kit?

[Whatisthis]

You need to get your power outlet checked. There is no way you should be frying that much stuff without having something wrong with your house power.

 

If you have a multimeter you can test voltages. You should read 230v across the hot wire and neutral. It should also read 230v across hot and ground. Neutral and ground should read 0. If you have any deviation there, you have an issue that you will want to contact an electrician to fix.

 

You can also get a power conditioner. It filters the power and stabilizes voltage. It’s like a surge protector on steroids. It fixes pretty much anything but bad wiring. 

[Lobster84]

12 minutes ago, Whatisthis said:

You need to get your power outlet checked. There is no way you should be frying that much stuff without having something wrong with your house power.

 

If you have a multimeter you can test voltages. You should read 230v across the hot wire and neutral. It should also read 230v across hot and ground. Neutral and ground should read 0. If you have any deviation there, you have an issue that you will want to contact an electrician to fix.

 

You can also get a power conditioner. It filters the power and stabilizes voltage. It’s like a surge protector on steroids. It fixes pretty much anything but bad wiring. 

Thank you for your responds. I do suspect the extension outlet has something to do with it. In my previous responds to CommanderAlex I said that the monitor turned off when inserting the power cable for the system. Monitor could be turned on by reinserting the adapter.

 

We had a new fuse box installed last November because we got solar panels now. The electrician, who worked for the electricity company providing power to us, made some extra groups to future proof our in-house power grid. We have no other appliances or lights doing weird stuff. Other computers are fine too.
I do plan to buy a watt meter to see if there is a spike in wattage on the outlet the system is connected. A voltage meter I have and already did a test and I got 230v.

 

I think there is a combination of stuff happening, after talking to the people that responded.

- Bad extension outlet, even if it is only about 3 months old. Going to replace this one. In the meantime I got another new extension outlet without surge protection. New one is on the way.

- Unstable working components, which I will check when tweaking settings as Maxmaster said.

- Check the GPU to see if it causes problems under more load

- Monitor voltages with OCCT and HWMonitor, to see if there is a spike with a component

- Check power on the grid with watt meter, already ordered one.

- Possibly get new components like RAM and keep in contact with the shop in case I need a new components that still got warranty on them.

- Stresstest the system a bit more with MEMTEST86+ and probably some others too. Depending on what I find.

 

The CMOS clearing has made it possible to get past the DRAM error and use the system.

[problemsolver]

8 hours ago, Lobster84 said:

I used a new cable for the PSU with this test too. When inserting the cable in the outlet, arching happened and disabled the monitor that is connected to my laptop.

The monitor works again after reinserting the adapter in the outlet.

Arcing (from the outlet to the plug) is normal... however, having another device turn off because you plugged something in is not normal.

 

Sounds like you've got a problem for an electrician (as @Whatisthis was mentioning) or there's some piece of critical information I'm missing.

[maxmaster]

9 hours ago, Lobster84 said:

By the way, how do I find out if it is a Samsung or Micron Kit?

Usually the samsung has same values like 36 36 36, there is code of the ram and the first digit is the producent like 5 or 3 etc but also micron has 40 40 40. The hynix a-die it depends, there is 40 40 40 5200mhz but on 1.1v , normally it has like 30 40 40, or 32 42 42 etc, and hynix m die. But honestly it depend kits tend to differ by a lot, so the producent might be as well putting there hynix a die , u have 1.25v at 5600mhz, but if thats a micron it might be bad xD

[Lobster84]

19 hours ago, problemsolver said:

Arcing (from the outlet to the plug) is normal... however, having another device turn off because you plugged something in is not normal.

 

Sounds like you've got a problem for an electrician (as @Whatisthis was mentioning) or there's some piece of critical information I'm missing.

I read into arcing and found that it is not a real problem indeed. Unless there are more effects like fire, smell of burning, discoloring of the outlet etc. Well that is not the case. I do think the extension outlet is bad and changed it out to another. That one doesn't give me arcing and is not turning my monitor off.
I ordered a new one with surge protection, should arrive in a couple days.

I bought the extension outlet in September. Going to try to get my money back since it has a warranty of 3 years.

 

As I said in my last post, I do think there is a combination of things going on right now. Maybe a defective extension outlet has damaged a component in the case. To determine if something is wrong with any component I have to run tests.

 

The situation now is as follow:

 

- I changed the extension outlet, no arcing or turning off monitors when I plugged it in.

- PC is starting after clearing CMOS and completing a memory training.

-- First restart after memory training was done, was slow. It hang for about 60 seconds in DRAM check.

-- Second restart (since I was testing peripherals), was faster.

-- Every restart made the startup faster.

 

This morning (December 28th) I started the pc again and it ran through checks including DRAM as normal. The only weird thing was that after the last, fourth check (I think it is called post check or boot check or something) I got two black monitors with a glow. Took about 40 - 60 seconds and then I got the Windows screen.

After an update that needed a restart, the system started as normal. No delays.

Something is happening and it is not consistent enough, for me, to pinpoint the issue. That is why I think there are multiple things happening.
The first step was changing the outlet, the second would be monitoring the system and read into tweaking settings like @maxmaster said. Maybe it is unstable because of that.

I am planning to do a memtest tomorrow.

[Lobster84]

14 hours ago, maxmaster said:

Usually the samsung has same values like 36 36 36, there is code of the ram and the first digit is the producent like 5 or 3 etc but also micron has 40 40 40. The hynix a-die it depends, there is 40 40 40 5200mhz but on 1.1v , normally it has like 30 40 40, or 32 42 42 etc, and hynix m die. But honestly it depend kits tend to differ by a lot, so the producent might be as well putting there hynix a die , u have 1.25v at 5600mhz, but if thats a micron it might be bad xD

Okay, still a little bit gibberish for me but I am trying to understand and look into the details. I get what you say but can't place it in context to the ram details.
I was looking in CPU-Z and OCCT. Can those programs show me the values you mentioned?
Cause I see values like 40-40-40-77 (memory tab - CAS# latency (CL) 40 | RAS# to CAS# Delay (tRCD) 40 | RAS# Precharge (tRAS) 40 | Cycle Time (tRAS) 77)
 

The SPD tab shows me a timings table where the same numbers are mentioned in the last column, the first colum (JEDEC) has lower numbers -> 36 37 37 70

The SPD tab also tells me for DRAM Manuf. SK Hynix.

Tell me if I am looking at the wrong stuff haha.

[maxmaster]

10 hours ago, Lobster84 said:

Okay, still a little bit gibberish for me but I am trying to understand and look into the details. I get what you say but can't place it in context to the ram details.
I was looking in CPU-Z and OCCT. Can those programs show me the values you mentioned?
Cause I see values like 40-40-40-77 (memory tab - CAS# latency (CL) 40 | RAS# to CAS# Delay (tRCD) 40 | RAS# Precharge (tRAS) 40 | Cycle Time (tRAS) 77)
 

The SPD tab shows me a timings table where the same numbers are mentioned in the last column, the first colum (JEDEC) has lower numbers -> 36 37 37 70

The SPD tab also tells me for DRAM Manuf. SK Hynix.

Tell me if I am looking at the wrong stuff haha.

Well i jedec is what all producents did test, and is working. But under the XMP in the picture u uploaded, there are 2 values, Low latency suport and EXPO high bandwidth support, make them enabled. It is a thing of am5 platform, your motherboard will suit the subtimings better.

[Lobster84]

6 hours ago, maxmaster said:

Well i jedec is what all producents did test, and is working. But under the XMP in the picture u uploaded, there are 2 values, Low latency suport and EXPO high bandwidth support, make them enabled. It is a thing of am5 platform, your motherboard will suit the subtimings better.

I will try that. Do I need to do a test after that to see if it runs stable?

[Lobster84]

I did a test yesterday with memtest and OCCT (power, ram, vram & cpu) and got no errors.
It all seemed stable in temperature, voltage and load.

[maxmaster]

7 hours ago, Lobster84 said:

I did a test yesterday with memtest and OCCT (power, ram, vram & cpu) and got no errors.
It all seemed stable in temperature, voltage and load.

Good expo and latency helps a lot if your ram is not listed on mb site. It is usually quicker also. You should really check what memory u got. If it ismicronlave it as it is, is working, cool. If samsung than it will work at 6000mhz 100%. If hynix m-die it should also easly reach 6200mhz, but if it is hynix a-die, than learn how to oc, watch videos etc, it might work 7400-8000mhz with 1:2. 

[Lobster84]

New update:
I brought the computer to the shop so they can check it out.

December 31st, I got the same issue again. No video output, Click in PSU and DRAM error. PC stays on, rgb of gpu is on, fans are on too. No excessive sounds, just as a stable run.

It is not the extension outlet, since the issue occurred again with the new outlet.

Same goes for the cable that powered the PSU.

There were no weird watt readings while using the pc and when the issue happened. So my guess is it is something in the components that was happening.

 

GPU seems to work fine. I have no glitches and monitoring it with HWmonitor and OCCT didn't show any weird results. I have monitored it when doing Heaven benchmark, while playing games for a minimum of 30 and max 60 minutes on different settings. Played games like Cyberpunk, Starfield and some indi sim/survival games.

 

Stresstesting didn't really show any weird results, except for one slightly higher temp reading on Ryzen 9 7900X CCD#0, in HWMonitor the max value was 96C and it was in red. First and only time I saw this was after a couple tests in OCCT.

MEMTEST86 had no errors

 

I was going to test the bios settings, like @maxmaster said, but decided to first bring it to the shop.
 

The only occurrence that happened before both times I got DRAM, click in PSU and no video is that the mouse lights stayed on while the pc is turned off.

Speakers have power too when pc is off, as if the powerstate would be set to keep things on. I checked the bios and it should be completely off.
 

[maxmaster]

On 1/3/2024 at 2:25 PM, Lobster84 said:

New update:
I brought the computer to the shop so they can check it out.

December 31st, I got the same issue again. No video output, Click in PSU and DRAM error. PC stays on, rgb of gpu is on, fans are on too. No excessive sounds, just as a stable run.

It is not the extension outlet, since the issue occurred again with the new outlet.

Same goes for the cable that powered the PSU.

There were no weird watt readings while using the pc and when the issue happened. So my guess is it is something in the components that was happening.

 

GPU seems to work fine. I have no glitches and monitoring it with HWmonitor and OCCT didn't show any weird results. I have monitored it when doing Heaven benchmark, while playing games for a minimum of 30 and max 60 minutes on different settings. Played games like Cyberpunk, Starfield and some indi sim/survival games.

 

Stresstesting didn't really show any weird results, except for one slightly higher temp reading on Ryzen 9 7900X CCD#0, in HWMonitor the max value was 96C and it was in red. First and only time I saw this was after a couple tests in OCCT.

MEMTEST86 had no errors

 

I was going to test the bios settings, like @maxmaster said, but decided to first bring it to the shop.
 

The only occurrence that happened before both times I got DRAM, click in PSU and no video is that the mouse lights stayed on while the pc is turned off.

Speakers have power too when pc is off, as if the powerstate would be set to keep things on. I checked the bios and it should be completely off.
 

Well thats sad, feel sorry 4u, mann there might be a lot of things wrong. But the gigabyte mb i think they have some issue. I have been manage to stabilize 7750mhz witl cl32, also i found out that 8000 mhz is working perfectly stable, until.... until u go to the bios or sometimes turn the computer off. The frequencies that were rock solid doesnt work any more, like literally u didnt make any changes in bios but the computer loose the stability after 5s comming errors in tm5. It is pretty anoying imagine how hard is to find the right frequencies and timings if u dont know why even on checked settings it doesnt work. Like u have 6 working(checked) profiles in bios and non of it is working if u load them just like that. Insane. I would say even that they will fix your problem (making computer temporary stable) which i am having while going into bios on stable state causing inexplicable instability thet incline me to think that the gigabyte motherboards has some tensions palpitation on SoC or somewhere else on the chipset. I discovered that by my self and also i founded the video that somehow seems to confirm this fact. Here is the link https://youtu.be/O2n4rOWehtQ?si=63W981pZUGVZVllA 

 

I even changed the psu to msi meg atx 3.0 pcie 5 1000W ,0 improvement*

[Lobster84]

On 1/5/2024 at 6:50 AM, maxmaster said:

Well thats sad, feel sorry 4u, mann there might be a lot of things wrong. But the gigabyte mb i think they have some issue. I have been manage to stabilize 7750mhz witl cl32, also i found out that 8000 mhz is working perfectly stable, until.... until u go to the bios or sometimes turn the computer off. The frequencies that were rock solid doesnt work any more, like literally u didnt make any changes in bios but the computer loose the stability after 5s comming errors in tm5. It is pretty anoying imagine how hard is to find the right frequencies and timings if u dont know why even on checked settings it doesnt work. Like u have 6 working(checked) profiles in bios and non of it is working if u load them just like that. Insane. I would say even that they will fix your problem (making computer temporary stable) which i am having while going into bios on stable state causing inexplicable instability thet incline me to think that the gigabyte motherboards has some tensions palpitation on SoC or somewhere else on the chipset. I discovered that by my self and also i founded the video that somehow seems to confirm this fact. Here is the link https://youtu.be/O2n4rOWehtQ?si=63W981pZUGVZVllA 

 

I even changed the psu to msi meg atx 3.0 pcie 5 1000W ,0 improvement*

Thank you, that was somewhat informative for me haha. It made me think about me resetting CMOS, reseating RAM and restarting the pc. Maybe something is going wrong with the voltages, which could potentially damage components. I have send this information to the shop so they can check it out. If this is what's happening, any replacement wouldn't help at all. The pc won't ever run stable unless I change out Gigabyte motherboard to something else.

I will keep you guys posted if there is any update.

[Lobster84]

Finally, I found the problem. It were multiple things actually.

PC was running without problems at the shop. The engineer that looked into it contacted me personally to discuss my setup and the problem.

 

- Displayport cables are replaced with new ones. Not sure if those were part of the problem but the engineer said he experienced similar problems in the past where the dp cables were the problem.

- Extension outlet is replaced with a new one. I asked the engineer what he would think would be a good outlet and he ordered one for me. Like discussed before in this forum. Although, I didn't find weird readings with the power outlet in the wall and the extension outlet, the outlet didn't seem to work properly (with connecting a power cable and turning off the monitors).

- PSU and bios: The psu was replaced. After discussions with the engineer it seemed that the old psu looked to be the problem since there were changes in the voltage the bios would get. I didn't know this, only after the message of Maxmaster I looked into this and asked the engineer to look into it too. He told me this was indeed a known problem and not only limited to Gigabyte. He also told me that the first couple changed motherboards were because of the voltage changes the bios made. To be sure they changed out the psu. The engineer made sure to set the voltage to 1.2V.

- Peripherals: I asked if peripherals could be the problem and the answer was basically "yes, but changes are very small". After discussion with the engineer it seems one or more of my peripherals could be the problem.

I mentioned my mouse keeping power after I shut down the pc and after that the pc would turn on and get stuck in DRAM. I am still in testing mode while using the pc as I normally would.

After I got my pc back from the shop, which was January 17th, it ran fine. Only yesterday I got a DRAM again. I reset the CMOS, so I can start the pc again.

I had powered on my printer while starting the pc. This never resulted in problems before and I am not 100% sure it was the issue for the DRAM but it seemed fitting.

The printer is connected to the network, wireless. It was also connected to my pc with USB. It could be power leading back into the USB port from the printer cable.

 

So basically, I think one of my peripherals in combination with one or more of the other things could be the problem and a good start to check out if someone has similar issues. I will Consider this topic as solved.

For anyone looking at this topic to find solutions, read through the topic to see if anything could be helpful.

 

Thanks @CommanderAlex, @maxmaster, @Whatisthis and @problemsolver for your help.

[maxmaster]

On 1/31/2024 at 10:23 AM, Lobster84 said:

Finally, I found the problem. It were multiple things actually.

PC was running without problems at the shop. The engineer that looked into it contacted me personally to discuss my setup and the problem.

 

- Displayport cables are replaced with new ones. Not sure if those were part of the problem but the engineer said he experienced similar problems in the past where the dp cables were the problem.

- Extension outlet is replaced with a new one. I asked the engineer what he would think would be a good outlet and he ordered one for me. Like discussed before in this forum. Although, I didn't find weird readings with the power outlet in the wall and the extension outlet, the outlet didn't seem to work properly (with connecting a power cable and turning off the monitors).

- PSU and bios: The psu was replaced. After discussions with the engineer it seemed that the old psu looked to be the problem since there were changes in the voltage the bios would get. I didn't know this, only after the message of Maxmaster I looked into this and asked the engineer to look into it too. He told me this was indeed a known problem and not only limited to Gigabyte. He also told me that the first couple changed motherboards were because of the voltage changes the bios made. To be sure they changed out the psu. The engineer made sure to set the voltage to 1.2V.

- Peripherals: I asked if peripherals could be the problem and the answer was basically "yes, but changes are very small". After discussion with the engineer it seems one or more of my peripherals could be the problem.

I mentioned my mouse keeping power after I shut down the pc and after that the pc would turn on and get stuck in DRAM. I am still in testing mode while using the pc as I normally would.

After I got my pc back from the shop, which was January 17th, it ran fine. Only yesterday I got a DRAM again. I reset the CMOS, so I can start the pc again.

I had powered on my printer while starting the pc. This never resulted in problems before and I am not 100% sure it was the issue for the DRAM but it seemed fitting.

The printer is connected to the network, wireless. It was also connected to my pc with USB. It could be power leading back into the USB port from the printer cable.

 

So basically, I think one of my peripherals in combination with one or more of the other things could be the problem and a good start to check out if someone has similar issues. I will Consider this topic as solved.

For anyone looking at this topic to find solutions, read through the topic to see if anything could be helpful.

 

Thanks @CommanderAlex, @maxmaster, @Whatisthis and @problemsolver for your help.

Gr8, i havent had any problems like that at 6000mhz, the problem occured over ddr 7600mhz. But at 7800mhz cl34 is stable right now, but i guess it is on the edge of stability. Like i made the 7960mhz stable but after week the stability dissappeared. So it is like that when the system is just bearly capable. You can check the stability by downloading the "test mem 5" so called tm5 and loading the extreme1@anta777 config. It is like 1hour stability test, which is one of the quickest & hardest. 

 

*Actually it isnt stable anymore, after i reboot few times it loses the stability. So it might be gigabyte motherboard issue with ulmck/2. Been running 8hours aida strees test then reboot and failiure after 2minutes.