Seriously STUMPED by this motherboard problem

[Energycore]

Hey Guys! It's a long time no see for a lot of you, hi!

 

I'm actually here because I have reached the end of my patience with my new Ryzen 5600X/B550 motherboard combo. Let me be more specific.

 

A month or so ago I got my CPU, motherboard and RAM upgrade. I've had a slew of issues that were frustrating but were eventually fixed, except for one.

 

Ever since the upgrade, my computer randomly shuts down, but the fans keep spinning (sometimes they go to 100% but not always). This can happen ANY time, even while in BIOS and seems to be random. Sometimes my computer goes a full 12 hours just fine, sometimes I have to start it up 5-6 times because it just shuts down right after entering windoms, or even right after POST.

 

Edit: Might be useful to add that there's no power cycling. I can boot up the computer just fine after it breaks.

 

Here's my system specs

- Ryzen 5 5600X

- AsRock B550 Phantom Gaming 4

- 16GB Crucial Ballistix 3200 C16 kit

- Old Toshiba SATA SSD for Windows

- New Crucial P5 with fresh Windows Installation

- EVGA GTX 1080 Ti Hybrid

- Phanteks AMP 550W (same internals as the Seasonic Focus GX)

 

Here's what I have already tried

- Different BIOS versions

- Different motherboard (tried an MSI B550 board)

- Clearing CMOS and running without XMP etc

- Ditching the old Windows install and using only the Crucial P5

- Reseating CPU/RAM/PSU Cables
- Different PSU (I also have a Corsair CX450M that I got from warranty and has pretty much been unused since I got it)

 

None of this has solved the problem. I have NO IDEA what the issue is and frankly I'm just about to pull out my hair over this.

 

Do you guys have any suggestions? It would be much appreciated

 

~Energycore

[Kilrah]

Done a memory test?

[Energycore]

Just now, Kilrah said:

Done a memory test?

Let me fire that up. memtest86 still the preferable test?

[Kilrah]

Believe that's just fine, yes...

[Levent]

4 minutes ago, Energycore said:

Let me fire that up. memtest86 still the preferable test?

I would suggest you to use either Prime95's memory test or Testmem5 using antra777's config. Other than that, this also could be your boot drive, as I also saw a lot of issues with Crucial P5 drive.

[raleighc]

Try setting a power limit on your CPU just below its peak transient consumption so it doesn't peak so high.  This helped me with the problems I was having with my 5800x.  Just an idea.

[da na]

9 minutes ago, Kilrah said:

Done a memory test?

Second this.

Prime95's mem test as well as memtest86 and Windows's built in are all reliable 

[BiotechBen]

One thing to remember with memtest is you need to loop it for a long time. Seen it report all good multiple times with bad gpu VRAM Christmas treeing the screen.

[Energycore]

1 hour ago, Levent said:

I would suggest you to use either Prime95's memory test or Testmem5 using antra777's config. Other than that, this also could be your boot drive, as I also saw a lot of issues with Crucial P5 drive.

1 hour ago, Kilrah said:

Believe that's just fine, yes...

1 hour ago, Mel0nMan said:

Second this.

Prime95's mem test as well as memtest86 and Windows's built in are all reliable 

Ok so I'm back after an hour or so of memtest86. I had 2 random shutdowns, but no errors reported. There was no pattern to the random shutdowns, one was 7 minutes in and the other was 45 minutes in. I'll take your guys' advice and run it for longer it's ok.

 

1 hour ago, Levent said:

I would suggest you to use either Prime95's memory test or Testmem5 using antra777's config. Other than that, this also could be your boot drive, as I also saw a lot of issues with Crucial P5 drive.

Shouldn't be related to the P5 drive, this also happens when I use my SATA Drive's Windows installation. I'm writing this from that drive with the P5 disconnected and put away.

[Kilrah]

Maybe try a single RAM stick at a time, in different slots too...

If that doesn't help I'd start suspecting the CPU.

[Energycore]

34 minutes ago, Kilrah said:

Maybe try a single RAM stick at a time, in different slots too...

If that doesn't help I'd start suspecting the CPU.

Sorry I forgot to mention this, I did try one single stick in all slots. Let me run a CPU stress test and see if it turns up any errors.

[Energycore]

So no errors while running Pridme95 but I can't keep the test running for very long because it'll just shut down. I've found it as likely to shut down while idle as under load and temps were extremely low for both CPU and Motherboard when I did the Prime95 test.

[Guest]

16GB Crucial Ballistix 3200 C16 kit 

 

What's the part number??

 

Only but a couple/few suggestions for you. 

Gimmie a minute to type some stuff.

[Guest]

If running XMP (which you probably are and should) be sure to find the SPD timings in bios and take any timings off of auto and match them to SPD.

 

In timings menu, disable "power down mode".

[Energycore]

5 minutes ago, ShrimpBrime said:

If running XMP (which you probably are and should) be sure to find the SPD timings in bios and take any timings off of auto and match them to SPD.

 

In timings menu, disable "power down mode".

I was actually not running XMP and running the RAM at the stock 2666MHz with Default timings. Let's see if I can get more stability if I turn on XMP

 

Edit: Nevermind it shut down seconds after posting xD

 

What are the SPD timings? I couldn't find them on the memory timings area at a glance. Power Down Mode is disabled by default.

[Guest]

13 minutes ago, Energycore said:

I was actually not running XMP and running the RAM at the stock 2666MHz with Default timings. Let's see if I can get more stability if I turn on XMP

 

Edit: Nevermind it shut down seconds after posting xD

 

What are the SPD timings? I couldn't find them on the memory timings area at a glance. Power Down Mode is disabled by default.

In mine (Asus) it's under the "tool" tab. Should be the same for yours I believe.

 

It should list a few secondaries. 

 

Some screen shots here and there too. Windows reports this as error 41 sudden power loss?

 

___

The sudden shut down at XMP enabled has me thinking cpu actually or IMC/Board don't like something about your memory. My vengeance does the same crap. So I don't use them without manually setting them up. I noticed the board sets some of the secondaries incorrectly when on auto. 

[Energycore]

Just now, ShrimpBrime said:

Windows reports this as error 41 sudden power loss?

How do I find out?

Just now, ShrimpBrime said:

The sudden shut down at XMP enabled has me thinking cpu actually or IMC/Board don't like something about your memory. My vengeance does the same crap. So I don't use them without manually setting them up. I noticed the board sets some of the secondaries incorrectly when on auto. 

So far I've got sudden shut down both with and without XMP. Do you think it's a good idea to try with another memory kit before buying a CPU?

(I can get a replacement 5600X tomorrow with Amazon so I kinda just wanna buy a new one and I can spare the money at least for now, but I don't want to ALSO buy another memory kit).

 

The memory kit's part number is BL2K8G32C16U4B. It's not on the QVL for this motherboard now that I checked.

[Guest]

10 minutes ago, Energycore said:

How do I find out?

So far I've got sudden shut down both with and without XMP. Do you think it's a good idea to try with another memory kit before buying a CPU?

(I can get a replacement 5600X tomorrow with Amazon so I kinda just wanna buy a new one and I can spare the money at least for now, but I don't want to ALSO buy another memory kit).

 

Type in search "event viewer" (Under Critical drop down)

 

Edit: Yes, another CPU would help diag the memory controller vs the memory. 

 

 

Quote

The memory kit's part number is BL2K8G32C16U4B. It's not on the QVL for this motherboard now that I checked.

And because it's not on the QVL, the auto timings are wrong. (I'm 90% sure at least 2 timings) 

 

I am not sure if your board automatically memory trains or not. I'd still check the SPD values in bios both at 2667mhz and XMP frequency. 

 

The CPU SOC on auto should be 1.08v (with XMP enabled.. Disabled .9v) You can confirm this with HWinfo64 (click sensors only). If not, set it manually. You will not need more, you won't be at those high frequencies to make it necessary. So 1.1v if you see vdroop below 1.08v during loads specifically read and write to memory. 

[Guest]

And lastly, if all of this still doesn't help, instead of running the 2667mhz Jedec, manually set that to 2133mhz. If auto timings are not 15-15-15-35, set them manually, dram voltage at 1.2v (set all defaults first, exit tab in bios I believe)

[Energycore]

Just now, ShrimpBrime said:

Edit: Yes, another CPU would help diag the memory controller vs the memory. 

Fair enough. Let's see if the new CPU does it.

 

2 minutes ago, ShrimpBrime said:

And because it's not on the QVL, the auto timings are wrong. (I'm 90% sure at least 2 timings) 

 

I am not sure if your board automatically memory trains or not. I'd still check the SPD values in bios both at 2667mhz and XMP frequency. 

 

The CPU SOC on auto should be 1.08v. You can confirm this with HWinfo64. If not, set it manually. You will not need more, you won't be at those high frequencies to make it necessary. So 1.1v if you see vdroop below 1.08v during loads specifically read and write to memory. 

Let me go into the BIOS and try this out.

 

Looks like SoC voltage is 1.094V. Not sure how to interpret that.

 

I also found the error 41 in the Event Viewer

[Guest]

1 minute ago, Energycore said:

Looks like SoC voltage is 1.094V. Not sure how to interpret that. 

 

I also found the error 41 in the Event Viewer

 

1.094 is good and fine. Even if running Jedec speeds, although it doesn't need too....

 

And there's the infamous error. You can google your brains out.

 

But the only fix I could come up with at home was manually setting up the memory timings for the board. Or slap in my B-Dies. One of the two. 

 

The online fix was RMA the Cpu on some occasions. 

[Energycore]

8 minutes ago, ShrimpBrime said:

1.094 is good and fine. Even if running Jedec speeds, although it doesn't need too....

 

And there's the infamous error. You can google your brains out.

 

But the only fix I could come up with at home was manually setting up the memory timings for the board. Or slap in my B-Dies. One of the two. 

 

The online fix was RMA the Cpu on some occasions. 

Ok so I was unable to find the SPD values in the Tool tab or the DRAM Timings menu. Here's a couple BIOS screenshots.

Spoiler

For now I am booted in with 2133MHz 15-15-15-35 @1.2V and I'll let you know if I get crashes.

[Energycore]

Just now, ShrimpBrime said:

Or slap in my B-Dies.

If only I'd just ponied up the additional money for B-Die RAM LOL

[Guest]

16 minutes ago, Energycore said:

Ok so I was unable to find the SPD values in the Tool tab or the DRAM Timings menu. Here's a couple BIOS screenshots.

  Hide contents

Snip pic

For now I am booted in with 2133MHz 15-15-15-35 @1.2V and I'll let you know if I get crashes.

You can try zen timings or any app that will list the correct secondaries for the memory. Your timings look ok to me, but obviously I don't know what they are supposed to be lol. 

 

Hope 2133mhz does ok. fingers crossed. Would be better than a system that randomly shuts down.

[Energycore]

Just now, ShrimpBrime said:

You can try zen timings or any app that will list the correct secondaries for the memory. Your timings look ok to me, but obviously I don't know what they are supposed to be lol. 

 

Hope 2133mhz does ok. fingers crossed. Would be better than a system that randomly shuts down.

So far no shutdown! If the solution is to manually set timings then I'm gucci and I don't need a new CPU.

 

EDIT: Hopefully I can get it stable above 2133MHz LOL otherwise I'll get pretty gimped performance