X870 will not post after switching itself off

Jerome said:
Two faulty 9800X3Ds in a row though is surely reaching beyond improbable.
Click to expand...
They are kind of "edge case" CPUs, at the bleeding edge of technology and in high demand (which might tempt them to make them faster and more of them, than perhaps is justifiable).

Though, one could argue AMD have had more than enough time to make these properly, because the X3D has existed for a long while.

Was EXPO enabled for the duration in both cases?
 
Tetras said:
Was EXPO enabled for the duration in both cases?
Click to expand...

Yes, EXPO enabled both times. I always used the latest (non-beta) BIOS versions.

If I recall correctly, EXPO was also enabled for the vast majority of the time I had the 7600X as well. It's also enabled now with the new 7800X3D installed.

I hadn't noticed any strange thermal or voltage issues with periodic monitoring on the latest 9800X3D. It could also be purely coincidental that they each lasted roughly one week.

One key point to note is that both seemingly died while the computer was completely idle. In both cases I was out of the room and came back to it. In the first case, the computer had turned itself off completely and in the second case, yesterday, the screens didn't wake when I pressed buttons on my keyboard/mouse but the PC itself was still running. I rebooted it via the reboot button and it subsequently failed to post.

I decided to enable EXPO again with the 7800X3D as I want to replicate the conditions which led to the failure of the 9800X3Ds. If it occurs with the 7800X3D then it can be concluded the motherboard is the major suspect and I can then justify spending out on getting a different model.

On the plus side, fitting a new CPU/cooler was a daunting task as a first timer but I feel mighty comfortable by this stage. :D

I've seen quite a few posts around the internet about 9800X3Ds failing here and there but I haven't seen any 'two in a row' cases besides my own as of yet. Time will tell.
 
Jerome said:
I decided to enable EXPO again with the 7800X3D as I want to replicate the conditions which led to the failure of the 9800X3Ds. If it occurs with the 7800X3D then it can be concluded the motherboard is the major suspect and I can then justify spending out on getting a different model.
Click to expand...
I'm only asking because the use of EXPO was part of the early burning issues with the 7800X3D and AMD introduced a voltage cap that was pushed out in the fixed BIOS of first gen boards. I would think it unlikely that this issue has occurred again, since any second gen boards should maintain those fixes, but different boards/manufacturers do have different behaviour.
 
Tetras said:
I'm only asking because the use of EXPO was part of the early burning issues with the 7800X3D and AMD introduced a voltage cap that was pushed out in the fixed BIOS of first gen boards. I would think it unlikely that this issue has occurred again, since any second gen boards should maintain those fixes, but different boards/manufacturers do have different behaviour.
Click to expand...

Makes sense. I also thoroughly checked the pins/contacts on the second failed CPU and no visible signs of scorching or damage. I know there was talk a couple of months ago about an MSI board burning out an 9800X3D but that was deemed to be an incorrect seating which lead to it.
 
Jerome said:
Makes sense. I also thoroughly checked the pins/contacts on the second failed CPU and no visible signs of scorching or damage. I know there was talk a couple of months ago about an MSI board burning out an 9800X3D but that was deemed to be an incorrect seating which lead to it.
Click to expand...
It does sound like the CPU had a problem internally, rather than the board killing it.

FYI: the burning issue did not apply to any non-X3D CPUs.
 
I've been emailing back and forth with AMD over the last few days trying to get them to accept the RMA.

I've sent all kinds of information such as dxdiag, the msinfo NFO file, the model of cooler I use, the power supply, etc.

After taking at least 24 hours to reply each time they're now asking me to "could you please connect the power supply using a 2x8 pin power connector and check if the CPU boots?" which is odd considering that's the only visible way to power the CPU and I'm powering a 7800X3D to write them the email.

Considering swapping the processor immediately rectified the issue, clearly the CPU has a fault.

It's an absolute nightmare. All I want them to do is accept the RMA so they can make me wait another 2 months to get a replacement. Again.
 
Last edited:
not sure if its been mentioned but ive noticed high vsoc with stock expo settings on asus boards from memory it sets auto to 1.3 and appears to spike to 1.45 from memory, i think its to cover those people buying expo 6400 kits. in my case i set a manual vsoc of 1.1 from memory with 6000, wonder if in some cases the cpu is already on the edge the vsoc maybe killing them.
 
Last edited:
not sure as i dont have 1 to test but if asus are doing it maybe asrock are too? but you could check or maybe someone else with an asrock board can tell us what they see, ive noticed it after prolonged sessions.
 
Last edited:
Jerome said:
I've been emailing back and forth with AMD over the last few days trying to get them to accept the RMA.

I've sent all kinds of information such as dxdiag, the msinfo NFO file, the model of cooler I use, the power supply, etc.

After taking at least 24 hours to reply each time they're now asking me to "could you please connect the power supply using a 2x8 pin power connector and check if the CPU boots?" which is odd considering that's the only visible way to power the CPU and I'm powering a 7800X3D to write them the email.

Considering swapping the processor immediately rectified the issue, clearly the CPU has a fault.

It's an absolute nightmare. All I want them to do is accept the RMA so they can make me wait another 2 months to get a replacement. Again.
Click to expand...

It is understandable considering they sent you one and you killed it within a day. AMD I would think are well within there rights to point at your hardware and say you have a problem.
 
It took around 7-8 days in each case, never within a day.

I appreciate they'd like to troubleshoot but giving me braindead instructions such as asking me to connect the power supply using the only possible way to power the CPU in the first place is frustrating. It's like asking me if I've tried plugging my computer power cable into a power socket.

If their troubleshooting was actually anything meaningful I'd be more than happy to go through any steps they'd ask.

This same hardware powered a 7600X for almost a couple of months (and a 7800X3D over the last few days) with zero errors or issues.

I'd also be happy to swap my other hardware (such as mobo/psu) but no retailer or manufacturer is going to accept my request if everything works fine and the only consistent failing device is this particular CPU.

Don't mind me though, I'm just moaning. :p
 
I noticed on your original photos you are using 2 x EPS cables, could this be the problem.

I have read that it does not matter if 1 or 2 is used as the PC will only draw the power it needs, but what if a glitch of some sort sent extra to the CPU.

May be best to just use 1 cable when you get the new 9800X3D.
Jerome said:
"could you please connect the power supply using a 2x8 pin power connector and check if the CPU boots?"
Click to expand...
They probably thought you only used 1 connector and want you to try both. (assuming they have not been told you already have)
 
freddie64 said:
I noticed on your original photos you are using 2 x EPS cables, could this be the problem.

I have read that it does not matter if 1 or 2 is used as the PC will only draw the power it needs, but what if a glitch of some sort sent extra to the CPU.

May be best to just use 1 cable when you get the new 9800X3D.

They probably thought you only used 1 connector and want you to try both. (assuming they have not been told you already have)
Click to expand...

Interesting, thanks as I didn't know that. TIL.

I had presumed both needed connecting due to both ports existing on the motherboard and the RM1000x came with the appropriate CPU cabling, so I opted to just plug in both right away.
 
Jerome said:
Interesting, thanks as I didn't know that. TIL.

I had presumed both needed connecting due to both ports existing on the motherboard and the RM1000x came with the appropriate CPU cabling, so I opted to just plug in both right away.
Click to expand...
First off let me apologise for what you are going through. With the two dead 9800x3d and AMD RMA service. So here is food for thought. The two EPS 12v should be plugged in if you are overclocking but you can run it on one.

So on Reddit there have been an upturn in the number of people who have had 9800x3d dying on them similar to your story. Now while there is no conclusive evidence at the moment they share similar traits. Most of the 9800x3d dying either use Asus motherboards or ASrock motherboards. Now while it has been known for a while that Asus likes to pump extra voltage though chips usually the Vsoc voltage aka 7800x3d burnout syndrome. There has been an increase if people with ASrock motherboards. There are two other reports of other manufacturer one is MSI and the other I can;t remember who it was. But a majority of them are Asus/ASrock motherboard owners. What do the two have in common? they are the same company, sister companies if you will. The name ASrock stands for Asus rock...and they are both owned by the same parent company Pegatron. That is not megatron evil twin pegging autobots.

i have been a lifelong user of Asus motherboards and the kind of voltages not just the 7800x3d or 9800x3d they shove on Auto would make you cringe. I have owned the 12,13,14th gen and on auto volts it not unusual to see it pumping close to 1.5v though a chip which means it not gonna last long. Asus motherboards traditionally like to pump higher system voltage than other manufacturers which is similar to AMD Vsoc voltage.

And I learnt a long time ago never to use XMP and I don't use Expo on AMD. I always manually set it. Now I am not sure if you plan to continue down the road with the ASrock motherboard that you have but if you do decide to then manually set your Vsoc. Set it to 1.2v or 1.25v. At 1.3v and above I am not sure how much the chip can handle before it burns out but not to the point like when Asus were pumping stupid Vsoc voltages though the first batches of 7800x3d.

I am also not saying everyone should ditch their Asus or Asrock boards either. Like I said earlier there is no hard conclusive evidence at the moment. But when you look at the current reports at face value especially on Reddit then one must exhibit caveat emptor. There is no guarantee this will happen to you if you use those boards just like there is no guarantee it won't.

If you decide you are no longer willing to proceed with this manufacturer then you can check out MSI or Gigabyte. I do hope you find resolution on this matter swiftly so that you can enjoy your setup finally.
 
Thanks very much for such an informative reply. I do hope it's occurring enough to warrant someone investigating it further and hopefully coming to a conclusion. Your explanations make perfect sense.

Expo is enabled although I've attached some photos of my bios to highlight the voltages that it's set. If someone more educated than I could provide guidance on if these numbers look correct, that would be amazing. I notice a few of the options are set to Auto and I'm not sure if they should be changed, although any recommendations would be much appreciated.

I do actually quite like this motherboard compared to the one I tested as a replacement back when this occurred the first time (ASUS PRIME X870-P WIFI). If it could be concluded that it's the board and not the CPU then I'd be happy to change it however, probably to another brand (as you suggested MSI or Gigabyte) entirely.

LdGnNn0.jpeg


F1AC9lg.jpeg


ja3iSwv.jpeg
 
@Jerome thanks for the pics just looking at that it basically the same as mine although I don;t use ASrocks I use Asus predominantly. So I was at work tonight and I was pondering why your board has killed 2 9800x3d but works. Then it dawned on me but it can;t explain the other users getting dead CPu's Please bear with me it a bit long winded.

This is what I gather from your OP and subsequent posts if I have any of the details wrong or mixed up then please correct me. It has occurred to me that you did get two initially working CPU's and then both after a week you said around the same time frame they died. It doesn't;t make sense for any CPU past or present to die after a set period of time under the same circumstances does that seem illogical? If it was one died one day and then another 2, 5, 10 whateverdays that seems more likely than stopping dead at pretty much the same point in time UNLESS while your motherboard work I have a sneaking suspicion that some sort of SMT has dislodged, knocked off, Died through either transportation or installation. The damage is no critical enough to stop the motherboard from working. But is enough to slow burn your 9800x3d does that seem more likely? And because it is a certain component that had dropped off or died then the damage caused to your CPU is done at the same constant rate.

Now I read you tried the 7700x and 9600x with the mother board and no issues right and they were fine yes? The part that is damaged is the part that access the 3d cache and kills it, that's why the 7700x and 9600x is unaffected.

So what are you options? Well I would strongly advise you from using this motherboard any further when you next get the 9800x3d. You can get another ASrock or asus or whatever you feel like. I don;t know how well your RMA would go if they ask the reason and you said some random internet guy says it killing my cpu's will go down well and unlikely to get RMA approval.

With all that you been though for the sake of that particular board I would take the L on it. Move on with a MSI or Gigabyte. Personally if it were me when you get this third 9800x3d I would not even risk it. I would also like to add while I do have a little electronics background it not qualified enough to say that this is definitively what is happening. I haven;t set foot in the electronics world for over 25 years and it has been that long since it was my career.

Good Luck.
 
It sounds sensible enough and I wonder if it could be the cause. I'm using a 7800X3D this time around (instead of the 7600X previously) and I've tried putting that theory to the test over the last couple of weeks, mentioned below.

Some small updates:
  • AMD have accepted and acknowledged the second 9800X3D is dead and will replace the unit.
  • Over on the ASRock Reddit page, there's now a megathread discussing roughly ~40 individual posts which mention dead 9800X3Ds at varying intervals and from varying motherboard manufacturers (although ASRock seems disproportionately represented). Similar circumstances of dying at random days/weeks after installation.
  • They claim both AMD and ASRock are aware of the issue but no specific information as to the cause/fix are available as of yet.
I've been using a 7800X3D for roughly 12 days now with no issues. I've left my PC running overnight for a few nights extra to expedite the running time and no issues have appeared.

I contacted my retailer to replace the X870 board but they haven't replied to my ticket so I may end up ringing them.

As always, everyone's input has been very much appreciated and I will continue to post updates.
 
Hello everyone, just another update to check in.

I have used the 7800X3D for just over 3 weeks now and no issues. EXPO enabled, I've been leaving my computer on overnight to extend the runtime, etc. No issues.

Today I've bitten the bullet and installed the replaced 9800X3D which I received a week ago.

Over on Reddit around 40 people have reported their 9800X3Ds dying with the vast majority being ASRock boards, although it has also happened with many other manufacturers too so it can't be conclusive. ASRock released a new BIOS, version 3.20, which has been recommended for those with certain boot issues. They didn't go as far as to take any responsibility for the processors dying but they are certainly aware of it. I believe it reduces the voltages ever so slightly on the memory when EXPO is enabled, which I have enabled again.

Let's see what happens. Round 3, ding ding.
 
I've got a x870e tachi lite and I'm on bios 3.2 now. Didn't have any boot problems. I updated to 3.16 when I got the board. I've been following the reddit threads too.
 
You must log in or register to reply here.
Back
Top Bottom