20k paperweight... Motherboard issues with Zenith II Extreme alpha

[Guest]

1 hour ago, theskaz said:

didnt do anything new. the screen is stuck at the BIOS logo, and the OLED is stuck on Detect HDD. 

 

On this mobo, you can switch BIOSs. like from A to B. On one bios, I just see a non-blinking cursor, the other one is this. I wonder if there is a way to flash both BIOSs to repair them....

Install a blank sata drive if you have one. May have to populate a drive somewhere to get past the detect hdd. Doesn't have to be an operating system I don't believe.

[theskaz]

i have tried various drives. seems to be getting worse. I opened a ticket with ASUS and they are RMAing it

[Guest]

1 hour ago, theskaz said:

i have tried various drives. seems to be getting worse. I opened a ticket with ASUS and they are RMAing it

Well you did all you could on your end. Now it's their turn.

Take a picture of the stickers on the board to see if you get the old one back or not. 

[theskaz]

ah, good catch

[theskaz]

@ShrimpBrime

 

Got the replacement yesterday. 

 

hooked some stuff up, not all, and smoke tested it. It made it to "Load VGA Bios" which is a good sign. turned it off because I needed to work on something else and came back later and turned it back on. the OLED showed "CHECK NVRAM" so I did a bios reset. Now I get "Check CPU" left and right. sometimes it does it at code 63, 24, once at 00. I updated the Bios using the flashback method and still the same error. I have reseated the CPU a couple times and ensured that it is torqued correctly. This is a replacement CPU

 

I do not have another threadripper to throw in there for testing. I have been troubleshooting this system since Nov 2020. If I had hair I would pull it out. 

[Guest]

1 hour ago, theskaz said:

@ShrimpBrime

 

Got the replacement yesterday. 

 

hooked some stuff up, not all, and smoke tested it. It made it to "Load VGA Bios" which is a good sign. turned it off because I needed to work on something else and came back later and turned it back on. the OLED showed "CHECK NVRAM" so I did a bios reset. Now I get "Check CPU" left and right. sometimes it does it at code 63, 24, once at 00. I updated the Bios using the flashback method and still the same error. I have reseated the CPU a couple times and ensured that it is torqued correctly. This is a replacement CPU

 

I do not have another threadripper to throw in there for testing. I have been troubleshooting this system since Nov 2020. If I had hair I would pull it out. 

I suspect these systems just don't like your memory kit after going back and reading the previous page to the thread.

 

f4-3600c18q2-256gtzn

 

Do you have any DDR4 memory that has smaller modules? Like a 2x8gb kit from a desktop or office PC or something>?

[theskaz]

23 minutes ago, ShrimpBrime said:

I suspect these systems just don't like your memory kit after going back and reading the previous page to the thread.

 

f4-3600c18q2-256gtzn

 

Do you have any DDR4 memory that has smaller modules? Like a 2x8gb kit from a desktop or office PC or something>?

I have a 3200 kit that i havent sold yet. I will give it a shot

[theskaz]

@ShrimpBrime

 

Tried a 3200 stick and same error. it seems to be mainly after CPU Code: 63 or 06

[theskaz]

also tried a 16GB 3200 sick, same issue

[Guest]

48 minutes ago, theskaz said:

@ShrimpBrime

 

Tried a 3200 stick and same error. it seems to be mainly after CPU Code: 63 or 06

All cpu errors. Stops/hangs at micro code DXE loading... I can't imagine this kind of terrible luck. 

 

I can only think to reseat the Cpu until the damn thing posts. You've already done everything else.

 

[theskaz]

2 hours ago, ShrimpBrime said:

All cpu errors. Stops/hangs at micro code DXE loading... I can't imagine this kind of terrible luck. 

 

I can only think to reseat the Cpu until the damn thing posts. You've already done everything else.

 

So i reseated it 5 times. The last time i didnt put the heatsink on. It booted up. Then i pushed the heatsink on with my hand. Booted up. Installed the heatsink all the way. No boot. I have tried different pressures and there seems to be nothing safe that works....

[Guest]

1 hour ago, theskaz said:

So i reseated it 5 times. The last time i didnt put the heatsink on. It booted up. Then i pushed the heatsink on with my hand. Booted up. Installed the heatsink all the way. No boot. I have tried different pressures and there seems to be nothing safe that works....

Ah finally some progress. So with a tight mount you're getting some board flex.

 

Screwing the cooler on makes it no post even if the cooler is not tightened down?

[theskaz]

1 hour ago, ShrimpBrime said:

Ah finally some progress. So with a tight mount you're getting some board flex.

 

Screwing the cooler on makes it no post even if the cooler is not tightened down?

correct. if it is as loose as possible it will post, but im scared to keep it that loose.

 

and there is a massive metal plate on the back.. man i hope i dont have to take it apart yet again...

[theskaz]

@ShrimpBrime 

 

I pulled everything out of the case and put it on a test bench. I bought an air cooler to see if it was the waterblock. could never get it working. Finally bit the bullet and RMAd the CPU and got this:

"

This is a confirmation notice from AMD Global Customer Care with

    regards to Cross Ship Warranty Request RMA# xxxxxxxxx. for your AMD

    retail packaged Processor(s) in a Box."

 

"Following up from our previous emails. Your replacement processor is

    ready to ship."

 

Looks like the CPU was bad too... Ive replaced both of these twice... what gives lol. 

[Guest]

4 hours ago, theskaz said:

@ShrimpBrime 

 

I pulled everything out of the case and put it on a test bench. I bought an air cooler to see if it was the waterblock. could never get it working. Finally bit the bullet and RMAd the CPU and got this:

"

This is a confirmation notice from AMD Global Customer Care with

    regards to Cross Ship Warranty Request RMA# xxxxxxxxx. for your AMD

    retail packaged Processor(s) in a Box."

 

"Following up from our previous emails. Your replacement processor is

    ready to ship."

 

Looks like the CPU was bad too... Ive replaced both of these twice... what gives lol. 

They probably didn't even test it and just OKed the RMA being these chips have known whea error issues. So they just send you a new one and be done with it.

 

But so far, no good luck on your end. 

I hope this time it all goes well and boots without issues. 

 

But If you still have problems, we'll have to continue the trouble shooting. Hope not for you though. 

 

And thans for the update btw. I was wondering about it over the weekend actually.

[theskaz]

Here we go again

 

[theskaz]

@ShrimpBrime

on the bench, 

 

if i reset the bios, and boot, it POSTs. i can get into the bios. i hit F10 and reboot and the memory says "C0" and then says "Memory Test". Ive tried a few different configurations, and nothing seems to be working. if i reset the bios, it will boot again. ive flashed 2 different bios's so far thinking its a bios issue.

[Guest]

1 hour ago, theskaz said:

@ShrimpBrime

on the bench, 

 

if i reset the bios, and boot, it POSTs. i can get into the bios. i hit F10 and reboot and the memory says "C0" and then says "Memory Test". Ive tried a few different configurations, and nothing seems to be working. if i reset the bios, it will boot again. ive flashed 2 different bios's so far thinking its a bios issue.

If you have an operating system installed, skip bios setup and see if it posts straight to the operating system. Then do a regular restart, go all the way back to the OS again. 

Then check for those whea errors with a stock (hopefully) running system ignoring the bios for the time being. Don't change hardware configuration either. leave it as you need it.

 

One thing you can double check for is good grounding at the wall outlet as well. I know it kind of sounds far fetched, but it's worth a shot to at least inspect it. 

This may require yo uto turn off the breaker for that room and pull the outlet. Be sure the screws are tight and the green wire is grounded. 

Not sure if you have a line conditioner, that may be worth a shot as well.

How would this effect the PC? Reduction in power ripple. Need nice smooth current.

 

And the AMI code C0 = nothing. You would use the diag speaker at this point. The system hung on a unlisted post code which usually are " reserved for future AMI codes". 

[theskaz]

@ShrimpBrime

 

Ok, ive been testing a lot and here are some findings

 

1. I changed the circuit that the server was on from a normal 20amp to a 30 amp behind an APC UPS. thats the best im going to get. the wire from that outlet to the panel is about 6 feet. I have a whole home surge protector in place too.

 

2. I have been able to populate 7 sticks of ram, and still get it to get past C0. i think i have a bad stick of ram causing it to hang at C0

3. 10G Nic will turn off and on during the POST and ulltimately remain off once OS boots.

 

if channel 4 has memory in it, the BIOS screen will not pop up. the POST results will, and the OS (Unraid) will boot, but I cannot get into the bios. Ive flashed 4 different bios and the all show the same behavior. I cannot get the memory up to its rated speed at all. if I go into the bios (3 channels of ram or less) and do ANYTHING, it locks up. also, if the 4th channel is populated, the "Load Video Bios" screen stays on for about 2 minutes and then will finally boot. again, not seeing the bios screen, just the POST image (asus logo). While the OS is loading, I dont see any errors. Im going to run a mem test overnight to see if anything shakes out.

 

what is the possibility of being a bug or something due to my hardware or something? I do have a SAS controller connected that has all my harddrives on it and 5 NVME drives on the motherboard.(2TB each)

[theskaz]

Old issue came back up. So the first and second set of cpu/mobos would cause the 10G nic to temporarily knock out my router. it just happened again, and causing a Kernel panic and having to reset the Bios, otherwise I get "Check CPU". 

 

this is very repeatable. if the 10G nic is plugged in, any port on my switch or router (ubiquiti) it will knock out my router and cause a kernel panic. if it is not plugged in, it will work fine. 

 

so 3 different cpus, all the same model, 3 different motherboards, all the same model, 3 different sets of RAM, and 2 different routers. all go on the fritz when the 10G nic is plugged in.

 

WTF

[theskaz]

now i get Kernel panics everytime i try to connect to the server through the web interface... will try to install windows, see how that goes

[OCNewbee]

jesus man. I am never getting a threadripper

[Guest]

5 hours ago, theskaz said:

 

so 3 different cpus, all the same model, 3 different motherboards, all the same model, 3 different sets of RAM, and 2 different routers. all go on the fritz when the 10G nic is plugged in.

 

WTF

Ok, so the NIC card is bad?

[theskaz]

4 hours ago, ShrimpBrime said:

Ok, so the NIC card is bad?

its the onboard nic.... I think its more than that.3 onboard nics causing the same issue?

[Guest]

6 hours ago, theskaz said:

its the onboard nic.... I think its more than that.3 onboard nics causing the same issue?

I got nothing.... lol

 

So thinking on it... you can enter bios. F10 save and it doesn't post. Uhg where do we start.

 

RSVD switch be sure is disabled position. Should be by default. 

 

Ideas to try:

 

• When PC doesn't post after F10, try a cold post after you shut it down. But don't clear cmos.
• Hit the "try again" button when the system hangs at the C0 post code instead of clearing cmos.
• Try the "safe boot" during the C0 hang up. It will default the board without having to clear cmos. Basically the same thing, but won't wipe passwords.
• The Flex Key button you can use to restart the board as well. In bios you can designate Aura lighting for example, like a hot key.

 

Lastly, it seems pretty obvious the chip doesn't like the memory for whatever stupid reason. I find this one to be most concerning, but you have a system that posts, it's just not in useable condition if you can't run all the memory. I'm stumped frankly, but can only recommend to run the memory at Jedec speeds which is 2667mhz or less, often times 2133mhz being the most stable. As a temporary thing for diagnostics of course. But I don't expect this chip to run that much memory at say 3200mhz. I'm pretty sure that's a little past the memory controller's capability to set up for daily work use. 

 

Unfortunately you are half a country away, I'd probably be of better help with hands on at this point. But I'll keep plucking away with ideas for you as long as symptoms keep changing, we are making some sort of progress.