X9DRD-EF won't boot anymore...

[nephri]

Hi,

I installed a X9DRD-EF into a CSE-815TQ-600CB chassis (1U).
With it, i have 2x E5-2670V1 and 16x 8Gb Micro DDR3 1600Mhz.

This week-end i was able to boot this server with 14 DIMMs and i installed Windows 10 sucessfully (for test the hardware).
I had issue with 2 dimms slots and i contacted supermicro support for that.

They asked to me to make few tests.
Today, i'm going to make theses tests but the server don't want to boot anymore....

When i power on the server, the power go on for around 1 second. Not enough for the fans to go to full speed. After this second, the power off for around 3 seconds. And it try again to power on.... but after 1 second it power off and it repeat this schema indefinitely.

I removed the motherboard for check short circuit (no success)
I removed all dimms.
I removed heatsink for seeing the "state" of both CPU.
I checked all power coord (24 pins, 8 pins, 4 pins, ...) and all are in place...

I don't see what i can do for resolving it !!
I replied to supermicro support about this, and i will see what they says.


I'm really bored and I'm having a hard time sleeping tonight....

If someone have some advices for helping, i would really appreciate it !!!

Best regards,
Desesperate Séb....

 

[nephri]

I'm trying an RMA from SuperMicro EU...
I will have the answer tomorrow... I'm very unsure it will be accepted.
The motherboard came from US, was a second hand (at least), i dont know the official reseller and i don't know if the motherboard is always under warranty.

So i test even if i have small chances the RMA will be approved by SM. Wait & see.

If the RMA will be rejected, i will see to open a dispute from the ebay seller...

It has few chances this is a large waste of money :-/

 

[frogtech]

Did you try another power supply,?

 

[nephri]

Yes i did...
I tried another PSU, it remains the same. I tried also cleared the CMOS.

 

[nephri]

As suspected, SuperMicro don't support the motherboard. they says to request RMA to my reseller directly.

 

[nephri]

I cleared again the CMOS (i think the first time i failed it) and the server boot again.
Now i have the "I" red button on the chassis that light up
I'm installing IPMI for seeing what is this alert because temperatures seems good.

 

[Fritz]

Reminds me of the time I ordered RDIMMs for my Lenovo S20 and had to send them back because it can only use UDIMMs. I hate gotchas.

 

[nephri]

the red blink was "fan failure". It seems replug a fan on FAN1 and the light off.
At idle fans are at 11000 RPM that is quite loud.

I ran a Prime95 and fans keeps to 11000 RPM but cpu2 temps reach 80°C
Do you think it's abnormal (it's a 1U chassis with passive heatsinks and an airshroud.

5 minutes idle after the Prime95, the cpu1 is to 39°C and the cpu2 to 53°C

I found the cpu2 a bit hot at idle...

 

[Diavuno]

Those chips are oct cores, so I expect some heat.

CPU2 normaly runs warmer, but not THAT much.

Try swapping heatsinks and applying new thermal paste, remember less is more.

I've seem too much/poorly applied paste add 15*C

 

[nephri]

i changed the CPU, removed the thermal paste and reaply it. Checking the airflow through the airshroud....

The CPU2 is very hot comparing the CPU1.

at 100% load, CPU1 is stable to 60°C ~ 65°C and the CPU2 is to 80°C ~ 85°C...

I did'nt know what i can do for avoid that.