Voltage outside of allowable limit (Starts up when IDRAC is removed) | PowerEdge R910

Hi SidK77,

sounds that there is something going on on the motherboard. The iDRAC itself is not mandatory to pass the POST, so the next steps I recommend are the Minimum-To-POST configuration.

This means removing all components except:

• Two Processors to POST and boot to DOS or OS (can use one installed, but system emits an error message for improper CPU configuration and will not hand off to boot)
• Two DDR3 RDIMMs in Slots 1 and 2 of Memory Riser_A
• I/O riser installed
• One power supply
• 2.5" SAS backplane (4-drive or 16-drive) installed and connected to the system board
• Control panel connected to the system board
• Power Distribution Board connected to the power supply and system board

As you can see, there is no iDRAC needed. If the system passes the POST you may test again with just adding the iDRAC.

If the no POST is back, it's very likely the iDRAC but can also be the slot on the motherboard. 

You may try to continue the re-installation of the other components but leave the iDRAC out. If the issue does not come back, well you might think of getting a replacement iDRAC then.

I hope that helps, let me know the results.

Best regards

Stefan

Behavior with no I/O Riser attached:

https://youtu.be/D6gaK40uSPY

I didn't show the PSU led in the video but it turns amber once the motherboard turns off. Also, I mentioned 3 Green LEDs in my previous reply but it's actually 2 and not 3.

With I/O Riser Attached:

https://youtu.be/gkOUnPu6S38

With I/O Riser removed and PSU attached to Slot 1:

https://youtu.be/YSB5r4SFztU

Please also note that when all PSUs are connected to the server, the behavior is the same and depends on what PSU the power cord is connected to i.e if the power cord is connected to PSU 1, the server behaves as shown in the attached video to this reply. If it's connected to a different PSU, the 15 secs power behavior is shown by the server.

If 2 power supplies are powered and attached (One at Slot 1 and one at Slot 2/3/4) while I/O Riser is removed, The server shows the following behavior:

https://youtu.be/-NRLWg-2Qmw

When I unplug the power cord from PSU at slot 1 while the power at slot 2 is connected, the server powers up and then powers down after exactly 15 secs.

Also tried swapping the 4 PSUs when was only using 1 PSU for minimal to post config.

When the I/O Riser is connected the light on the PSUs stay green indefinitely.

Hello,

I'd like to bring to your attention that when I mentioned IDRAC in my post, I was referring to the I/O Riser (The board that IDRAC attaches to, the removal of the actual IDRAC module from the I/O Riser doesn't make a difference)

I performed the diagnostics as explained, and found the following results:

When the I/O Riser is removed and a single PSU is attached to Slot 1 of the Power Distribution Board and all the other slots are left empty, the server doesn't perform POST regardless of what components are attached.

When the I/O Riser is removed and the server is in Minimum-To-POST configuration, and PSU is attached to Slot 2/Slot 3/Slot 4, the fans on the server starts up with no red LEDs on the board and 3 green LEDs. After exactly 15 seconds the fans turn off and one green LED on the mainboard dies as well (only 2 solid green lights are left on), the LED on the PSU is green for the 15 secs duration but turns amber once the server shuts down. There is no display during the time when the server turns on. There is no display/backlight on the LED.

When the I/O Riser is installed, the server shows the error message "Voltage outside allowable limit" on the display and the fans don't turn on regardless of what slot/number of PSUs connected.

I installed all the components and the results were the same. The only thing that gives me different results is the slot of the PSU and the presence of the I/O Riser. I also removed the actual IDRAC module from the I/O Riser but that didn't make a difference.

Hello SidK77,

Does the LCD show any error number for Voltage outside allowable limit message?

Something like VLT####?

Nope, There is no error code on the LCD. Just the message.

Hi SidK77, 

The blink code error your getting isn't matching up with the documentation I've been able to locate exactly, but it may be that the information I have to reference is for a different CPLD revision than what you're on. To resolve the error, I'd be looking at replacing the motherboard (MB) and power distribution board (PDB), based on what's been described in the thread. 

The blink pattern most closely matches a failsafe lockout and the steps you've taken with the NVRAM clear should have cleared that condition, if we could clear it manually. Hazarding a guess here - but I'd think the issue may be with one of the onboard DC buck converters.

The continued behavior at the min to POST config reinforces PDB and MB to me, and is what I'd be looking at replacing in a warranty situation. If you have alternate known good parts available to isolate the PDB and MB (such as memory and processors) you could try to further confirm this idea. However, it is something I'd be cautious about, because if there is a power conversion issue and too much voltage is somehow applied to one of those components, another failure could happen. With a warranty, that's less concerning because we could just send more parts, but if it's out, you may want to take that into consideration, as well. 

Hopefully, this is of some use to you, but please let us know if you have additional questions.