X9SRL-F does not boot while inside a chassi, but boots when outside the chassi

[ulf]

Hi, I am building two almost identical systems.

Both use the
* Supermicro X9SRL-F motherboard
* E5-2650 v2 processor
* Noctua U12DX i4 fan
* 48 GB DDR3 PC3L-12800R

I have a Supermicro SC743 chassi. The combination works in this chassi.
I also have Coolermaster CM Force 500 chassi, with a Coolermaster V750 semimodular power supply.
The chassi has a fan at the back. The board does not work in this chassi.

I have read that the Supermicro boards are sensitive to the power supplies.
The power supply has a 4+4 connector which I connect to the motherboard as well as the 20+4 connector.
The theory was that the power supply was the problem. The datasheet does not specify specifically EPS12V.

I then tried an EVGA 850W P/W which states 2xEPS12V but no luck.

I then tried connecting the power supply of the Supermicro chassi to the board in the Coolermaster chassi.
This did not work. So I changed my theory, and now clues were pointing at a faulty board.

After thinking a while, I decided to try the "faulty" board in the Supermicro chassi.
Suddenly the board boots....

I tried the working board in the Coolermaster chassi, and it does not work.
The LEDs turn on, but not the CPU.

I then try the Coolermaster chassi approach again.
I place it on a piece of cardboard on top of the chassi,
and reconnect the wires (ATXx24,EPS12Vx8, USB and 2x4 front panel connectors, a header to use the internal sound)

Then the board boots.

The board boots when in a Supermicro chassi, when using the Coolermaster power supply.
========================
How can a chassi affect the board?

As I see it, I have the following connections.
1. 8 screws for the mounting holes on the motherboard.
2. 8 distances between the chassi and the motherboard.
3. Backplate grounds the board - except the chassi is painted.
4. Id Button turning on a Blue Led at the back
================
5. ATXx24
6 EPS12Vx8
7 Front Panel connectors (2x4)
8 USB to front panet.

When the board is outside the chassi, the difference is 1-4.

=================
Why would that make a difference?


I tried booting the board while the ID button is pressed, without success.
It does not appear to affect the boot.

Very strange problem where I would appreciate fresh ideas.

Edit:
The backplate seems isolated from to the chassi since both the inside and the outside of the chassi is painted all black.

 

[Nick2253]

First off, are you sure you're doing everything the same: same headers hooked up? Same drives? Same PSU? Same memory/CPU? etc.

I would say your problem is that you are either not doing something the same, and that different thing is causing the issue. Or, more likely given what you wrote, is that your motherboard is in fact grounding out on the case. Just because it's painted, doesn't mean you can't ground out against it.

My suggestions are as follows:

1. Get to a completely minimal working example. What's the least hardware you need to get the system to boot?
2. Try that hardware installed outside of the chassis. Confirm that it boots!
3. Move that hardware into the chassis. Change nothing else! Plug in nothing else!
4. Try it. Does it boot?
5. If not, to rule out the chassis grounding out the motherboard, set the motherboard on a piece of cardboard inside the chassis. It should boot now if that theory is correct.
6. If that theory is correct, get longer standoffs for the new chassis.
7. ????
8. Profit!

 

[JaimieV]

Yes, grounding issues. The piece of cardboard making it work is strong evidence!

It may not need longer standoffs, it may need one taken out or moved to the correct position.

 

[ulf]

Yes, that was the issue. I asked Supermicro, and they suggested the same thing.
The chassis has marking for all the standoffs for ATX and mini-ITX etc.

A B C
D E F
H
I
J K L
I mounted all of them according to the chassi instructions initially,
but quickly discovered that there was no hole for "E" so I moved that standoff to "I".

The X9SRL-F has the following holes:
A B C
D E F
H
I
J K L

But I forgot the standoff remaining at "B", which is right under the DDR3 memories.
Have not tested that yet, but I think that should solve the problem.
Sorry about bothering people about basic problems, but I thought it was either a motherboard
problem or a power supply problem, and when I finally decided it was a chassi problem
it was late and I was tired.