I built a new UnRaid server with Supermicro X9-SCM motherboard / 16GB ECC RAM / Xeon E3-1230v2 CPU and 3 LSI HBA cards. This was built into a 2nd hand 24 bay server case (this was as new as it had never been used) with a new 850W PSU.

I bought the motherboard / CPU and memory as a bundle on eBay with a screenshot of the BIOS from a private seller who had used it for a UnRaid server for a number of years. I mounted them in the case with PSU and HBA cards, it would not respond to the power switch and when shorting the motherboard header it would POST but have no video.

I returned the motherboard / CPU and memory to the seller for a refund

I bought a another Supermicro X9-SCM motherboard from a recycler on eBay sold as tested, a new unopened Xeon E3-1230v2 CPU and 16GB ECC RAM.

I had this running with just the motherboard / CPU and memory on a bench for 24 hours OK

I've then installed it in the server case with the HBA cards, again the motherboard won't start using the case's power switch. I've shorted the motherboard header and I get garbled video as the board starts up.

The LED's on the HBA card are flashing OK

I've tried resetting the BIOS, replacing the CMOS battery and resetting the BIOS, and after number of attempts the board will not longer POST at all. There is no power LED lit and the IPMI LED does not flash either.

I am at my wits end and I just can't see how installing the motherboard in the server case can cause the motherboards to fail twice in a similar way?

I have checked that the motherboard standoffs are in the right location, so that is no shorting to the underside of the motherboard.

Perhaps the power button is faulty, but could the connections for the power button / NIC LED's / reset switch and HDD LED be causing damage to the motherboard, I can't see it though?

I've returned the motherboard for a refund

Could it just be coincidence that 2 motherboard have failed in a similar way, they are old the last one was build in 2012 and may have had a hard life in a data centre.

I've tested the PS are OK but just don't know what to do next???

I'm tempted to buy a new motherboard / CPU and memory from one supplier, but wary about installing the server case?

Whilst on test the BIOS reported the -12v was -63V, but the motherboard booted and worked OK on soak test.

I have another UnRaid server with the same motherboard and all voltages are shown OK

Could I have a duff PSU, on a dumb cheap PSU tester only 1 of the 2 12v LED's are lit?

I am looking at a more expensive tester like the Thermaltake Dr. Power III ATX Power Supply Tester that does ATX12v3.1

Before installing another replacement board, I would have the PSU and its cabling checked. Two similar failures in the same build make the shared hardware worth investigating first.

Did the successful bench test use the same 850 W PSU and exactly the same cables? What are the PSU and chassis models? Were any modular cables reused from another PSU?

Modular cables can have different pinouts, even within one brand. Use only cables approved for the exact PSU model.

For the next controlled test:

1. Use a known-good PSU with its own cables.

2. Test the board outside the chassis on a nonconductive surface, with CPU, cooler, and one compatible DIMM.

3. Leave HBAs, drives, backplane, front USB, and the complete front-panel harness disconnected.

4. Connect both the 24-pin motherboard supply and the correct 8-pin CPU supply.

5. Confirm the exact board model before identifying the power-switch pins. The X9SCM family manual specifies JF1 pins 1–2.

Disconnect AC before every hardware change, and wait until the standby LED goes out.

If that setup works, inspect chassis contacts and trapped wiring before installation. Then reconnect items individually, with a startup test after each change.

Was the unlit “power LED” the case LED or the onboard standby LED? That distinction matters.

The reported −63 V is an unverified BIOS reading, not proof of the actual PSU output.

It's a brand new Corsair RX850x PSU and I used the same modular cables to feed each motherboard. The enclosure is a Inter-Tech 4U IPC 4U-4724 Storage Server Case. I also tried an old Antec 500W PSU that have in my spares box too.

I've bought a PSU tester which I should have tomorrow to test that the PSU is working OK.

After the first board, I had the 2nd board installed with just the CPU / 2 sticks of RAM installed on a non conductive surface , and it worked faultlessly just showing the BIOS screen. The board power LED was on and the IMPI LED was flashing. Once the board would no longer POST both LED's were no longer lit.

The issue started only after the board was mounted in the case, front panel wiring was connected and HBA's were installed, although the video was garbled the LED's were flashing OK on the HBA cards.

So the only common components between the boards are PSU / HBA cards and case.

I've had HBA cards fail before but it simply caused errors with reading disks, I can't see how they could effect the board unless the board were faulty and couldn't supply enough power to the cards? The HBA cards are used, but bought from a eBay seller who has flashed them to IT and tested before sale.

I'll check the case out with a mutimeter to see if I can discover any faults with the front panel wiring.

I am tempted to try another board, soak test with just CPU & memory. Then soak test with HBA cards, adding one at the time. Then install in the case but depending on what I find on testing, not use the front panel wiring - I'll wire my own power button.

The attraction with this motherboard is that it has 4 PCIe slots which I can used for 3 HBA cards and a NvME card for pool drive.

I already have another Unraid server using the same hardware, which has run for 10 years with 20 hard drives and I have only had to replace some HBA cards after they started to report faults after unclean shutdowns after continued power cuts in a storm, even though I have a UPS.

Thanks, that clarifies that the onboard LEDs stopped working. If AC remained connected and the PSU was switched on, that makes the standby-power path worth investigating.

When you tried the Antec, was the board outside the chassis with the HBAs and front-panel wiring disconnected? Did either onboard LED return, or did it POST?

A faulty HBA can prevent startup through an electrical fault. Its LEDs flashing does not rule that out, but nothing yet identifies an HBA as the cause.

Your proposed staged testing makes sense, with one useful change: test the bare motherboard inside the chassis before adding the HBAs. Otherwise, mounting and card installation remain mixed together.

After checking the PSU, cables, and chassis:

1. Establish a working minimal bench setup with a known-good PSU and its own cables.

2. Install that same minimal setup in the chassis, leaving HBAs, backplane, and front-panel wiring disconnected.

3. If it still starts normally, add one HBA at a time, initially without its storage cables.

4. Connect the backplane and front-panel wiring in separate stages, with a startup test after each change.

Disconnect AC before every hardware change. Stop at the first abnormal result.

When the video showed as garbled after the board was installed in the server case , I took the board out of the case and removed the HBA cards.

I then put it back on the test bench with just CPU and RAM, with the new PSU.

Initially it would POST and but away a few attempts, the garbling was getting worse.

I then tried to clear CMOS and the board wouldn't POST. I took out the CMOS battery, cleared the BIOS and put in a new battery, this had no effect.

I then tried swapping with the Antec PSU, which I have tested OK with my basic PSU tester, but this had no effect either.

Thanks, that clarifies the sequence. The video corruption continued to worsen on the bench with the Corsair PSU, before you tried the Antec.

That makes a temporary chassis short an incomplete explanation. However, it does not rule out damage that started during installation. The unsuccessful Antec test also cannot establish what caused the original failure.

Were all Corsair modular cables supplied with that exact PSU? When you tried the Antec, did you use its own cables throughout?

Please share the new tester results and the exact Corsair model from its label. A basic tester pass alone does not establish reliable operation under system load. Corsair also documents misleading −12 V and power-good readings with some newer supplies and testers.

Since you returned the board, has the seller confirmed whether it works with their own test setup? That would be useful independent evidence before you risk another replacement.

I would have the PSU and cables checked before proceeding with the staged tests discussed earlier.

I have spotted that the X9-SCM motherboard needs -12v on pin 14, from reading forums I understand that Corsair PSUs, ATX 3.0 and 3.1 do not support the -12v rail.

I wonder if that is my issue, out the case at testing stage BIOS was reporting -12v at -63v. My working board with an older RX850x PSU reports -12v as OK

I have only just returned the 2nd board today so I await the supplier's comments.

The Corsair PSU has been used with the modular cables that it was supplied with. The Antec PSU is an older design with fixed cables

I'll report back once I have tested the PSU with my new PSU tester.