SOLVED X10SLL Won't Turn On

[GreyKatana]

Hi guys, finally received my Supermicro X10SLL-F (after receiving a X10SLM+F with a bent serial port), built the whole system tonight, power led is on and heartbeat led is blinking, everything's looking good.
I push the power button... nothing. Not a beep, no fans start spinning, absolutely dead silent.

So I check the front panel connections and everything seems good. I even tried bridging the power button pins (as suggested in the Supermicro manual) and nothing. There's power going to the board otherwise the LED wouldn't light up, but nothing else works.

Do you guys have any ideas? At this point I'm guessing bad mobo. This is my first Supermicro product (well second after the broken X10SLM) and if this board is DOA, this might very well be my last.

Full specs just in case (but since it doesn't turn on I don't think it matters anyway):
Supermicro X10SLL-F
Pentium G3450
8GB Crucial 1600MHz ECC Memory
Seasonic G-360
Fractal Design R4
2 x 1TB WD Red Drive
2 x 32GB Sandisk Ultra Fit Boot Device
And a whole bunch of disappointment

Thanks for the help.

 

[DrKK]

I know your problem, I think.

You probably got a board with the version 1.0 BIOS.

The V1 BIOS that came with that board maybe doesn't support the G3450. I can't remember if the G3450 is one of the "new"/"refreshed" pentiums (pentia?). If it is, the symptom with the X10 series motherboard is that absolutely *NOTHING* will happen when you hit the power button. I am 99.11293% sure this is your problem.

There are two ways out of this:

1) Update the BIOS from your mobo. But in order to do this, you'll need to boot/post on a legit CPU SKU, which you don't have.

2) Update the BIOS from IPMI. But in order to do this, you'll have to pay SuperMicro for a "product key" for that feature. Which I believe is like $25.

or

1a) Borrow a CPU from a buddy. Any of the original/old school 1150's will work fine, except maybe the i5's and i7's.

 

[DrKK]

There's almost no chance it's a bad mobo.

Odds are 99.9% it's that CPU does not work with the v1 BIOS. I just did some checking.

 

[DrKK]

Also, in fairness sir, there are several places in this forum where we tell users that if they are getting Haswell refresh CPU's, that if their X10 board is v1 BIOS, that they will be in this pickle.

Maybe you can return that CPU, and get a G3220. There will be absolutely no discernible difference in performance whatsoever.

 

[Ericloewe]

Yeah, that's a Haswell Refresh CPU.

If it is the BIOS versions, there are two options:

Update the BIOS with a supported non-Haswell Refresh or Broadwell CPU (and keep it or sell it or whatever)

OR

Fork over 20 bucks to Supermicro for an IPMI BIOS update license. This method flashes the BIOS ROM using the BMC, bypassing the need for the system to boot.

 

[GreyKatana]

Thanks for the info guys.

Also, in fairness sir, there are several places in this forum where we tell users that if they are getting Haswell refresh CPU's, that if their X10 board is v1 BIOS, that they will be in this pickle.

I had read those posts, but IIRC the were all mostly dated from 2014, I guessed that SM would have updated their BIOS since then. Also in my build thread, nobody flagged anything about the CPU. Everyone said that it was a good buy.

OK, on with more positive thoughts. If it's the BIOS version, and it could be updated through IPMI. Even without the code I guess I should be able to access the IPMI interface right?

 

[DrKK]

Well, it's not SuperMicro, it's the vendor you bought it from. Sometimes they have old stock. If you boguht the board TODAY from SUPERMICRO directly, it would not have v1 BIOS.

You can access the IPMI interface, sure. Just figure out what the IP is. Look in your DHCP table, or download "fing" for Android, or something. But, the BIOS updater in the IPMI will *NOT* start without the product key from Supermicro, and I believe it is $20-$25.

 

[GreyKatana]

OK but at least this way I could confirm the board is not dead and possibly see the BIOS version.
Now to figure out how IPMI works...

 

[GreyKatana]

Well, it's not SuperMicro, it's the vendor you bought it from. Sometimes they have old stock. If you boguht the board TODAY from SUPERMICRO directly, it would not have v1 BIOS.

I bought from NewEgg Canada but the board shipped from Los Angeles. Would have hoped the revision to be good as you said here:
https://forums.freenas.org/index.ph...0slm-f-bios-to-old-for-cpu.35756/#post-218571

 

[DrKK]

I bought from NewEgg Canada but the board shipped from Los Angeles. Would have hoped the revision to be good as you said here:
https://forums.freenas.org/index.ph...0slm-f-bios-to-old-for-cpu.35756/#post-218571

It's reasonable sir. I'm not saying you weren't reasonable on this. I would have bet $50 that you would have been fine. Turns out not, I guess.

Now, you *HAVE* gone into the IPMI (which should be running fine), and verified that BIOS version, right?

 

[DrKK]

It's very simple, you figure out what the IP address is for the IPMI (you can use "fing" for android, connect it via wi-fi, and see what IPs you have), then just punch that into your web browser.

 

[GreyKatana]

Thanks DrKK for following this thread and for giving me a hand.

I just tried accessing IPMI since I couldn't reach the machine during the day.
Connected the IPMI interface to my router, checked the routing table and dhcpd logs to find out that the IPMI interface was given out IP 192.168.1.7.
Tried using IPMIView, IPMIPing or just plain IP in my browser and nothing.
Everything says "Can't connect", "Packet dropped" or a variation on the same subject.

I did a "netstat -rn" to confirm that the route is set, see the Use counter increase as I try to ping it, but I can't connect.

 

[DrKK]

Do not connect the IPMI interface. This board comes out of the box in "failover" mode, meaning, that your IPMI will be on the same ethernet jack as your main network. 98% of us run it in this way. Just connect your ethernet cable to the main "igb0" port, not the special IPMI one.

 

[GreyKatana]

Interesting, will try it out tonight. Thanks!

 

[GreyKatana]

OK, some new info. IPMI works through the first network interface. Reports BIOS as version 3.0!
So the problem is not the Haswell Refresh.

Trying to Power On through IPMI, nothing happens.
I'm beginning to think that either I messed up something or the motherboard and/or CPU is really dead.

 

[DrKK]

OK, some new info. IPMI works through the first network interface. Reports BIOS as version 3.0!
So the problem is not the Haswell Refresh.

Trying to Power On through IPMI, nothing happens.
I'm beginning to think that either I messed up something or the motherboard and/or CPU is really dead.

I KNOW WHAT'S WRONG.

Sorry brother, I didn't think of this before.

The X10SLL will not turn on if the CPU fan is not attached to the FAN A header. Not Fan 1, not fan 2,
the CPU FAN **MUST** be attached to the FAN A header, or it will not turn on.

That's definitely what's wrong.

Sorry bro, based on the symptoms I should have thought of that sooner.

 

[DrKK]

Very interesting.

The manuals for the X10SLL+-F and the X10SLM+-F are quite clear that FAN A is the required header for the CPU.

However, this is less clear for the X10SLL-F and X10SLM-F, and we have one guy in the IRC that claims his is running fine without a CPU fan in FAN A.

In any case, the CPU fan should be on FAN A.

Definitely report back if that fixes it. And confirm if you have an X10SLL+-F or an X10SLL-F

 

[Ericloewe]

The manuals for the X10SLL+-F and the X10SLM+-F are quite clear that FAN A is the required header for the CPU.

But Supermicro's support has said that FAN A is meant for the PCI-e card area.

In fact, I think I moved my CPU fan from FAN A to FAN 2 or something, leaving FAN A empty (or at least I planned to...). I'll have to check when I get home.

Besides, it's quite common to have only a passive cooler on the CPU, with the rackmount chassis' crazy powerful fans providing all the airflow.

 

[DrKK]

But Supermicro's support has said that FAN A is meant for the PCI-e card area.

In fact, I think I moved my CPU fan from FAN A to FAN 2 or something, leaving FAN A empty (or at least I planned to...). I'll have to check when I get home.

Besides, it's quite common to have only a passive cooler on the CPU, with the rackmount chassis' crazy powerful fans providing all the airflow.

I agree with you. But let me just tell you, there have been AT LEAST TWO people in the past year with X10 boards that would not start, and the issue was nothing was attached to the FAN A header.

It seems clear, however, that unlike what I thought, this does not necessarily apply to all the models.

 

[GreyKatana]

OK will try it out tonight. I have a X10SLL-F.
I hope it is something like that because if it is a component, I don't have spare hardware to confirm with.

Thanks again for all the help guys. I'll owe you a beer after.