Gryphon Z87 -- Intermittently hangs at BOOT_DEVICE_LED during POST

[EddyKilowatt]

I have a new Gryphon build that only makes it through a cold-boot POST about 1 time out of every 3 or 4. I get the expected rapid sequence of CPU, MEM, VGA, and BOOT_DEVICE LED's, but... it hangs after a couple of seconds in the BOOT_DEVICE phase, my monitor stays dark, and I usually hear a small audible 'click' from any attached HDDs or optical drive. The BOOT_DEVICE_LED never goes out.

When I *do* make it through POST, the monitor lights up and I get the BIOS splash screen. I can either enter the BIOS, or let Windows (Win7 64-bit) load. As long as I make it through POST, the whole system works great and has for over a week -- I was able to install Windows to my SSD, copy my data files off old HDDs, install applications, and get back to work with Lightroom and Premiere. (Sorry, no games yet... hope you guys don't mind me hanging in your Forum.)

Once I've made it through POST, it seems like I can warm-boot, Reset from the BIOS, or Restart from Windows as many times as I want. Sleep and wakeup from Windows works every time as well. It's just powering down, and then pressing the power button back on, that seems problematic.

Things I've tried, none with any repeatable effect:
1) Tried both SSD and DVD drives as boot device
2) Reset CMOS with battery out (RTC reset to 1979 or something)
3) Re-flashed BIOS from stock (think it was 1007) to 1504
4) Updated Chipset, VGA, LAN, USB, and SATA drivers to latest from ASUS Support page
5) Changed PSUs from a 450W Delta OEM to a 430W Thermaltake TR2

Other than changing boot devices, I've been keeping the BIOS at all Default settings; in fact my memory is still running at 1333 MHz since I haven't enabled XMP settings yet while I figure this out.

Since this only affects cold boot, I am inclined to think it's a power sequencing or regulation issue, interacting with the Mobo somehow. But that's just a guess, one I'm not inclined to test with a ~$100 PSU purchase without further evidence. I'd love to be able to get some more insight into what's going on, if anyone knows where to find diagnostic tools for this board. I'm open to suggestions for other tests to try or components to check.

Or if anyone knows that BIOS 1603 addresses this issue, that'd be great too... ASUS is pretty cryptic about what each update addresses.

thanks,
Eddy

[HiVizMan]

Remove all hard drives from your system please and see how quick or slow it is to get into the bios screen.

[EddyKilowatt]

Looks like about 7 seconds from power-on till the BIOS screen first appears -- on those times when it does appear. The majority of the time it still hangs with BOOT_DEVICE LED on. This is with all SATA cables unplugged from their drives, although it's also ~7 seconds with the SSD plugged in.

[HiVizMan]

That is not how it should be - RMA the motherboard.

[Raja]

What memory are you using and which USB devices? Might be prudent to list that as well

Before the BIOS screen appears, the board performs a series of memory and PCIe training routines after performing power good checks. The memory training part can loop round a few times if the system is borderline stable during the training process - this also extends time to POST.

POST from AC power cycle or S5 is more "difficult" than warm reset as initial system training incorporates more routines. I

BOOT halt can also occur on unstable systems - it just manifests that way and can sometimes be misleading. Just things to bear in mind...

Could be what Vizbo said above, or what I said here. Good to know these things before going on a goose chase..

-Raja

[EddyKilowatt]

Memory is in my sig, 16 GB (2x8) G.Skill Ripjaws F3-1600C9D-16GXM, but still clocked at default 1333 (no XMP yet). No change when I drop that to one stick (8 GB) in the A2 slot.

USB complement is one generic mouse and one generic keyboard, everything else (card readers, ports, thumb drives) unplugged.

Just as a point of interest, the 7 seconds of POST that I observe breaks down into something like 1/2 second of CPU_LED, 1 second of DRAM_LED, 2 seconds of VGA_LED, and the remainder BOOT_DEVICE_LED. It's pretty much been that way since the first time I booted the board. I haven't measured, but I don't have the impression that the timing is much different between a cold power cycle and a warm boot, other than of course those occasions when I've reset the CMOS and it does a kind of double-boot thing and then goes to the American Megatrends screen.

What would a normal healthy set of LED blink behaviors look like for this board?

Assuming the board's running into trouble during the training routines... does it set any codes or flags that would be visible later, to help diagnose what's going on? Are there diagnostics I could run to probe things?

Once it makes it out of POST this system seems very stable... with SDDs and HDDs hooked up, I've moved hundreds of GBs of files around between the drives, rendered videos, run Prime95 while tuning fan settings, all without any hint of instability. It's just getting past that first "seven seconds of doubt" that's the problem...

thanks,
Eddy

[Raja]

LED behaviour is correct and POST time is correct I think.

Check the CPU socket for any misaligned pins - pay special attention to lower side (south).

Out of interest did you ever have any self powered hubs or drive bays plugged into the board?

-Raja

[EddyKilowatt]

Negative on the USB drive bays or hubs -- don't own any -- Just a USB stick for software installs. My Dell monitor USB was plugged in for awhile, it has a card reader, but I assume it's responsible about its behavior. (Haven't actually read any cards with it.)

I can check CPU pins but it'll take a few days as I have to scare up some heatsink paste first, not sure it is sold retail in my small town.

Note I am still on 1504 BIOS, not 1603... any chance that's an issue?

[Raja]

Try 1603 and see what happens. Use USB BIOS flashback to go back to 14** as well and see what happens.

The Dell USB hubs - I have encountered one situation in the past where these leak DC back into the board IIRC... (that's not a good thing to happen).

[EddyKilowatt]

As I mentioned in #6 above, this board has been failing POST, intermittently, since the first time I powered it up... before the monitor (or any other) USB port was connected. But if you'd like to outline a test procedure, I can check leakage current, impedance, source voltage, etc... I can get tools as needed.

I'll try the BIOS flashes using Flashback and report what happens.