GA-890FXA-UD5 (rev 2.0) & X6 1090T: no POST, 10 short beeps

[cp8086]

Hi!

In June I bought the components to build a new PC.

The "core system" was a Thuban 1090T with a Gigabyte GA-890FXA-UD5, 2 dual-channel kit (4x2GB: G.Skill DDR3 1600Mhz PC12800 ECO CL7 F3-12800CL7D-4GBECO - on motherboard's Memory Support List!), a Powercolor HD 5830 PCS+ and a pair of Caviar Black "FAEX" in RAID0 (plus a Samsung HD203WI, a PCI TV card, three optical drives, a floppy, some fans, etc.)

The system ran fine until September 15th, then it didn't pass POST: the speakers do ten 'short' beeps, pause, and so on (while nothing is displayed on screen).

On the user manual, such a beep sequence is not described.

I wrote to http://ggts.gigabyte.com.tw, but I didn't get useful info.


- I removed all external peripherals but keyboard (mouse, printer and so on)
- I replaced my USB keyboard with a PS/2 model
- I removed the only non-graphic card (a TV Tuner)
- I replaced my graphic card (HD5830 PCI Express) with an old S3 Virge (PCI)
- I removed all disk drives (floppy disk, hard disks, optical drives) (data and power cables)
- I temporarily reconnected some disk drives to SB850 or GSATA2 (one at a time)
- I removed all case fans (leaving CPU fan only)
- I checked CMOS battery with a digital multimeter and it seems good (2.99V)
- I replaced the original CMOS battery with a new (Duracell)
- I repeatedly cleared CMOS (both removing battery and with "CMOS_SW quick button")
- I temporarily removed keyboard also!
- I've checked PSU 5V and 12V via a digital multimeter on a peripheral Molex connector and they're OK (5.01V, 12.11V)
- I replaced the PSU (Enermax Modu87+ 600W with a Corsair CX400)

In all these circumstances, the PC did ten short beeps.

Later I removed all four DIMM (two consecutive dual channel kit: each module is numbered xxx59, xxx60, xxx61, xxx62) then I placed the last one in the first slot.
The PC did eleven short beeps!

I shut down and switched on several times, the PC always did eleven short beeps.

Then i restored the four DIMM in their sequence and the PC again did ten short beeps.

Last I removed my 1090T: it looked OK, then I reinstalled it and the Coolermaster HyperTX3 cooler. The PC again did ten short beeps.


Then I shipped my GA890FXA-UD5 to my supplier for RMA.

According to my supplier (www.f2store.it), my motherboard 890FXA-UD5 was good (tested with a Phenom II X4 955 Deneb), so I spent about fifty EUROs (including shippings) for nothing...

But when I got back my motherboard, the problems were the same, so I shipped my CPU Phenom II X6 1090T to his (another) supplier for RMA.

After three weeks, today I got another new CPU (9H09173F00xxxxx instead of the old 9E83403Exxxxx), but the problems were exactly the same!

Could be that 1090T "Thuban" (6 core, Turbo) uses some "resources" (e. g. some pin) not used by an older X4 Deneb?

Id est, could be that the motherboard works with Deneb but is faulty for Thuban???

Should I do again RMA for my Gigabyte GA890FXA-UD5?

Thank you!

[absic]

Hi there,
there, shouldn't be a problem with the 1090T on this board and it has been supported from the board's first release.

Although it isn't mentioned in the Manual the beeps that you are hearing, point towards a memory problem.
Have you been able to test the RAM with Memtest?

Have you cleared CMOS and tried re-setting BIOS? If you manage to do this, try starting the PC with only 1 stick of RAM, what happens?

[Dark Mantis]

Yes as absic has said 5-11 short beeps indicate a memory error.

[cp8086]

Hi, I thak you for answers.

<<shouldn't be a problem with the 1090T on this board and it has been supported from the board's first release.>>
The PC ran fine with 1090T for about three months (and I upgraded BIOS from F1 to F3 then to F4)
Then from Sep. 15th no more "signs of life" (no POST, nothing on display, beeps!)

<<the beeps that you are hearing, point towards a memory problem.>>
According to manual: 1 long 1 short: Wrong in DRAM or M/B

<<Have you been able to test the RAM with Memtest?>>
Yes, when the PC ran OK I found that the four DIMM (8GB) could work at 1600MHz with 8-8-8-24 latencies @ 1.41V (Memtest OK for about 4 hours!) (each RAM kit, on Intel LGA1156, should work at 1600 MHz with 7-8-7-24 @ 1.35V)

<<Have you cleared CMOS and tried re-setting BIOS?>>
I repeatedly cleared CMOS (both removing battery and with "CMOS_SW quick button")
As the PC doesn't pass POST, I can't set nothing in BIOS (frequency, timings, etc.)

<<try starting the PC with only 1 stick of RAM, what happens?>>
With one DIMM: 11 short beeps!
Without DIMM: 12 short beeps!

Are these other info any useful?

[absic]

Hello again,

If you are trying to run your RAM at 1600Mhz you may have a problem with the CPU. I have 8 Gig of RAM rated at 1600 Mhz but am only running it at default setting of 1333Mhz because of the known issues with the Memory Controller on the AM3 CPU's.

These CPU's can seem to be quite happily running with the RAM at 1600 Mhz and then fall over for no apparent reason and it seems as if this may be the case for you. I would recommend RMA'ing the CPU to AMD and, when you get the replacement, load Optimized Defaults in BIOS. That will make your RAM run at 13333Mhz but your CPU won't be at risk.

If you haven't already read through this article: http://forum.giga-byte.co.uk/index.php/topic,2515.0.html you might like to, as it does explain the situation and relationship between the RAM and CPU.

[Dark Mantis]

The chances are you have burnt out the memory controller on the CPU if youn have been running it at 1600 mhz. 1333 is considered max safe speed for AMD chips


Damn I must be getting slow. Thats twice now I've been beaten to the post!

[cp8086]

Hi, I thak you again for answers.

<<you may have a problem with the CPU>>
<<The chances are you have burnt out the memory controller on the CPU>>

I sent my first 1090T for RMA and I got a new one, which didn't work even a second!
(on the same GA890FXA-UD5 sent for RMA and got back as working - according to my supplier - with a Deneb 955!)


I read the very interesting post at http://forum.giga-byte.co.uk/index.php/topic,2515.0.html
However, my G.Skill RAM is on my motherboard's Memory Support List for 1600MHz and ran in un-ganged mode.
(Gigabyte tests only one kit/ 2 DIMM, perhaps with 2 kits/4 DIMM I shouldn't pass 1333MHz...even if Memtest was OK and PC ran OK for three months)

After I repeatedly cleared CMOS, when I installed the new 1090T the RAM should work at 1333MHz, but the PC didn't start even a second!

For these reasons, I suppose that maybe some "resource" of the motherboard (i.e. some AM3's pin) are faulty but used only by Thuban, not by Deneb.

Is that possible???

[absic]

Sorry but I don't have a concise answer for you with regard to pin assignment on the Thuban or Deban CPU's and I can't explain why the motherboard tested OK with a Deneb CPU but it won't work with the 1090T.

I think, under the circumstances, you might do better asking GTTS for the advice on the matter. You can start a Support Ticket here: http://ggts.gigabyte.com.tw/tech.asp?ClassID=2&Country=Italy&SourceWeb=B2C

[cp8086]

Thanks.

I began to write to http://ggts.gigabyte.com.tw from Sept. 15th but I didn't get any useful info (not even meaning of beeps not present on manuals!)

I point out that PSU and CPU fan work regularly.

I also tried to plug only ATX 4 pin from PSU and check the 12V line on the other 4 PIN (EPS12V: 4+4 PIN)  via a digital multimeter: OK

[absic]

I know how frustrating this must be for you but unfortunately I don't think there is an easy answer. As for Gigabyte Tech Support, you do have to persevere I'm afraid, they are not the quickest or best when it comes to replies.

The Memory controller is now built into the CPU and not the motherboard and the only way of knowing where the problem lies is to get all of the parts tested together.

My experience leads me to think there is a problem with the CPU or RAM but, I could be mistaken. Without having your actual components, on a bench in front of me I can only surmise as to the problem.

Have you tried using a different brand of Memory? I know you have tested your own but if possible try 1 stick of RAM produced by a different manufacturer.

[cp8086]

Thanks.

I have only 4 DDR3 DIMMs... I swapped them, I checked them one at a time...

And I have only one AM3 CPU, so I can't do "cross-tests"

I'd seek for a friend of mine with a recent AM3 PC... but it's difficult!

Unfortunately I bought CPU, motherboard and RAM from three different online stores!

I did RMA for motherboard (OK with Deneb according to my supplier and not replaced) and then for CPU (replaced with a new one), should I RMA my RAM now???

[absic]

Do you have local PC repair Shop who could test items for you?

I know this would cost some money but it would save you the trouble of RMA'ing the RAM if it turns out to be something else

[cp8086]

Local PC Shops (in my country) are very disappointing: they are very expensive (e.g. a 500GB hard disk could cost more than a 2TB on line!), they seldom have "new" products (released in the last 9-15 months), they have very limited range of products...

Therefore I bought components on-line since 10 years ;-)

I'd seek some friends for check at least the DIMMs (even on not-AM3 motherboard)

However thank you again.

[Tiger]

Have you tried using one stick of RAM in first slot and two sticks of RAM in the first two slots without using the 400 watt Corsair ps and S3 video card?
If not, then reinstall the original power supply and video card and see if it posts without using all 4 sticks of RAM.

[cp8086]

I plugged again Modu87+ and tested S3 Virge PCI and HD5830 PCI-E one at a time
Then i did many trials with four DIMM in various order, then with only two DIMM and also with one DIMM at a time
Then i removed all video cards: beeps were always the same.

4 DIMM: 10 short beeps!
1 DIMM: 11 short beeps!
0 DIMM: 12 short beeps!