GA-990FXA-UD5 memory errors

[absic]

What Voltage have you set for these tests?

Have you tried 1333 MHz at 10,10,10,27 with Voltage at 1.5V?

[Nihth]

For the other tests I did 1.6V at 1600MHz and 1.5V at 1333MHz

I just tried the 1333 MHz at 10,10,10,27 with Voltage at 1.5V and it dident work.

The reason why memtest is running at 200 MHz DDR400 is that it cannot detect the SPD values due to some SMBUS or something.

[absic]

Well, as the things you have tried don't seem to be helping, all I can think of now, is waiting to see what help/advice Gigabyte Tech Support can give.

[Nihth]

Okey, I will contact them.

Thanks for your help!

[Nihth]

I filled in a form at Gigabyte, but they wont open before 2nd Jan, will see what happens then. Thanks again!

[JasonPW]

I just received a RMA'd rev 1.0 GA-990FXA-UD5 from Gigabyte.  The symptoms I was getting were:

PC 100% stable AMD 1100T over clocked to 3.8Ghz. 4xPC1600 DDR3 16GB.

Upgraded CPU to AMD 8350.  Memory errors, freezes, failed memtest, frozen memtest, unable to run at PC1333 speeds reliably. Tried new memory (2x8GB from Crucial and Corsair) same issue.  Finally got it stable by disabling turbo core and manually setting voltage higher to compensate for vdroop and memory at 1600.  I was convinced it is a vdroop issue on the rev 1.0 so I got hold of a rev 1.1 version of the board with LLC control in BIOS and it ran all previously failing memory at 1866 speeds and the 8350 over clocked to 4.6Ghz (Turbo disabled).

Now I haven't tested this returned board (Staying with my 1.1) which is my original rev 1.0 (Checked S/N).  But I can see it has had a resistor changed near the voltage controller next to the CPU socket.  I know there was a DIY fix (If your good at soldering and have very good eyesight) that replaced a resistor on the UD7 which disabled vdroop on it's rev 1.0 so I suspect they have simply replaced the same resistor on the controller to do the same here.

I recommend RMA'ing it if it's under manufacturers warranty to get this rework.  If not the rev 1.1 sorted it here.

[Beetle]

I have also a GA-990FXA-UD5 but rev. 1.1 and I also have problems with 8GB DIMMs (Corsair Dominator Platinum 8GBx2 CMD16GX3M2A1866C9).
 I have run the memory both in 1333 & 1866 with both 1.50v and 1.65v but memtest86+ finds error. The memory works in GA-F2A85X-UP4.

The answer I got from Gigabytes support was:
Well, since the DIMMs are not officially supported on this motherboard that might be the case why they aren't working in your system.
Even if you try to use them in lower clockspeed and different voltages it does not necesarly make them compatible, since the DIMMs are factory clocked.

[JasonPW]

I'm still sure I had to adjust the LLC calibration setting for my 8350 CPU to get the memory stable even on my rev 1.1 at default clocks.  At 4.6Ghz I defiantly have (I cannot remember settings off the top of my head)

The problem on the rev1.0 I was chasing memory errors where none existed.  The memory went from working with the AMD 1100T CPU to not working with the 8350.  I think the symptoms I was getting was because the CPU was under volted too much by default on the rev 1.0 (vdroop) with the 8350 CPU and as the memory controller is within the CPU I could get memory errors or freezes.  I changed the CPU back to the 1100T and the problems went away.  I RMA'd the 8350 and the same fault recurred with the new one.  The only thing left to try was the main board and the rev1.1 was the fix for me. (The modified rev 1.0 has vdroop disabled so in theory should be OK as a well)

For Reference I am now running GEIL 8GBx2 DDR3 at 1866 dual channel with no problems. 9-9-9-24 1T at 1.60V

[Marco76]

Hi all!

I'm having the exact same problem as Nihth....

2x8GB Kit Venegance LP Corsair DDR3...

Memtest86+ test # 6 errors....

Tryed hundreds of combinations in BIOS settings with no result...

Can someone help??

Thanks

Marco

[Nazra7]

rev 1.1 here and it made it past #6 memtest +86 v4.2 with these settings.  I would love to see someone test these further since I want to game tonight!



4.2ghz on an 8120fx all garbage c1e ect turned off including turbo.

1.56v to ram this might be too much
.25 to cpu
.75 to cpu-nb

9.33x  1866 w/ 2/4 sticks installed (I split the kit with my nephew to give the controller the stability for OCing)

[Nazra7]

Nope error'd on test #7... called Corsair Tech got a canned if it doesn't work at 1333 in memtest send it RMA.   SHA as if the new kit would fix it.   So I nope'd that set DOCP to profile 1.  1.055v dram it show'd,  error'd so I tweak'd it a little,  changes:

10-11-10-30  with 160ns(dont think that changes anything) also set the 7's to 8's to loosen em up some.
went to 4.0ghz overclock for testing and passes beyond test #8 in memtest (YAY)
for the 3.0revision boards make sure to set Tfaw at 28!  that did the trick for the nephew's build.

So it looks like voltage is a bad thing in this set.

Obviously something in the XMP fixed the stability enough to find a comfortable spot but I'm not an expert just a push this and see what happens dude.

If someone out there has this setup as well and would like to share please do so...


(I exported the BMP from overdrive if it helps)

[Nazra7]

Proof well at least one pass...

[EvilNucca]

Had the same issue with same board and same ram contacted corsair after getting errors in memtest86 and windows memory diagnostic the tech at Corsair said the memory should work fine in my board tested all 4 sticks 1 at a time got errors in two took the good sticks  put those in the 1 and 2 slots everything worked perfect RMA to corsair waiting on the new sticks to see if they work with all four. Seems that Corsair has issues with the Vengeance Ram I know of two other incidents where the ram failed or had errors and after getting new ram from corsair everything worked fine. Hopefully I will get decent Ram and it all works fine if not I will Ebay the sticks off and get some samsung Ram 8gb sticks that are a lower speed.