Cant POST at 1600mhz anymore

Max1s

Distinguished
May 24, 2011
1,050
1
19,460
Hello all,

Important specs:
Intel i5 2500k
2x2GB G.Skill Ripjaws X DDR3 1600mhz CL6
ASUS P8Z68-V

I have a strange problem... once again.
Last night everything in my pc was running fine. When I turned it on this morning, it would not post and the "DRAM-LED" was red. Turns out I can only post and boot with running 1333mhz. After testing a few things I have established this:

1. Any timings or voltages at 1600mhz will not post.
2. I can run my XMP voltages and timings at 1333mhz without problems.
3. The RAM is not faulty. (At 1333mhz CL6 I've tryed individual sticks, different slots, 4 passes of memtest, some prime95...)

I have reseated my RAM and CPU, I have played with X.M.P versus manual settings, VTT voltages, and I think my RAM isn't at fault here....

FYI: I had my CPU overclocked to 4ghz but now I brought back default clocks hoping to rule that out.

Whats strange is that its not "faulty" in the traditional sense, it either works or dosn't work....
Any ideas? Faulty mobo?
 

raytseng

Honorable
May 15, 2012
666
0
11,060
If it was working before. Might be worth a try to do a cmos/bios reset. (The hard one via the mobo jumper/button/battery or whatever.

But ya, second the idea to check for bios update.
 

Max1s

Distinguished
May 24, 2011
1,050
1
19,460
No, I have the most recent bios. I reset the RTC + settings via a jumper but same problem.

Also, I told you that I brought my overclock back down to normal clocks? Well the turbo multiplier option is completely missing from the bios! (I noticed this before the RTC reset.)

I "updated" the bios to the same version, but I dont think it actually flashed, it just reset settings. I cant flash an older version because it wont let me.
 
My best assumption here, i.e. less than a guess, is that you degraded your CPU by using, being honest about it, foolish voltages. There are a few voltages that no doubt about it will permanently degrade a CPU. Otherwise hopefully you have a bad stick, try the Rated settings with (1) one stick.

Now for guesses, F3-12800CL6D-4GBTD with very high DRAM and VCCSA, and chances might include excessive VTT and/or PLL voltages. Heck perhaps a very high LLC with too high Phase controls sprinkle in a high vCore.

Q - What RAM kit?
Q - What were ALL of your OC BIOS settings? ; please list them all.
 

Max1s

Distinguished
May 24, 2011
1,050
1
19,460


Well thanks for bringing something new to the table but I doubt my OC of 4ghz at under 1.2v degraded my cpu in a matter of months... :p Other than the CPU offset vcore I hadn't touched my RAM voltages. Left all the VTT and PLL stuff on auto... LLC was also only on 50%.

4ghz Overclock Info
LLC - High (50%)
VRM Spread Spectrum - Disabled
Phase Control - Standard
Duty Control - Extreme (Current balance as opposed to thermal balance)
CPU Current - 120%
Offset voltage -0.17volts, bringing us to around 1.15 under load 100% at around 50 degrees celcius.


My RAM kit is the 2x2GB G.skill Ripjaws X 1600 CL6, designation F3-12800CL6D-4GBXH. It only needs 1.5 volts, so the 1.65 SB problem is out of the picture.
http://www.newegg.com/Product/Product.aspx?Item=N82E16820231444
 

Maxx_Power

Distinguished
Actually, the DRAM sticks degrade over time. DDR's were known to degrade from 400 max mhz to 333 max mhz. In the DDR2 generation, a lot of sticks certified to run at 1066 (above 800mhz) eventually had to settle at 800 mhz. In the DDR3 generation, same thing happens. A lot of the low latency 1600 ram eventually becomes degraded enough to no longer operate correctly on ANY board with ANY CPU after about a year of use. A year of use seems to be what it took for DDR2 as well. Some memory silicon dies degraded quicker (Micron G-die), some slower (Samsung TCCD), some doesn't seem to (Winbond), all of those were DDR examples. DDR2's I stopped tracking who made the die.

What you can try is to manually set timings instead of the XMP profile. So that the timings all match with the XMP profile (check on the sticker on the RAM), then increase the ram voltage up one notch, from what it is now. However, if the ram voltage is already at 1.65Volts, you have no more room to increase voltage. Doing so will damage the memory controller (over prolonged use) at >1.65 Volts. You can also try to increase its CL one notch.
 
Degrading a CPU can be done in a lot of different ways. I do not recommend an LLC -> High with a 4GHz OC, and it's not the OC it's the variety of Voltages. Example, I can set my vCore -> 1.60v and set the CPU to Stock or 3.5GHz; my CPU will become degraded in days. Frankly, for a 4GHz the LLC doesn't need to be touched and if you have instability adjust the Offset. Many 'guides' were dead wrong to encourage LLC changes for that low of an OC, and Duty Control use defaults e.g. T.Probe.

In the BIOS first Load Defaults (F5) now look at the VCCSA and perhaps the PLL, next AI Overclock Tuner -> XMP. Q - What are the VCCSA, DRAM and PLL Voltages?

/Edit - I find it rather disturbing 'why' the F3-12800CL6D-4GBXH were discontinued so quickly?! Many of the newer kits that have been discontinued set bad voltages. I recall a similar Patriot kit that set the VCCSA -> 1.35v which is (was) way too high and it was damaging CPUs.
 

Max1s

Distinguished
May 24, 2011
1,050
1
19,460
Aha, so you are saying either some part of my cpu (IMC?) degraded or my RAM degraded? Its only been 2 years with this ram, and 1 year with this 25k @4ghz.
Auto vs XMP vs manual voltages are all the same:
VCCIO (VTT) = 1.056
CPU PLL = 1.8
PCH = 1.081

What you both are saying makes sense, but if my memory degraded to the point where it couldn't run 1600mhz CL6 @ 1.5volts, it would be able to run 1600mhz CL9@1.5v right?
 
I'd be willing to bet the issue is your CPU **IF all else is the EXACT SAME including BIOS versions. A clear indicator of a Degraded CPU is that the SAME voltages no longer work BUT HIGHER do correct the problem(s). My experience in a degraded RAM (VERY RARE) is if in the process of exceeding Rated you excessively increased the DRAM Voltage +0.30v or higher, RAM has a toleration of ±0.20v; not always on the -0.20v but certainly on the +0.20v.

Let's try the following:
F5
/Change Intel SATA -> RAID if you have RAID or leave it
AI Overclock Tuner -> XMP ; by itself it appears to fail so add the following:
VCCIO -> 1.20v ; max safe voltage
DRAM Voltage -> 1.60v ; +0.10v
F10 ; Save & Exit = Yes
 

Max1s

Distinguished
May 24, 2011
1,050
1
19,460
I tryed running XMP @ 1.6 volts with 1.2 volts for VCCIO in RAID mode, same problem.... how likely is it that I degraded the cpu, even with LLC? How can we rule out the motherboard? I dont have any other DDR3 sticks lying around, although I might be able to borrow some in a few days...
 
Well first with settings that do work, I'd test the RAM with Memtest and see if any errors are cropping up first. Now IF you're getting errors then RMA the RAM; they're bad.

MOBO - IF you couldn't boot at all then maybe it's the MOBO.

A lot of things can affect the PC you name the component including 'clean power' from both the PSU and wall socket.

I assume you tried to Clear the CMOS (properly). Unplug the PSU for 5 minutes (drain all power for its capacitors), then using the Jumper method move the Jumper for 5 seconds or longer, place back; see video - http://www.youtube.com/watch?v=rdHH9KrceR0

Also, you might want to try re-seating the RAM a couple times then boot her up and try XMP again.
 

Maxx_Power

Distinguished
Like jaquith said, test your ram with Memtest, but if you have the opportunity, test the RAM/CPU on a different board (not both at the same time, one at a time) to see if the RAM is still at SPEC according to a different CPU/Board, and if the CPU is still at SPEC according to a different RAM/Board. This would isolate the component that is no longer able to operate at 1600mhz.

I know you said you increase the DRAM voltage, can you also dial back the timings a bit ? May have to do manual adjustments. RAM degradation is rare with modern high CL DDR3 (9 or up), but still common with low CL (<9) ram, because of the way they are binned.
 

Max1s

Distinguished
May 24, 2011
1,050
1
19,460
I've already reseated RAM, did a jumper reset and did several passes with Memtest.
Yeah my PSU isnt the best and could have knocked something out... :-0

I'll try clearing the CMOS "properly."
 

Maxx_Power

Distinguished


What is your PSU ? Did you update BIOS lately ? If you are going to update BIOS, be sure to set the RAM at the settings that your computer was perfectly stable at (not 1600).
 

Max1s

Distinguished
May 24, 2011
1,050
1
19,460
I tried clearing the CMOS properly as well as runing 1600mhz CL9 @ 1.6v, still no POST.

My PSU is a two and a half year old CoolerMaster GX750w. I read reviews where a rail (I think it was the 3.3v) was fluctuating badly, and other reviews where they said it was a great PSU.

I updated my bios a month or two ago, but the error occured a few days ago. Remeber that I cant POST at all at 1600, just a black screen and the DRAM_MEM LED is lit, so I wont be doing any BIOS update running 1600.
 
Couple BIOS tries, and all are safe:

BIOS2:

F5
/Change Intel SATA -> RAID if you have RAID or leave it
AI Overclock Tuner -> Manual
DRAM Timing Control /Enter
(enter) 11-11-11-30-2N ; those are as close to fail-safe as you can get

VCCIO -> 1.20v ; max safe voltage
DRAM Voltage -> 1.60v ; +0.10v

F10 ; Save & Exit = Yes

BIOS3:

F5
/Change Intel SATA -> RAID if you have RAID or leave it
AI Overclock Tuner -> Manual
DRAM Timing Control /Enter
(enter) 11-11-11-30-2N ; those are as close to fail-safe as you can get

VCCIO -> 1.20v ; max safe voltage
DRAM Voltage -> 1.60v ; +0.10v
CPU PLL Voltage -> 1.90v
F10 ; Save & Exit = Yes

DIMM Slots:

Normally use the Blue DIMM slots but just for, if nothing else, curiosity try the Black DIMM slots.
 
Regarding power, yeah sure no doubts if there's oddball or dirty power being delivered then sure all sorts of oddball problems can occur.

I know you know DDR3-1600 CAS 6 is fringe at best, and a tiny nudge can make it all unstable.

I had a UPS that took a dump and I was ready to replace my monitor because it was only working IF I restarted the system, but was failing when the PC woke from hybrid sleep. Go figure.
 

Max1s

Distinguished
May 24, 2011
1,050
1
19,460
Tryed the different settings to no avail. Its like theres some brick wall at 1600mhz. No matter what timings, no matter what voltages.

Dont know what you refer to as a "UPS" But I also had a wakeup problem on this board that fixed through a bios update. Thats the only reason I dont completely hate ASUS boards, they keep releasing BIOS updates that actually fix things.

Thank you for your help, I guess my next step should be taking my RAM over to my friends house and trying it out over there, or vise virsa. (There aren't any other 1155 or DDR3 systems in this house, excluding a Mac. Makes troubleshooting sooo much harder...)
 

Maxx_Power

Distinguished


Keep us posted on what happens with the swap try.
 

Max1s

Distinguished
May 24, 2011
1,050
1
19,460
If anyones still here, my friend FINALLY got around to testing my ram, (school just started and hes a lazy one) and he booted right up on XMP settings. Didnt do any prime runs, but he didnt have any problems, which compared to my black screen I would call "working fine."
So is it my cpu or mobo that it shot? Since the cpu multiplier option has disappeared from the bios, I'm starting to think my mobo is crapping out.
 

Maxx_Power

Distinguished


Could very well be the mobo then. Can you try your ram in your friends computer ?
 

TRENDING THREADS