Ghost in the system - Error Code 55 etc.

[melloj]

Hello there,

Long time listener, first time caller.

My system was as follows:

ASROCK p67 Extreme 4 Mobo
i7-2600K (running at 4,000mhz, so slight overclock from stock 3,400mhz)
Cooler Master heatsink
2xGigabyte GTX 770 Windforce in SLI
4x4GB Corsair Vengeance DDR3 1600mhz
Intel SSD boot drive
Corsair SSD secondary drive
WD HD for storage.
Creative fatality sound card
Corsair 850 PSU

I basically had this system since 2012, though upgraded graphics card and hard drive over the years. Was rock solid (though sometimes computer would reset during windows boot, but never thought much about it as it was rare).

About 2 weeks ago, started getting increasing BSODs in Windows 7. BSOD messages were varied, and occurred during random computer activities (internet, game, word processor). Then started getting just straight up hard locks - no BSOD or anything, just system crash. Then started getting trouble booting up with the Mobo show an error code 55 on the Debug panel on the Mobo itself. Error code 55 is an issue with the RAM according to the ASROCK mobo users manual. So tested each RAM stick individually using memtest86 and no issues were found. Tested the PSU with a Dr.Power device - all checked out.

BSODs and boot failure with Code 55 kept happening. Took off the overclock on the CPU back to stock speeds, same issues persisted.

Finally decided to take things apart and put them back together. When I removed the CPU, I noticed that there were some bent pins on the mobo socket and figured this must be the source of the problem (though hadn't been a problem for 2-3 years).

So got a new mobo, the ASROCK z77 Extreme 4. MOBO came yesterday, so I installed everything again. Boot the system - goes into error code 55!!! Put in only two RAM sticks at A2 and B2 slots (as suggested in the user manual), system boots fine and goes through some Prime 95. Put in all 4 ram sticks, system boots fine and goes through some Prime 95. Re-introduce overclock to 4,000mhz on the CPU, system boots fine and goes through some Prime 95. Temps are always more than OK. Go to sleep with a tentative (though confused) grasp on apparent success.

Boot system this morning - Error Code 55!!!! System won't boot.

So, we know it's not the Mobo. Is it the CPU or the RAM? I'd like to think it's the RAM because that's the cheaper fix, but I am wondering if the CPU is failing; perhaps the bent pins created a slow burn for it to eventually malfunction.

Any ideas to help me decide the next step here?

 

[xorbe]

My 2500k system went down the same way -- an increase of BSOD and hard locks after 2-3 years of use. No proof, but I suspect I cooked the CPU. Error Code 55 points to RAM though from a quick web search, for AsRock. But it could be the cpu memory controller too ...

 

[melloj]

Yeah, but from what I've seen reading around on forums, the error code 55 comes up for people when it actually isn't the RAM causing the trouble. I did forget to mention that at one point I started getting the error code for CPU initialization error (can't remember the exact code).

I suspect I'm going to have to get a new CPU here...

 

[Caring1]

How many times did you test each stick with Memtest?
You should do each stick for at least 6 runs.
Have you got another CPU you can put in the system to test it?

 

[melloj]

I think the default setting in Memtest it 4 runs, so 4 runs each. I also changed it so that each core of the CPU was being used one after the other for each test during the runs (I think the default is using only 1 core).

Unfortunately no spare CPU, which would be ideal of course.

I guess I could order the RAM off of Amazon and simply return it if the problem persists, certain then that the issue is with the CPU.

 

[sneekypeet]

Try voltages before buying more gear. Specifically a bump to the CPU voltage or the IMC, maybe even the memory. Little jumps of course, but it could be that one of the three requires a bit of love now that things have gotten older.

 

[melloj]

I assume by IMC you mean Intel Memory Controller? If so, is there a separate voltage setting for that within BIOS? I'm not in front of the computer right now, but can't recall seeing that anywhere.

 

[jsfitz54]

In addition to post #6 try running the IPDT: https://downloadcenter.intel.com/download/19792/Intel-Processor-Diagnostic-Tool-64-bit-

Yes to post #7

 

[melloj]

In addition to post #6 try running the IPDT: https://downloadcenter.intel.com/download/19792/Intel-Processor-Diagnostic-Tool-64-bit-

Yes to post #7

Thanks - yeah that was one of the first things I did; no issues came back. Have run it multiple times. That's been the most frustrating part, the fact that no tests have indicated any failures.

 

[jsfitz54]

So give the memory voltage a bump, one step at a time. Start there due to error 55.
What is stock voltage of those Corsair memory sticks 1.5, 1.6, 1.65?

If 1.5 try 1.6.

 

[bubbleawsome]

I assume by IMC you mean Intel Memory Controller? If so, is there a separate voltage setting for that within BIOS? I'm not in front of the computer right now, but can't recall seeing that anywhere.

If I recall correctly it means integrated memory controller. My motherboard does not have a voltage for that IIRC.

 

[melloj]

So give the memory voltage a bump, one step at a time. Start there due to error 55.
What is stock voltage of those Corsair memory sticks 1.5, 1.6, 1.65?

If 1.5 try 1.6.

Yeah its 1.5.

I've got the CPU set to 1.21 or so at the moment after I put in the OC. Might bump that up as well once I get home to closer to 1.3 and see how things go.

 

[Cybrnook2002]

Try voltages before buying more gear. Specifically a bump to the CPU voltage or the IMC, maybe even the memory. Little jumps of course, but it could be that one of the three requires a bit of love now that things have gotten older.

I think this....

Don't forget, just because 55 is a Memory error code. That does not necessarily mean the issue is on your memory sticks or Ram slots. Intel chips now have the IMC (integrated into your CPU), like has been pointed out. So if you have been running an OC on this chip for a few years (600 mhz is a good oc, not that modest) you might have been grinding at that IMC especially if your voltages were off, could have been starving other components.

 

[cadaveca]

This might also be due to updating BIOS, and changes made by the board maker to pre-defined timings for your sticks. "55" usually indicates a timing/setting problem, not a real hardware problem unless the memory stick itself is bad (and won't work at lowered speeds at all).

Also, this can be caused by a bad heatsink install, or the board warping over time. A heatsink that pulls on the motherboard can bend the board in such a way that the CPU does not seat properly in the socket. Continued thermal cycles can cause the board to form a permanent bend which affects how the CPU retention mechanism holds the CPU in place.

 

[zo0lykas]

try look here, maybe this link help's you

55 code

 

[OneMoar]

as dave hinted at I would pull the cooler and cpu .. again and re-seat it and then tighten it down in a cross-pattern
followed by a cmos clear and see what happens
over tight is worse then not tight enough
bump the LCC/ring voltage a notch or two as well

 

[melloj]

Alright, so have bumped up DRAM voltage to 1.6 and put in an offset on the CPU voltage to +0.01v (effectively around 1.3v at full CPU load).

Like I said above, it's a new motherboard just installed two days ago (as thought the issue was bent pins on older motherboard), so I doubt it's mobo warping here.

So far system seems to be running. However if as has been suggested, this is a symptom of the CPU/RAM nearing the end of its shelf life, I guess things will get worse again not too far in the future even with the increased voltage.

Tomorrow I might pull cpu/cooler again (even though I just put it in yesterday), and make sure it's properly attached. I usually tighten down until first feel resistance on the screw and not past that, which I assume is correct. Same with attaching backplate.

Thanks to all - will update if anything changes.

 

[melloj]

<sigh>

In the morning power up the system and Error Code 55. Then it just keeps rebooting itself without going into POST.

The way I get it past this is to remove 2 of the DIMMS, leaving 2 behind and then the system boots fine.

So I ran Memtest86 again (as I realize that the last time I ran it was before I put in the new mobo) with two of the older ram sticks (ver 2.12) in only and it crashed after completing the first pass (though didn't show any errors). Then I tried it again with all four sticks in and it would crash without even getting past 20% of the first pass.

Took out the two ver 2.12 sticks and left the newer ver 8.16 sticks and system seems to boot fine. I've logged into windows to post this update and will next try memtest on just the 8.16 sticks.

Seems to be however that the system gets unstable when all 4 sticks are in. I would suspect its the DIMM slots themselves, but like I said this is a new Mobo and the old mobo was doing the same thing, so I doubt I have such bad luck that my previous mobo developed a DIMM slot malfunction and the new board came with one. Like I said in the original post, on the old mobo I ran memtest on each stick individually and didn't get any errors or crashes.

Have some new RAM coming in next few days and will see if that resolves the issue. If not, I suspect it may be the CPU memory controller, but not sure how to test that on its own.

 

[Caring1]

Individually the sticks of Ram may be fine, but there may be an incompatibility issue using the old and new sticks together.
I'd say at least one of the old sticks is faulty.
Run the full test for the recommended amount of runs on one stick at a time.

 

[jsfitz54]

Do you use the XMP memory setting or do you manually input the settings in the Bios?

Do the different memory versions have different ram voltage, ie old set 1.65v vs new at 1.5v?

 

[melloj]

No, they're the exact same ram sticks, just bought with a few months apart. I generally let the BIOS determine the ram stick settings automatically, but I have verified that it is showing up as 9-9-9-24 in the BIOS screen, which is the recommended setting for this ram.

So I had inserted the two ver 8.16 ram sticks on their own and left Memtest to run while I went out. Came back home and noticed that the computer had hard locked at the start of the second pass, the same as with the ver 2.12 sticks, so that's another theory out the window.

I can run each stick on its own again to see, though I had tried that before when I had the old Mobo and nothing seemed amiss.

As it doesn't seem that the ram itself is amiss, and I can't imagine the old mobo and the new mobo both have faulty ram slots, that seems to point at the CPU doesn't it?

Though I have to say once I actually boot up, things seem to be fairly stable. Haven't had a BSOD or a hard lock since I bumped the voltage....just that it stops booting once I have all 4 sticks in.

 

[OneMoar]

IMC has had it ...

 

[melloj]

Yep. Will test it with the new ram once it arrives. I suspect if the issues persist I'm going to be looking at a new ivy bridge cpu. Probably the i7 3770k if I can find one kicking around.

 

[jsfitz54]

for giggles, have you tried to loosen the timing on the ram?

Corsair stock #'s:
9-10-9-27 for 1866.

9-11-10-30 for 2133
9-11-11-31
11-11-11-27

My thought is that with all 4 Ram slots populated, with the overclock you were running, that the memory needed a bump in voltage to maintain good working order at the start. Being under volted (at stock v) has lead to ram failure (starved v).???

 

[fullinfusion]

code 55 from all my experiences has lead to a bad memory stick.. I had a set of Muskin rams sticks and spent an hour on the phone with Shawn and with all the tests he asked me to do and timings to try it was a bad stick.

Sent the kit back and got the replacements, they were flawless after that.