Batwing~SPARTA~ Posted September 17, 2011 Share Posted September 17, 2011 Hi guys Today, out of the blu, my PC decided to crash with no apparent reasons 4 times in a a row. It happened twice under load (gaming) and twice on stanfdard desktop working (surfing internet or opening windows) After a preliminary check about temps (they didn t show up as "danger zone"), I started making a stress test with prime95. Under stress no issue, curiously it crashed whenI decided to close the test. I was closing all windows and i got the crash. Nature of this crash is Video goes out at first, got the DVI no Data on monitor, computer still seems running and after few seconds, final total failure - reboot needed. Today is the first time is doing this stuff. There were no sensible changes between yesterday and today, no updates, thunderstorms, nothing. I thought about possible PSU going failing, so on the screenshot you will see also power readings, but honestly nothing wrong there at first look. I decided to lower down a nitch my OC settings from CPU @ 3.67 to 3.57 and RAM from 17xx to 13xx and tested again. At this very moment seems very steady and after 25 minutes had no crash (yet). Take a look here, this is the Steady ongoing testing: Any thoughts is welcome Link to comment Share on other sites More sharing options...
Skaz~SPARTA~ Posted September 17, 2011 Share Posted September 17, 2011 check HDD to see if its corrupt. Link to comment Share on other sites More sharing options...
NoScream~SPARTA~ Posted September 17, 2011 Share Posted September 17, 2011 welcome to over clocking, instability when you least expect it..... I've run 2 years with W7 and it's solid but with XP I had many issues, it transpired most were MEMORY running over clocked & driver issues. I would always recommend getting high grade memory & the best motherboard you can afford, not for features but for quality. As you over clock, it stresses all data channels....... I know many wanting 4.0Ghz+ but I'm happy sub 4G with no crashes. 3.57 is a good speed. Link to comment Share on other sites More sharing options...
Donziboy2 Posted September 17, 2011 Share Posted September 17, 2011 Run memtest, also see if windows created a dump file, it may be of some use. Link to comment Share on other sites More sharing options...
Batwing~SPARTA~ Posted September 17, 2011 Author Share Posted September 17, 2011 I am starting thinking it was the RAM a bit too much OCed.. actually it ran solid for 5 hrs witn no crashes after I lowered the RAM a bit.. however.. it has been working perfectly for months.. oh well it works now, I don t complain Link to comment Share on other sites More sharing options...
Hajimoto Posted September 18, 2011 Share Posted September 18, 2011 (edited) I 100% agree with Viiiper that you are experiencing OC Pains. The thing about OC'ing is that the additional loads placed on a chip will have an effect overtime, what the effect will be, is a crap shoot. It could be a slow degradation of circuits which eventually start to timeout, the component could be fine for 5000 hours and then catastrophically fail or you could run the sucker at full OC for 5 years and experience no change what so ever, it truly is a crap shoot. All of us OC'ers know that when we consider start an OC project, that we have to research the hell out of the manufacturing locations and production batch numbers in an attempt to minimize flaws in the chip for a more stable OC. These flaws have zero effect on a chip that will be performing within the manufacturers specifications but the flaw will start to rear its head if pushed beyond its intended use. AGE As electronics age, the molecular structure of the materials they are made from change which have an effect on their ability to deal with voltage/current. These changes effect the components ability to pass electrons at the same efficiency as when it was first produced. This inefficiency is translated in a couple of ways, one being heat, another is failure in timings because the voltage/current is not within the expected range and the circuit fails or times out. In order for the circuit to perform at the SAME load a solution is to adjust voltages to compensate for the degradation of components. Of course this increase in voltage produces more heat which requires better cool and the circle continues. Another way to stabilize the circuit is to lower the load (reduce the OC) which is what you have done, but I urge you to get more information as to what the variables are that are degrading to assist in the next stage as to what the degradation is. Example: If you calculate the change in systems instability and determine it is directly related to RAM voltage, you are in a better position to diagnose the system should there be another change in the future as the new calculated baseline can instantly be referred to. Simply reducing the OC does not tell you which component has changed, and for some I guess it is not important, but for me, as an engineer, it is most important to determine the cause or else I run the risk of catastrophic failure in the future without a complete understanding of why and therefore I am condemned to repeat failure. HEAT It has been my experience that one of the most common factors for the sudden change to an otherwise stable build is ................. Wait for it............changes in the efficiency of cooling. Take a good look at all of your air moving and heat transfer components for dust and build up which substantially reduce cooling. Below is my check list of items to confirm right out the gate : (as they relate to cooling) Air intakes and exhaust filter media clean. Fan blades clean and rotating smoothly. Heat exchanger air pathways clear and unobstructed. Fan RPMs within 10% of baseline build numbers. Ambient temperature within 5-8 degrees of the baseline numbers. CFM counters on air exchanges through the case within 85% of baseline numbers. Idle and active temperatures ranges compared to baseline numbers. Hope this diatribe helps.... LOL Edited September 18, 2011 by Hajimoto Link to comment Share on other sites More sharing options...
Batwing~SPARTA~ Posted September 18, 2011 Author Share Posted September 18, 2011 Hey there Haji I think you are spot on about degradation, possible heat issues and voltages. Lowering a little notch the OC (actually 3.57 is still a respectable speed) is giving again a very steady performance. No crashes in the last 24 hrs. However is important to remember I know I have some unidentified issue with my MB. This is a plague on my system since quite long time. Every once in a blu moon, this MB reminds me its inettitude. I am not taking a direct action at the moment becauseI am waiting for a MB full failure and then I will evaluate a full system upgrade. Now I am running this system (MB, CPU and RAM) OCed with full pleasure since 2 years. I think it is serving pretty well. I think would prefer waiting at least another year before thinking of new upgrades. At this moment, reducing a bit of stress, I gained 5 - 7 degrees less heat on CPU but most importantly I am stressing the MB circuits a bit less and I honestly think I was pushing the RAM a bit too high. I am thinking around Xmas a small adjustment getting the Corsair CWCH50-1 http://www.compusa.com/applications/SearchTools/item-details.asp?EdpNo=5267687# and some new fast RAM Cas Latency 6 http://www.newegg.com/Product/Product.aspx?Item=N82E16820231352 This adjustment will not solve the issue on a possible failing MB but at least can be recicled as soon as the MB fails and I ll get another one Link to comment Share on other sites More sharing options...
Halli~SPARTA~ Posted September 18, 2011 Share Posted September 18, 2011 Why not Not overclock till you get a new system? Link to comment Share on other sites More sharing options...
EBE Posted September 18, 2011 Share Posted September 18, 2011 did you check for dust i know you will get there and fix this. Link to comment Share on other sites More sharing options...
Recommended Posts