Im planing to change capacitors on my motherboard to a more powerfull. Example 1000uf 25v into 1000uf 35v. So im wondering will my cpu benefit with bigger clocks and better stability?
To be honest, I don't think it would make any difference at all, unless you installed a higher quality capacitor (which is more of a case of being able to live up to the specs long term under load). The maximum voltage along that capacitor line is probably only a couple of volts anyway, the higher voltage of the capacitor just means the capacitor can stand 35V - something which you won't want on a line meant for only a few volts. To get any possible change, you might like to try a slightly higher capacitance, because that will make a difference!... just install one with slightly higher capacity, say 1100uF 25V. You can also go 35V etc, but like I said that shouldn't make any difference, and they just cost more I said it will make a difference, and that is true, but whether it leads to a higher or more stable overlcock or not, I somehow doubt it That said, installing a higher capacity capacitor can lead to a more stable voltage if that is an issue, and it is that which will give you benefit (a capacitor acts as a voltage stabiliser). If you install a capacitor with a too high of a capacity, it will take longer to 'respond' and the end result is the concept won't work. A change from 1000uF to 1100uF should be fine though.
Finally some1 replied. Ty very much for ur reply. I will try it tomorrow if i have time. And ofc reply back with results. Oh yes one more thing which i almost forgot. CPU sometimes boot oc'ed to 2.6 and sometimes not. For instance it will boot on that clock 5 times in a row but 6th will fail. Most of the time helps if i just unplug psu cable and plug it in and it boots up again. Any thoughts on that matter?
Zaeb, I would recommend not doing this. You do not know the schematic for the motherboard or what exactly that capacitor is doing. A 35V capacitor is not "stronger." It just can handle higher voltage, which is not something you need. If the capacitor you are replacing is rated for 25V, that means its purpose is to make sure the voltage regulator does not produce too much noise on the 12 V line. Making this capacitor larger will not help your overclocking. Also, you run the risk of damaging your motherboard since it may be difficult to solder/desolder these large capacitors, especially if they are using unleaded solder which is typical these days. Edit: With regards to your 2.6 Ghz question, that just means your computer is not stable at that overclock. This is what happens when you're at the limit of your overclock, sometimes it posts, sometimes it won't post. Most likely, your computer will always crash at 2.6 ghz when you give the CPU 100% load with a program like prime95.
Wont help at all unless the caps you are replacing have a fault that is causing you a problem. Higher value/rated caps will not give you higher overclocks than a normally working motherboard.
thank u for the reply guys. this helped me allot. u conviced me i wont do it :/ regarding my 2.6 oc it is not true what u sed xl_ence. im stressing it 100% for a few hours and it didnt twinch a bit. it maybe that the mobo is just to weak and to obsolete even for this also old cpu.
That's interesting. What program are you using to stress it 100%? Perhaps it is a memory overclock problem then. You're memory is probably OK if you're using prime95 to stress it 100% though since that programs stresses memory and CPU.
i didnt used any programs i encoded some video and cpu was at 100% for some time at least an hour. memory can stand up to 1066mhz but i lovered it to 800 so when i oc its not above 850. there is one strange thing though about my memories. i have 2xkingston hyper x 1066 and 2xocz flex xlc 1100. my mobo supported 800mhz from the box but thrue bios updates it supported 1066. when i put only kingston and leave it to bios to set mhz it sets it at 800 so i manualy adjust it to 1066. but when i put ocz bios sets it at 667. and i hed problems running both memories at 1066 but eventualy i managed them to work at 1066. its just strange that when i boot to 2.6 pc can run for days without problems. meaning runing latest games and stressing cpu in allmost any condition. but sometimes when i restart pc wont boot. when it doesnt want to boot it helps when i increase voltage on ht link and chipset. cpu is on 1.250 default when oc 1.300 ht link and chipset 0.50.
You need to do tests that specifically push components to the max or you may spend a long time locating the problem.
just started to run prime. all 4 cores at 100%. i forgot to mention this. when i put kingston in first 2 slots then ocz i can boot at 1066 but if i put ocz first then kingston there is no way im booting at 1066.
The default test on Prime95 is "blend". This is a good test for Sandybridge CPUs but isnt as good a CPU test for AMD kit. Use small FFTs for AMD CPUs. P95 blend is not as good as a memory test as "memtest" which requires booting from a CD or USB pen to use. Given your last post, its worth running memtest overnight.
i never do that. first 2 slots r red and second 2 r yellow. so i follow color guidelines to place ram meaning kingston in red and ocz in yellow. i will try memtest just to be shure its not ram so i can narrow down the problem. prime is runing for an hour now and no errors. ram usage is increasing 3.1 at the moment. cpu 71 celsius.
The only time I have ever messed with capacitors were a few weeks ago when my neighbor gave me a 22" Samsung 225bw LCD display that when turned on, showed a picture for a fraction of a second then went dark. I took the display apart and on the power supply board were bulging capacitors which I replaced with new ones that were a better brand from a repair kit i found on Ebay. Cost $12 to fix the free display. Works great.
Mixing and matching means brands also. The Kingston may not be playing nice with the OCZ brand or vice versa. It is best to by matched sets, using 4 sticks, pick a brand and buy a package that contains 4.
Blend is best for long testing because it tests a "blend" of small and large FFTs, hence the name.. The blend test contains both small FFTs which fit inside the CPU cache and large FFT which are too big and must be stored in system memory.
With Sandybridge (and probably IB too), P95 "blend" works best for finding your stable overclock (and memory stability to some degree), but with other chips I found that "small FFTs" was best. For testing memory properly, there is no substitute for memtest.
