Hard drive failures and the somewhat unlikely culprit
I've been through several different powersupplies, motherboards, Memory DIMMS, CPUs, Hard drives, etc... throughout my 7 years DIY pc building. In that time...only the latter half of probably 2 years have I noticed what I thought was hard drive failure. If you read my post (last from the bottom on this thread )
So I had two more drive failures on the same computer where I cranked my Vcore to 2 volts on my 8k9a3+. I have an antec true power 430 powersupply and thought this would be sufficient with 2 hard drives, 2 optical drives, 1 barton mobile cpu, a fx 5200 graphics card, 10/100 NIC, USB card, and a soundblaster live. I was racking my brain because both drives were giving me the click of death (the WD 250GB especially with about 20 clicks in 2 minutes). My 10,000 RPM 74GB raptor was no exception. It took me a nap to see the correlation because the symptoms after my overclock didn't appear for 2 weeks. My testing was as follows:
1st, I noticed that when running spinrite, my 250 GB drive, which had formerly been incessently clicking, was now running like a well oiled top. I then put that drive on it's own stand alone "curb feeler" power supply. That fixed the problem with that drive. For my feat at fixing the raptor, I dropped my vcore and overclock. This fixed that problem.
So now I pose some questions to you, my brethren of overclockers...how can I line condition a stand alone power supply (not necessarily ATX but it must not be a fire hazzard) for my 5 V rail and be sure my 12 V rail is adequate? I would like to do this at a modest price point. I do have a Smart-UPS 1500 and I'm not sure if this "cleans" up some of the power before it gets to my tower. Is there a way to keep an ATX power supply on constantly and safely? I was thinking of using an ATX tester, but it has warning labels that discourage full time use. Surely if a 430 Watt power supply isn't enough, then neither will a 550, which is why I was thinking of a 550 for my PCI cards, CPU, RAM, and the 430 for my optical drives and hard drives.
Please provide me with feedback, advice and opinions as I'm in electrical purgatory and would like to take some action so I can start using this computer at it's full potential.
I'm thinking either line noise in my house's old wire (the part that is still on the old fuse box) or my UPS is adding noise and preventing my system from overclocking to a 400Mhz memory and 200Mhz bus when it worked fine in the basement with a differant CPU (but same memory and PS). The basement is all new wires on a breaker box. The CPU is not the limitor, but I can't get stable at only 180Mhz bus now. It also might be some external EM interferance as the basement is likely better sheilded from that stuff.
The VCore should not effect your hard drive and your PS should not either unless it is defective. I have an Antec 480TB and 430 TP and have only had 1 hard drive ever die on my and it was a 6 year old WD 4.3GB drive I use EXTREMELY hard for at least 4 of those years.Maybe it needs to be on a differant line. Something on that line could be causing issues. The only thing I can think of that overclocking could cause is extra heat (which is usually what ultimitly cause the "click of death" on most drives) or extra EMI inside the case. Maybe better hard drive cables would help.
Raising the bus speed though might cause truble if the PCI bus is not locked. This with overclock the IDE interface and some drives are VERY sensative to this, but I though it would just not recognise the drive or cause data errors.
AMD Phenom II x4 945 3Ghz | ASUS M4A77TD | 2X WD 1TB SATA 2 hard drive | 2x2GB Corsair XMS3 | nVidia GeForce 8800 GTS | ATI TV Wonder Theater Pro 550 | Antec P-160 case | Antec 650w Earth Watts | LG Blu-ray Super Drive | LG DVD RW | Windows 7 Pro
Todd A, you brought up some excellent points with EMI and cables. I'm only using serial ATA for both drives, and I suppose that could cause it too.
I don't think the actual overclock affected it because I was only gonig by the multiplier and I underclocked with the clicking still present. Only when I dropped my vcore did the raptor clicking subside.
I tried taking it off the UPS as well with no difference (new cord to a different outlet too).
I'm going to borrow a 300Watt PWS tomorrow from a buddy and use it just for my hard drives and optical drives. I'm then going to try my old vcore and up my speed by 500mhz to where I was. I'll let you know my findings. I may still need a better PWS for my MB because Sisoft estimated my CPU Core power at 150 Watts when I was running full open.
Now before I get started, it may well be like you say. However, here are two other possible explanations:
1. If I was to take a wild guess for a click-of-death HDD failure, my first guess would be temperature. That's how mine failed.
Especially with 10,000 RPM drives, heat production is significantly higher.
It also has the perverse side effect of not appearing (A) shortly after boot-up, and (B) with an open case (until much later). Hence, what you think to be an effect of trying with another power supply might be just better ventilation.
2. If it's indeed because of overclocking, it's IMHO far more likely to be because of overclocking the PCI bus, than because of the VCore.
ATA Hard drives tend to accept only a narrow range of frequencies for their input signal. E.g., if the drive thinks it works in ATA-133 mode, it expects the data to come at 133 MB/s, or close enough.
SATA drives tend to be even more fussy about it. (And the Raptor is SATA.)
Basically it boils down to: can your mobo lock the PCI bus at 33 MHz regardless of overclocking, or does the PCI bus get overclocked too? Intel boards tend to have that lock, most boards for XPs didn't, and the K8T800 boards for the A64 didn't either.
Let me start off with #1. This is what I first thought, however, my hdd temps were around 29 degrees c and when I began getting the clicking, I bought a blower fan to further cool them to 25 degrees c (this didn't help either). As a matter of fact, I thought the 250 was so dead, I froze the bastard overnight. One entire day in an anti-static bag covered in Ice located in my freezer. I hooked it up the next day and the same clicking, I couldn't get any data back. So I tried what I mentioned earlier in the thread with putting it on it's own power supply, and viola...it's still working and hasn't clicked once. I'm surprised the son-of-a-***** is still working because of all the condensation that came off it after the freeze. So you were thinking the same as I with the temps...but yet, this wasn't the case.
Everybody seems to like the idea of the PCI bus, but all I do is change the multi. I leave it at the 166FSB. I've got an unlocked barton core.