Sudden OC instability/broken cdrw. HELP!

[superslug2]

System Specs

Sempron 3100 (paris)
DFI NF3 250GB LanParty
2x512 Twinmos SpeedPremium UTT
eVga 6600GT

So originally I only had one stick of the Twinmos Ram and i came home, set everything up, cranked the vcore to 1.6, lowered the LDT to 3x and upped my FSB to 268. Instant 2.4ghz with a divider. Then went on to raise the vdimm voltage to 2.9 and timings to 2-2-2-10 (1T) Stable... memtested 24 hrs, Primed on max priority for 22 hours, rock solid

Then my problems begin. the system begins to lose integrity as some errors begin to show up while playing World of Warcraft. At that time, i thought what the heck, i'll just reformat. At that point things start going downhill. I install Windows on a CDRW drive, but once i install the mobo controllers, windows hangs. I cant get it to start and it just sits there forever. Mind you I INSTALLED on this drive too and there were no errors. Funny thing is, during my first system build, everything was flawless. The only thing that changed was the second stick of ram but i memtested and it was fine.

When i pull out the drive (disconnect the ide cable from the mobo) windows boots fine which leads me to think that my drive is fried.... i have no idea how this happened. i was told that my OCing might have done it since it could be the case that the frequency locks on my mobo wernt working properly... i dont know

I took out the CDRW and redid everything with another generic CD drive i had around, everything is alright... except now sometimes on coldboots, the system restarts. And or does not recognize my video card and tries to install it again saying its new hardware. Further restarts alleviate this problem.... is this a power supply problem?? Im using the stock PS 380w trupower from a sonata case

Id like to know what might be causing these problems.... ram... powersupply... bad mobo

any help would be great

ty

 

[superslug2]

bumpage

 

[imported_whatever]

did you add another 512 stick of RAM at some point? s754 freqently will not work with 2*512.

 

[superslug2]

yeah, i added another stick... but now when i take a stick out... its still pretty much messed up. and the cd drive still causes windows to hang.... powersupply maybe?

 

[redhatlinux]

I can't be sure that this is the exact same situation, but here goes:
Many moons ago, when Ocing was much less of an art than it is now I had problems with an AMD chip, that was sold to me as
Overclocks great, passes tests 100%. It did but at some point into the installation of some product, maybe an early version of Office, CD drive would fail. A few RMA's on CD drive later and still failure. I've seen this problem on other Oc'd chips as well.
Just today, I'm looking at CPU and MOBO combos for a friend and ask if the Semperon is a "Hottie" like my old Duron and sure enuf, yep, need a pretty big HSF. Power Supply shouldn't be a problem 300 watts should be OK, unless your intending on putting in a mega graphics card, says a very reliable techie I know. just 2 cents from THE DOC

 

[superslug2]

hey there, thanks for the reply

as far as my CPU temps go, im completely fine. im running a zalman 7000b with as5 and my temps idle at under 30 and at max on a burning hot day maybe hit 50

 

[redhatlinux]

SLUG... its not the temp that you read thats the problem, internal core temps are the big deal. I'll try to make it simple to explain better.
Transistors are a basic binary switch, off/on. The ON state used be the hot-spot. With gate thickness of 90nm .. 60nm etc the Off state is not really off, its like a fosset that drips, muliply that by the uber # of transistors in a cpu and the internal leakage and hence the heat is the big deal. Trust me... I've been into the innards of cpu's for over 30 years... heat, voltage, EMI, transient noise etc etc will cause problems. I'm not sure why CD readers are the first to show problems, I think that its either timing/noise/voltage .... on the ATA bus that HD's either don't see, or retry on the fly and never even report an error. Even moving the power for the ZALMAN off the mobo, and on to a rigged up PS connector could solve you problems.

 

[redhatlinux]

SLUG... its not the temp that you read thats the problem, internal core temps are the big deal. I'll try to make it simple to explain better.
Transistors are a basic binary switch, off/on. The ON state used be the hot-spot. With gate thickness of 90nm .. 60nm etc the Off state is not really off, its like a fosset that drips, muliply that by the uber # of transistors in a cpu and the internal leakage and hence the heat is the big deal. Trust me... I've been into the innards of cpu's for over 30 years... heat, voltage, EMI, transient noise etc etc will cause problems. I'm not sure why CD readers are the first to show problems, I think that its either timing/noise/voltage .... on the ATA bus that HD's either don't see, or retry on the fly and never even report an error. Even moving the power for the ZALMAN off the mobo, and on to a rigged up PS connector could solve you problems.

 

[superslug2]

hmm, i think i see what you mean

thanks =)

 

[redhatlinux]

A few more 'ideas'. Semperon's run hot because AMD basically OC'd them, they squeeze all the speed they can. Low end cpu's are cheap but they sell a whole bunch. The 'headroom for user OC is therefore small, to have a stable system. Next, AS products in general are pretty difficult to apply correctly(Intel and AMD never use it, only PCTC ... pads). AS has been known to dry up, the carrier for the silver dries up and leaves gaps. Large HS's with big fans kinda 'rattle the compound and it seperates a bit.
Possibly the real explanation your CD-R/W and HD are set as a DMA device's, Direct Memory Access, thru the chipset. You should NEVER install OS software with your System OC'd. See, most of the time your HD is working thru its own cache, but OS install's are so demanding that the HD cache is pretty useless and data comes straight to/from the heads. With an OC bus, data is written slightly off sync, then when you read back OC'd, the off sync effect is doubled. When you added extra memory, you changed the memory timings slightly, 2 chips on the same bus always do, now you have more sync slip and .... errors. It's real hard to describe if you have never looked at a data bus and clock using a 'scope and seen the slip.